VISVESVARAYA TECHNOLOGICAL UNIVERSITY

Jnana Sangama, Belgaum- 590 018

A Project Report On

“DETECTION OF AUTISM SPECTRUM DISORDER IN EARLY

STAGE USING MACHINE LEARNING MODEL”
Submitted in the partial fulfilment of the requirements for the award of the Degree of

Bachelor of Engineering in Computer Science and Engineering

submitted by
Rakshitha V [1DB18CS104]
Sirisha M N [1DB18CS138]
Sneha H R [1DB18CS140]
Tejaswini D S [1DB18CS149]

Under the guidance of

Mr. Mohammed Kaleem
Assistant Professor
Department of CSE

Don Bosco Institute of Technology

Mysore Road, Kumbalgodu, Bengaluru-560074
2021-2022
 VISVESVARAYA TECHNOLOGICAL UNIVERSITY
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
DON BOSCO INSTITUTE OF TECHNOLOGY
Mysore Road, Kumbalgodu, Bengaluru-560074

CERTIFICATE

This is to certify that the project Phase-II entitled “ Detection Of Autism Spectrum
Disorder In Early Stage Using Machine Learning Model” is bonafide report carried out
by Rakshitha V [1DB18CS104], Sirisha M N [1DB18CS138], Sneha H R [1DB18CS140]
and Tejaswini D S [1DB18CS149] students of Don Bosco Institute of Technology,
Bangalore in partial fulfillment for the award of the degree of Bachelor of Engineering in
Computer science and Engineering of Visvesvaraya Technological University, Belagavi
during the academic year 2021-22. The project Phase -II report has been approved as it
satisfies the academic requirements in respect of the Project Phase-II prescribed for the
Bachelor of Engineering Degree.

Mr. Mohammed Kaleem Dr K B Shiva Kumar Prof. B S Umashankar

Guide, Department of CSE H.O.D., Department of CSE Principal, D.B.I.T

External Viva
Name of Examiner Signature with date
 VISVESVARAYA TECHNOLOGICAL UNIVERSITY
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
DON BOSCO INSTITUTE OF TECHNOLOGY
Mysore Road, Kumbalgodu, Bengaluru-560074

DECLARATION

We, Rakshitha V [1DB18CS104], Sirisha M N [1DB18CS138], Sneha H R

[1DB18CS140] and Tejaswini D S [1DB18CS149] students of eighth semester B.E, at the
department of Computer Science and Engineering, Don Bosco Institute of Technology,
Bengaluru declare that the project phase-II entitled “Detection Of Autism Spectrum
Disorder In Early Stage Using Machine Learning” has been carried out by us and
submitted in partial fulfillment of the course requirements for the award of degree of
Bachelor of Engineering in Computer Science and Engineering discipline of Visvesvaraya
Technological University, Belagavi during the academic year 2021-22.

Place: Bengaluru RAKSHITHA V [1DB18CS104]

Date: SIRISHA M N [1DB18CS138]
SNEHA H R [1DB18CS140]
TEJASWINI D S [1DB18CS149]
 ACKNOWLEDGEMENT

The satisfaction and euphoria that successful completion of any project is incomplete without
the mention of people who made it possible, whose constant guidance and encouragement
made our effort fruitful.

First and foremost, we ought to pay our due regards to Don Bosco Institute of Technology,
which provided us a platform and gave an opportunity to display our skills through the
medium of project work. We express our heartfelt thanks to our beloved principal Prof. B S
UMASHANKAR, DBIT, Bangalore for his encouragement all through our graduation life
and providing us with the infrastructure.

We express our deep sense of gratitude and thanks to Dr. K B SHIVA KUMAR, Head of
the Department, Computer Science and Engineering for extending his valuable insight
and suggestions offered during the course.

It is our utmost pleasure to acknowledge the kind help extended by our guide Mr.
Mohammed Kaleem, Assistant Professor, Department of Computer Science and
Engineering, for guidance and assistance which consequently resulted in getting the project
phase II work completed successfully.

Last but not the least I would like to thank teaching and non-teaching staff for their
cooperation extended during the completion of the project Phase-II.

Rakshitha V [1DB18CS104]

Sirisha M N [1DB18CS138]

Sneha H R [1DB18CS140]

Tejaswini D S [1DB18CS140]
 ABSTRACT

Autism spectrum disorder is a neurological developmental disorder that affects a person’s

interaction, communication and learning skills. Although diagnosis of autism can be done at any
age, it’s symptoms generally appear in the first two years of life and it develops through time.
Autism patients face different types of challenges such as difficulties with concentration, learning
disabilities, motor difficulties, sensory problems, mental health problems such as anxiety,
depression, and many others. In addition to that they also exhibit a complex collection of behaviors
which makes it difficult for the trainers to identify the methodology to be adapted for training them.

