Stock Value Prediction Accuracy Enhancement Using CNN and Multiple Linear Regression for NIFTY

Stock market investment is the common man’s interest nowadays. In fact, after the pandemic years, people have realized the potential of money earnings available in the investment and trading of the stock market. Though it gives profit to the public, but only when the choice of the stock and the time of investment is right. Before investment, the investor should assess the performance of the stock based on past behaviour and try to forecast the value of the stock in the coming future. If the prediction goes wrong, it will be a lossful investment and it can cost a lot to the investors. Hence, it is important to have the right prediction for investment and trading. There are many stock value prediction systems have implemented, but the accuracy of the model is still questioned and is not suitable in today’s volatile market. Here, the attempt is made to design and develop the stock value prediction system with higher accuracy and faster response for intraday as well as long-term investments and trading. The most popular and effective deep learning algorithm of convolutional neural network (CNN) is used for developing the model along with multiple linear regressions for predicting stock value. The model is trained and tested by using the Indian market NSE NIFTY dataset by Kaggle datasets and presented the improved accuracy numbers for forecasting of NIFY index of NSE. The results have shown that prediction accuracy is improved as compared to existing approaches in less time.