Machine Learning Projects

This repository contains a collection of machine learning projects completed. Each branch represents a different project focusing on various machine learning concepts and implementations.

Project Branches

1. knn: Implementation of K-Nearest Neighbors algorithm and performance evaluation procedures
2. decision-tree-and-model-evaluations: Implementation of Decision Tree algorithm and model evaluation procedures
3. Tweets sentiment analysis: Tweet Sentiment Analysis for Stock Market Performance Prediction
4. Perceptron-model: Perceptron Implementation and Model Comparison Study

Branch Details

K-Nearest Neighbors and Data Preprocessing

• Implementation of KNN classifier
• Data preprocessing techniques
• Model comparison with Naive Bayes
• Performance visualization and analysis

Decision Tree and Model Evaluations

• Implementation of Decision Tree Classifier for heart disease prediction
• Evaluation using various metrics (accuracy, precision, recall, F-measure, etc.)
• Detailed performance analysis and model evaluation

Tweets Sentiment Analysis

• Analysis of tweet impact on stock market performance
• Feature extraction from text data using TF-IDF and binary representations
• Implementation of various feature selection techniques
• Comparison of different linear regression approaches (closed-form, LASSO, SGD)
• Large-scale data processing and optimization

Perceptron Model

• Custom implementation of Perceptron algorithm
• Comprehensive model comparison study including:
  • Logistic Regression (with different regularizations)
  • Decision Trees
  • K-Nearest Neighbors
  • Neural Networks
• Implementation of grid search and random search for hyperparameter tuning
• ROC curve analysis and performance metrics comparison

Author

Jinghan (Summer) Sun jinghan.sun@emory.edu