CROP RECCOMENDATION SYSTEM

USING MACHINE LEARNING

Submitted impartial fulfillment of the requirements for The award of Bachelor of

Science in COMPUTER SCIENCE

By-

MANISHA PAYAL- COMP2100021

(Reg:-010505210019)

MRINAL BARMAN - COMP2100013

(Reg:-010505210012)

PRONOY SAHA- COMP2100004

(Reg:-010505210004)

DEPARTMENT OF COMPUTER ANF INFORMATION SCIENCE

RAIGANJ UNIVERSITY,

RAIGANJ, UTTARDINAJPUR, WEST BENGAL

Pin-733134

1
 DECLARATION
I hereby declare that the project report entitled Crop Recommendation
System Using Machine Learning submitted in partial fulfillment of the
requirements for the degree of Bachelor of Science in Computer Science at
RAIGANJ UNIVERSITY, is my original work, and has not been submitted
earlier either to this institution or to any other institution for the fulfillment
of the requirement of any course of study.

I further declare that the information presented in this report is true to the
best of my knowledge and belief. Wherever contributions of others are
involved, every effort is made to indicate this clearly, with due reference to
the literature and acknowledgment of collaborative research and
discussions.

I also certify that this work has been conducted in accordance with the
guidelines and regulations of RAIGANJ UNIVERSITY

Manisha Payal
COMP2100021
Department of Computer and Information Science.
Raiganj University
Year:- 2021-2024

2
 ACKNOWLEDGEMENT

I would like to express my deepest gratitude to all those who have supported
me throughout the development of the “Crop Recommendation System using
Machine Learning” project.

First and foremost, I would like to thank Dr. SUPRIYAN SEN ,my project
advisor, for their invaluable guidance, insightful feedback, and continuous
encouragement. Their expertise and dedication were instrumental in
shaping the direction of this project.

I am also grateful to my team mates who provided thoughtful discussions

and suggestions, which helped me refine my approach and methodology.
Additionally, I extend my appreciation to
RAIGANJ UNIVERSITY for providing the necessary resources and
environment conducive to research and development.

Finally, I would like to acknowledge the authors and researchers whose

work provided a foundation for this project, as well as the open-source
community for offering tools and datasets that were crucial in building this
system.

Thank you all for your contributions and support.

MANISHA PAYAL.

3
 INTRODUCTION

The Crop Recommendation System project aims to enhance agricultural

efficiency by providing tailored crop suggestions based on specific
environmental and soil conditions. The system utilizes machine learning
algorithms to analyze various inputs, such as soil characteristics, climate
data, and historical crop performance, to generate recommendations for the
most suitable crops.

Key Components:

Data Collection: Gathers data on soil properties (e.g., pH, nutrient levels),
weather patterns, and previous crop yields.

Data Analysis: Employs machine learning models to identify patterns and

correlations between the input variables and crop success.

Recommendation Generation: Provides actionable crop

recommendations based on the analyzed data to optimize yield and resource
use.

User Interface: Features an intuitive interface for users to input their data
and receive recommendations.
By integrating these components, the Crop Recommendation System helps
farmers make informed decisions, leading to better crop selection and
increased agricultural productivity.

4
OBJECTIVES OF THE PROJECT

The Objectives of the Crop Recommendation System using Machine Learning

are:

Optimize Crop Selection: Use machine learning algorithms to recommend

the most suitable crops for a given set of environmental and soil conditions
to maximize yield and efficiency.

Enhance Decision-Making: Provide data-driven insights to help farmers

make informed decisions regarding crop selection based on accurate and
relevant data.

Improve Resource Utilization: Reduce wastage of resources such as water,

fertilizers, and pesticides by recommending crops that are well-suited to the
local conditions.

Increase Agricultural Productivity: Boost overall crop yield and farm

productivity by suggesting crops that are more likely to thrive under specific
conditions.

Adapt to Changing Conditions: Utilize real-time data and machine learning

models to adapt recommendations as environmental and soil conditions
change.

Simplify the Recommendation Process: Create an easy-to-use interface

that allows farmers to input their data and receive actionable crop
recommendations with minimal effort

5
SCOPE

The scope of the Crop Recommendation System using Machine Learning

encompasses several key areas to ensure comprehensive support for
farmers and agricultural stakeholders. These areas include:

1.Data Integration: Collection and integration of diverse datasets such as

soil properties, weather patterns, historical crop yields, and market trends.
Use of real-time data feeds for continuous updating of recommendations.

