Overview
This study presents a systematic investment strategy utilizing machine learning algorithms to predict equity returns, leveraging a blend of technical and fundamental market insights. The strategy hinges on a comprehensive dataset that includes macroeconomic indicators, stock fundamentals, and technical signals, focusing on the S&P 500 Equal Weight Consumer Staples Index.
The data is meticulously prepared and transformed for algorithmic analysis. Feature engineering is then applied to extract predictive signals, which are analyzed using Support Vector Machine (SVM) algorithms with poly and RBF kernels. These models, targeting different forward return periods, are rigorously tested for predictive accuracy using metrics such as precision and recall.
The research acknowledges the inherent limitations of its data and methods, offering a realistic evaluation of the model's capabilities and suggesting areas for future enhancement. The results highlight the potential of machine learning in financial decision-making, paving the way for further advancements in quantitative finance strategies.
Repo Navigation
-
The project's code is split into four jupyter notebooks labeled by topic and ordered by number (starting with 1_price_data_collection). Each notebook is thoroughly commented
-
The final report is available for download via final_report.docx
-
All data used is available in the data directory