Dealing with seasonality by narrowing the training set in time series forecasting with kNN. (1st August 2018)

• Record Type: Journal Article
• Authors: Martínez, Francisco; Frías, María Pilar; Pérez-Godoy, María Dolores; Rivera, Antonio Jesús
• Is Part Of: Expert systems with applications. Volume 103(2018)
• Page Start: 38
• Page End: 48
• Publication Date: 2018-08-01
• Subjects: Time series forecasting -- kNN regression -- Seasonal time series; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2018.03.005
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A new scheme for dealing with seasonality in time series forecasting. It is based on building a different model to forecast every season. Experimental results with nearest neighbor regression enhance classical approaches. Abstract: In this paper, a new strategy for dealing with time series exhibiting a seasonal pattern is proposed. The strategy is applied in the context of time series forecasting using k NN regression. The key idea is to forecast every different season using a different specialized k NN learner. Each learner is specialized because its training set only contains examples whose targets belong to the season that is able to forecast. This way, the forecast of a specialized k NN learner is an aggregation of target values of the same season, reducing the likelihood of misleading forecasts. Although the strategy is applied to k NN, we think that other computational intelligence approaches could take advantage of it.