This repository contains the datasets, R code, and final report for making predictions about the future value of the EPS (Earning Per Share) index, which measures net income expressed in monetary terms with reference to each share.
The "Russel_3000_foundamental_enlarged" dataset has 2493 records and a total of thirteen variables and contains data on the largest U.S. companies by market capitalization.
The process that led to providing an accurate estimate of the index in question starts with a preliminary exploratory analysis of the data through a feature correlation examination, then proceeds with the implementation of regularized regression models, Lasso and Elastic Net specifically, to the implementation of the linear regression model. In choosing features, considerations are reinforced with any additional models, Stepwise Selection, Backward and Forward Selection, to justify the final choice.
Ulteriori informazioni sui risultati ottenuti e interpretazioni personali sono disponibili nel report finale!