DEVELOPMENT OF A YAM DISEASE DIAGNOSIS EXPERT

SYSTEM

BY

COMPUTER SCIENCE GROUP 1

REGISTRATION NUMBERS (001 – 022)

DEPARTMENT OF COMPUTER SCIENCE

FACULTY OF PHYSICAL SCIENCES

NNAMDI AZIKIWE UNIVERSITY

AWKA

SEPTEMBER 2024.
DEVELOPMENT OF A YAM DISEASE DIAGNOSIS EXPERT
SYSTEM

BY

COMPUTER SCIENCE GROUP 1

REGISTRATION NUMBERS (001 – 022)

A PROJECT SUBMITTED TO THE

DEPARTMENT OF COMPUTER SCIENCE, FACULTY OF

PHYSICAL SCIENCES, NNAMDI AZIKIWE UNIVERSITY,

AWKA IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE DEGREE OF

BACHELOR OF SCIENCE

SEPTEMBER 2024
CREDIT

GROUP ONE MEMBERS INCLUDE:

1. NWANKWO ONYINYECHUKWU VIVIAN – 2020514001
2. CHUKWUDI MARVELLOUS CHUKWUZURUM – 2020514002
3. OKOYE-CHINE KENECHUKWU DANIEL – 2020514003
4. ONWUDIWE ALVIN MORRIS – 2020514004
5. ANAZONWU CHIKE JEFFREY – 2020514005
6. UKAEGBE SOMTOCHUKWU CHUKWUDIKE – 2020514006
7. NWAFOR CHIEMERIE GODWIN – 2020514007
8. EMEASOBA CHINAZAEKPERE CHINYERE – 2020514008
9. OBIAJULU TOBIAS EKENE – 2020514009
10. UCHENNA CHUKWUEBUKA MARTIN – 2020514010
11. ORAEMESI OLUEBUBE JANE - 2020514011
12. CHRISTIAN EMMANUEL IKECHUKWU – 2020514012
13. OKPALA JEREMIAH TOCHI – 2020514013
14. AROTU NKIRU MARIAN – 2020514015
15. EBONYI EBUBECHUKWU VICTOR – 2020514016
16. AMAKOM RITA-MARY NGOZI – 2020514017
17. IGBOELI CHUKWUEBUKA RAPHAEL – 2020514018
18. NWOSU DIVINE EBUBE – 2020514019
19. LARRYSON CHIJIOKE GIFT – 2020514020
20. JIDEANI CHUKWUEMEKA CELESTINE – 2020514021
21. IGWE VALENTINE CHIMOBI - 2020514022
DEDICATION
ACKNOWLEDGEMENTS
ABSTRACT
TABLE OF CONTENTS
TITLE PAGE ..............................................................................................................................................................
COVER PAGE.............................................................................................................................................................
CREDIT (GROUP ONE MEMBERS)........................................................................................................................
DEDICATION.............................................................................................................................................................
ACKNOWLEDGEMENTS.........................................................................................................................................
ABSTRACT.................................................................................................................................................................
TABLE OF CONTENTS.............................................................................................................................................
CHAPTER ONE..........................................................................................................................................................
INTRODUCTION............................................................................................................................................
1.1. BACKGROUND OF THE STUDY.........................................................................................................
1.2. STATEMENT OF THE PROBLEM........................................................................................................
1.3. SIGNIFICANCE OF THE STUDY..........................................................................................................
1.4. SCOPE OF STUDY..................................................................................................................................
1.5. AIM AND OBJECTIVES OF THE STUDY............................................................................................
1.6. DEFINITION OF TERMS........................................................................................................................
1.7. LIMITATIONS OF THE STUDY............................................................................................................
CHAPTER TWO..........................................................................................................................................................
LITERATURE REVIEW.................................................................................................................................
2.1. THEORETICAL REVIEW.......................................................................................................................
2.1.1 SYSTEM IN AGRICULTURE...................................................................................................
2.1.2. LEARNING APPROCHES IN DISEASE DIAGNOSIS..........................................................
2.1.3. CHALLENGES IN DEVELOPING EXPERT SYSTEM FOR
AGRICULTURE..................................................................................................................................
2.1.4. ROLE OF EXPERT SYSTEMS IN SUSTAINABLE AGRICULTURE..................................
2.1.5. FUTURE TRENDS IN AGRICULTURAL EXPERT SYSTEMS............................................
2.2. REVIEW OF RELATED WORKS...........................................................................................................
2.3. SUMMARY OF LITERATURE REVIEW AND KNOWLEDGE GAP.................................................
CHAPTER THREE......................................................................................................................................................
METHODOLOGY AND SYSTEM ANALYSIS............................................................................................
 3.1. METHODOLOGY OF SYSTEM ADOPTED.........................................................................................
3.2. ANALYSIS OF THE EXISTING SYSTEM............................................................................................
3.2.1. DATA FLOW OF THE EXISTING SYSTEM..........................................................................
3.2.2. WEAKNESSES OF THE EXISTING SYSTEM.......................................................................
3.3. ANALYSIS OF THE PROPOSED SYSTEM..........................................................................................
3.3.1. ADVANTAGES OF THE PROPOSED SYSTEM....................................................................
3.4. HIGH-LEVEL MODEL OF THE PROPOSED SYSTEM.......................................................................
CHAPTER FOUR........................................................................................................................................................
SYSTEM DESIGN AND IMPLEMENTATION........................................................................................................
4.1 OBJECTIVES OF THE DESIGN..............................................................................................................
4.2 CONTROL CENTRE/MAIN MENU........................................................................................................
4.3 THE SUBMENUS/SUBSYSTEMS..........................................................................................................
4.4 SYSTEM SPECIFICATIONS...................................................................................................................
4.4.1 DATABASE DEVELOPMENT TOOL......................................................................................
4.4.2 DATABASE DESIGN AND STRUCTURE..............................................................................
4.4.3 MATH SPECIFICATION...........................................................................................................
4.4.4 PROGRAM MODULE SPECIFICATION.................................................................................
4.4.5 INPUT/OUTPUT FORMAT.......................................................................................................
4.4.6 ALGORITHM.............................................................................................................................
4.4.7 DATA DICTIONARY................................................................................................................
4.5 SYSTEM FLOWCHART..........................................................................................................................
4.6 SYSTEM IMPLEMENTATION...............................................................................................................
4.6.1 PROPOSED SYSTEM REQUIREMENTS................................................................................
4.6.1.1 HARDWARE REQUIREMENTS...........................................................................................
4.6.1.2 SOFTWARE REQUIREMENTS.............................................................................................
4.6.2 PROGRAM DEVELOPMENT...................................................................................................
4.6.2.1 CHOICE OF PROGRAMMING ENVIRONMENT...............................................................
4.6.3 DOCUMENTATION..................................................................................................................
CHAPTER FIVE..........................................................................................................................................................
SUMMARY, CONCLUSION AND RECOMMENDATION........................................................................
 5.1. SUMMARY..............................................................................................................................................
5.2. CONCLUSION.........................................................................................................................................
5.3. RECOMMENDATION............................................................................................................................
5.3.1 APPLICATION AREAS.............................................................................................................
5.3.2 SUGGESTION FOR FURTHER RESEARCH..........................................................................
REFERENCES.............................................................................................................................................................
 CHAPTER ONE
INTRODUCTION

1.1. BACKGROUND OF STUDY

Yam (Dioscorea spp.) is the common name for a monocotyledonous tuber-producing vine plant
with several species (approximately 600) (Mondo et al., 2020). It is a staple crop in many parts of
the world, especially in West Africa, where it plays a crucial role in food security and the economy
(Obidiegwu and Akpabio, 2017). However, yam cultivation faces significant challenges due to
various diseases that can severely impact yield and quality. The root of the problem lies in the
difficulty of diagnosing these diseases accurately and timely, especially for farmers who may not
have easy access to expert knowledge (Kalu et al., 2023). Diseases like anthracnose, yam mosaic
virus, and soft rot are common, but they often go untreated or are misdiagnosed, leading to
substantial crop losses. This issue underscores the need for a reliable and accessible method of
diagnosing yam diseases to help farmers take prompt and effective action.

Building on the need for accurate diagnosis, the development of expert systems in agriculture has
gained attention over the years as a way to address challenges like these. Expert systems are
computer programs that mimic the decision-making abilities of a human expert, providing
solutions based on a set of rules derived from expert knowledge. In the context of yam disease
diagnosis, an expert system can serve as a valuable tool for farmers, offering them guidance on
identifying diseases and suggesting appropriate treatments. Previous research has shown that
expert systems can be particularly effective in environments where access to human experts is
limited (Sirait, 2023). By applying this technology to yam cultivation, it is possible to bridge the
knowledge gap and provide timely assistance to farmers, ultimately improving crop management
and reducing losses.

