Time Series Forecasting System for Stock Market Data

The forecasting of stock market has been popular and hottest research and it has the issue in the complexity and volatility. The consideration of stock has the nature of dynamic as the domain of financial. Predictive analytics is provided by big data with machine learning approaches for the extraction of relevant information through huge volumes of data and provides more adorable efforts. Moreover, time series prediction system is the vital and important research area in today. Therefore, there is a critical requirement in forecasting methods to be effective and efficient utilization of large amount of market data for the analysis of future forecasting in stock price movement. In this paper, deep learning-based prediction system is proposed for next day stock market prediction analysis using Multilayer Perceptron (MLP) and data preprocessing techniques such as weighted moving average, min-max normalization, Box-Cox transformation are used for feature engineering. In the system evaluation, weighted moving average with multilayer perceptron model is best model for time series data analysis system.