Early diagnosis of ASD helps to have a quality life by providing thorough care and therapy.
However, in many developed countries, Diagnosis takes too much time. Besides, a trained medical
expert or specialists are required to identify autism spectrum disorder as there are no direct medical
tests. At present many techniques are used to evaluate autism in general, without identifying their
uniqueness or specific characteristics. In this project work Machine Learning algorithms such as
Naïve-Bayes classifier and SVM (Support Vector Machine) classifier have been utilized for
detecting autism. We have achieved the highest accuracy of 100% on SVM classifier and 95%
accuracy using naïve bayes classifier. Therefore, prediction model that can reliably detect ASD is
proposed.
 TABLE OF CONTENTS
Acknowledgement I
Abstract II
Table of Contents III
List of Figures IV

Sl.No. CHAPTERS Page No

1. INTRODUCTION 1
1.1 Overview 1

1.2 Existing System 2

1.3 Proposed System 2

1.4 Objectives 2
1.5 Problem Statement 3
2. LITERATURE SURVEY 3-4
3. DESIGN AND IMPLEMENTATION 5

3.1 System Architecture 5

3.2 Detailed Design 9

3.3 Data Flow Diagrams 11

3.4 Use-Case Diagram 13

3.5 Sequence Diagram 13

3.6 Class Diagram 14

 3.7 Software Requirements 14
4. RESULT AND SNAPSHOTS 15

5. CONCLUSION 20

REFERENCES 21

Figure No. TITLE Page

No
3.1 System Architecture 6

3.1.1 Data from Autism Dataset 7

3.1.2 Code for testing the data 8

3.1.3 (a) Implementation of the SVM classifier model 8

(b) Implementation of the Naïve-bayes model 8

3.1.4 (a) Assessing performance of the SVM model 9

(b) Assessing performance of the Naïve-bayes model 9

3.3 Data Flow Diagrams 11

3.4 Use Case Diagram 13

3.5 Sequence Diagram 13

3.6 Class Diagram 14

4.1 Performance Evaluation of Naïve-bayes and SVM 15

4.2 Naïve bayes and SVM classifier Accuracy Comparison 15

4.3 Home Page 16

4.4 Registration Page 16

4.5 Login Page 17

4.6 Prediction Page 17

4.7 Suggestion Page 18

4.8 About Page 18

4.9 Chatbot Conversation Page 19

LIST OF FIGURES
DETECTION OF AUTISM SPECTRUM
DISORDER IN EARLY STAGE USING
MACHINE LEARNING MODEL
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

CHAPTER 1
INTRODUCTION
1.1 OVERVIEW

Autism spectrum disorder is a neuro developmental disorder that affects a person’s interaction,
communication and learning skills. Although diagnosis of autism can be done at any age, its
symptoms generally appear in the first two years of life and develops through time. Autism patients
face different types of challenges such as difficulties with concentration, learning disabilities,
mental health problems such as anxiety, depression etc., motor difficulties, sensory problems and
many others.

Current explosion rate of autism around the world is numerous and it is increasing at a very high
rate. According to WHO, about 1 out of every 160 children has ASD. Some people with this
disorder can live independently, while others require life-long care and support.

Diagnosis of autism requires significant amount of time and cost. Earlier detection of autism can
come to a great help by prescribing patients with proper medication at an early stage. It can prevent
the patient’s condition from deteriorating further and would help to reduce long term costs
associated with delayed diagnosis. Thus, a time efficient, accurate and easy screening test tool is
very much required which would predict autism traits in an individual and identify whether or not
they require comprehensive autism assessment.

The objective of this work is to propose an autism prediction model using ML techniques and to
develop a mobile application that could effectively predict autism traits of an individual of any age.
In other words, this work focuses on developing an autism screening application for predicting the
ASD traits among people of age groups 4-17 years and for people of age 18 and more.
Chatter Bot: A chatbot is software application used to conduct an on-line chat conversion via text
or text-to-speech, in lieu of providing direct contact with a live human agent. Designed to
convincingly simulate the way a human would behave as a conversational partner, chatbot systems
typically require continuous tuning and testing, and many in production remain unable to
adequately converse or pass the industry standard Turing test.

