Deep Learning Based Model to Forecast the Direction of Stock Exchange Market Using social media: A Review

Abstract

A company\'s stock exchange price is not just a measure of the present worth of the value of its shares, but an indicator of its expected future performance as well. The company is an incorporated organization with the Regulatory Body and shares trading members are registered with SEBI. Given that stock market statistics are extremely time-specific and usually nonlinear in nature, it is very difficult to forecast future stock values. The forecast offers details about the present state of the flow of inventory values. This will also be used to decide whether to purchase or sell the individual shares of the particular stock for consumers. Many scientists were conducted with different data mining techniques to forecast stock market prices. This paper attempts to forecast the stock price of companies in the national stock exchange index LIX15 using Neural Network technology. For designing and training the models, the past data of the chosen stock will be used. To verify the consistency of the model, the effects of the model are used for comparison with actual data.

Introduction

I. INTRODUCTION

Over the course of time, one will note that mankind has wanted to eliminate the suffering from his life[1]. The presumption in society is that making more money brings security and luxury, so it is reasonable that there has been a lot of focus paid to trying to predicting the stock rates. Various strategies, hypotheses, and metrics, all with their own properties, have been attempted with differing degrees of effectiveness. Although it was attempted using a number of methods, no one was successful in solving the problem. The field of artificial intelligence researchers and investors alike are expecting that neural networks will help them figure out the underlying principles of consumer behavior. A stock market is a public venue where shares of stock and derivative securities are exchanged at a given price.

This is the type of market in which securities listed on a stock exchange can still trade. The corporation has a governing authority whose shares are traded on the stock exchange [3][4].

It is often referred to as the secondary market because it takes place between two participants and influences the primary market [6]. The Financial Exchange puts the two disparate people together to trade in their securities. Pricing in the stock market is determined by supply and demand. When stocks that are popular fall out of favour rise in price, on the market, but stocks that are being dumped on the market will decrease. Listed companies are described as any organization that has publicly-filed shares that are traded on the stock exchange.

Investors like to make the most money by buying and selling when their shares are at a point when they are both at their maximum. It is, however, very difficult to correctly forecast the future price of a commodity using stock market statistics, which are usually non-time invariant and subject to chaotic, nonlinear trends. Stock market data has grown rapidly connects people from around the globe and has become a challenge to use as a forecasting tool because of the changing technologies of finance and globalization [7] [8].

A. Motivation

Investors look to profit from equity portfolios by purchasing and selling their holdings at the opportune moment of maximum or minimum potential profit. Unlike traditional, physical statistics, the future price of a stock can be difficult to forecast using stock market information since the stock market is extremely variable and nonlinear. the rapid growth in economic globalization and the international markets, and emerging information technology are all resulting in new ways to expand the reach of our research[9].

II. RELATED WORK

DarmadiKomo et al.algorithmic models to get and price estimates, respectively, each trained on separate sets of data were applied and used the radial basis function (RBF) and multilayer percept (MLP) In these tests, actual data from Wall Street Journal (Dow Jones Industrial Average) were used as a reference. The models' proposed Dow Jones index gained a notable level of accuracy, generating index funds that tracked over 80\% of the average return on each month. In reality, the results show that the RBF network is much superior to the MLP network.

D. VenugopalSetty et al.There was an in-depth study of the effectiveness of the different approaches of data mining for business results. This is provided for more details on the basics of the Indian stock market, such as how important data mining is to the field of prediction, and other relevant data mining techniques are detailed in the article. In other words, there is an increasing distance between stronger storage and more efficient retrieval systems. There is a whole new order of discovery which should be put in place to use to enhance end-user information layout and resolution that would be necessary.

Dase R.K. et al. expanded his expertise by conducting a literature review on the use of neural networks for stock prediction They found that the results of their investigation showed that with time series analysis it was not possible to reliably predict indexes, however it seems that an artificial neural network may be sufficient for this purpose. Neural networks are capable of using an impressive amount of knowledge from a vast amounts of data. Based on the previous research, they find that an Artificial Network model of the market is effective in forecasting the world's financial markets. They argued, with evidence that this is a novel area of application for Artificial Neural Networks and that there are strong expectations for their application in accurate stock market index analysis.

Akhter Mohiudd et al., In his study, he used a neural network-based approach to forecasting stock price fluctuations. a neural network was employed to allow an estimation of possible stock returns a variety of different approaches have been used to test the potential of the indicator to deliver incorrect results In actual, controlled experiments, real data from the National Stock Exchange of India (NSE) was used to test the accuracy of this process This data, from the 02/started up on 2nd January, 2007 to the 03rd March 2010: TCS, Wipro, Axis Bank, Maruth and Tata Steel. Neural PERSISTEMBLE was in his opinion did not meet the norm but he also advanced new neural system design and training strategies to minimise the error in potential predictions.

