PROJECTS

In our project, we utilized Random Forest and Support Vector Machine (SVM) algorithms to classify the T-Stage of breast cancer based on a dataset of 4024 patients. Through extensive analysis and model tuning, we achieved a test set accuracy of approximately 77% using Random Forest, demonstrating its effectiveness in predicting cancer progression. Despite these promising results, further refinement and exploration of alternative techniques are needed to enhance the model's performance and address challenges such as overfitting. This work highlights the potential of machine learning in improving early cancer detection and treatment planning.

This project involved analyzing daily closing stock prices at five-minute intervals from January 1, 2018, to January 1, 2022, for seven US stocks: AAPL, RTX, JPM, CSCO, GE, NKE, and IBM. The primary objective was to develop a model that predicts the future prices of AAPL, RTX, JPM, CSCO, and GE using a Multilayer Perceptron (MLP) neural network. The data was sourced from the NYSE, cleaned, normalized, and structured for modeling. Despite initial experiments with a single-layer MLP yielding moderate accuracy, attempts to improve predictions using a two-layer MLP led to overfitting and increased training time. The study highlights the challenges of stock market prediction due to external influences and the necessity of balancing model complexity with performance and training efficiency.  

 Developed an interactive visualization describing the impact of 2008 financial crisis on US foreign aid in Tableau.