Machine Learning Models

This repository contains implementation for various Machine Learning models. Following is the list of contained models:

• Perceptron

  • Path - ml.algo.perceptron.Perceptron.py
  • Model outputs the following:
    • Training and validation accuracies over epochs
    • Confusion Matrix

• Single Layer Neural Network with a single hidden layer containing N number of hidden units

  • Path - ml/algo/neural_net/SingleLaterNeuralNet.py
  • Report containing results for the following experiments performed on the MNIST dataset:
    • Experiment #1: Find training accuracies and plot confusion matrix for the neural network trained with a varying number of hidden units
    • Experiment #2: Train the neural network with a varying number of training samples
    • Experiment #3: Train the neural network with a varying number of momentum values (.25, .5 and .95)
  • Model outputs the following:
    • Training and validation accuracies over epochs
    • Confusion Matrix

• Naive Bayes classifier

  • Path - ml/algo/naive_bayes/naive_bayes.py
  • Classification accuracy of the model on the yeast_test.txt dataset = 44.0083%

• K-means clustering algorithm to cluster and classify the OptDigitsdata dataset

  • Path - ml/algo/k_means/KMeans.py
  • Model outputs the following:
    • Average mean square error
    • Mean square separation
    • Mean entropy
    • Accuracy

Installing Dependencies

• Using pip - Install all the necessary requirements specified in the requirements.txt file by running pip install -r requirements.txt
• Using pipenv [preferred] - Install all the necessary requirements specified in the Pipfile file by running pipenv sync