Novel Greedy Heuristic Optimized Multi-instance Quantitative for the Prediction of Stock Price
Abstract
In recent information technology relates to advanced computing, a high amount of information accumulated constantly. Mainly in the field of finance related computing technologies because all these computing technologies generate real time, tremendous data which consists different transactional records. Because of randomness and complexity in stock market related computing systems the stock price prediction is a hot concept and challenging the task. As information of web content improves in stock markets, researchers and investors usually extract different indicator factors, i.e., sentiments and events from prediction related stock market real time financial data. Because of the present scenario in financial and unknown factors in the stock market arena, prediction of stock price is a challenging task although traditional authors worked on neural networks to improve the prediction of stock prices in different financial areas. To improve the index based composite stock market movement’s prediction in multi instance quantitative data, in this paper, propose a Novel Greedy Heuristic Optimized Multi-instance Quantitative (NGHOMQ) approach to explore required data from factors and discarding their parameter relations. It can be used to combine sentiments and events and evaluate quantitative information in the comprehensive manner, use novel heuristic calculation to represent successive stock price related events. To prevent stock price prediction according to optimized statistical performance in heuristic modes with the multi instance use Pareto optimization. In addition to that our proposed approach is able to identify input of data to making predictions in stock market price in financial computing technologies. Experimental results of Indian stock market data describes the effectiveness of NGHOMQ compared to traditional neural networks related frameworks/approaches.