Exploring Deep Learning to Determine the Optimal Environment for Stock Prediction Analysis

Abstract

Time series data and its analysis is a challenging task as data kept changing continuously and based upon the new data arriving, the previous analysis might often seem obsolete. Time series data is time-ordered datasets, which is a more advanced area of data analysis. When evaluating a time series, many aspects must be considered, that can be used to help the explain time series. To analyse time series data accurately and rapidly, deep learning is quite helpful. Further, deep learning is also abiding with a couple of challenges, as there are several activation functions, loss functions, optimizers, number of deep layers. In this paper, experimentally, the various parameters of deep learning would be testing upon time series data to determine the optimal environment for stock prediction analysis.