碩士 / 國立臺灣科技大學 / 電機工程系 / 104 / In this study, a novel forecasting method for the temperature is proposed. The
proposed method combines a learning idea into the two-factor fuzzy time series model
proposed in the literature [1]. The considered factors for temperature prediction are
two factors, the values of temperature and the cloud density. The idea is that when
more historical data are used, the accuracy of forecasting model may increase. It is
assumed that the weather temperatures similar to past temperature histories. After the
fuzzy time series for prediction was proposed, based on it, many researchers proposed
different models used in many various fields. The important ingredients of fuzzy time
series forecasting model are definition and partition of the universe of discourse,
definition of fuzzy sets, establishment of fuzzy logical relationship and
defuzzification. The proposed model combine the fuzzy time series with the method
of classification of Logistic Regression (LR). The method of LR is to extract fuzzy
logical relationships from historical data and can be regarded as a learning mechanism.
From this study, it can be found that the method of LR can get more accuracy and
stability than LVQ and another method in [1].
Identifer | oai:union.ndltd.org:TW/104NTUS5442156 |
Date | January 2016 |
Creators | Zheng-Xun Lin, 林政勳 |
Contributors | Shun-Feng Su, 蘇順豐 |
Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
Language | en_US |
Detected Language | English |
Type | 學位論文 ; thesis |
Format | 49 |
Page generated in 0.0142 seconds