Project Overview

A machine learning project to predict autism spectrum disorder based on input features. This repository contains data, notebooks, trained models, and supporting scripts for building, evaluating, and deploying the classifier.

Repository Structure

Autism-Detection/
├── train.csv
├── test.csv
├── AutismDetection.ipynb
├── best_model.pkl
├── encoders.pkl
└── README.md

🧠 Project Overview

This project aims to develop a machine learning model that can detect autism (or likelihood thereof) given input variables (demographic, clinical, behavioral features). The pipeline involves:

1. Data loading & exploration
2. Data preprocessing / feature engineering
3. Model training & hyperparameter tuning
4. Evaluation (accuracy, recall, precision, confusion matrix)
5. Serialization of the best model + encoders
6. Optionally, inference / deployment

Prerequisites

• Python 3.7+
• Required libraries

  numpy
  pandas
  scikit-learn
  matplotlib / seaborn
  joblib / pickle
  jupyter
  
  

Steps to run

# Clone the repo
   git clone https://github.com/mehershiri/Autism-Detection.git
   cd Autism-Detection

# Install dependecies if needed
   pip install -r requirements.txt

# Start Jupyter Lab/ Notebook using the following commands: 
   jupyter notebook
   or
   jupyter lab

# Open and run AutismDetection.ipynb step by step.
   -The notebook loads train.csv and test.csv
   -It preprocesses features, encodes categorical variables
   -Trains multiple models, chooses the best one
   -Saves the best model & encoders (best_model.pkl, encoders.pkl)
   -Evaluates performance on the test set