CS109 Final Project – Predictive Analytics for Equities
- Helen Kidane
- Pascal Bernard
- Steven Vasilakos
The zip file of this Github project can be downloaded here.
Exploring the Decomposition of Monthly Stock Returns
CS109 FINAL PROJECT – PREDICTIVE ANALYTICS FOR EQUITIES
How might we decompose, model, and thus, predict monthly stock returns using Lasso Regression to reduce feature dimensionality ?
In this project we gather monthly fundamental factor information about a given set of US stocks over the past five (5) years to explore potential models to explain returns (price change) for the current and possibly subsequent monthly period. For the current period we may view this as a model to decompose the price returns for the current month, and for the subsequent monthly period the goal is to model expected price returns given information from the current period.
This project scope will focus on the former problem, however, the decomposition of current returns may give insight into modeling future returns.
No causation is being assumed here, only a model to explain returns given the data set.