Despite the potential of expert systems in agriculture, the specific application to yam disease
diagnosis is still relatively underexplored. While there have been advances in developing expert
systems for diagnosing plant diseases in crops like rice, maize, and wheat, yam has not received
the same level of attention (Kalu et al., 2023). This lack of focus may be due in part to the regional
nature of yam cultivation, primarily in developing countries where resources for advanced research
are often limited. However, the need for such a system is pressing, given the economic importance
of yam and the significant losses farmers face due to disease. Previous research efforts have been
concentrated on general disease identification in plants, but there remains a gap in targeted
solutions for yams, especially those that are accessible to farmers in rural areas.

Recognizing this gap presents an opportunity for innovation and problem-solving. The
development of a yam disease diagnosis expert system could fill this void by providing a
specialized tool tailored to the needs of yam farmers. Such a system would not only assist in
identifying diseases but could also suggest preventive measures and treatments, thereby reducing
the time and resources spent on trial-and-error methods (Akinlolu et al., 2023). Furthermore, by
incorporating local knowledge and adapting to the specific environmental conditions where yams
are grown, the system could offer more accurate and relevant advice than generalized plant disease
diagnosis tools. This focus on yam-specific challenges underscores the importance of developing
solutions that are not only technically sound but also contextually appropriate for the farmers who
will be using them.

However, the development of a yam disease diagnosis expert system is not without its challenges.
One of the primary limitations is the availability of comprehensive and reliable data on yam
diseases. For an expert system to function effectively, it requires a robust knowledge base built
from expert input, field observations, and scientific research (Expert Systems in Artificial
Intelligence - Javatpoint, n.d.). Unfortunately, detailed studies on yam diseases are less prevalent
compared to those on more globally cultivated crops. This scarcity of data makes it difficult to
develop a system that can cover all possible scenarios a farmer might face. Additionally, according
to Okon et al. (2022) there are variations in how yam diseases manifest in different regions, further
complicating the creation of a universally applicable system. These challenges highlight the need
for ongoing research and collaboration with local experts to gather the necessary information for
building a reliable and effective system.

Addressing these challenges requires a multifaceted approach. First, there is a need for increased
research efforts focused specifically on yam diseases, particularly in regions where yam is a major
crop. Collaborative projects involving agricultural researchers, local experts, and farmers could
help in gathering the essential data needed to build the knowledge base for the expert system.
Additionally, leveraging advancements in technology, such as machine learning and image
recognition, could enhance the system’s ability to diagnose diseases even when data is limited.
These technologies can help in identifying patterns and symptoms that may not be well-
documented, thereby improving the accuracy of the system. By combining traditional expert
knowledge with modern technological tools, it is possible to create a more comprehensive and
adaptable system that can meet the diverse needs of yam farmers.

Despite these efforts, gaps still remain in the existing research on yam disease diagnosis. Many
studies have focused on specific diseases or regions, leading to fragmented knowledge that may
not fully address the broader challenges faced by yam farmers. Additionally, while there has been
some progress in developing diagnostic tools, these are often limited in scope or accessibility,
particularly for small-scale farmers who may not have the resources to implement complex
systems. This is where the development of a comprehensive yam disease diagnosis expert system
becomes important. By integrating and expanding upon existing research, this system has the
potential to provide a unified platform that addresses the full range of challenges associated with
yam cultivation. Moreover, by making the system accessible through mobile platforms or other
easily available technologies, it can reach a wider audience, including those in remote or
underserved areas.

The impact of developing such a system extends beyond disease diagnosis. It has the potential to
improve overall agricultural productivity and sustainability by providing farmers with timely and
accurate information. This can help reduce crop losses, enhance yield quality, and strengthen food
security in regions where yam is a crucial food source(Akinlolu et al., 2023). Additionally, data
collected through the system could contribute to larger agricultural databases, offering valuable
insights for future research and policy-making. Thus, this expert system not only supports
individual farmers but also contributes to broader efforts in addressing food security challenges in
yam-producing regions.

1.2. STATEMENT OF THE PROBLEM

Yam is a crucial crop in many parts of West Africa, but it faces serious problems from diseases
that can greatly reduce its yield and quality. Farmers struggle to identify these diseases accurately
because traditional methods are often unreliable and prone to mistakes. Common yam diseases,
such as anthracnose, yam mosaic virus, and soft rot, are often missed or misdiagnosed, leading to
ineffective treatments and significant crop losses. Many farmers, especially those in rural areas,
lack access to expert advice and resources to properly manage these diseases. This situation
highlights the urgent need for an easy-to-use and reliable tool that can help farmers correctly
diagnose and treat yam diseases. Without such a tool, farmers continue to face difficulties, which
affects not only their crops but also food security and the local economy. Developing a reliable
yam disease diagnosis system is essential to help farmers improve their crop management and
ensure better harvests.

1.3. SIGNIFICANCE OF STUDY

The Development of a Yam Disease Diagnosis Expert System is a project of paramount
significance, particularly in the context of enhancing agricultural productivity and ensuring food
security in regions where yam is a staple crop, such as West Africa. Yam is not only a vital source
of carbohydrates but also plays a crucial role in the livelihoods of millions of farmers. However,
the cultivation of yam is often threatened by various diseases, which can lead to substantial yield
losses if not diagnosed and managed promptly. This expert system aims to leverage advanced
technologies, particularly Application Program Interface (APIs), to create a mobile application that
enables farmers to accurately identify yam leaf diseases through image recognition. By providing
real-time, accessible, and user-friendly diagnostic support, the system empowers farmers to take
timely action against diseases, thereby minimizing crop loss and maximizing yield potential.
Furthermore, the project addresses the pressing issue of limited agricultural extension services in
many rural areas, where farmers may lack access to expert advice or timely information. By
democratizing access to disease diagnosis, the expert system not only enhances the farmers' ability
to manage their crops effectively but also contributes to sustainable agricultural practices and
improved economic resilience. Ultimately, the successful implementation of this project could lead
to increased yam production, better food security, and enhanced livelihoods for farming
communities, making it a vital initiative in the ongoing efforts to modernize agriculture and
combat food scarcity in vulnerable regions.

1.4. SCOPE OF STUDY

The scope of the Development of a Yam Disease Diagnosis Expert System encompasses a
comprehensive approach to creating an innovative diagnostic tool specifically designed for yam
farmers, particularly in regions where yam cultivation is prevalent and economically significant.
This project will involve several key components, starting with the collection of extensive datasets
comprising images of yam leaves affected by various diseases, along with healthy specimens for
comparative analysis. The development phase will include the application of Application Program
Interface API, to train the system for accurate disease identification based on visual symptoms. The
expert system will be designed as a mobile application, ensuring accessibility for farmers in rural
areas who may have limited access to technology and agricultural extension services. The project
will focus on user interface design to ensure that the application is intuitive and easy to use, even
for individuals with minimal technological expertise. The scope will also include the integration of
educational resources, such as guidelines on disease management and prevention strategies, to
enhance the farmers’ understanding of yam diseases and promote best practices in crop
management.
Ultimately, the scope of this project is not only to develop a diagnostic tool but also to foster a
sustainable agricultural ecosystem that empowers farmers with the knowledge and resources
needed to combat yam diseases, thereby improving crop yields, enhancing food security, and
supporting the economic stability of farming communities.

