Autism Detection using Machine Learning

This project builds a machine learning model to detect autism based on text data, using a dataset titled autism_data.csv. The model is trained and optimized using Intel OneAPI extension for scikit-learn, which increases processing speed by 1.73x.

Dataset

The dataset used for training the model is autism_data.csv, which contains features such as:

• Age
• Gender
• Family History
• Social Skills
• Communication Behavior
• Others related to autism detection

Libraries and Tools

• Python
• Scikit-learn (with Intel OneAPI extension for speed optimization)
• Intel OneAPI

Key Steps

1. Data Loading: Load the dataset autism_data.csv.
2. Data Preprocessing: Handle missing values, normalize the data, and split it into training and testing sets.
3. Model Building: Train a logistic regression model to predict autism.
4. Speed Optimization: Apply Intel OneAPI extension to accelerate the training time by 1.73x.
5. Evaluation: Measure model performance (accuracy, precision, recall) and compare with baseline models.

Results

The model achieves a performance improvement in both accuracy and speed, achieving 87% accuracy and a 1.73x faster training time with Intel OneAPI.

Files

• Autism_Detection_with_ML.ipynb: The notebook for building and training the machine learning model.
• autism_data.csv: The dataset used for training the model.

Conclusion

This notebook demonstrates how Intel OneAPI can be effectively used to accelerate machine learning models and improve the detection of autism.

Deep Learning for Autism Detection using CNN

This project develops a deep learning model using Convolutional Neural Networks (CNN) to classify images for autism detection. The model is enhanced with Intel OneAPI for TensorFlow for optimized training performance.

Model Overview

The deep learning model is built using a CNN architecture to classify whether an image indicates autism or not. The model is then saved as weights to be used in a separate application file.

Dataset

• Image dataset containing features related to autism, preprocessed for input to the CNN.

Libraries and Tools

• Python
• TensorFlow (with Intel OneAPI extension)
• Intel OneAPI

Key Steps

1. Data Preprocessing: Load the image dataset, resize, normalize, and split into training and validation sets.
2. Model Building: A CNN is built to learn image features related to autism detection.
3. Speed Optimization: Leverage Intel OneAPI for TensorFlow to optimize training time and model performance.
4. Training: The model is trained using the preprocessed data and optimized for classification accuracy.
5. Saving Weights: After training, the model’s weights are saved for use in the application (app.py).

Results

The CNN model shows excellent performance in detecting autism with image data. The Intel OneAPI extension increases training speed significantly.

Files

• deep_learning.ipynb: The deep learning notebook implementing the CNN.
• app.py: Application that uses the trained CNN model’s weights to classify images.
• Model Weights: Saved weights from the trained CNN model for further use.

Conclusion

The CNN model, optimized with Intel OneAPI, performs well in detecting autism from images, enabling its integration into real-time applications for autism detection.