Autism Screening using Various Machine Learning Models

• 

  Machine Learning Models Comparison

Inspiration

Autistic Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by impaired communication, cognitive, and social skills and abilities. Existing screening tools for detection of autism are expensive, cumbersome, time-intensive, and sometimes fall short in predictive value. This low cost, quick and easy to use diagnostic test will help healthcare professionals and individuals assess if they should pursue formal treatment options.

What it does

This model can be used as initial screening test before doing more expensive diagnostic tests. People can take this test from the comfort of their home, on their computer or mobile phone for initial assessment.

How we built it

Data preprocessing was done by cleaning and removing irrelevant fields. Then, data was transformed using label encoding technique which converts non-numerical labels to numerical labels and normalizes the data. For evaluating the models, 10-fold cross validation technique was used in which the data is partitioned into ten equal sizes and nine samples were used for training and one for validation, repeating the process ten times with different sample for validation each time. Each model’s accuracy score, and confusion matrix that describes performance of the model were generated.

Challenges we ran into

Data Preprocessing and finding which attributes/fields are actually relevant to prediction

Accomplishments that we're proud of

The models developed have achieved average accuracies of 96.7%(LR), 90.7%(DT), 95.4%(NB), 92.3%(SVM) and 97.4%(NN) with standard deviations of 0.023(LR), 0.036(DT), 0.034(NB), 0.035(SVM) and 0.01(NN) respectively.

What we learned

Dropping gender and age from the input feature list improved accuracy which means they are not useful for predicting ASD. Accuracy of models drop if only response to questions i.e. behavioral traits are used for training. So, other factors i.e. if the person was born with jaundice, family history and ethnicity are important factors to consider along with behavioral traits for ASD screening

What's next for Autism Screening using Various Machine Learning Models

By collecting more data and using that for training, we can further improve the accuracy of the model.

Built With

• kaggle
• keras
• matplotlib
• numpy
• pandas
• python
• scikit-learn