Stock Price Prediction System

Abstract

To determine the future stock value of a company is the main purpose of stock price prediction there is a continuous change in the price of stocks which is affected by different industries and market conditions. The high dimensionality of data is a challenge for machine learning models because highly correlated dimensions/attributes may exert influence on precision of the model. PCA is used to reduce dimensionality to fit linear regression algorithm for future stock price prediction. Linear regression algorithm is used prior to and subsequent to implementation of Principal Component Analysis on the Tesla stock price data. Results manifest that production of machine learning models can be boosted by PCA, reducing the correlation and appropriate selection of principal components for high redundancy of data. Root mean square value and R-square value is used for assessment. Keywords: Principal component analysis, Linear regression, Root mean square error, r square value.