Adaptive MLELM-AE model for efficient prediction of stock market data

Abstract The stock market makes a mention of public markets that contains buying, issuing, and selling shares which trade on a stock exchange. The aim of stock market is to confer capital to companies that they can utilize for funding and spreading their businesses also to serve investors. But it is elusive to prepare right decision for the companies in particular trading of stocks because of dynamic and intermediate nature of the share price. The charge of funding and commercial enterprise possibilities within the inventory market can boom if an efficient algorithm could be developed to predict the price of an individual stock. There are many deep learning algorithms available in which Extreme learning machine (ELM) is one of the most efficient technique for training single layer feed-forward neural networks (SLFNs). Integrating ELM with auto encoder has gotten another viewpoint for extracting features using unlabeled data. This paper attempts to focus on predicting stock market five days ahead by using a new variant of deep neural network i.e multilayer extreme learning machine with auto encoder (MLELMAE). This model is applied on YES, SBI, and BOI datasets there by the performance of the proposed model is measured and compared with other Deep Learning (DL) techniques like Radial Basis Function Neural Network (RBF), Back Propagation Neural Network (BPNN), and ELM in terms of Mean Absolute error (MAE), Mean Square Error (MSE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). The results also show that the proposed model outperforms best over other DL techniques.