2.Machine Learning Model Development: Development and training of

machine learning models to analyze and predict the best-suited crops for
specific conditions. Continuous refinement of models using new data and
feedback to improve accuracy.

3.User Interface Design: Creation of an intuitive and user-friendly interface

for farmers to input their data and receive recommendations. Support for
multiple languages and accessibility features to cater to a diverse user base.

4.Geographic Coverage: Expansion to cover various geographic regions

with different climates, soil types, and agricultural practices. Adaptation of
models to regional specifics and local crop varieties.

5.Integration with Agricultural Practices: Recommendations for crop

rotation and diversification to maintain soil health and sustainability.
Suggestions for best practices in planting, irrigation, and fertilization based
on crop recommendations.

6.Scalability and Performance: Ensuring the system can handle large

volumes of data and provide recommendations quickly. Scalability to
accommodate increasing numbers of users and expanding datasets.

6
7.Education and Support: Providing training materials and support to help
farmers understand and utilize the system guidance on interpreting
recommendations and implementing suggested practices.

8.Impact Assessment: Monitoring and evaluating the impact of the system

on crop yields, resource usage, and economic benefits for farmers. Gathering
user feedback to continuously improve the system.

9.Collaboration and Partnerships: Collaborating with agricultural

research institutions, partners to enhance data quality and system
effectiveness. Partnering with local agricultural organizations to facilitate
system adoption and reach.

7
PROJECT DETAILS

By addressing these areas, the Crop Recommendation System aims to

become a comprehensive tool that supports sustainable and efficient
agricultural practices, ultimately benefiting farmers and contributing to food
security.
Data Collection: Gather relevant data from various sources, including soil
properties, weather data, historical crop yields, and market trends.

Data Preprocessing: Clean and preprocess the collected data, ensuring it is

normalized and relevant features are extracted for analysis.

Machine learning model Development: Develop and train machine l

models using appropriate algorithms. Validate and optimize the models to
ensure accuracy.

Model Integration: Integrate the trained machine learning models into the
recommendation system for real-time analysis and decision-making.

User Interface: Design an intuitive interface for users to input their data,
receive recommendations, and visualize results.

Recommendations: Provide tailored crop suggestions, best agricultural

practices, and resource management tips based on the analysis.

8
BLOCK DIAGRAM

dataset

Data pre-processing

Training Data Testing data

GaussianNB Algorithm Model

al

Crop Recommendation

Fig :- 1.1

9
Fig :- 1.2

10
COMPONENTS

The Crop Recommendation System using Machine Learning comprises

several essential components, each playing a critical role in ensuring the
system’s effectiveness and reliability. These components are:

Data Preprocessing: Data Collection:

• Soil Data: Information on soil type, pH level, moisture content, and

nutrient composition.

• Weather Data: Historical and real-time weather conditions, including

temperature, rainfall, humidity, and seasonal patterns.

• Data Cleaning: Removing noise, handling missing values, and

correcting inconsistencies in the data.

• Normalization: Scaling data to a uniform range to ensure accurate

model training.

• Feature Extraction: Identifying and selecting relevant features that

impact crop growth and yield.

Machine Learning Model Development:

• Algorithm Selection: Choosing appropriate machine learning

algorithms (e.g., decision trees, random forests, neural networks) for
the recommendation system.

11
 • Model Training: Training the models using the preprocessed data to
learn patterns and relationships.

• Model Validation: Evaluating model performance using techniques

like cross-validation to ensure accuracy and robustness.

• Hyper parameter Tuning: Optimizing model parameters to improve

prediction accuracy.

Model Integration:

• System Integration: Integrating trained machine learning models

into the recommendation system.

• API Development: Developing APIs for seamless communication

between the user interface and the backend models.

User Interface (UI):

• Data Input Forms: User-friendly forms for farmers to input soil data,
weather conditions, and other relevant information.

• Dashboard: Interactive dashboard displaying crop

recommendations, visualizations, and analysis results.

12
 • Localization: Support for multiple languages and regional
customization to cater to diverse user bases.

Recommendation Engine:

• Real-time Analysis: Providing real-time crop recommendations

based on current data inputs.
• Best Practices: Suggesting agricultural best practices, such as
irrigation methods, fertilization schedules, and pest control measures.

