Projects

This educational pursuit led me to devise various machine learning projects in Python. My undertakings ranged from creating algorithms capable of image recognition—such as identifying cats in photographs—to discerning patterns within scientific datasets. Captivated by the potential of these technologies, I resolved to design a sophisticated trading algorithm.

My development efforts have yielded a robust program capable of:

• Aggregating and processing financial data for over 500 stocks spanning three decades, involving data cleansing, organization, and normalization of key variables.
• Implementing a pair of distinct Random Forest Classifier (RFC) machine learning models to ascertain optimal times for stock transactions, all the while mitigating against prospective look-ahead bias. This curated dataset now serves as a benchmark for training future generations of machine learning algorithms.
• Conducting exhaustive tests with a variety of machine learning models against this training dataset, tweaking parameters to optimize for maximum return on investment. My preliminary results, derived from a high-risk simulation scenario, indicate an impressive average annual return of 34%

Moving forward, my objective is to bifurcate the outcomes into two distinct categories—short-term and long-term investments—to discern which algorithms exhibit superior performance in volatile, short-term market conditions, and for modest trades, and which are more suited to predict long-term market appreciations. This is all a learning process but I am really enjoying this process of trying new things.