Dept of CSE, DBIT 2021-2022 Page 1

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

1.2 EXISTING SYSTEM

Classification is a supervised learning method in machine learning and statistics in which computer
programme learns from the data input supplied to it and then applies that learning to categorise
fresh observations. Machine learning techniques such as Naive Bayesian probabilistic approach,
Decision Tree Classifier, K Nearest Neighbor Classifier, and Support Vector Machine Classifier
are used to solve the multiclass issue. The ability of these approaches to simulate knowledge as
well as categorise an input even if the number of samples is minimal with high accuracy is one of
the reasons for their use.

1.3 PROPOSED SYSTEM

In this proposed work Machine Learning algorithms such as Naïve-Bayes classifier and Support
Vector Machine classifier are utilized for identifying autism traits accurately in peoples of different
age groups and to detect whether one has autism are not.
 A portal is developed for prediction of Autism based on the input provided.
 Algorithms such as Naive-Bayes and SVM are used to predict the presence of autism.
 Along with predicting autism the proposed system also identifies the Autism stages and
suggestions are provided based on the diagnosis.
 A chatbot to assess the condition of the patient based on the input provided.

1.4 OBJECTIVES
The objective of this work is to propose an autism prediction model using ML techniques that
could effectively predict autism traits of an individual of any age. In other words, this work focuses
on developing an autism screening application for predicting the ASD traits among people of age
groups 4-11 years, 12-17 years and for people of age 18 and more.
 To propose an autism prediction model using ML Techniques, that could effectively predict
autism traits of an individual of any age group.
 Learning the characteristics of children with autism.
 Early identification of the stages of autism.
 Diagnosis using improvement of autism child.

Dept of CSE, DBIT 2021-2022 Page 2

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Dept of CSE, DBIT 2021-2022 Page 3

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

1.5 PROBLEM STATEMENT

Autism spectrum disorder is a neuron developmental disorder that affects a person’s interaction,
communication and learning skills. Although diagnosis of autism can be done at any age, its
symptoms generally appear in the first two years of life and develops through time. Autism patients
face different types of challenges such as difficulties with concentration, learning disabilities,
mental health problems such as anxiety, depression etc., motor difficulties, sensory problems and

many others. However, in many countries, it is too late to diagnose and involves lengthy
procedure.
To simplify this procedure, Autism prediction model has been created using Machine Learning
algorithms such as Naïve-bayes and SVM classifiers.

Dept of CSE, DBIT 2021-2022 Page 4

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

CHAPTER 2
LITERATURE SURVEY

Guannan Li et al. [2] has proposed a volume-based analysis of the amygdala and hippocampal
subfields of the infant subjects with risk of ASD at around 24 months of age. Specifically, to
address the challenge of low tissue contrast, we propose a novel deep-learning approach, i.e.,
dilated-dense U-Net, to automatically segment the amygdala and hippocampal subfields.

Fitrilia Susanti et al. [3] Children with autism have communication disorder that affects the
children face difficulty interacting and communicating with their environment both verbally and no
verbally. To facilitate it, learning visual communication was needed because children with autism
were better at receiving information visually than orally. The application of visual communication
learning for children with autism developed in digital media used the Picture Exchange
Communication System (PECS) method.

Leslie Mertz et al. [4] Trained clinicians are able to diagnose autism by the time when a child is
two years old, and recognized beneficial therapies for the children can then be pursued at once.
Unfortunately, the author points out that the average age of diagnosis of autism in the United States
is 4.3 years. The late diagnosis is owed in large part to a lack of accomplished workers to conduct
the qualitative assessment and evaluate children accurately, including the ones who have milder
forms of ASD.

Osman et al. [5] The classification technique for ASD diagnosis was utilised in this study on
children aged 4 to 11. For cataloguing, the K Nearest Neighbor algorithms and Linear Discriminant
Analysis are used. Thirty percent of the data set was chosen as test data and seventy percent was
chosen as training data to test the algorithms. As a consequence of the work, the LDA method has a
90.8 percent accuracy, while the KNN approach has an 88.5 percent accuracy. The sensitivity and
specificity values for the LDA algorithm are 0.9524 and.08667, respectively. For the KNN method,

Dept of CSE, DBIT 2021-2022 Page 5

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83
these values are 0.9762 and 0.80, respectively, while the F-measure values for the LDA algorithm
are 0.9091 and for the KNN algorithm are 0.8913.

