Stock Price Prediction with Random Forest

The Problem

Stock price prediction is a regression problem: estimate a continuous closing price from the day's other observable market data. The exercise is as much about honest evaluation as about the model — a plot of predictions against actuals reveals overfitting that a single accuracy figure hides.

Approach

A Random Forest Regressor learns the relationship between intraday market features and the closing price.

Workflow

• —Data acquisition from the Yahoo Finance historical dataset
• —Preprocessing and feature preparation
• —Random Forest Regression training
• —Prediction of closing prices on held-out data
• —Evaluation with MAE, MSE and R²
• —Visualization of actual against predicted price series