• Resource Management: Advising on optimal use of resources like

water, fertilizers, and pesticides.

Feedback Loop:

• User Feedback Collection: Gathering feedback from users on the

effectiveness of recommendations and overall system performance.

• Model Updates: Continuously updating and retraining models based

on new data and user feedback to improve accuracy and relevance.

Scalability and Performance:

• Cloud Infrastructure: Utilizing cloud services for scalable storage,

processing power, and deployment.

13
 • Performance Optimization: Ensuring the system can handle large
datasets and provide recommendations quickly.

Security and Privacy:

• Data Security: Implementing measures to protect user data and

ensure confidentiality.

• Compliance: Adhering to relevant data protection regulations and

standards.

Education and Support:

• Training Materials: Providing educational resources and guides to

help users understand and utilize the system effectively.

• Customer Support: Offering support channels for users to get

assistance and resolve issue.

By integrating these components, the Crop Recommendation System aims to

provide accurate, driven crop suggestions, helping farmers make and
enhancing agricultural productivity

14
RESULTS

Fig :- 1.3

Fig :- 1.4

15
Fig :- 1.5

16
ADVANTAGES

Advantages of Crop Prediction Using Machine Learning

Implementing machine learning for crop prediction offers numerous
advantages, significantly improving agricultural practices, decision-making,
and resource management. Here are some key benefits:

1. Improved Accuracy and Reliability

- Data-Driven Predictions: Machine learning models leverage

large datasets, including historical crop yields, weather
patterns, soil conditions, and more, to make highly accurate
predictions.

- Adaptability: Models can be continuously updated and

retrained with new data, improving their accuracy over time
and adapting to changing conditions.

2. Enhanced Decision-Making

- Timely Insights: Farmers and agribusinesses can receive

predictions well in advance, allowing for better planning of
planting, irrigation, fertilization, and harvesting schedules.

- Risk Management: By predicting potential risks like droughts,

pests, or diseases, machine learning models help in taking
preventive measures, reducing crop losses.

17
 3. Optimization of Resources
Efficient Resource Allocation: Machine learning helps optimize the
use of resources such as water, fertilizers, and pesticides by predicting
where and when they are most needed.

- Cost Reduction: By optimizing inputs and reducing waste,

farmers can lower their operational costs and improve
profitability.

-Yield Optimization:
By analyzing various factors influencing crop growth, machine learning
models can suggest optimal planting strategies, leading to higher yields.

- Precision Agriculture: Machine learning enables precision

farming practices, where interventions are tailored to specific
areas of the field, leading to more efficient and productive
farming.

4. Early Detection of Problems

- Pest and Disease Prediction: Machine learning models can

predict outbreaks of pests or diseases based on environmental
conditions, allowing for early intervention.

- Climate Adaptation: The models can help farmers adapt to

climate change by predicting how different climate scenarios
will affect crop growth and suggesting adaptive strategies.

18
5. Scalability

- Wide Applicability: Machine learning models can be scaled to

different regions, crop types, and farming practices, making
them useful for a wide range of agricultural settings.

- Integration with IoT: When integrated with Internet of Things

(IoT) devices, machine learning can process real-time data from
sensors in the field, further enhancing its predictive capabilities.

6. Sustainable Agriculture

- Environmental Impact: By optimizing the use of inputs like

water and fertilizers, machine learning helps reduce the
environmental footprint of farming.

- Resource Conservation: Efficient use of resources leads to

conservation of water and soil health, contributing to long-term
agricultural sustainability.

7. Economic Benefits

- Increased Profits: With higher yields, optimized resources, and

reduced losses, farmers can achieve better financial outcomes.

- Market Forecasting: Machine learning can also be used to

predict market demand and prices, helping farmers make

19
 informed decisions about crop selection and marketing
strategies.
8. Customization and Personalization- Tailored Recommendations:
Machine learning models can provide personalized recommendations
based on specific farm conditions, crop types, and farmer preferences.

- Localized Insights: Models can be trained with local data to

provide insights that are specific to a region’s climate, soil, and
other environmental factors.

9.Continuous Improvement

- Learning from Data: Machine learning models improve over

time as they process more data, becoming increasingly accurate
and reliable.

- Feedback Loops: Farmers can provide feedback on the

predictions, which can be used to refine and improve the
model’s performance.