1.5. AIM AND OBJECTIVES OF STUDY

The aim of this study is the development of a yam disease diagnosis expert system
Objectives Include:
1. To provide users with precise identification of yam diseases through analysis of photos and
text descriptions, enabling farmers and agricultural professionals to take timely and
appropriate action.
2. To educate farmers and users about various yam diseases and their symptoms, thereby
improving their ability to identify and manage these issues independently.
3. To contribute to the overall health and productivity of yam crops by providing early
detection and preventive measures for common and uncommon yam diseases.
4. To create an easy-to-use and accessible expert system that requires minimal technical
knowledge, ensuring that farmers in remote or under-resourced areas can benefit from the
tool.
5. To design an intuitive and user-friendly application interface that allows users to easily
upload images and descriptions, and quickly receive diagnosis and recommendations.
1.6. DEFINITION OF TERMS
1. API : An API, or Application Programming Interface, is a set of rules and protocols that
allow different software applications to communicate with each other. It defines the
methods and data structures that developers can use to interact with a software component,
web service, or other software tools.
2. Application: An application, often referred to as an "app," is a software program or a
collection of programs designed to perform specific tasks or functions for a user.
Applications can run on various devices, including computers, smartphones, tablets, and
other digital systems. They are the tools that allow users to interact with a computer system
to perform activities ranging from basic tasks, like word processing, to more complex
functions, like video editing or managing databases.
3. Artificial intelligence: Artificial Intelligence (AI) refers to the simulation of human
intelligence in machines that are programmed to think, learn, and perform tasks typically
requiring human intelligence. AI systems are designed to mimic cognitive functions such as
learning, reasoning, problem-solving, perception, language understanding, and decision-
making
4. Comparative Analysis: Comparative analysis is a method of comparing two or more
items, concepts, datasets, or systems to identify similarities, differences, trends, and
relationships. It is commonly used in various fields, including business, research,
economics, and social sciences, to gain insights, evaluate options, and support decision-
making.
5. Datasets: A dataset is a structured collection of data, typically organized in a way that
makes it easy to access, manage, and analyze. It consists of individual data points (or
records) that are organized into a format such as tables, spreadsheets, or databases. Each
data point or record in a dataset usually includes various attributes or variables, which are
the specific pieces of information being tracked.
6. Development phase: The development phase is a critical stage in the software
development lifecycle (SDLC) or any project management process, during which the actual
creation or coding of the software product, system, or project takes place. This phase
involves converting design specifications into a functional product by writing code,
integrating systems, and building features.
7. Diagnosis : ¹Diagnosis refers to the process of identifying a disease, condition, or problem
based on the symptoms, signs, and other relevant information. It is a critical step in medical
 practice, but the term can also be applied in other fields such as engineering, computer
science, and psychology.
8. Expert knowledge: Expert knowledge refers to the deep, specialized understanding and
insight that an individual possesses within a particular field or domain. This type of
knowledge is typically acquired through extensive education, experience, and practice over
a significant period of time. Experts are recognized for their ability to apply this knowledge
effectively to solve complex problems, make informed decisions, and provide authoritative
advice.
9. Knowledge: Knowledge refers to the understanding, awareness, or familiarity gained
through experience, education, or learning. It encompasses facts, information, and skills
acquired over time and can be applied to various situations and contexts.
10. User interface: A user interface (UI) is the point of interaction between a user and a digital
system, such as a computer, software application, or website. It encompasses all the
elements through which users can interact with the system, allowing them to perform tasks,
access information, and control the system's functions.

1.7 LIMITATIONS OF STUDY

When conducting a project to diagnose yam disease, several limitations may impact the study's
outcomes, reliability, and generalizability. Identifying and addressing these limitations is
crucial to ensure that the findings are interpreted correctly and that future research can build on
the results effectively.
1. Limited Sample Size: A small number of yam plants or samples could reduce the study’s
statistical power and the reliability of the findings.
2. Lack of Diversity: If the study only includes samples from a specific region or variety of
yam, the results may not be generalizable to other regions or varieties.
3. Geographical Differences: Environmental factors such as soil type, climate, and farming
practices vary widely and can influence disease prevalence and symptoms, making it
challenging to generalize the findings.
4. Seasonal Variability: Disease symptoms and pathogen behavior can change with seasons,
affecting the consistency of the results.
5. Limited Resources: Constraints in funding, time, or access to advanced diagnostic
equipment could limit the depth and scope of the study.
6. Expertise Availability: The availability of experts to accurately diagnose and interpret
disease symptoms could affect the study’s outcomes.
7. Mutation and Evolution: Pathogens may evolve or mutate, leading to variations in disease
presentation that the study might not fully capture.
 CHAPTER TWO

LITERATURE REVIEW

2.1 THEORETICAL REVIEW

In this Chapter, an in-depth analysis and exploration of existing knowledge form the foundation for
the development of a yam disease diagnosis expert system.

2.1.1 Systems In Agriculture

Expert systems are a branch of artificial intelligence that use a knowledge base and inference
engine to solve complex problems in specific domains. In agriculture, these systems are used to aid
farmers in decision-making by diagnosing diseases, recommending treatments, and predicting crop
yields. For yam disease diagnosis, an expert system can integrate expert knowledge about various
yam diseases, symptoms, and environmental factors, and use this to diagnose diseases with
accuracy. This reduces the need for extensive laboratory testing and provides timely advice to
farmers, ultimately enhancing crop yield and food security.

2.1.2 Learning Approaches In Disease Diagnosis

Machine learning algorithms have become pivotal in the development of expert systems,
particularly in disease diagnosis. These algorithms can analyze large datasets, identify patterns, and
learn from them to improve diagnosis accuracy over time. For yam disease diagnosis, machine
learning can help in the identification of diseases based on image recognition, symptom analysis,
and environmental data. Incorporating machine learning into the expert system enhances its ability
to generalize and improve its performance over time.

2.1.3 Challenges In Developing Expert Systems For Agriculture

The development of expert systems for agriculture, including yam disease diagnosis, faces several
challenges. These include the complexity of modeling expert knowledge, the variability in
symptoms due to environmental conditions, and the limited availability of high-quality data.
Moreover, expert systems must be user-friendly and accessible to farmers, who may have varying
levels of technological expertise. Overcoming these challenges requires collaboration between
computer scientists, agricultural experts, and farmers.

2.1.4 Role Of Expert Systems In Sustainable Agriculture

Expert systems play a crucial role in promoting sustainable agriculture by optimizing resource use
and minimizing the impact of diseases. In yam cultivation, timely and accurate disease diagnosis
can prevent the overuse of pesticides and reduce crop losses, thereby contributing to sustainability.
Expert systems can be integrated with other technologies, such as Geographic Information Systems
(GIS) and IoT devices, to provide holistic solutions for disease management.

2.1.5 Future Trends In Agricultural Expert Systems

The future of expert systems in agriculture, including yam disease diagnosis, is likely to be shaped
by advancements in AI, big data, and cloud computing. These technologies will enable the
development of more sophisticated systems that can process vast amounts of data in real-time,
provide more accurate diagnoses, and deliver personalized recommendations to farmers.
Additionally, the integration of expert systems with mobile platforms will make these tools more
accessible to small-scale farmers, thus democratizing access to agricultural technology.

2.2 REVIEW OF RELATED WORKS

A 2020 study by Rafi Surya and Elliana Guatama delves into cassava leaf disease detection in
Indonesia using Convolutional Neural Networks (CNN) under Deep Learning. According to data
from the Indonesian Central Statistics Agency in 2015, cassava production in Indonesia was
21,201,415 tonnes per year. Apart from being a food ingredient, it is also used as an industrial raw
material and animal feed. The use of this technique came to play after in 2016, the production of
cassava reduced due to disease affecting the cassava crops. This network is built with the
assumption that the input used is an image. This network has a special layer called the convolution
layer, where in this layer an input image will be processed based on a predefined filter. From each
layer, it will produce a pattern from several parts which will be easier to classify, thus making the
propagation function more efficient to implement. This technique can make the image learning
function more efficient to implement. Therefore, the study takes advantage of the advantages of
CNN, namely being able to classify an object intended for image data, so that the CNN model will
be used as an introduction to the four types of healthy cassava leaf and cassava leaf diseases that
can be found in Indonesia. The study applies the CNN method with the MobileNets architecture
and the TensorFlow library which includes the package in the Google Collabs Python software
version 3.7 to recognize images of cassava leaves that focuses on finding the best classification
accuracy in detecting diseases in cassava leaves. By using the TensorFlow library, the results of
model trials and evaluation cassava leaves show an accuracy of 0.8538 for training and 0.7496 for
data validation. So, it can be concluded that the implementation of Deep Learning with the
Convolutional Neural Network (CNN) method can detect cassava disease images.

In a research done by Selvaraj et al, an AI-powered banana diseases and pest detection was looked
into. Worldwide, banana production is affected by numerous diseases and pests. Novel and rapid
methods for the timely detection of pests and diseases will allow to surveil and develop control
measures with greater efficiency. The aim of this research being to carry out disease detection
using DCNN. Large datasets of expert-prescreened banana disease and pest symptom/damage
images were collected from various hotspots in Africa and Southern India. To build a detection
model, three different CNN architectures were retrained using a transfer learning approach. A total
of six different models were developed from 18 different classes (disease by plant parts) using
images collected from different parts of the banana plant. The study also showed ResNet50 and
InceptionV2 based models performed better compared to MobileNetV1. These architectures
represented the state-of-the-art results for banana diseases and pests detection with an accuracy of
more than 90% in most of the models tested. With a future view to run these detection capabilities
on a mobile device, the performance of SSD (Single Shot Detector) MobileNetV1 was evaluated.
Performance and validation metrics were also computed to measure the accuracy of different
models in automated disease detection methods.

