Time Series Data Mining: An Experiment Using Large Data Analytics-Based Method

Abstract

Time series data is frequently found in data sets and is becoming a popular topic for study. By mining time series data, it is possible to realize time series predictions and get the time series' reflection of the social economic phenomena' regularity and growth process, and extrapolate to forecast its course of development. In the big data age, time series prediction has received an increasing amount of interest. Accurate trend prediction is the fundamental use of time series prediction. We provide many time series autoregressive (AR), moving average (MA), and ARIMA models in this work. AR and MA are then merged to create the ARIMA model. The ARIMA is used in conjunction with particular situations to anticipate risk for the National SME Stock Trading (New Third Board) time series in general scenarios. The case studies reveal that the findings of our study are often congruent with the real scenario, which has tremendously aided in the forecast of financial hazards.