Making Healthcare Smarter via Machine Learning: A smart ASD (Autism Spectrum Disorder) detection…

Corresponding GitHub Project: https://github.com/ritabratamaiti/Autism-Detection-API

Also check out RapidML documentation at: https://ritabratamaiti.github.io/RapidML

Also read my Medium Article on Universal Basic Income: If your job can be automated, it most probably will be; and why that’s a good thing….

The National Institute of Mental Health (NIMH) defines Autism spectrum disorder (ASD) as so:

Autism spectrum disorder (ASD) is a developmental disorder that affects communication and behavior. Although autism can be diagnosed at any age, it is said to be a “developmental disorder” because symptoms generally appear in the first two years of life.

Furthermore, the following symptoms are factored when making diagnosis:

• Difficulty with communication and interaction with other people
• Restricted interests and repetitive behaviors
• Symptoms that hurt the person’s ability to function properly in school, work, and other areas of life

(Source: NIMH)

This article is a comprehensive guide to the creation and usage of the Autism detection API, and a more rigorous discussion of Autism can be found at https://www.nimh.nih.gov/health/topics/autism-spectrum-disorders-asd/index.shtml

The API design can be divided into three sections:

• Generating the appropriate machine learning model, and training the model on our data. This is done with RapidML

• Using the helper module, make predictions upon receiving inputs from the API’s URL, from the actual Flask API

Important:

Find the notebook here: https://github.com/ritabratamaiti/Autism-Detection-API/blob/master/RapidML_ASD_Prediction.ipynb

(Contains data descriptions in various formats)

API Demo:

Visit: http://ritabratamaiti.pythonanywhere.com/query?ip=0,0,0,0,0,0,0,1,0,1,2,30.0,m,White-European,no,no,Ireland,no,Self,NO

You will get a value of 0 on your browser, indicating that the person does not suffer from ASD.

The values that are assigned to ip:(0,0,0,0,0,0,0,1,0,1,2,30.0,m,White-European,no,no,Ireland,no,Self,NO) indicate the various features as per dataset description. You are free to change the API parameters to explore different results. 1 = ASD present; 0 = ASD absent

Application Demo:

The above API can be used in a wide range of applications, ranging from web-apps, scripts, and mobile applications for ASD detection. Here, I have used the API to make an android app that gets user input, and then make predictions in the cloud, and delivers the output.

You can take the app for a test run here: https://github.com/ritabratamaiti/Autism-Detection-API/tree/master/Android%20App%20Based%20on%20API

About this project:

I am really glad that you checked out this project. As an Artificial Intelligence researcher, it has always been my dream to apply AI and general machine learning in the field of healthcare and medicine. This is one of the several projects that I am working on currently, which explores machine learning in healthcare.

I would really appreciate if you could support my work by

• Checking out RapidML at: https://rapidml.readthedocs.io/
• Starring my GitHub project repository: https://github.com/ritabratamaiti/Autism-Detection-API
• Visiting my AI for medical imaging proposal website: https://ritabratamaiti.wixsite.com/aimedicine
• Providing your valuable suggestions/inputs about the AI for medical imaging via this form: https://docs.google.com/forms/d/e/1FAIpQLSfdtiBknS4OWv3pQW9eBex2sGoSUKO_PZzM4DL5Q7ixM4JeLQ/viewform
• Check out pythonanywhere, an amazing solution for hosting python web applications. Also, if you wish to upgrade your account, follow this link: https://www.pythonanywhere.com/?affiliate_id=003915da (Note that this is an affiliate link.)
• Providing any amount of monetary support for continued development via paypal.me/ritabratamaiti