ESforRPD2: Expert System for Rice Plant Disease Diagnosis was a study done by Agus et al.
Rice plant diseases in Indonesia are increasingly rampant because of farmers lacking knowledge of
the early symptoms of rice plant diseases and the lack of experts. Data of rice plant pests and
diseases in 2016 were taken from Samarinda, East Kalimantan, Indonesia using an in-depth survey,
and rice experts from the Department of Food Crops and Horticulture of East Kalimantan Province
were recruited for the project. ESforRPD2 was developed based on the pest and disease
experiences of the rice experts and uses a Waterfall Paradigm and Unified Modeling Language
which consists of use case diagrams, activity diagrams and class diagrams. The ES architecture
consists of three parts: the user interface, the inference engine and the knowledge base as proposed
by Lucas and van der Gaag. Two types of user case diagrams were created namely: “Use case for
user” consisting of four cases (Article, Consulting, Choose System and Consulting Result) and
“Use case for Expert” consisting of three cases (Symptoms, Diseases and Relation). The
application uses four datasets (disease- and symptoms-data, knowledge base, and symptoms-
disease-weight decision table). This System can detect 48 symptoms and 8 types of diseases of rice
plants from 16 data tests with a sensitivity of 87.5%. This system can also provide
recommendations for the treatment of unidentified diseases. ESforRPD2 is the first version of ES
(in a bilingual interface that combines both English and Indonesia language) to make a user-
friendly for Indonesian users. Users sign use a consultation page to choose the symptom of the rice
plant. The ES performs the calculation process to obtain the trust level using the Dempster-Shafer
method. The user page is the main webpage for users without logging in. The consultation page
starts the user consultation about the disease of rice plants. The ES will output as a display showing
the symptoms, diagnosis of disease and the confidence level. The ESfor RPD2 is developed using
CPU with specifications of Intel Corei3, 4GB RAM and 300GB HDD. The same specification of
CPU is needed to operate the application. The ESforRPD2 was tested applying symptom-data
inputs by clicking the symptoms selected. In a single test using the case of four symptom-data
inputs selected, namely (i) Spots on leaf midrib, (ii) Little spots are dark brown or slightly purple
rounded shape, (iii) Spots on oval shaped leaves and evenly distributed on the leaf surface, (iv) The
size of spots is 2-10mm long and 1mm wide, a display of diagnosis page will appear, following
clicking of the “submit diagnose” button. The diagnosis page shows the confidence level, the ES
giving the sensitivity of disease type detection of 91%.

A work carried out by Quintero-Dominguez et al, presents an Expert System for the presumptive
diagnosis of fungal diseases in crops, called Fungi. Countries all over the world have experienced
the negative impact that phytopathogenic fungi have on food security. A fast diagnose of the kind
of fungus that affect the crops is essential to stop disease proliferation and so minimize the losses.
The expert system was developed using SWI-Prolog for the creation of the knowledge base and
Java for the graphic user interface. This system allows a fast and reliable diagnosis of the kind of
fungus that affect crops, the specific crops that can be diagnosed with Fungi are: rice, tobacco,
bean, banana, garlic, onion, corn, coffee and cocoa. Fungi was developed under free software
license. Besides, the comparison of human experts diagnoses and Fungi diagnoses allowed to
validate the knowledge base of the system. The application of black box tests and specialist
criterion allowed corroborating the correct operation of the proposed expert system.

A research work proposed by Laxmi Rananavare and Sanjay Chitnis looks into Vrikshayurveda
(An ancient Indian science of plant life) and the development of an expert system to diagnose plant
diseases using its principles. Vrikshayurveda includes complete plant-life knowledge
compendicum of plant physiology, horticulture, pathology, and treatment. The manuscript
containing this is not easily accessible to farmers or researchers who want references for specific
topics without having to read the whole book, and so, the knowledge in this manuscript was
converted to form an expert system which can provide the answers to specific queries from the
farmers and agricultural stakeholders. It is a rule based ES using backward chaining. Its database
containing ten diseases and evaluation done for all dataset. The results are compatible with the
expert’s diagnosis. Thus the users can get comprehensive information on Vrikshayurvedic
expertise on all elements of disease and plant protection. The ES is created using Python, it is
implemented using decorators, and two library modules, that is, Expertia and Flask. The expert
system creates a hypothesis about the disorder in a hierarchical format, that is, the hypothesis
categorises different disorders under it. The hypothesis is further classified as internal and external.
The research identifies different disorder categories based on Ayurvedic principles and expert
knowledge. For example, “vata” may be associated with symptoms such as dryness, roughness,
and mobility issues while “pitta” may manifest as symptoms such as inflammation, heat, and
irritability. Hypotheses are then identified for each disorder category. The formulated hypotheses
are integrated into the Expertia program as predicates, allowing the system to reason about the
presence of specific disorders based on observed symptoms. The system prompts the user with a
series of yes/no questions about the presence of symptoms in the plant. These questions correspond
to the symptoms associated with each disorder category. The user’s responses are recorded for
further evaluation. The system evaluates the user’s responses and checks whether the observed
symptoms align with the symptoms associated with the hypothesized disorder category. It uses
logical inference rules defined in Expertia to reach conclusions about the presence or absence of
the hypothesized disorder. Based on the evaluation and inference process, the system updates its
belief statement regarding the presence of a particular disorder in a plant. If the observed
symptoms match the symptoms associated with a particular disorder, the system concluded the
plant likely has that disorder. The developed system is evaluated using sample plant data and
known disorders. The accuracy of the system in correctly classifying plant disorders is assessed.
Validation techniques, such as cross-validation or expert evaluation, can be employed to ensure the
system’s reusability.

A paper published by Fatima Salman and Samy S. Abu-Naser, reviewed an expert system created
for the diagnosis of castor diseases. The castor bean is a large grassy or semi-wooden shrub or
small tree. Any part of the castor plant can suffer from a disease that weakens its ability to grow
and eliminates its production. The ES aids the agricultural engineers to detect and diagnose
diseases of castor such as seeding blight, alternaria blight, cercospora leaf spot, powdery mildew
and wilt. CLIPS Shell and Delphi XE 10.2 languages are used for designing and implementing the
proposed system. The expert system accomplishes diagnosis for five castor diseases, it can be
applied by displaying all symptoms in the list and selects it to analyse the disease. The ES will ask
the user to choose the symptoms that appear on castor plant from the list, then click “analyse”
button to diagnose the castor disease name, survival and speed, favourable conditions and snapshot
of the disease.

A study by Ajoh et al proposed a combined activation function to enhance the learning

performance of Convolutional Neural Networks (CNNs) specifically for yam disease classification.
CNNs are effective in image classification tasks. The proposed combined activation function
integrates the Rectified Linear Unit (ReLU) and the Exponential Linear Unit (ELU) activation
functions. ReLU is known for his simplicity and computational efficiency, while ELU addresses
the issues of the “dying ReLU” problem and the negative saturation. To evaluate the system,
experiments were carried out on a dataset consisting of images of diseased and healthy yam plants.
The combined activation function achieved superior results compared to models using individual
activation models. The results indicate that the CNN model with the combined activation function
learns discriminative features that are highly representative of yam disease patterns enabling
accurate classification.

2.3 KNOWLEDGE GAP

Despite significant advancements in plant disease diagnosis systems for crops like cassava, banana,
and rice, there remains a notable gap in the development of specialized expert systems for yam
disease diagnosis. Current research, such as the work by Ajoh et al., primarily focuses on CNN-
based image classification for yam diseases but does not fully explore the integration of multi-
modal data or the incorporation of expert knowledge into a comprehensive diagnostic system.
Unlike existing systems for other crops, which combine CNNs with rule-based approaches and
user-friendly interfaces, yam disease diagnosis lacks such a unified and accessible solution.

Additionally, there is a need for localized disease databases and models specifically tailored to yam
diseases, as well as mobile-compatible systems that can be used by farmers in remote areas. The
absence of a holistic approach that integrates advanced image analysis, expert knowledge, and
user-friendly, mobile-ready interfaces for yam disease diagnosis highlights a significant
opportunity for further research and development. Creating such an expert system would greatly
enhance the accuracy and accessibility of yam disease management, addressing a critical gap in the
current agricultural technology landscape.
 CHAPTER THREE

METHODOLOGY AND SYSTEM ANALYSIS

3.1 METHODOLOGY ADOPTED

In developing the Yam Disease Diagnosis Expert System, a structured methodology was followed
to ensure that the system meets the requirements of its users and provides accurate and reliable
disease diagnosis. The methodology consists of six key phases which includes Requirements
Gathering, System Analysis, System Design, Implementation, Testing and Validation, and
Deployment and Maintenance.

3.1.1 Requirements Gathering

The first phase in the development of the expert system involved gathering requirements from
various stakeholders, including farmers, agricultural experts, and extension officers. This phase
was critical in understanding the specific needs and challenges faced by users in diagnosing yam
diseases.

