TIME SERIES ANALYSIS AND PREDICTION

Abstract

Nowadays, in the world of digitalisation that promotes ease to do any task as well as in more inferior time. Also, by utilising the latest techniques, there is a provision to execute things accurately in the present by taking responsibility for the future. Various fields-maintained records, these records required to analysed timely, such time-series data set now analysing using various Machine learning models. When we have a time-series data set, the various models can be used for forecasting. Now, the challenge is to choose the most appropriate when we have non-constant and inconsistent data sets. In this project, different kinds of time series prediction and forecasting models have applied on the same sets of data in order to identify the best. Also, before applying any model, test dataset for the time series components such as trend and seasonality. This project is done in several stages and also, removing the inconsistency present in the acquired data sets before applying. The three data sets selected are agriculture value-added percentage of GDP is data set 1, India's monthly inflation rate is dataset 2 and Revenue generation of a restaurant is data set 3. The selected data sets are essential and secure relevance place in this analysis model. For this purpose, chosen most generalised time series models Autoregressive integrated moving average (ARIMA), Exponential Smoothing Models, Holt Linear Trend model, Seasonal Decomposition, Cross-validation moreover, and Neural Network model and examine the best suitable model when the data set contains trend, or seasonality or both.