Open AccessRedundancy-test-based hyper-parameters selection approach for support vector machines to predict time series
Author(s) -
Yanfang Yu,
Junde Song
Publication year - 2012
Publication title -
wuli xuebao
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.199
H-Index - 47
ISSN - 1000-3290
DOI - 10.7498/aps.61.170516
Subject(s) - residual , redundancy (engineering) , computer science , series (stratigraphy) , benchmark (surveying) , selection (genetic algorithm) , algorithm , support vector machine , model selection , artificial intelligence , paleontology , geodesy , biology , geography , operating system
The selection of hyper-parameters is a crucial point in support vector machine modeling. Different from previous method of choosing an optimal model by using basic statistics of residuals in, the new approach selects hyper-parameters by checking whether there is redundant information in residual sequence. Furthermore, omni-directional correlation function (ODCF) is used to test redundancy in residual, and the accuracy of the method is proved by theoretical analysis and numerical simulation. Experiments conducted on benchmark time series, annual sunspot number and Mackey-Glass time series, indicating that the proposed method has better performance than the recorded in the literature.