This is a Streamlit web application that allows users to train a Linear Regression model with customizable data preprocessing, model selection, and hyperparameter tuning. It provides interactive visualizations and performance metrics to help users understand their data and model.

• CSV File Upload: Accept .csv file uploads.
• Robust CSV Parsing: Users can specify CSV delimiter (comma, semicolon, tab) and encoding (UTF-8, Latin-1, ISO-8859-1).
• Missing Value Strategies: Choose from 'Drop Rows', 'Mean Imputation', or 'Median Imputation'.
• Outlier Handling: Option to 'Remove Outliers (IQR)' for numerical columns.

• Data Preview: Show a preview of the first few rows of the uploaded dataset.
• Optional Visualizations: Users can select to view:
  • Outlier Visualizations (Box Plots): Box plots for selected numeric columns to identify outliers.
  • Feature Distribution Plots (Histograms): Histograms for selected numeric columns to understand data distribution.
  • Correlation Matrix Heatmap: A heatmap showing the correlation between numeric columns.

• Auto-detection: Automatically detects numeric and categorical columns.
• User Selection: Allows users to select:
  • One or more feature columns (X) (numeric and one-hot encoded categorical).
  • One target column (y) (numeric only).
• Categorical Feature Handling: Selected categorical columns are automatically one-hot encoded.

• Model Selection: Choose between 'Linear Regression', 'Ridge', and 'Lasso' models.
• Hyperparameter Tuning: For Ridge and Lasso, a slider is available to adjust the alpha regularization parameter.
• Cross-Validation: Implements k-fold cross-validation (default k=5) for robust performance evaluation.

• Performance Metrics: Displays key metrics in a table format:
  • Coefficients
  • Intercept
  • R² score
  • Mean Absolute Error (MAE)
  • Mean Squared Error (MSE)
  • Cross-Validation R² (Mean and Standard Deviation)
• Visualizations:
  • Predicted vs Actual Values Plot: Scatter plot with a regression line.
  • Residual Plot: Scatter plot of residuals vs predicted values.
  • Feature Importance Bar Chart: Bar chart showing the magnitude of coefficients for each feature.

• Download Trained Model: Download the trained model as a .pkl file.
• Download Predictions: Download the model's predictions on the test set as a .csv file.

• Frontend/UI: Streamlit
• Backend/Logic: Python
  • scikit-learn for machine learning models.
  • Pandas for data manipulation.
  • Plotly Express for interactive visualizations.
  • NumPy for numerical operations.
  • statsmodels for trendline in plots.

1. Clone the repository:

   git clone https://github.com/Mani212005/LRml.git
   cd LRml

2. Create and activate a virtual environment:

   python -m venv venv
   # On Windows:
   venv\Scripts\activate
   # On macOS/Linux:
   source venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

4. Run the Streamlit application:

   streamlit run app.py

   This will open the application in your web browser.

   

• More advanced outlier detection methods.
• Support for other regression models (e.g., RandomForestRegressor, XGBoost).
• More sophisticated hyperparameter tuning (e.g., GridSearchCV, RandomizedSearchCV).
• Saving/loading models directly within the app.
• User authentication and persistent storage.