Traditional statistical as well as artificial intel- ligence techniques are widely used for stock market forecasting. Due to the non linearity in stock data, a model developed using the tra- ditional or a single intelligent technique may not accurately predict results. The term stock trend refers to the movement of the stock go- ing up if the price increases versus the stock going down if the price decreases.Exploitation of the efficient search optimization technique of genetic algorithms combined with a long- term and short-term memory neural network (LSTM)-based prediction model can help in identifying an optimal subset of stock techni- cal indicators along with the time-steps to look back for accurate stock prediction. The main research question we aim to explore is how can feature engineering benefit model prediction accuracy?
To optimize the run-time, parallelization was implemented as can be seen in the diagram below.
