Comparative study between LR and LSTM using stock market dataset for prediction using regression metrics . Stock markets are difficult to monitor and require plenty of other variables when trying to interpret the movement and predict prices. In this work one Recurrent neural network approach LSTM and a basic machine learning technique Linear Regression have been compared in terms of their prediction and performance on stock market price movements. Comparing the results trained on a daily basis, for the Reliance Industries . LSTM in particular is able to model financial time series better than Linear Regression. One of the main feature of LSTM is that each node is a memory cell capable of storing contextual information. Hence LSTMs perform better as they are able to keep track of the context-specific temporal dependencies between stock prices for a longer period of time while performing predictions for the future. An analysis of the results also indicate that LSTM gives better accuracy when the size of the dataset is big. With more data, more patterns can be extracted out by the model, and the weights of the layers can be better adjusted whereas Linear regression smaller the data more the accuracy it reflected. A huge dataset lowers the accuracy of Linear Regression.
Linear Regression Model implemented on 21 years of data
Linear Regression Model implemented on 1 year of data