Elizabeth Stevens et al. [6] they have done the cluster study on the sample of 2,116 children with
ASD in order to find the patterns of stimulating behaviours observed in centre and home-based
clinical settings. The results indicates that while the presence of multiple challenging behaviours is
common in most of the cases as a leading behaviour.

Airi Tusji et al. [7] made studying and creating dynamic interpersonal distance models and real-
time
recognition systems (ASD). In precise way we demonstrated the quantitative quantities of chasing
behaviours observed during therapy. Chasing behaviours are a extremely social activity since the
children has to guess the movement of a partner (therapist).

Cheol-Hong et al. [8] A structure to detect, record and label different behavioural patterns of kids
with autism spectrum disorder (ASD) have been established. The structure integrates 2 diverse
sensor platforms which are both wearable and inert.

Elena Pattani et al. [9] The study intended to examine the variations about compassion and coping
in mothers of kids (3 to 11 age group) separated in two distinct groups: an “ASD group” consisting
of mothers of kids detected with autism spectrum syndrome and a “control group” including
mothers of children with typical development.

ZhiZheng et al. [10] They proposed ASOTS, a closed-loop independent computer system for
developing social alignment skills for teenagers with ASD, in this work. This technology is thought
to be able to detect and track a child's attention in response to social alignment offers, as well as
assisting the adolescent in achieving correct social alignment when necessary.

Dept of CSE, DBIT 2021-2022 Page 6

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

CHAPTER 3

DESIGN AND IMPLEMENTATION

3.1 SYSTEM ARCHITECTURE:

System architecture is a conceptual model that defines the structure and behavior of the system.
Architecture comprises of the system components and the relationship. Describing, how they
work together to implement the system as whole. The proposed system architecture is shown in
Fig.1.

Fig.3.1. System Architecture

The research was carried out in four phases:

1. Data Collection with Synthesization
2. Training and Testing Dataset
3. Developing the Prediction Model
4. Evaluation of the Prediction Model.

Dept of CSE, DBIT 2021-2022 Page 7

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

A. Data Collection and Synthesization:

Data synthesis, in the context of information and data, refers to analysing what we've learnt from
our data collection and making a choice, as well as developing a program, strategy, or method
based on the information analysed. To put it another way, data synthesis is the process of
combining ideas from various sources to produce something new that suits the demands of the
community and population we're working with. Any service's function revolves around data
collection and analysis. Good data is information that is accurate, dependable, and reflects what is
truly going on in the service, the subject you're researching, or the community as a whole. For
data to be valuable, we must have faith in its accuracy. Another crucial feature of a good
measurement tool is its validity. Validity refers to the tool's ability to measure what it claims to
measure. The process of validating a data gathering tool or scale is determining the level of
confidence with which we make judgments about persons using the scale.

Fig 3.1.1 Data from the Autism dataset

B. Testing and Training Dataset:

The training dataset often consists of pairs of an input vector (or scalar) and the corresponding
output vector (or scalar), where the answer key is commonly denoted as the target (or label).
Based on the result of the comparison and the specific learning algorithm being used, the
parameters of the model are adjusted. The test dataset is a dataset used to provide an unbiased
evaluation of a final model fit on the training dataset. It is independent of the training dataset, but
Dept of CSE, DBIT 2021-2022 Page 8
that follows the same probability distribution as the training dataset.
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Fig.3.1.2. Code for testing the data

C. Developing the Prediction Model:

Prediction models are intended to aid healthcare professionals and patients in making decisions
about diagnostic tests, therapy initiation and discontinuation, and lifestyle changes. While they
cannot replace clinical experience, they can provide objective information about a person's illness
risk and help to avoid some frequent biases in clinical decision-making. Biases in the way data is
obtained or filtered for use by the model, on the other hand, might add other forms of biases,
therefore the underlying data and cohort selection are critical. Here, we have used two algorithms
in order to develop the prediction model:
 Bernoulli Naive Bayes algorithm
 SVM algorithm

Fig.3.1.3. (a). Implementation of the SVM classifier Model

Dept of CSE, DBIT Fig.3.1.3. (b). Implementation

2021-2022of the Naïve-bayes Model Page 9
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

D. Evaluating the Prediction Model:

A classification problem or a regression problem are the two types of challenges that a
performance evaluation model can address. We will need to utilise different metrics to evaluate
our model depending on which group the project belongs to.

Fig.3.1.4. (a). Assessing performance of SVM Model

Fig.3.1.4. (b). Assessing performance of the Naïve-Bayes Model

3.2 DETAILED DESIGN

The System design phase involves two sub phases:

• High Level Design.

• Low Level Design.

