Stock Market Prediction using Optimised LSTM Model

Abstract

In today’s world there have been many technologies evolved that can efficiently predict the stock market data for a given period of time, the need to accurately predict the stock market data can also be achieved using various Machine Learning algorithms or classifiers, such as - Decision Tree Classifier, SVM, K-nearest Neighbors, Logistic Regression, etc. The problem with these types of Machine Learning algorithms is that, although they give a pretty good result on the training dataset, however for the testing dataset the accuracy is not up to the mark, that is often times they have high variance. Apart from the training model used for the classification, the type of dataset used also plays an important role in determining the accuracy of the prediction, the stock market prediction dataset is a combination of various features (Close Stock Price, Open Stock Price, Highest Stock Trading, Lowest Stock Trading, Volume Traded, Dividends, Stock Splits) which helps in determining whether the stock price will go up or go down on the following day; in addition to these features, there are numerous factors affecting stock prices these factors include - Financial News related to the company, Newsletters of the organization, Annual Report of the organization, dividends, launches, and the current market scenario. Natural disasters and epidemics can also affect stock prices.