D. Ashok kumar et al. General ideas about time series were addressed, a need for market indexes, the implications of applying an ANN to time series, and a review of previous work was conducted to explore models using neural networks for time series forecasting. NIFICS performs between the stock market index (MIDCAP) and neural network market model (BSE). In the results, the finding shows that the score is somewhat above average. According to the outcome of their research, the best performance is obtained with an optimum weighting factor of 0.28, with a momentum of 0.5, and a best epoch of 2960. The model achieved a lower-than-predicted fit in the industry, and could be applied to any kind of stock results.

AdityaNawani et al. The comparison of data mining techniques and market forecasting methods can be explored to use in developing a trading firm's market capitalization models. their thesis investigates how neural networks are employed along with the Graphical User Interface for the MATLAB Graphical Digital Toolbox to achieve findings that are trustworthy. When the qualified method is applied, it can be used to make forecasts about the parameters involved in the supply and demand in a given sector.

III. OPEN ISSUES

Trading in shares is big business in many economies. Based on the information on their websites, Stockbrokers do not seem to have any intelligent tool that can help them advise clients on which stocks are suitable for any buy or sale trade. These websites provide information that points to use of fundamental, technical and time series analysis methods. These prevalent methods show a trend on future movement and not the likely trade price for any stock in future. It is therefore 7 desirable to have a tool that does not just point a direction of price movement, but also indicates the most likely price value of the stock itself. An ANN model that is well tuned with the appropriate parameters can be used to develop such a predictive tool.

Conclusion

Investors are very popular with forecasting the stock price because investors want to know their return on their investments. Technical analysts and traders historically used stock market forecasts on the basis of past rates, amounts, price dynamics and fundamental trends. The forecast of stock prices is now very high. Not just the financial state but also the complex when equity markets are impacted Nation economics, political environment and natural disasters, etc. Back on the financial exchange In nature, the standard methods also cannot be predicted accurately. Investors are very popular with forecasting the stock price because investors want to know their return on their investments. Technical analysts and traders historically used stock market forecasts on the basis of past rates, amounts, price dynamics and fundamental trends. The forecast of stock prices is now very high Not just the financial state but also the complex when equity markets are impacted Nation economics, political environment and natural disasters, etc. Back on the financial exchange in nature, the standard methods also cannot be predicted accurately. The future stock price is predicted by the automotive regressive model. We are really common with the model. Investigate quite precisely the estimation of stock prices. This process of forecasting investment returns would help financially In such unpredictable situations, institutions and share brokers will forecast the future price.

References

[1] DarmadiKomo, Chein-I Chang, Hanseok KO, “Neural Network Technology for Stock Market Index Prediction”, International Symposium onSpeech, Image Processing and Neural Networks, 13-16 April 1994 [2] D. VenugopalSetty, T.M.Rangaswamy and K.N.Subramanya, “A Review on Data Mining Applications to the Performance of Stock Market”,International Journal of Computer Applications, (0975 – 8887) Volume 1 – No. 3, 2010 [3] Dase R.K. and Pawar D.D., “Application of Artificial Neural Network for stock market predictions: A review of literature”, InternationalJournal of Machine Intelligence, ISSN: 0975–2927, Volume 2, Issue 2, pp-14-17, 2010 [4] AkhterMohiuddin Rather, “A prediction based approach for stock returns using autoregressive neural networks”, IEEE, 978-1-4673-0126-8,2011 [5] D. Ashok kumar and S. Murugan, “Performance Analysis of Indian Stock Market Index using Neural Network Time Series Model”,International Conference on Pattern Recognition, Informatics and Mobile Engineering (PRIME), IEEE, 978-1-4673-5845-3, 2013 [6] AdityaNawani, Himanshu Gupta, Narina Thakur, “Prediction of Market Capital for Trading Firms through Data Mining Techniques”,International Journal of Computer Applications (0975 – 8887) Volume 70– No.18, May 2013 [7] Chi Kin Chow, Tong Lee, “Construction of multi-layer feed forward binary neural network by a genetic algorithm Neural Networks”, IJCNNProceedings of the 2002 International Joint Conference, Honolulu, HI, USA, 2002 [8] Cao Q, Leggio KB, SchniederjansMj., “A comparison between Fama French’s model and artificial neural network s in predicting the Chinese stock market”, Computers & Operations Research, Vol. 32, pp. 2499-2512, 2005. [9] SatyajitDhar, Tuhin Mukherjee, Arnab Kumar Ghoshal, “Performance Evaluation of Neural Network Approach in Financial Prediction:Evidence from Indian Market”, Proceedings of the International Conference on Communication and Computational Intelligence, pp.597-602, 2010.

Copyright

Copyright © 2022 Miss. Prajakta Pradeepkumar Shah, Prof. Gujar S. S.. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.