Event Forecasting for Thailand’s Car Sales during the COVID-19 Pandemic

Simple item page
dc.contributor.authorLeenawong, Chartchai
dc.contributor.authorChaikajonwat, Thanrada
dc.date.accessioned2026-08-06T10:37:20Z
dc.date.available2026-08-06T10:37:20Z
dc.date.issued2022-07-01
dc.description.abstractThe COVID-19 pandemic that started in 2020 has affected Thailand’s automotive industry, among many others. During the several stages of the pandemic period, car sales figures fluctuate, and hence are difficult to fit and forecast. Due to the trend present in the sales data, the Holt’s forecasting method appears a reasonable choice. However, the pandemic, or in a more general term, the “event”, requires a subtle method to handle this extra event component. This research proposes a forecasting method based on Holt’s method to better suit the time-series data affected by large-scale events. In addition, when combined with seasonality adjustment, three modified Holt’s-based methods are proposed and implemented on Thailand’s monthly car sales covering the pandemic period. Different flags are carefully assigned to each of the sales data to represent different stages of the pandemic. The results show that Holt’s method with seasonality and events yields the lowest MAPE of 8.64%, followed by 9.47% of Holt’s method with events. Compared to the typical Holt’s MAPE of 16.27%, the proposed methods are proved strongly effective for time-series data containing the event component.
dc.identifier.citationData, 7(7), 2022
dc.identifier.doi10.3390/data7070086
dc.identifier.issn23065729
dc.identifier.other2-s2.0-85133262958
dc.identifier.urihttps://dspace.kmitl.ac.th/handle/123456789/13315
dc.sourceData
dc.subjectCOVID-19
dc.subjectevent component
dc.subjectHolt’s method
dc.subjectseasonality
dc.subjectThailand’s car sales
dc.titleEvent Forecasting for Thailand’s Car Sales during the COVID-19 Pandemic
dc.typeArticle

Files

Collections

Research