Several methods were employed to gather requirements:

- Interviews and Surveys: Conducted with farmers and agricultural extension officers to identify
the most common yam diseases, observable symptoms, and current diagnostic practices.

- Literature Review: Extensive review of scientific literature and agricultural reports was carried
out to compile information on yam diseases, their symptoms, causes, and treatments. This included
reviewing existing databases and digital tools to understand the current landscape of disease
diagnosis.

- Expert Consultations: Engaged with agricultural scientists and plant pathologists to

determine the critical indicators and decision criteria for diagnosing specific yam diseases.

This comprehensive approach to requirements gathering ensured that the system would be built on
a solid foundation of expert knowledge and real-world needs.
3.1.2 System Analysis

The second phase focused on analyzing the existing methods and systems used for

Diagnosing yam diseases. The goal was to identify the strengths and weaknesses of these methods
and use this analysis to inform the design of the proposed system. The existing methods can be
broadly categorized into manual diagnosis and basic digital tools:

 Manual Diagnosis: This method relies heavily on the experience and visual inspection
skills of farmers or agricultural officers. Typically, they observe symptoms on the yam
plants, compare them with known disease characteristics, and attempt to make a

diagnosis. If they are uncertain, they might consult an expert or reference agricultural guides.
However, this process is time-consuming and prone to errors due to limited knowledge and
subjective judgment.

 Basic Digital Tools: Some digital tools are available that provide basic information on yam
diseases. These tools often allow users to input observed symptoms and receive a list of
potential diseases. However, these tools generally lack a comprehensive knowledge base
and advanced diagnostic capabilities, leading to inaccurate or incomplete diagnoses.

3.1.3 System Design

The design phase involved creating a robust and user-friendly expert system tailored to the needs
of the users. The key components of the system were carefully designed to ensure functionality,
accuracy, and ease of use.

 Knowledge Base: A comprehensive database was designed to store detailed information on

various yam diseases, including symptoms, causes, environmental conditions, and
recommended treatments. This knowledge base was developed in collaboration with
agricultural experts and built using structured data models to ensure easy access and
updates.
 Inference Engine: The core of the expert system is its inference engine, which uses rule-
based logic to simulate expert decision-making. The engine was designed to process user
inputs (symptoms) and match them against the knowledge base to suggest potential
 diagnoses. Advanced algorithms were integrated to handle uncertain or incomplete data,
improving the system’s diagnostic accuracy.
 User Interface: The user interface was designed to be simple and intuitive, allowing users
to easily input symptoms and receive diagnostic results. The interface was developed for
both web and mobile platforms to ensure broad accessibility. Features such as drop-down
menus, search functions, and symptom checklists were included to enhance usability.

3.1.4 Testing and Validation

Testing and validation were essential to ensure the accuracy and reliability of the expert system.
This phase involved several types of testing:

Unit Testing: Individual components, such as the knowledge base and inference engine, were
tested to ensure they functioned correctly. This testing phase involved verifying that each
component met its design specifications and produced accurate outputs.

Integration Testing: After individual testing, the components were integrated, and the system was
tested as a whole. This testing ensured that all parts worked together seamlessly and that data
flowed correctly between components.

3.1.5 Deployment and Maintenance

The final phase involved deploying the expert system and providing ongoing support and updates
to ensure its continued effectiveness.

Deployment: The system was deployed as an app based platform to ensure scalability and
accessibility. Users could access the system through the dedicated mobile app, making it available
to farmers in remote areas.

3.2 ANALYSIS OF THE EXISTING SYSTEM

The existing systems for diagnosing yam diseases are primarily manual or rely on basic
digital tools that are not optimized for accurate and quick diagnosis. A thorough analysis of these
existing systems revealed several limitations that the proposed expert system aims to address.

3.2.1 Data Flow of the Existing System

fig(i) Here is a data flow diagram for the yam disease diagnosis process

The diagram illustrates both the manual diagnosis process and the digital tool workflow, showing
how information flows between different steps, entities, and data stores.

1. Diagnosis Process: Flows from observation through interpretation, consultation, and

finally treatment, involving farmer’s knowledge, agricultural manuals, and local
experts.
2. Digital Tool Workflow: Starts with user input, goes through a search in the digital
database, outputs potential diseases, and concludes with the user’s action based on the
output
Procedures

1. Manual Diagnosis Process:

 Observation: Farmers observe visible symptoms on yam plants, such as discoloration,
wilting, or spotting.
 Interpretation: Based on their knowledge and experience, they interpret these symptoms
to hypothesize potential diseases.
 Consultation: In cases of uncertainty, farmers consult with local experts or reference
agricultural manuals to confirm their diagnosis.
 Treatment: Based on the diagnosis, they apply treatment methods, which may or may
not be effective due to potential misdiagnosis.

2. Tool Workflow:
 Input: Users enter observed symptoms into a digital tool or application.
 Search: The tool searches its limited database for matching symptoms and suggests
possible diseases.
 Output: A list of potential diseases is provided, often without clear guidance on the
accuracy or certainty of the diagnosis.
 Action: Users decide on a course of action, often with minimal guidance on effective
treatments or next steps.
fig(ii) Subsystem interaction diagram

Subsystem Interaction Diagram

This diagram details the interaction between different subsystems within the Yam Disease
Diagnosis System

Traditional Diagnosis Subsystem represents the manual, experience-based approach involving

farmers and local experts.

Digital Diagnosis Subsystem represents a technology-driven method utilizing a digital tool and a
database for disease identification.

Digital Database is a repository within the digital subsystem that stores information about
diseases and symptoms.

3.2.2 Weaknesses Of The Existing System

Knowledge Gaps: Many farmers lack the specialized knowledge needed to accurately diagnose
diseases, leading to frequent misdiagnoses and ineffective treatments.
Limited Accessibility: In remote areas, access to agricultural experts or reliable digital tools is
limited, which hampers timely and accurate diagnosis.

Inefficient Processes: Manual diagnosis is slow and labor-intensive, often requiring multiple
consultations and revisits, which delays treatment and increases crop loss risks.

Inconsistent Results: Reliance on subjective judgment and limited knowledge results in

inconsistent and unreliable diagnostic outcomes.

Data Limitations: Existing digital tools often have outdated or incomplete databases, limiting
their ability to provide accurate diagnoses or treatment recommendations.

3.3 ANALYSIS OF THE PROPOSED SYSTEM

The proposed Yam Disease Diagnosis expert system aims to revolutionize the way farmers
identify and manage yam diseases by leveraging advanced technologies such as artificial
intelligence (AI) and machine learning (ML). Unlike traditional diagnostic methods that rely
heavily on the subjective interpretation of symptoms by farmers and local experts, the proposed
system will use a structured, data-driven approach to accurately identify

diseases. By analyzing a wide range of symptoms and patterns in the data, the system can
provide more precise and reliable diagnoses, reducing the risk of misdiagnosis that often leads to
ineffective treatment and crop loss.

One of the core components of the proposed system is its comprehensive database, which
contains extensive information on yam diseases, symptoms, environmental conditions, and
effective treatment methods. This database is continuously updated through contributions from
agricultural experts, ongoing research, and real-time data collection from the field. The expert
system utilizes this vast repository of knowledge to compare input symptoms against known
disease profiles, enhancing its diagnostic accuracy.
Furthermore, the integration of environmental data such as soil quality, moisture levels, and
weather patterns enables the system to consider external factors that may influence disease
development, providing a more holistic diagnostic process.

To enhance user experience and accessibility, the proposed system will be developed as a user-
friendly mobile application. This application will feature an intuitive interface where farmers can
easily input symptoms through text or images. The system will employ image recognition
technologies to analyze visual symptoms on yam plants, such as leaf

discoloration, wilting, or spotting. In addition, the application will support multiple languages
and include voice input options to accommodate farmers with different literacy levels and
preferences, ensuring that the technology is accessible to a broad audience.

Another significant feature of the proposed system is its ability to provide actionable
recommendations. After diagnosing a disease, the system will offer specific treatment plans
tailored to the severity of the infection and the farmer’s available resources. It will suggest
appropriate fungicides, insecticides, or cultural practices such as crop rotation or soil treatment.
The system will also include educational resources to help farmers

understand disease prevention and management strategies better, fostering a proactive approach
to crop health and sustainability.

3.3.1 Advantages Of The Proposed System

The proposed expert system offers several advantages that significantly enhance yam disease
diagnosis and management:

Enhanced Diagnostic Accuracy: By using AI and ML algorithms, the system can analyze vast
amounts of data and identify diseases with a high degree of accuracy. This reduces the risk of
misdiagnosis that is common with traditional methods, which rely heavily on farmers’ subjective
assessments and experience.