In the high-level design, the proposed functional and non-functional requirements of the
software are depicted. Overall solution to the architecture is developed which can handle those need
Dept of CSE, DBIT 2021-2022 Page 10
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

3.2.1 HIGH LEVEL DESIGN

There are several design considerations issues that need to be addressed or resolved before getting
down designing a complete solution for the system.
The main assumptions and dependencies identified are as follows:

• Python Version 3.8 and PyCharm Community has to be installed.

• There shall not be any firewall or other engines that prevents the remote requests from the
portal.
• There should not be any permission related issues on any cluster. The host OS should take
of permitting all the requests to the cluster from the interface layer.

Naïve Bayes Classifier: In machine learning, naive Bayes classifiers are a family of simple
"probabilistic classifiers" based on applying Bayes' theorem with strong (naive) independence
assumptions between the features. Naive Bayes has been studied extensively since the 1960s. It
was introduced into the text retrieval community in the early 1960s, and remains a popular
method for text categorization, the problem of judging documents as belonging to one category
or the other with word frequencies as the features. With appropriate pre-processing, it is
competitive in this domain with more advanced methods including support vector machines. It
also finds application in automatic medical diagnosis. Naive Bayes classifiers are highly scalable,
requiring a number of parameters linear in the number of variables (features/predictors) in a
learning problem. Maximum- likelihood training can be done by evaluating a closed-form
expression, which takes linear time, rather than by expensive iterative approximation as used for
many other types of classifiers.

Support Vector Machine (SVM Classifier): In machine learning, support-vector machines

(SVMs, also support-vector networks) are supervised learning models with associated learning
algorithms that analyze data used for classification and regression analysis. Given a set of training
examples, each marked as belonging to one or the other of two categories, an SVM training
algorithm builds a model that assigns new examples to one category or the other, making it a non-
probabilistic binary linear classifier. An SVM model is a representation of the examples as points
in space, mapped so that the examples of the separate categories are divided by a clear gap that is
as wide as possible. New examples are then mapped into that same space and predicted to belong
to a category based on the side of the gap on which they fall.

Dept of CSE, DBIT 2021-2022 Page 11

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

3.2.2 LOW LEVEL DESIGN

During the detailed phase, the view of the application developed during the high-level design is
broken down into modules and programs. Logic design is done for every program and then
documented as program specifications. For every program, a unit test plan is created. The entry
criteria for this will be the HLD document. And the exit criteria will the program specification
and unit test plan (LLD).

3.3 DATA FLOW DIAGRAMS

A dataflow diagram is a graphical representation of the "flow" of data through an information

system, modeling its process aspects. A DFD is often used as a preliminary step to create an
overview of the system without going into great detail, which can later be elaborated. DFDs can
also be used for the visualization of data processing.

A DFD shows what kind of information will be input to and output from the system, how the data
will advance through the system, and where the data will be stored.

I. DFD – 0

Figure 3.3.1 Data flow of DFD0 for Autism prediction mode

Dept of CSE, DBIT 2021-2022 Page 12

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

II. DFD – 1

Figure 3.3.2 Data flow of DFD1 for Autism prediction model

III. DFD – 2

Figure 3.3.3 Data flow of DFD2 for Autism prediction model

Dept of CSE, DBIT 2021-2022 Page 13

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

3.4 USE-CASE DIAGRAM

A use case diagram at its simplest is a representation of a user's interaction with the system that
shows the relationship between the user and the different use cases in which the user is involved.
A use case diagram can identify the different types of users of a system and the different use
cases and will often be accompanied by other types of diagrams as well.

While a use case itself might drill into a lot of detail about every possibility, a use case diagram
can help provide a higher-level view of the system. It has been said before that "Use case
diagrams are the blueprints for the system". They provide the simplified and graphical
representation of what the system must actually do.

Figure 3.4 Flowchart of the Use-case diagram for Autism prediction model

3.5 SEQUENCE DIAGRAM

Sequence Diagrams are interaction diagrams that detail how operations are carried out. They
capture the interaction between objects in the context of a collaboration. Sequence Diagrams are
time focus and they show the order of the interaction visually by using the vertical axis of the
diagram to represent time what messages are sent and when.

Figure 3.5 Sequence Diagram for Autism prediction mode

Dept of CSE, DBIT 2021-2022 Page 14

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

3.6 CLASS DIAGRAM

Class diagram is a type of static structure diagram that describes the structure of a system by
showing the system's classes, their attributes, operations (or methods), and the relationships
among objects. The class diagram is the main building block of object- oriented modeling. It is
used for general conceptual modeling of the structure of the application, and for detailed
modeling translating the models into programming code. Class diagrams can also be used for
data modeling.

