machine-learning-intro

repo for some of my first machine learning projects using python

Perceptron

Implemented a single layer perceptron neural network using only numpy for emotion classification. Dataset used: https://www.kaggle.com/praveengovi/emotions-dataset-for-nlp

Backpropagation

Implemented a three-layer Neural Network that updates weights with backpropagation for handwritten digit recognition (MNIST dataset).

• Part 1: only used Pytorch to implement model described above (used this website to understand how to implement backpropagation from scratch: https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/)

• Part 2: used scikit-learn framework

Clustering and PCA

• Implemented a Principal Component Analysis (PCA) network to extract one sound from an audio file using only numpy.
• Implemented a K-Means clustering algorithm using only numpy. Dataset used: https://www.kaggle.com/datasets/parulpandey/palmer-archipelago-antarctica-penguin-data

Phishing URL Detection (CNN-LSTM)

Implemented a CNN-LSTM hybrid model to detect malicious phishing URLs from benign URLs using the keras framework. Dataset used: https://www.kaggle.com/datasets/taruntiwarihp/phishing-site-urls/code?resource=download