Seasonal patterns in time series data are periodic and recurring patterns caused by certain factors such as weather, holidays, repetition of promotions, or changes in the economic climate. Good data forecasting is very important for making decisions in the business sector, such as retail prices, marketing, production and other business sectors. There are several approaches that can be taken to analyze time series data that has a seasonal or trending pattern. Among them is the classical approach which decomposes seasonal and non-seasonal factors, then forecasts with certain assumptions. Then there is also an approach using artificial intelligence, in this case a more flexible feed-forward neural network is used as a tool for forecasting time series data. In this study the data used is data with a regular seasonal pattern 12. For data with a pattern like this SARIMA (1,1,1)(0,1,1)12 with a MAPE of 1.775% gives better results than FFNN 12-10-1 which produces a MAPE value of 7.5226%.
Copyrights © 2023