Figure 4.6 Class Diagram

3.7 SOFTWARE REQUIREMENTS

 OS: Windows 7 and above

 Back end: Python 3.7.5
 Front end: HTML, CSS
 IDE: Jupyter Notebook and PyCharm

Dept of CSE, DBIT 2021-2022 Page 15

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

CHAPTER 4

SYSTEM TESTING

4.1 TESTING

The goal of testing is to find mistakes. Testing is a technique for determining how well a software
product works. The process of attempting to find every possible flaw or weakness in a work
product is known as software testing.

It allows you to test the functionality of individual components, sub-assemblies, assemblies,

and/or a whole product. It is the process of testing software to ensure that it meets its
requirements and meets user expectations, and that it does not fail in an unacceptable way. By
running test data and live data, the system has been confirmed and validated. To assess the
network's breakpoint, we performed manual testing as well as stress testing in our system. Manual
testing was carried out with the help of Selenium software, while stress testing was carried out
with the help of hundreds of nodes rented from an online server.

The first testing was performed in the first module, Data Preprocessing, to confirm that the
dataset was free of missing or unknown values. Data cleansing is completed successfully using
the original CSV file as input.

The second and third tests are performed in the second module, Feature Extraction, to lower the
dataset's dimensionality. To obtain the reduced feature dataset, the preprocessed CSV file is used,
and PCA and Random Forest are successfully applied individually.

4.1.1 MANUAL TESTING AND AUTOMATED TESTING

Manual software testing is a sort of testing in which test cases are manually run by a tester
without
the use of any automated technologies. The goal of manual testing is to find bugs, faults, and
flaws
in a software product. Manual software testing is the most basic of all testing methods, and it aids
Dept of CSE, DBIT 2021-2022 Page 16
in software applications.
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Before any new application can be automated, it must first be manually tested. Manual software
testing takes more time and effort, but it is required to determine whether automation is possible.
Manual testing concepts do not necessitate familiarity with any testing tool. "100% Automation is
not achievable," says one of the Software Testing Fundamentals. This necessitates manual testing.

Before software is launched into production, automated testing is used to ensure that it is working
properly and fulfilling requirements. This type of software testing employs predefined sequences
that are carried out by testing tools. Automated testing tools conduct software tests, report the
results, and compare the findings to previous test runs.

Automated tests can be used for a variety of scenarios, including unit, API, and regression testing.
The fundamental advantage of automated software testing is that it reduces as much manual effort
as possible into a series of scripts.

If unit testing, for example, takes a significant portion of a quality assurance (QA) team's
resources,
this procedure should be considered for automation.

Automated tests can be conducted multiple times throughout the day. This method is consistent
with continuous testing, continuous integration (CI), and continuous delivery (CD) software
development approaches, all of which strive to get code changes into production without
requiring
manual intervention.

4.1.2 UNIT TESTING

Modules' unit testing entails creating test cases to ensure that the program's underlying logic is
working properly and that program input produces valid outputs. Validation should be performed
on all decision branches and internal code flow. It is the testing of separate software units of an
application after they have been completed and before they are integrated. This is an invasive
structural test that relies on knowledge of its structure. Unit tests are used to test the business
process application and/or system configuration at the component level. Unit tests confirmed that
Dept of
each CSE,
unit DBIT
route of a business process followed2021-2022
the published specification and containedPage 17
clearly
defined input
 DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

and output.

4.1.2.1 USER INPUT

In the user interface, data is entered via a graphical user interface (GUI) and then tested. Each
element is checked for valid and invalid data ranges.

4.1.2.2 ERROR HANDLING

We attempted to address all issues that happened when running the GUI forms in this system.
Reading an empty record and displaying a compiler message are two common issues we've seen.
Exception handling has been used to deal with errors. Many faults have been handled with the
help of python's try catch blocks. The majority of the failures happened while committing data to
the database, and they were handled with extreme caution.

