Machine Learning Model Selection

A user-friendly web application developed as part of the CS351 - Artificial Intelligence course at GIKI.

This project simplifies the process of selecting and evaluating machine learning models, focusing on both regression and classification tasks, enabling users to preprocess data, experiment with models, and make informed decisions. Selecting the most suitable machine learning model for a dataset can be challenging, especially for individuals with limited technical expertise. This project addresses this challenge by providing an intuitive, no-code platform for preprocessing datasets, selecting models, and evaluating their performance.

Deployment

Deployed at: https://mlmodelselection.streamlit.app/

Setup

1. Clone this repository:

   git clone https://github.com/JunaidSalim/ML_Model_Selection.git

2. Navigate to the project directory:

   cd ML_Model_Selection

3. Install dependencies:

   pip install -r requirements.txt

4. Run the application:

   # Without CatBoost Model
   streamlit run app.py 
   
   # With CatBoost Model
   pip install catboost
   streamlit run main.py 

Functionality

1. Upload Dataset

Easily upload datasets in .csv format. The interface ensures the file is ready for processing.

2. Dataset Preview

Preview the uploaded dataset, including statistical summaries of numerical features.

3. Data Preprocessing

• Handle missing values.
• Normalize data.
• Apply one-hot encoding for categorical variables.
• Select features and target variables for modeling.

4. Task Selection

The app detects the task type (regression or classification) and allows manual adjustments if necessary.

5. Hyperparameter Tuning

Optimize model performance by adjusting parameters interactively.

Hyperparameters for Regression

Hyperparameters for Classification

6. Results and Model Comparison

View and compare evaluation metrics for various models to determine the best fit for your dataset.

License

This project is licensed under the MIT License. See the LICENSE file for details.