ML Implementations from scratch

Small Implementation of linear regression, logistic regression and Decision tree model from scratch using Python.

Introduction This code offers the following functionalities:

• Custom r2_score function to calculate the R-squared value (coefficient of determination).
• LinearRegression class for linear regression:
  • Configurable learning rate and number of iterations.
  • fit method to train the model on a dataset.
  • predict method to make predictions.
• Testing script demonstrating usage:
  • Creates synthetic data.
  • Trains and fits the model.
  • Makes predictions and evaluates performance.
  • Visualizes results.

Usage

1. Import and initialize:

   from regression import LinearRegression
   
   model = LinearRegression(learning_rate=0.001, num_iterations=1000)

2. Use the Fit and predict Methods

   X_train, y_train = ...  # Your training data
   model.fit(X_train, y_train)
   predictions = model.predict(X_new)

Dependencies

• numpy
• sklearn (used for testing)
• matplotlib (used for testing)