Table 4.1 Unit Testing Registration Specifications

4.1.3 INTEGRATION TESTING
An integration plan is used to connect the various parts of a system. The integration plan lays out
how modules will be joined and in what order to complete the system. The partially integrated
system
Dept of is tested
CSE, DBITafter each integration phase. Integration
2021-2022testing's main goal is to check thePage 18
module's interface.
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

The module dependency graph is a key feature that influences the integration strategy. The order
in which various modules call one other is represented by the module dependency graph. A
module
dependency graph is represented by a structure chart. As a result of looking at the structure chart,
an integration plan can be created using any of the following methods:
1. Big-bang approach
2. Top-down approach
3. Bottom-up approach
4. Mixed approach

Table 4.2 Integration Testing Login Specifications

4.1.3.1 BOTTOM UP INTEGRATION TESTING

Each subsystem is tested separately first, and then the entire system is checked. A subsystem may
be made up of several modules that communicate with one another via well-defined interfaces.

The fundamental goal of each subsystem's testing is to check the interfaces between the numerous
modules that make up the subsystem. Both the control and data interfaces are put to the test
Bottom-up
Dept of CSE,integration
DBIT testing has the advantage of allowing multiple disconnected subsystems
2021-2022 Page 19to
DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

be tested at the same time. The complexity that arises when a system is made up of a large
number of little subsystems is a disadvantage of bottom-up testing.

The fundamental goal of each subsystem's testing is to check the interfaces between the numerous
modules that make up the subsystem. Both the control and data interfaces are put to the test.
Bottom-up integration testing has the advantage of allowing multiple disconnected subsystems to
be tested at the same time. The complexity that arises when a system is made up of a large
number of little
subsystems is a disadvantage of bottom-up testing.

We tested all of the individual programs first in the Main module, and after receiving successful
results in the individual program testing, we went on to the integration.
We integrated a few programs and then put them to the test, with positive results.
4.2 TEST CASES

Dept of CSE, DBIT 2021-2022 Page 20

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Dept of CSE, DBIT 2021-2022 Page 21

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Dept of CSE, DBIT 2021-2022 Page 22

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

CHAPTER 4
RESULT AND SNAPSHOTS
Naïve Bayes algorithm gives 95 % accuracy in the detection of autism from dataset whereas
Support Vector Machine gives highest accuracy of 100%. Once the user enters the data answering
AQ-10 questions and other individual characteristics details model predicts whether a person has
autism or not using better performing and more accurate algorithm and displays it to the user.
Suggestion page is also included along with the prediction page to provide user a better knowledge
about the prediction made. Comparison of performance of both the algorithms are shown in Fig.5.
This shows that SVM Classifier gives the highest accuracy among the Naïve Bayes and SVM
classifier algorithms.

(a). Evaluative metrics of Naïve-bayes (b). Evaluative metrics of SVM

Fig.4.1. Performance evaluation result

Fig.4.2. Naïve bayes and SVM classifier Accuracy Comparison

Dept of CSE, DBIT 2021-2022 Page 23

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

SYSTEM SNAPSHOTS
Figure 4.3 shows the home page of the application for the user.

Figure 4.3. Home page

Figure 4.4 shows the registration page which allows the new user to register into the portal and get
their diagnosis done with their provided details.

Figure 4.4. Registration page

Dept of CSE, DBIT 2021-2022 Page 24

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83
Figure 4.5 shows the log in page which allows the user to log to take a test for diagnosis of Autism.
If there comes a new user, they can click the register button and can get themselves registered.

Figure 4.5. Login page

Figure 4.6 shows the predict page determines whether the user has Autism or not, based on the
inputs provided.

Figure 4.6. Prediction page

Dept of CSE, DBIT 2021-2022 Page 25

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Figure 4.7 shows the Suggestion page which helps user to follow the course of treatment or the
activities stated, so one can get clarified with the information provided here.

Figure 4.7. Suggestion page

Figure 4.8 shows the About page where we can get an idea of what Autism is and how to deal with
it.

Figure 4.8. About page

Dept of CSE, DBIT 2021-2022 Page 26

 DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

Figure 4.9 shows the compose section of the Chatbot page. This allows the registered users to
compose and send text to chatter bot.

Figure 4.9. Conversation page

Dept of CSE, DBIT 2021-2022 Page 27

 DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83
CHAPTER 5

CONCLUSION

Autism is a neurological developmental disability that hampers normal brain development affecting
communication, social interaction, cognition and behavior. Autism is known as a spectrum disorder
because its symptoms and characteristics appear in a variety of combinations that affect children
because of which they have to face severe challenges.

A portal is developed for prediction of Autism based on the input provided. Algorithms are used to
predict the presence of Autism. Naive Bayes algorithm, Support Vector Machine algorithm,
Random Forest algorithm etc. are the algorithms used to predict the presence of Autism along with
its stages.

A Chatbot is built to assess the condition of the patient based on the input provided. An interactive
session is involved in order to analyze the patient’s condition and recommend suitable medicines.

