IDENTIFICATIONS OF AUTISM SPECTRUM DISORDER

(ASD) IN ADULTS USING VARIOUS MACHINE

LEARNING ALGORITHMS

Team Members:
1. Saran G [20CS117]
2. Sugavanaesh S [20CS146]
3. Tharun kumar C [20CS154]

Guided By:
Mr.Yuvaraj.S
AP/CSE
1
OUTLINE
▸ Abstract
▸ Introduction
▸ Objective of the System
▸ Existing Scenario
▸ Proposed Solution
▸ Project Modules
▸ System Requirements
▸ Explanation of Each Modules
▸ Conclusion
▸ References
ABSTRACT
▸ The paper comprehensively examines Autism Spectrum Disorder (ASD) in
adults.
▸ It emphasizes the scarcity of clinical or screening-oriented behavioral datasets
for ASD, hindering accurate analyses and improved screening processes.
▸ The paper proposes a novel dataset with 20 unique traits, specifically designed
for ASD screening in adults, addressing this deficiency.
▸ This dataset has potential to enhance ASD research, aiding in identifying
crucial autistic traits and improving patient classification for more effective
screening.

3
 INTRODUCTION
▸ Autism Spectrum Disorder (ASD) affects social interactions, communication,
learning, and behavior, appearing in early childhood but diagnosable at any age.
▸ Recent studies emphasize understanding and assisting individuals with ASD
throughout their lives, shifting from viewing it solely as a childhood issue to
recognizing its impact on all ages.
▸ Early identification of ASD is crucial for prompt treatment and improved
quality of life, with machine learning showing promise, especially in children.
▸ However, diagnosing ASD in adults presents challenges due to the diverse
autistic community, necessitating tailored approaches considering variations in
symptoms and life events.

4
OBJECTIVES OF THE SYSTEM
▸ The main objectives of the proposed system are to enhance the accuracy and
reliability of identification of Autism Spectrum Disorder (ASD) in adults
using an impressive unsupervised classifier boosting Algorithm to classify
and predict with greater accuracy and AUC score

▸ We can achieve this by using hybrid model technique which comprises of

integrating more than one supervised classifier algorithms. Which helps in
boosting the efficiency and specificity of the model.

5
EXISTING SCENARIO
▸ Traditional ASD detection where not consistent and less accurate.

▸ Early identification for adults was a tedious task

▸ Machine learning and Deep learning algorithms have been applied to

Autism Spectrum Disorder.
▸ These algorithms can learn and identify the disorder with great accuracy
provided the dataset has all necessary parameters to classify the patient.

6
PROPOSED SYSTEM
▸ As we have many ASD detection but they are less accurate.
▸ So we are going to test different type of datasets with different
algorithms.
▸ We are going to use classification algorithms, because identifying ASD
will come under classification with data.
▸ There are several ML algorithms used for algorithm, we choose SVM,
Logistic Regression and Naive Bayes.

7
 PROJECT MODULES
1.DATA PREPARATION:
➔ Data must first be prepared for use as an input by machine
learning algorithms by being cleaned, processed, and
occasionally modified. There are numerous incorrect or missing
records in the collection, and other attributes also need to be
modified. Almost all learning algorithms greatly benefit from this
preprocessing in terms of performance and ability to predict
future events.
➔ Cleaning the data is most crucial aspect of machine learning. The
level of data purification is as important as training the model
cause this data will only be used by the model to get trained.

8
PROJECT MODULES
2.FEATURE IMPORTANCE AND EVALUATION:
➔ Measuring a classification model's performance is vital.
➔ Metrics like accuracy, specificity, and sensitivity evaluate model
effectiveness.
➔ Various machine learning algorithms were tested for ASD screening in
adults, children, and adolescents.
➔ Results showed accuracy ranges for each dataset, with FNN consistently
performing the best.

9
SYSTEM REQUIREMENTS

HARDWARE REQUIREMENTS
▸ CPU type : Intel core i3 processor
▸ Clock speed : 3.0 GHz
▸ RAM size : 8 GB

SOFTWARE REQUIREMENTS
▸ Operating System : Windows 10 / Linux

10
CONCLUSION

▸ In conclusion the patient’s with ASD need more recovery time as it’s a long
term learning curve. It can be easily misdiagnosed in adults. So we used
various algorithms to classify them with great accuracy and efficiency. To
achieve this we have pre-processed the data to clean the non-clinical data to
train the model in a proper way.

11
REFERENCES
▸ S. M. Mahedy Hasan, M. P. Uddin, M. A. Mamun, M. I. Sharif, A. Ulhaq, and G.
Krishnamoorthy, “A Machine Learning Framework for Early-Stage Detection of Autism
Spectrum Disorders,” IEEE Access, vol. 11, pp. 15038–15057, 2023, doi:
10.1109/access.2022.3232490.
▸ S. J. J. Jui, R. C. Deo, P. D. Barua, A. Devi, J. Soar, and U. R. Acharya, “Application of
Entropy for Automated Detection of Neurological Disorders with Electroencephalogram
Signals: A Review of the Last Decade (2012–2022),” IEEE Access, vol. 11, pp.
71905–71924, 2023, doi: 10.1109/access.2023.3294473.
▸ N. A. Mashudi, N. Ahmad, and N. M. Noor, “Classification of adult autistic spectrum
disorder using machine learning approach,” IAES International Journal of Artificial
Intelligence (IJ-AI), vol. 10, no. 3, p. 743, Sep. 2021, doi: 10.11591/ijai.v10.i3.pp 743-751.
▸ J. Shin, Md. Maniruzzaman, Y. Uchida, Md. A. M. Hasan, A. Megumi, and A. Yasumura,
“Handwriting-Based ADHD Detection for Children Having ASD Using Machine Learning
Approaches,” IEEE Access, vol. 11, pp. 84974–84984, 2023, doi:
10.1109/access.2023.3302903.
▸ S. Raj and S. Masood, “Analysis and Detection of Autism Spectrum Disorder Using
Machine Learning Techniques,” Procedia Computer Science, vol. 167, pp. 994–1004, 2020, 12
doi: 10.1016/j.procs.2020.03.399.
THANK YOU!!!

13