Data Analysis and Classification of Autism Spectrum Disorder Using Principal Component Analysis.

Area of Science:

• Neuroscience
• Developmental Psychology
• Data Science

Background:

• Autism spectrum disorder (ASD) is a complex early developmental disorder affecting social interaction and communication.
• Current data analysis and classification of ASD remain challenging due to the wide spectrum of signs and symptoms.
• Understanding the neural underpinnings of ASD requires advanced analysis of responses to various stimuli.

Purpose of the Study:

• To develop and evaluate a method for classifying autism spectrum disorder (ASD) in both adults and children.
• To analyze neuroscientific data, specifically responses to audio and video stimuli, from individuals with ASD.
• To improve the accuracy and efficiency of ASD diagnosis through data-driven approaches.

Main Methods:

• The study employed practical component analysis (PCA) for data analysis and unsupervised classification.
• A three-stage methodology was implemented: dataset preparation, data analysis, and unsupervised classification.
• Neuroscientific data from adults and children with ASD, responding to audio-visual stimuli, were analyzed.

Main Results:

• The classification model achieved a sensitivity of 78.6% and specificity of 82.47% for adults with ASD.
• For children with ASD, the model demonstrated higher accuracy with a sensitivity of 87.5% and specificity of 95.7%.
• The results indicate the effectiveness of the proposed method in differentiating ASD in different age groups.

Conclusions:

• The practical component analysis method shows significant promise for the accurate classification of autism spectrum disorder.
• The developed approach offers a valuable tool for aiding in the diagnosis of ASD in both pediatric and adult populations.
• Further research can refine this method for earlier and more precise identification of ASD.