 In future, this work can be extended to work with different types of ASD data like:

 MRI scan data, EEG data and gene sequences data.

 Facial images and video sequences.

 Voice recognition method, sensors can be used to understand their actions, and the recorded
speech therapy can be added to this project.

 A discussion forum consisting of parents/guardians of autism patients can be evolved so as to help

each other to deal with patients. A video session with their guardians can be developed for effective
analysis of patient’s condition.

Dept of CSE, DBIT 2021-2022 Page 28

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

REFERENCES

[1] Base paper – “Machine Learning-Based Models for Early Stage Detection of Autism Spectrum Disorders” Tania
Akter , Md. Shahriare Satu, Md. Imran Khan , Mohammad Hanif Ali , Shahadat Uddin , Pietro Lió , Julian M.W.
Quinn , And Mohammad Ali Moni Received October 10, 2019, accepted October 30, 2019, Date of publication
November 11, 2019, Date of current version November 27, 2019 IEEE.

[2] Guannan Li, Meng-Hsiang Chen3, Gang Li, Di Wu4 , Quansen Sun1 , DinggangShen, Li Wang, “ A preliminary
volumetric MRI study of Amygdala and Hippocampal subfields in Autism during infancy”,16th International
Symposium on Biomedical Imaging, 2019 IEEE.

[3] FitriliaSusanti, DanangJunaedi, VeronikhaEffendy, “Communication Learning User Interface Model for Children
with Autism with the Goal-Directed Design Method”, 7th International Conference on Information and Communication
Technology, 2019 IEEE.

[4] Leslie Mertz, “New Quantitative Approach to Autism Diagnosis”, International Conference on Medical and Health
Informative 2019 IEEE.

[5] Osman Altay, Mustafa Ulas, “Prediction of the Autism Spectrum Disorder Diagnosis with Linear Discriminant
Analysis Classifier and K-Nearest Neighbor in Children”, 6th International Symposium on Digital Forensic and
Security,2018 IEEE.

[6] Elizabeth Stevens, Abigail Atchison, Laura Stevens, Esther Hong,DoreenGranpeesheh, Dennis Dixon, Erik Linstead,
“A Cluster Analysis of Challenging Behaviors in Autism Spectrum Disorder”, 16th IEEE International Conference on
Machine Learning and Applications, 2017 IEEE.

[7] Airi Tsuji, Satoru Sekine, Takuya Enomoto, Soichiro Matsuda, Junichi Yamamoto, Kenji Suzuki, "Modeling of the
Chasing Behaviors for Developmental Program of Children with Autism Spectrum Disorders” 16th international
conference on cognitive informatics and cognitive and computing, 2017 IEEE.

[8] Cheol-Hong Min, “Automatic Detection and Labeling of Self-Stimulatory Behavioral Patterns in Children with
Autism Spectrum Disorder”, 16th International Conference on Data Mining Workshop, Barcelona, Spain, Dec. 2016
IEEE.

[9] Elena Pattini, Dolores Rollo, “Response to stress in the parents of children with autism spectrum disorder”,
Instrumentation and Measurement Society prior to the acceptance and publication, 2016 IEEE.

[10] ZhiZheng, Qiang Fu, Huan Zhao, Amy R. Swanson, Amy S. Weitlauf, Zachary E. Warren, and NilanjanSarkar,
“Design of an Autonomous Social Orienting Training System (ASOTS) for Young Children with Autism”, IEEE
Transactions on Neural Systems and Rehabilitation Engineering 2016 IEEE.

Dept of CSE, DBIT 2021-2022 Page 29

DETECTION OF AUTISM SPECTRUM DISORDER 18CSP83

DETAILS OF THE STUDENT

BATCH – 2018-2022
NAME: Rakshitha V Sirisha M N Sneha H R Tejaswini D S

USN: 1DB18CS104 1DB18CS138 1DB18CS140 1DB18CS149

PHONE NUMBER: 9902320728 8050893900 6361912381 9380690449

EMAIL ID: rakshithavgowda2 Sirishamnnagaraju Sneharamkrish Tej555tejaswini

[email protected] [email protected] na242gmail.co @gmail.com
m

PLACED – IN: Legato Wipro - Mphasis

PERMANENTADDRE Mallathahalli, Maruthi Nagar, Madugiri, Vigneshwar

SS Bengaluru Tumkur Tumkur Nagar,
Bengaluru

Dept of CSE, DBIT 2021-2022 Page 30