Long term historical records of the stock marketsare widely used in technical research to define, understandand analyze stock market's time series trends and patternswhich can be used to generate huge profits during tradingsessions. Even though, technical analysis using differenttechnical measures have been shown to be helpful inforecasting market patterns, formulating specific tradingrule is a challenging task. In this research paper, we havetried to analyze investor's sentiments considering USpresidential elections and effects of Covid 19 as an explicitfluctuating factor affecting stock market performance.In addition to this , in this research work ,we have tried toidentify correct and better trading rules and trading points,technical indicators to be considered using mathematicalformulations, to determine when to buy or sellstocks.Thus, given dynamically varying stock marketbehaviour in high frequency tradingenvironment, it is important to integrate market sentimentsinto forecasting operations. This paper combines sentimentsinto stock forecasting model using the log bilinear (LBL)model for short term stock market's sentiment patternlearning and recurrent neural (RNN) for longterm sentiments pattern learning which achieves betterperformance then deep learning based stock priceforecasting existing methodologies.
展开▼