Overall, the use of machine learning in crop prediction revolutionizes

agriculture by making it more data-driven, efficient, and sustainable. This
leads to better outcomes for farmers, the environment, and the agricultural
industry as a whole.

20
LIMITATIONS

Limitations of Crop Prediction Using Machine Learning

While machine learning offers significant advantages in crop prediction, it

also has limitations that can affect its performance, applicability, and
reliability. Here are some key limitations:

1. Data Quality and Availability

- Dependence on Data: Machine learning models rely heavily on

the quality and quantity of data. Poor quality data, such as
incomplete, outdated, or biased datasets, can lead to inaccurate
predictions.

- Data Scarcity: In some regions, particularly in developing

countries, there may be a lack of historical data on crop yields,
soil conditions, and weather patterns, which can limit the
effectiveness of the model.

2. Complexity of Agricultural Systems

- Dynamic Environments: Agriculture is influenced by a wide

range of factors, including weather, soil, pests, diseases, and
human interventions, making it challenging for models to
account for all variables accurately.

- Non-Linear Relationships: The interactions between different

environmental factors are often non-linear and complex, which
21
 can be difficult for models to capture accurately without
sophisticated algorithms and large amounts of data.

3. Model Generalization-

- Overfitting: A model trained on a specific dataset may perform

well on that data but fail to generalize to new, unseen data,
especially if the new data comes from different regions or
conditions.

- Local Adaptation: Models trained in one geographical area may

not be easily applicable to another area due to differences in
climate, soil, and farming practices, requiring region-specific
adaptations.

4. Resource Intensity

- Computational Requirements: Training and deploying

machine learning models, especially complex ones like deep
learning models, can require significant computational
resources, which may be costly or inaccessible for smaller farms
or organizations.

- Data Processing: The preprocessing of large agricultural

datasets, including cleaning, normalization, and feature
selection, can be time-consuming and requires expertise.

22
5. Interpretation and Explainability

- Black Box Models: Many machine learning models,

particularly, deep learning models, function as “black boxes,”
providing predictions without clear explanations. This lack of
transparency can make it difficult for farmers to trust or
understand the decisions made by the model.

- Complexity: The complexity of machine learning models can

make it challenging for non-experts to understand or interpret
the results, which may limit their adoption in the agricultural
sector.

6. Sensitivity to Data Changes.

- Model Drift: Over time, changes in environmental conditions,

farming practices, or data collection methods can cause the
model’s performance to degrade, a phenomenon known as
model drift. Regular updates and retraining are necessary but
can be resource-intensive.

- Sensitivity to Outliers: Machine learning models can be

sensitive to outliers or anomalies in the data, which can skew
predictions and reduce accuracy.

7. Ethical and Privacy Concerns

- Data Privacy: Collecting and using data from farms, especially

if it includes sensitive information like land ownership or
financial details, raises concerns about data privacy and
security.

23
 - Bias in Data: If the training data contains biases (e.g., focusing
on large-scale farms rather than smallholders), the model may
perpetuate these biases, leading to unfair or inequitable
predictions.

8. Economic Barriers

- Cost of Implementation: The development, deployment, and

maintenance of machine learning models can be expensive,
making them less accessible to small-scale farmers or those in
developing regions.

- Access to Technology: Limited access to the necessary

technology, such as high-speed internet, modern computing
resources, or advanced sensors, can hinder the adoption of
machine learning in agriculture.

9. Dependence on External Factors

Reliance on Weather Data: Many crop prediction models heavily rely
on accurate weather forecasts. If the weather data is inaccurate or
unavailable, the model’s predictions can be significantly affected.

- Infrastructure Dependency: The effectiveness of machine

learning models can be dependent on the availability of
infrastructure like irrigation systems, which may not be present
in all farming regions.

24
10. Maintenance and Continuous Learning-
- Need for Continuous Updates: Agricultural models require
regular updates and retraining to incorporate new data and adapt to
changing conditions, which can be a continuous and resource-intensive
process.

- Lack of Expertise: Implementing and maintaining machine

learning models require technical expertise, which may be
lacking in some agricultural regions or organizations.

These limitations highlight the challenges of applying machine learning in

crop prediction, particularly in diverse and dynamic agricultural
environments. Addressing these limitations requires a combination of better
data management, more adaptable models, accessible technology, and
ongoing collaboration between data scientists, agronomists, and farmers.