Rapid Response and Timely Interventions: The system’s ability to provide quick and accurate
diagnoses allows farmers to take immediate action to address disease outbreaks. Early detection
and timely intervention are crucial in preventing the spread of diseases and minimizing crop
losses.

Cost-Effective Solution: The system reduces the need for costly consultations with agricultural
experts by providing automated diagnoses and treatment

recommendations. This makes advanced diagnostic tools accessible to smallholder farmers who
may not have the resources to seek expert advice regularly.

Empowerment and Knowledge Sharing: The system not only diagnoses diseases but also
educates farmers about the causes and prevention of yam diseases. By providing detailed
explanations and actionable advice, the system empowers farmers with knowledge and fosters a
proactive approach to crop management, improving overall farm productivity.

3.4 HIGH-LEVEL MODEL OF THE PROPOSED SYSTEM

The high-level model of the Yam Disease Diagnosis expert system consists of several integrated
components designed to work together to deliver accurate and timely

diagnoses. The model emphasizes seamless interaction between these components to ensure the
system is user-friendly, reliable, and effective in diverse farming contexts.

Components of the High-Level Model:

The proposed expert system will consist of the following components:

Knowledge Base: This component stores comprehensive information on yam diseases, including
symptoms, causes, environmental conditions, and treatment options. It is continuously updated
with new data and expert input treatment options.

Inference Engine: The inference engine is the core component of the expert system, responsible
for reasoning and drawing conclusions based on the input data and the knowledge base. It uses
techniques such as rule-based reasoning or case-based reasoning to identify potential diseases
and recommend appropriate treatments.
User Interface: The mobile application’s user interface is designed to be intuitive and
accessible, allowing farmers to input symptoms easily and receive clear diagnostic results. The
interface supports multiple input methods—text, images, and voice commands— ensuring it is
accessible to farmers of all literacy levels and technical abilities.

Knowledge Engineer: The knowledge engineer is responsible for acquiring and organizing
domain-specific knowledge about yam diseases, symptoms, and treatment options. They work
closely with agricultural experts to ensure the accuracy and completeness of the knowledge base.

Knowledge Acquisition: This component is responsible for gathering and structuring

information about yam diseases. It may involve interviews with experts, literature reviews, and
data collection from field studies.

Interpreter: The interpreter translates the system’s output into a format that is understandable to
the user. It presents the diagnosis and treatment recommendations in a clear and concise manner.

Dynamic Database: The dynamic database stores information about user interactions,
diagnoses, and treatment recommendations. This data can be used to improve the system’s
performance over time by identifying patterns and trends.
Fig(iii) high level model diagram
Overall Workflow:

The user interacts with the system through the user interface, providing information about the
yam plant’s symptoms.

The inference engine processes the input data and searches the knowledge base for matching
symptoms.

Based on the comparison, the inference engine identifies potential diseases and applies relevant
rules.

The interpreter presents the diagnosis and treatment recommendations to the user.

The system stores information about the user’s interaction, diagnosis, and recommendations in
the dynamic database.
 CHAPTER FOUR
SYSTEM DESIGN AND IMPLEMENTATION

4.1 OBJECTIVES OF THE DESIGN

The primary objectives of the system design are outlined to guide the development process.
These objectives serve as a foundation for creating a robust and user-friendly system. They aim
to produce the design specifications for the system that will satisfy the requirements defined
during the systems analysis. Here are the objectives considered while designing this system.

1. Accurate Diagnosis: To provide precise identification of yam diseases based on

symptoms or images uploaded by the user.
2. User-Friendly Interface: To design a simple and intuitive user interface for farmers or
agricultural experts to interact with the system.
3. Customization and Extensibility: To allow the system to be easily updated with new
diseases, symptoms, and treatment methods.
 4. Real-time Processing: To ensure fast response times when diagnosing diseases, even in
areas with limited connectivity.
5. Integration with Customized Gemini API: To leverage the power of a customized
Gemini API for data processing, machine learning, and disease recognition.

4.2 CONTROL CENTRE/MAIN MENU

The main menu of a software, also known as the primary or home menu, serves as the central
navigation hub for users to access various features and functions within the software. It is a
critical component of the user interface and plays a significant role in shaping the user
experience.
1. Home Screen: Displays options for diagnosing diseases, viewing disease information,
and accessing the database.
2. Diagnosis Tool: Option to input symptoms or upload images of affected yam plants for
diagnosis.

4.3 THE SUBMENUS/SUBSYSTEMS

Submenus are secondary menus that are accessed from a primary menu, providing a hierarchical
structure for organising and accessing content. They are commonly used in various interfaces,
such as websites, computer software, and mobile applications. The use of submenus can enhance
user experience by organising information into logical categories and reducing clutter on the
primary menu. Each submenu or subsystem within the system contributes to specific
functionalities, providing a detailed and organised structure.

Diagnosis Subsystem
Interfaces with the Gemini API to process symptoms or images and return a diagnosis.

Database Subsystem
Manages the information related to yam diseases, including symptoms, descriptions, and
treatments.
4.4 SYSTEM SPECIFICATIONS
System specifications refer to the formalization of the structural operational semantics, providing
programming and specification languages with an interpretation. They play a crucial role in
system design by allowing for the validation of system accuracy and aiding in the examination of
system-level characteristics such as time specifications, resource utilization, communication, and
non-functional aspects of the system.

4.4.1 Database Development Tool

1. Customization: Integration with Gemini API for seamless data exchange and processing.

4.4.2 Database Design and Structure

Structure:
1. Diseases Collection: Stores information on each disease (e.g., name, symptoms,
treatment).
2. Images Collection: Stores images of diseased yam plants, linked to relevant diseases.

4.4.3 Math Specification

1. Algorithmic Models: Uses machine learning models deployed via the Gemini API for
disease classification and symptom matching.
2. Statistical Analysis: Performs real-time data analysis to provide insights and predict
disease trends.

4.4.4 Program Module Specification

1. Diagnosis Module: Interfaces with Gemini API for processing user input and providing a
diagnosis.
2. User Interface Module: Provides the front-end interface for users to interact with the
system. A chat screen where users can send images and/or text descriptions while
chatting with the AI.

4.4.5 Input/Output Format

It is essential to have a well-designed input/output format. This is because the input requirement
includes all the tools that would be used for enrollment and proper record keeping, while the
output pertains to all the relevant data linked with student records, ensuring comprehensive and
accurate information management.

Input:
1. Symptoms: Text input or selection from predefined options.
2. Images: Upload of images in formats like JPEG, PNG.
3. Dual Input Support: Users can use either or both input methods (image and text) to get a
diagnosis.

4. Output:

Diagnosis: Disease name, description, and recommended treatment.

4.4.6 Algorithm
The algorithms used are necessary for the speed of retrieving and storing data in the diagnosis
system.
1. Symptom Analysis: Uses a rule-based system or machine learning models (via Gemini
API) to analyze symptoms and match them with known diseases.
2. Image Recognition: Employs image recognition models (via Gemini API) to identify
diseases based on visual patterns.
3. Decision-Making: Combines results from symptom analysis and image recognition to
provide a final diagnosis.

4.4.7 Data Dictionary

A data dictionary is a centralized depot of information about data such as meaning, data type,
relationships to other data and origin. It can also be referred to as a metadata repository.
1. Disease: Represents a yam disease entity with attributes like name, symptoms, and
treatment.
2. Symptom: Represents individual symptoms linked to diseases.
3. User: Represents users of the system with attributes like user ID, preferences, and
diagnosis history.
4. Image: Represents uploaded images linked to diseases.

4.5 System Flowchart

A flowchart is a visual representation of the flow of data, processes, and decisions within a
system. It typically uses standardized symbols to depict the various components of a system and
the flow of information between them.

1. User Input: User selects symptoms or uploads an image.

2. Data Processing: The system processes the input via the Gemini API.
 3. Diagnosis: The system returns the diagnosis and recommended treatment.

4.6 System Implementation

System implementation refers to the process of designing, installing, and integrating a new or
revised information system into an organization’s operations. The main goals of system
implementation are to complete the design as necessary, install and begin using the new or
revised information system, and ensure that the system meets the organization’s requirements
and quality standards.

4.6.1 Proposed System Requirements

The requirements for the implementation of this system is as follows;
1. Hardware Requirements
2. Software Requirements

4.6.1.1 Hardware Requirements

The term "hardware" refers to all of the mechanical and electronic components of a computer or
other type of communicating system. Hardware can also be used to denote the entire physical
component of a computer. Hardware requirements refer to the physical computer resources, also
known as hardware, needed for the use and operation of a system or software application. These
requirements can be divided into minimum and recommended system requirements, with the
latter usually providing better performance and resource usage. The minimum requirements for
running the yam disease diagnosis expert system are simply a mobile phone with stable internet
connection.

