Stock Price Prediction: A Comparative Study between Traditional Statistical Approach and Machine Learning Approach

Abstract

Stock market is one of the most important sectors of a country's economy. Prediction of stock prices is not easy since it is not stationary in nature. The objective of this paper is to find the best possible method to predict the closing prices of stocks through a comparative study between different traditional statistical approaches and machine learning techniques. Predictions using statistical methods like Simple Moving Average, Weighted Moving Average, Exponential Smoothing, Naive approach, and machine learning methods like Linear Regression, Lasso, Ridge, K-Nearest Neighbors, Support Vector Machine, Random Forest, Single Layer Perceptron, Multi-layer Perceptron, Long Short Term Memory are performed. Moreover, a comparative study between statistical approaches and machine learning approaches has been done in terms of prediction performances and accuracy. After studying all the methods individually, the machine learning approach, especially the neural network models are found to be the most accurate for stock price prediction.