Systematic Literature Review: Stock Price Prediction Using Machine Learning and Deep Learning

Abstract

This research was conducted using a literature review method to analyze various studies that will identify the type of analysis with the attributes used, the methods used, the methods most often used, and the methods that have the best performance. This study collected research from 2016 -July 2021, selected based on predetermined criteria and then collected 40 papers. This review found that there are four research topics, namely estimation, classification, clustering, and association. The findings of this study are four research topics, namely fundamentals, technicals, sentiment analysis, and even a combination of analyzes that use their respective attributes and datasets. There are thirty-one different methods found to be used in predicting stock prices. LSTM, MLP, RF, and SVM are the most widely used methods. In addition, MLP is a method that gives the best performance of 71.63% and LSTM of 70%. The use of combined machine learning methods with ensemble techniques, deep learning, and selection of input attributes in pre-processing is recommended for better accuracy.

Enhancing Bitcoin Price Volatility Estimator Predictions: A Four-Step Methodological Approach Utilizing Elastic Net Regression

“…Finally, a comprehensive literature review was undertaken by Dopi et al [24], which focused on the application of machine learning and deep learning techniques in the prediction of cryptocurrency market stock prices. Their study investigated the correlation between the stock and cryptocurrency markets, with a particular focus on the potential and extensive utility of predictive modelling tools in comprehending complex market dynamics.…”