4.6.1.2 Software Requirements

Software requirements are the description of features and functionalities of a target system,
conveying the expectations of users from the software product. They can be categorized into
functional requirements, non-functional requirements, and domain requirements. Functional
requirements describe what the system should do, while non-functional requirements specify
how the system should do it, including performance, security, and usability. Domain
requirements are specific to the application domain and are not user-specific. Gathering clear,
correct, and well-defined software requirements is fundamental to the entire software
development project.

1. Operating System: Android for the mobile app.

2. API: Customized Gemini API for processing and machine learning.
3. Development Environment: Flutter for mobile app development, Gemini API for backend
services.

4.6.2 Program Development

In the realm of program development, the focus lies on the systematic creation and enhancement
of software. This encompasses the entire software development life cycle, from
conceptualization to implementation and maintenance.

4.6.2.1 Choice of Programming Environment

Visual studio code was our choice for developing the program. It stands out as a popular and
versatile choice with its feature-rich and extensible nature makes it an ideal environment for
coding tasks. Additionally, it provides robust support for various programming languages and
frameworks.

4.6.3 Documentation
Documentation is a pivotal aspect of the student information management system (SIMS)
project, designed to empower users with comprehensive guidance on how to use the system. This
section presents a detailed user manual tailored for all users.

Usage

1. Home Screen: Start by creating a new chat in the home screen or opening one of
your chats in your chat history.
 2. Chat Screen: Send images and/or text descriptions while chatting with the AI.

3. Select an Image: Users can take a photo of the yam or select an existing image from
their gallery.

4. Describe the Yam: Alternatively, or additionally, users can describe what the yam
looks like, including any visible signs of disease.

5. Get a Diagnosis: Yam Doctor processes the input and responds with a probable
diagnosis of any diseases or confirms that the yam is healthy.

CHAPTER FIVE
SUMMARY, CONCLUSION AND RECOMMENDATION

5.1 SUMMARY

Yam (Dioscorea spp.), a staple crop in West Africa, plays a critical role in food security and the
economy. However, yam cultivation faces significant challenges due to various diseases such as
anthracnose, yam mosaic virus, and soft rot, which often go undiagnosed or are misdiagnosed by
farmers. This results in substantial crop losses, impacting both food security and the livelihoods
of farmers. The primary issue lies in the lack of accessible and reliable diagnostic tools,
particularly in rural areas where farmers may not have easy access to expert knowledge.

To address this gap, the development of an expert system for yam disease diagnosis offers a
promising solution. Expert systems are computer programs that mimic human decision-making
processes by using a set of rules derived from expert knowledge. In the context of yam disease
diagnosis, this system would guide farmers in identifying diseases and suggesting appropriate
treatments. Although expert systems have been successfully applied to other crops like rice,
maize, and wheat, yam remains under-explored.

The proposed expert system operates through a chatbot interface that leverages a customized
Gemini API for processing user inputs. The process involves the following steps: the user
provides input to the chatbot, either as text or an image of a diseased yam plant. The chatbot
processes this input and customizes it for a request to the Gemini API. The API, which integrates
machine learning and image recognition, analyzes the data and returns a diagnosis. Finally, the
chatbot sends the API's response back to the user as a reply. This approach ensures that farmers
receive accurate, timely, and accessible diagnostic information.

However, the development of such a system faces challenges, particularly in terms of the
availability of comprehensive and reliable data on yam diseases. A robust knowledge base is
essential for the system to function effectively, yet detailed studies on yam diseases are less
prevalent compared to other crops. Furthermore, variations in how yam diseases manifest in
different regions complicate the creation of a universally applicable system.

To overcome these challenges, there is a need for increased research efforts focused specifically
on yam diseases. Collaborative projects involving agricultural researchers, local experts, and
farmers could help gather the essential data required to build the system's knowledge base.
Additionally, advancements in technology, such as machine learning and image recognition, can
enhance the system's ability to diagnose diseases even when data is limited. By combining
traditional expert knowledge with modern technological tools, the system can meet the diverse
needs of yam farmers.

5.2 CONCLUSIONS

The development of a yam disease diagnosis expert system presents a critical opportunity to
improve yam cultivation by providing accurate and accessible diagnostic tools. This system can
help farmers identify diseases and take timely action, reducing crop losses and enhancing food
security. By leveraging modern technologies such as machine learning and image recognition
through a customized API, the system can offer precise diagnoses even in regions with limited
resources. Continued research, adaptation to local conditions, and ongoing support will be
essential in ensuring the system's effectiveness and long-term impact on yam production.
5.3 RECOMMENDATIONS

I. Increase Research Efforts: Focus on yam-specific diseases, particularly in regions where

yam is a major crop. Collaborative efforts between agricultural researchers, local experts, and
farmers will be crucial in gathering the necessary data to build a comprehensive knowledge base
for the expert system.

II. Leverage Modern Technology: Integrate machine learning, image recognition, and other
technological advancements to enhance the system's diagnostic accuracy. The customized
Gemini API is a key component in processing user inputs and providing accurate disease
diagnoses based on text and image data.

III. Tailor the System to Local Conditions: Ensure that the expert system incorporates local
knowledge and adapts to the specific environmental conditions where yams are grown. This will
make the system more relevant and effective for farmers in different regions.

IV. Enhance Accessibility: Make the expert system available through mobile platforms or other
easily accessible technologies, ensuring that it reaches a wider audience, including farmers in
remote or underserved areas.

V. Provide Continuous Updates and Support: Establish mechanisms for regularly updating
the system with new data and disease information. Additionally, offer ongoing support to users,
including training and assistance in using the system effectively.

By implementing these recommendations, the yam disease diagnosis expert system can become a
valuable tool for improving agricultural productivity, sustainability, and food security in yam-
producing regions.

5.3.1 Application Areas

The yam disease diagnosis expert system can be applied in various agricultural settings,
particularly in regions where yam is a staple crop. Its primary users include small-scale farmers
in rural areas who lack access to expert diagnostic services. Agricultural extension officers can
also use the system to provide accurate diagnoses and treatment recommendations during field
visits. Additionally, research institutions and universities may apply the system as a teaching tool
in agricultural programs, helping students understand plant disease diagnosis and management.
Here are some potential application areas:

1. Early Disease Detection and Prevention

• Rapid diagnosis: The system can quickly identify diseases based on user-provided symptoms,
potentially preventing further spread and damage.
• Preventive measures: Once a disease is diagnosed, the system can recommend appropriate
preventive measures, such as crop rotation, fungicide application, or cultural practices.

2. Agricultural Research and Development

• Disease surveillance: The system can contribute to disease surveillance efforts by tracking
the occurrence and spread of yam diseases in different regions.
New disease identification: It can assist researchers in identifying new or emerging diseases by
comparing symptoms to known disease patterns.

• Farmer Training and Education

Knowledge transfer: The system can be used to educate farmers about different yam diseases,
their symptoms, and effective management strategies.
Decision support: It can provide farmers with personalized advice based on their specific crop
conditions and disease concerns.

4. Pest and Disease Management Programs

Integrated pest management (IPM) on the system can be integrated into IPM programs to
optimize disease control strategies and minimize the use of chemical pesticides.
Policy formulation: It can inform policymakers about the prevalence and impact of yam diseases,
helping to shape effective agricultural policies.

5. Quality Control and Export

Product safety of the system can be used to ensure the quality and safety of yam products,
particularly those intended for export.
Compliance: It can help producers comply with international standards and regulations related
to plant health and disease control.
By providing accurate and timely disease diagnoses, a yam disease diagnosis expert system can
significantly improve the productivity and sustainability of yam cultivation, benefiting both
farmers and consumers. Finally, government agencies and NGOs working on food security and
agricultural development can use the system to support yam farming initiatives, enhancing
productivity and reducing losses.

5.3.2 Suggestions for Further Research

Further research is needed to expand the system’s knowledge base by gathering more data on
yam diseases, especially in different geographical regions. This includes studying less common
diseases and variations in disease manifestation due to environmental factors. Additionally,
research should focus on improving the system’s algorithms to enhance the accuracy of disease
diagnosis, particularly in distinguishing between similar symptoms caused by different diseases.
Investigating the integration of the system with other agricultural technologies, such as sensors
for early disease detection, could also be valuable.

