SEASONAL TIME SERIES MODELLING WITH INTERVENTION AND OUTLIER FACTORS

• Main Author: Ridwan
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/49717
• Institution: Institut Teknologi Bandung

Many time-based phenomena that occur every year, it causes seasonal events, such as the holiday, rainy season, etc. so that the time series process is not only influenced by previous observations, but there are other factors such as seasonality. Not only seasonality, sometimes there are also phenomena that cause drastic changes to the mean of observation or only changes the observation for a moment, such as the existence of cultural festival that causes many tourists to visit at that time, a natural disaster, or government policy towards observation. this is referred to as intervention and outlier. Therefore, this modelling will be carried out by seasonal time series with intervention and outlier factors. This research is expected to evaluate the effect of interventions and outlier on the seasonal time series process, so that the effects of both factors can be obtained. In addition, the duration of these effects are also reviewed. Specifically, Innovation Outlier (IO) also reviewed the influence of seasonal factors on the characteristics of the IO effect. The intervention used was tourism promotion by the Bali Tourism Office and also the eruption of Agung Mountain for the case of domestic tourist data in Bali, while for the case of inflation rate data in Indonesia, the intervention was covid-19. The eruption of Mount Agung did not have a significant effect on many domestic tourists in Bali, nor did covid-19 have a significant effect on the inflation rate in Indonesia.