/Regression-Models-Numpy-Lasso-Ridge-DT

Regression models(lasso, ridge, DT) using NumPy.

Primary LanguageJupyter NotebookMIT LicenseMIT

Ridge and Lasso Regression Models with NumPy

Objective

In this project, we will build regression models from scratch using NumPy on sports players, providing flexibility and control over the training process.

Data Description

The dataset contains information about sports players and aims to predict their scores. It comprises around 200 rows and 13 columns.

Tech Stack

  • Language: Python
  • Libraries: Pandas, NumPy

Approach

  1. Import required libraries and read the dataset.
  2. Data Pre-processing:
    • Remove missing data points.
    • Drop categorical variables.
    • Check for multicollinearity and remove highly correlated features.
  3. Create train and test data by random shuffling.
  4. Perform train-test split.
  5. Model Building using NumPy:
    • Linear Regression Model
    • Ridge Regression
    • Lasso Regressor
    • Decision Tree Regressor
  6. Model Validation:
    • Mean Absolute Error
    • R-squared

Modular Code Overview

  1. Input folder: Contains the dataset files.
  2. Src folder: Contains modularized code for pipeline - data processing, model building, and validation.
  3. Output folder: Stores the trained models for future use.
  4. Lib folder: Includes reference materials, such as notebooks and presentations.