Comparative Analysis of Artificial Neural Networks for Predicting Nifty 50 value in The Indian Stock Market

Comparative Analysis of Artificial Neural Networks for Predicting Nifty 50 value in The Indian Stock Market

연구 분야: Artificial Intelligence

학회: 2021 5th International Conference on Information Systems and Computer Networks (ISCON)

Nowadays, The Indian stock market is very promising and every trader wants to take advantage of this market. So, prediction plays an important role to predict the value of Nifty 50 in the stock market. Nifty 50 comes under the futures and options (F&O) segment in the market. Machine learning tools are used for better prediction purposes. In this paper, we have presented four neural network models (Feedforward neural network, generalized regression neural network, radial basis neural network, and exact radial basis neural network) and all are based on machine learning. The research work proposed four model to predict the accuracy of Nifty 50 value by comparing it with a minimum error rate. The data set samples of last nine years, Jan, 2013 to June, 2021 collected from the Yahoo Finance URL, has been used in order to carried out the research work. Total four features as initial opening value, day high, day low, and total volumes traded on that particular day are considered as input parameters and the closing value of the day taken as an output parameter for the Nifty 50 value of Stock Market. The experimental results reveals that the best model to predict the Nifty 50 stock market is radial basis neural network model provides 98.17% accuracy in contrast to the rest of the three models.