Publication

Abstract : Forecasts and analysis of demand or stock prices present the concern and future opportunities for investors as a challenge and area of interest to academia or any analyst because of its complication. It indicates the need to make accurate forecasts when it comes to defining trade trends as well as the capability of delivering proper predictions that are vital in making profitable and sound financial decisions. Therefore, to build a new enhanced model of stock market index prediction, in this paper, a combined strategy got introduced in involving both CNN and LSTM. As for the upcoming task of price prediction of stocks, the approach is based on the use of CNNs and LSTMs, meaning that it is required to have a stable base of an algorithm and work on increasing the accuracy of long-term predictions. This given the market is integrated and not isolated, makes the identification of stocks prices extremely challenging and therefore the need for complex methods. The algorithm developed here uses LSTM, as it is a form of recurrent neural networks, having the market data as input to predict the share prices and they have a comparatively higher accuracy of prediction as compared to existing sharing and prediction algorithms. Graphical results are in order in order to illustrate the proposed hybrid model after training and testing have been done in various numbers of input data. In this way, this approach enables not only the current and predicted figure of stock market as significant factors in the evolution of stock exchange but also have a comprehensive study of the stock market trends.