The development of a yam disease diagnosis expert system should focus on several key areas to
enhance the system’s overall effectiveness and applicability. One critical area is the expansion
and diversification of the database of yam diseases. This involves collecting comprehensive data
on a wider range of yam diseases, particularly those that are region-specific or less documented
in existing literature. Incorporating this information will enable the system to provide more
accurate and relevant diagnoses, especially in diverse farming environments. Additionally,
research should investigate how regional climatic and soil conditions affect disease
manifestation, as this will allow for the development of more localized and precise diagnostic
models.

Another area of focus is the enhancement of the system's underlying technologies, particularly in
terms of machine learning algorithms and image recognition techniques. Advanced machine
learning models, such as deep learning, could be employed to improve the system's ability to
differentiate between diseases with similar symptoms. Similarly, image recognition technology
can be refined to more accurately detect visual patterns associated with specific diseases, even in
images of varying quality. This technological improvement will be crucial for ensuring that the
system remains reliable and effective in real-world scenarios where image quality and symptom
descriptions may be inconsistent.

Exploring the integration of sensor-based early detection technologies with the expert system is
another promising area for future research. Sensors placed in the field could monitor
environmental factors and plant health in real-time, providing data that could be fed into the
expert system to predict and diagnose diseases at an earlier stage. This proactive approach could
significantly reduce crop losses by allowing farmers to take preventive measures before diseases
become widespread.

Furthermore, research into user experience and system design should be conducted to ensure that
the expert system is accessible and user-friendly for farmers of all literacy and technological
proficiency levels. Studies should examine how different interface designs impact usability and
gather feedback from farmers to continuously improve the system's functionality. This includes
considerations for mobile accessibility, language options, and culturally relevant features that
make the system more adaptable to various regions and farming communities.

Lastly, it is important to investigate the potential for multilingual and culturally sensitive
adaptations of the expert system. Since yam is grown in diverse regions with different languages
and cultural practices, the system must be tailored to meet these varying needs. Research into
how language and cultural factors influence disease management practices can inform the
development of more personalized and effective diagnostic tools that resonate with local farmers.
This will ensure that the system is not only technically sound but also relevant and practical for
the communities it aims to serve, user experience studies could provide insights into how the
system can be further optimized for accessibility and effectiveness in diverse farming contexts.
REFERENCES

OBIDIEGWU, J. E., AND AKPABIO, M. E. (2017). THE GEOGRAPHY OF YAM CULTIVATION IN

SOUTHERN NIGERIA: EXPLORING ITS SOCIAL MEANINGS AND CULTURAL FUNCTIONS. J. ETHN
FOOD 4, 28–35. DOI: 10.1016/J.JEF.2017.02.004

MONDO, J. M., AGRE, P. A., EDEMODU, A., ADEBOLA, P., ASIEDU, R., AKORODA, M. O., ET AL.
(2020). FLORAL BIOLOGY AND POLLINATION EFFICIENCY IN YAM (DIOSCOREA SPP.). AGRI
10:560. DOI: 10.3390/AGRICULTURE10110560

KALU, C., NNABUE, I., EDEMODU, A., AGRE, P. A., ADEBOLA, P., ASFAW, A., & OBIDIEGWU, J.
E. (2023). FARMERS’ PERSPECTIVE TOWARD A DEMAND LED YAM BREEDING IN NIGERIA.
FRONTIERS IN SUSTAINABLE FOOD SYSTEMS, 7. HTTPS://DOI.ORG/10.3389/FSUFS.2023.1227920
SIRAIT, H. (2023). BASIC ELEMENTS AND CHARACTERISTICS IN BUILDING AN EXPERT SYSTEM.
JOURNAL NEOSANTARA HYBRID LEARNING, 1(3), 227–248.
HTTPS://DOI.ORG/10.55849/JNHL.V1I3.337

AKINLOLU, A. O., ODEJOBI, O. A., AJAYI, F. O., & JIMOH, E. R. (2023). MOBILE-BASED DEEP
LEARNING FOR YAM DISEASE DIAGNOSIS. HTTPS://DOI.ORG/10.1109/SEB-
SDG57117.2023.10124483

EXPERT SYSTEMS IN ARTIFICIAL INTELLIGENCE – JAVATPOINT. (N.D.). WWW.JAVATPOINT.COM.

HTTPS://WWW.JAVATPOINT.COM/EXPERT-SYSTEMS-IN-ARTIFICIAL-INTELLIGENCE

OKON, N. I., MARKSON, A. A. A., OKoN, E. I., ITA, E. E., UYOH, E. A., ENE-OBONG, E. O. E., &
NTUI, V. O. (2022). CHARACTERIZATION OF SOME FUNGAL PATHOGENS CAUSING ANTHRACNOSE
DISEASE ON YAM IN CROSS RIVER STATE, NIGERIA. PLOS ONE, 17(6), E0270601.
HTTPS://DOI.ORG/10.1371/JOURNAL.PONE.0270601

MEENA, H. R., SINGH, K. M., & SWANSON, B. E. (2017). EXPERT SYSTEMS APPLICATIONS IN
AGRICULTURE: A REVIEW. AGRICULTURAL RESEARCH, 6(2), 173-177

OJHA, T., MISRA, S., & RAGHUWANSHI, N. S. (2017). WIRELESS SENSOR NETWORKS FOR
AGRICULTURE: THE STATE-OF-THE-ART IN PRACTICE AND FUTURE CHALLENGES. COMPUTERS
AND ELECTRONICS IN AGRICULTURE, 118, 66-84.

LECUN, Y., BENGIO, Y., & HINTON, G. (2015). DEEP LEARNING. NATURE, 521(7553), 436-444.

ZHANG, W., ZENG, X., & YAO, J. (2018). DEEP LEARNING FOR IMAGE-BASED CROP DISEASE
DIAGNOSIS: A SURVEY. COMPUTERS AND ELECTRONICS IN AGRICULTURE, 151, 234-243.

FOUNTAS, S., ESS, D., SORENSEN, C. G., & BLACKMORE, B. S. (2015). FARMER ACCEPTANCE OF
PRECISION AGRICULTURE TECHNOLOGY. PRECISION AGRICULTURE, 16(6), 767-791.

BALDI, P., & BRUNAK, S. (2001). BIOINFORMATICS: THE MACHINE LEARNING APPROACH. MIT
PRESS.

TILMAN, D., BALZER, C., HILL, J., & BEFORT, B. L. (2011). GLOBAL FOOD DEMAND AND THE
SUSTAINABLE INTENSIFICATION OF AGRICULTURE. PROCEEDINGS OF THE NATIONAL ACADEMY
OF SCIENCES, 108(50), 20260-20264.
PAREEK, A., & TRIVEDI, R. S. (2011). DEVELOPMENT OF EXPERT SYSTEMS

SURYA, R, & GAUTAMA, E. (2020). CASSAVA LEAF DISEASE DETECTION USING

CONVOLUTIONAL NETWORKS. 2020 6TH INTERNATIONAL CONFERENCE ON SCIENCE IN
INFORMATION TECHNOLOGY (ICSI TECH), 97-102, 2020

SELVARAJ, M, VERGARA, A, RUIZ, H, SAFARI, N, ELAYABALAN, S, OCIMATI, W, BLOMME, G.

(2019). AI POWERED BANANA DISEASES AND PEST DETECTION. PLANTS METHODS, 15, 1-11.

AGUS, F, IHSAN, M, KHAIRINA, D, PURNAWAN , K. (2024). ESFORRPD2: EXPERT SYSTEM FOR

RICE DISEASE DETECTION, F1000 RESEARCH, 7:1902

QUINTERO-DOMINGUEZ, L, RODRIGUEZ, SANCHEZ, D, AVILA, B. (2019). EXPERT SYSTEM FOR

THE PRESUMPTIVE DIAGNOSIS OF FUNGAL DISEASES IN CROPS, BERNARDO LEON AVILA’S LAB

LANANAVARE, L, CHITNIS, S. (2024). TECHNOLOGY FROM TRADITIONAL KNOWLEDGE:

VRIKSHAYURVEDA-BASED EXPERT SYSTEM FOR DIAGNOSIS AND MANAGEMENT OF PLANT
DISEASES, JOURNAL OF AYURVEDA AND INTEGRATIVE MEDICINE, 15(1):100853

SALMAN, F, ABU-NASER, S. (2019). EXPERT SYSTEM FOR CASTOR DISEASES & DIAGNOSIS,
KNOWLEDGE AND INFORMATION, SYSTEMS, 3(3):1-10

AJOH, S, OGAH, S, JIKA, T, DANIEL, A, AKPA, J. (2024). A COMBINED ACTIVATION FUNCTION

FOR LEARNING PERFORMANCE IMPROVEMENT OF CONVOLUTIONAL NEURAL NETWORKS FOR
YAM DISEASE CLASSIFICATION.