Premium
A novel two‐factor forecasting model for fuzzy time series
Author(s) -
Li ShengTun,
Lin SuYu,
Cheng YiChung
Publication year - 2007
Publication title -
pamm
Language(s) - English
Resource type - Journals
ISSN - 1617-7061
DOI - 10.1002/pamm.200700344
Subject(s) - computer science , series (stratigraphy) , reliability (semiconductor) , fuzzy logic , probabilistic forecasting , interval (graph theory) , salient , monte carlo method , time series , data mining , process (computing) , overhead (engineering) , variety (cybernetics) , artificial intelligence , machine learning , econometrics , mathematics , statistics , paleontology , power (physics) , physics , quantum mechanics , combinatorics , probabilistic logic , biology , operating system
The study of fuzzy time series has increasingly attracted much attention due to its salient capabilities of tackling vague and incomplete data. A variety of forecasting models have devoted to improving forecasting accuracy, however, the issue of partitioning intervals has rarely been investigated. Recently, we proposed a novel deterministic forecasting model to eliminate the major overhead of determining the k‐order issue in high‐order models. This paper presents a continued work with focusing on handling the interval partitioning issue by applying the fuzzy c‐means technology, which can take the distribution of data points into account and produce unequal‐sized intervals. In addition, the forecasting model is extended to allow process twofactor problems. The accuracy superiority of the proposed model is demonstrated by conducting two empirical experiments and comparison to other existing models. The reliability of the forecasting model is further justified by using a Monte Carlo simulation and box plots. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)