25
REQUIREMENTS

.The requirements for developing a Crop Recommendation System using

Machine Learning can be categorized into functional and non-functional
requirements:

Functional Requirements

1.Data Collection: Ability to collect soil data (e.g., pH level, moisture

content, nutrient levels).Capability to gather weather data (e.g.,
temperature, rainfall, humidity, seasonal patterns).Mechanism to obtain
historical crop yield data. Access to market data (e.g., current demand,
prices).

2.Data Preprocessing: Tools for data cleaning to handle missing values and
remove noise. Methods for normalizing and scaling data. Algorithms for
feature extraction and selection.

3.Machine Learning Model Development: Selection of appropriate

machine learning algorithms. Training models on the collected and
preprocessed data. Validation techniques to ensure model accuracy and
reliability. Hyperparameter tuning for model optimization.

4.Model Integration: Integration of trained machine learning models with

the recommendation system. Development of APIs for interaction between
the user interface and backend.

5.User Interface: friendly data input forms for farmers. Interactive

dashboard for displaying recommendations and visualizations. Localization
support for multiple languages and regional preferences.

26
6.Recommendation Engine: Real-time analysis and crop recommendation,
agement advice (e.g., optimal use of water, fertilizers).

Non-Functional Requirements

1.Scalability: System should handle large datasets and an increasing

number of users efficiently. Cloud infrastructure to support scalable storage
and processing.

2.Performance: Quick response times for data input, processing, and

recommendation generation. Optimization techniques to enhance system
performance.

3.Security and Privacy: Secure storage and transmission of user data.

Compliance with data protection regulations.

4 Usability: Intuitive and easy-to-navigate user interface. Comprehensive

help and support resources for users.

5 Reliability: High availability and uptime of the system. Robust error

handling and recovery mechanisms.

6 Maintainability: Modular and well-documented codebase for easy

maintenance and updates. Regular system updates and improvements based
on user feedback and new data.

7.Interoperability:Compatibility with various data sources and formats.

Ability to integrate with other agricultural tools and systems.

27
8.Education and Support: Training materials to help users understand the
system. Customer support channels for assistance and issue resolution.

By meeting these functional and non-functional requirements, the Crop

Recommendation System can effectively support farmers in making
informed crop choices, leading to improved agricultural productivity and
sustainability.

28
SUMMARY

The crop prediction project leverages machine learning techniques to

forecast agricultural outcomes, such as crop yields, based on various input
factors including historical data, weather patterns, soil conditions, and
farming practices. This project aims to enhance agricultural decision-
making, optimize resource use, and improve overall farm productivity.

Key Objectives

- Accurate Predictions: Develop a model that accurately

predicts crop yields or other agricultural outcomes based on
diverse datasets.

-Resource Optimization: Use predictions to help farmers allocate

resources like water, fertilizers, and pesticides more efficiently.

- Risk Management: Provide early warnings for potential risks

such as droughts, pests, or diseases, enabling timely
interventions.

Approach

- Data Collection: Gather extensive data on crop yields, weather

conditions, soil properties, and other relevant factors from various
sources.

- Data Preprocessing: Clean, normalize, and prepare the data for

training the machine learning model.

29
 - Model Development: Build and train machine learning models, such
as regression models, decision trees, or neural networks, to make
accurate predictions.

- Validation and Testing: Evaluate the model’s performance using

validation techniques and test datasets to ensure its reliability and
generalizability.

- Deployment: Implement the model in a real-world setting where it

can be used to make predictions and provide actionable insights to
farmers.

Benefits

- Enhanced Decision-Making: Farmers receive data-driven

insights that inform their planting, irrigation, and harvesting
strategies.

- Increased Yields: By optimizing farming practices, the project

aims to increase crop yields and improve food security.

- Sustainability: The model promotes sustainable agriculture by

optimizing the use of inputs and reducing waste.

Challenges and Limitations

- Data Quality: The model’s accuracy depends on the availability

and quality of data, which can vary by region.

30
- Model Generalization: Ensuring the model can generalize
across different environments and farming practices is a
significant challenge.

- Complexity and Cost: Developing and deploying machine

learning models require expertise and resources, which may be
a barrier for small-scale farmers.

31
CONCLUSION

The crop prediction project demonstrates the potential of machine learning

to revolutionize agriculture by providing precise, actionable insights that
can lead to increased productivity, resource efficiency, and sustainability.
However, it also highlights the importance of addressing challenges related
to data quality, model adaptability, and accessibility to ensure that the
benefits of this technology are widely realized.

