Improved Analysis of Stock Market Prediction: (ARIMA-LSTM-SMP)

Abstract

In the present scenario, stock trading is one of the most important activities carried out by many users directly or even indirectly. Due to the growing importance of stock trading among the vast publics, also Predictions on stock market prices is also gaining equal importance. Nevertheless, when dealt with this issue, it becomes a great challenge since the environment is hugely multifaceted and active environment. There are numerous educations from several zones motivated to carry on that task and application of Machine Learning methods play an important role in many of them. In the continuing research work, countless instances where Machine Learning procedures, can produce pleasing results when performing estimate-based analysis. The capability to forecast stock movement is critical for investors in stock market. Using everyday time series data, anyone can predict the inclination using simple moving average system. The auto regressive integrated moving average (ARIMA) representations is widely discovered for time series calculation. This work gives an insight into widespread usage of constructing stock price based analytical work by taking into consideration, the ARIMA model. Available typical data pertaining to stocks, is obtained from Tata Global Beverages are used for stock-based estimation. From the experimental results it is confirmed, the ARIMA model has a robust possible for quick period-based prophecy and will contest positively with existing techniques employed for stock price prediction. This work also concentrates on the convention of LSTM networks consequently, to guess upcoming movements of stock prices based on the past amount. For this objective, an estimation model is constructed, and a sequence of tests had been conducted and all outcomes investigated beside a quantity of metrics to judge if this algorithm works when compared to other methods related to Machine Learning domain. All these techniques were associated to find the best model for prediction. The results showed that LSTM achieved finest performance and projected the market with precision of 77%.