32
FUTURE ENHANCEMENTS

Future Enhancements for Crop Prediction Using Machine Learning

To further improve the crop prediction project and extend its capabilities,
consider the following enhancements

1. Integration with Real-Time Data

- IoT Sensors: Incorporate real-time data from IoT sensors

deployed in the field to monitor soil moisture, temperature, and
other environmental conditions continuously.

- Weather Forecasting: Integrate advanced weather forecasting

models to provide up-to-date and accurate weather predictions,
which can enhance the model’s forecasting accuracy.

2. Improved Model Accuracy

- Advanced Algorithms: Explore and implement advanced

machine learning techniques such as deep learning or ensemble
methods to improve prediction accuracy and handle complex
data relationships.

- Hybrid Models: Combine multiple machine learning models

(e.g., combining regression models with decision trees) to
leverage their respective strengths and improve overall
performance.

33
3. Enhanced Data Sources

- Satellite Imagery: Utilize satellite data to monitor crop health,

assess field conditions, and predict crop yields based on remote
sensing.

- Crowdsourced Data: Incorporate crowdsourced data from

farmers and agricultural organizations to enrich the dataset and
provide more localized insights.

4. Customization and Personalization

- Localized Models: Develop region-specific models tailored to

local conditions, crop types, and farming practices to improve
prediction relevance and accuracy.

- User-Friendly Interfaces: Create intuitive and user-friendly

interfaces or mobile applications that allow farmers to interact
with the model and receive personalized recommendations.

34
 5. Automation and Integration

- Automated Data Collection: Implement automated systems for

data collection and preprocessing to streamline the workflow
and reduce manual effort.

- System Integration: Integrate the crop prediction model with

other agricultural management systems, such as irrigation
controllers or crop management platforms, to automate
decision-making processes.

6. Advanced Analytics and Visualization

- Predictive Analytics: Develop advanced analytics capabilities to provide

not just predictions but also insights into potential future scenarios and
trends.

- Visualization Tools: Enhance data visualization tools to provide

clear and actionable insights, such as interactive maps and
dashboards that visualize predictions and recommendations.

7. Model Adaptation and Learning

- Continuous Learning: Implement mechanisms for continuous

learning and model retraining to adapt to new data and
changing conditions, ensuring the model remains accurate over
time.

- Adaptive Algorithms: Develop algorithms that can automatically

adjust to changes in data distribution and environmental
factors, minimizing the need for manual intervention.

35
 8. Addressing Data Privacy and Security

- Data Encryption: Ensure robust data encryption methods to

protect sensitive information and maintain user privacy.

- Compliance: Regularly review and update practices to comply

with data protection regulations and address any privacy
concerns.

9. Collaborative Research and Development

- Partnerships: Collaborate with research institutions,

universities, and agricultural organizations to stay at the
forefront of technological advancements and incorporate
cutting-edge research enhancementproject.

- Feedback Mechanisms: Establish feedback loops with users

(farmers) to gather insights on model performance and areas for
improvement.

10.Expansion and Scalability- Global Expansion- Extend the model’s

applicability to different regions and countries, adapting it to diverse
climatic conditions and farming practices.

- Scalable Infrastructure- Develop scalable infrastructure to

support large-scale deployment and handle increasing volumes
of data.

36
REFERENCES

[1] Manpreet Kaur, Heena Gulati, Harish Kundra, “Data Mining in Agriculture
on Crop Price Prediction: Techniques and Applications”, International
Journal of Computer Applications, Volume 99– No.12, August 2014.

[2] J. Meng, “Research on the cost of agricultural products circulation and its
Control under the new normal economic development,” Commercial Times,
no. 23,pp. 145147, 2016.

[3] A. Kaloxylos et al., “Farm management systems and the future Internet
era,” Comput. Electron. Agriculture., vol. 89, pp. 130–144, Nov. 2012.

[4] N. N. Li, T. S. Li, Z. S. Yu, Y. Rui, Y. Y. Miao, and Y. S. Li, “Factors influencing
Farmers’ adoption of new technology based on Logistic-ISM model-a case
study of Potato planting technology in Dingxi City, Gansu Province,” Progress
in Geography, vol. 33, no. 4, pp. 542-551, 2014.

37
.

38
39
.

40