DocumentCode :
2893542
Title :
Stream Prediction Model Based on Tendency Correction
Author :
Meng, Fanrong ; Zhuang, Peng
Author_Institution :
Dept. of Comput. Sci. & Technol., China Univ. of Min. & Technol., Xuzhou, China
fYear :
2009
fDate :
18-20 Sept. 2009
Firstpage :
189
Lastpage :
193
Abstract :
Linear regression model is widely used in data stream prediction processing. In order to eliminate the prediction deviation caused by small data set, curve tendency correction technique is used to increase the prediction accuracy. Firstly the weighted moving method is used to modify the prediction function parameters. This algorithm improves the predicting accuracy, but causes low efficiency of time and space. Based on this algorithm, the exponential smoothing method is proposed. It is proved that this algorithm can reduce the space and time complexity, and also improves the prediction accuracy.
Keywords :
computational complexity; curve fitting; data handling; prediction theory; regression analysis; smoothing methods; curve tendency correction technique; data stream prediction processing; exponential smoothing method; linear regression model; prediction accuracy; prediction deviation; prediction function parameters; space complexity; stream prediction model; time complexity; weighted moving method; Accuracy; Application software; Computer buffers; Computer science; Information systems; Intelligent sensors; Linear regression; Predictive models; Regression analysis; Sensor systems; Data Stream Prediction; Data stream; Regression model; Tendency Correction;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Web Information Systems and Applications Conference, 2009. WISA 2009. Sixth
Conference_Location :
Xuzhou, Jiangsu
Print_ISBN :
978-0-7695-3874-7
Type :
conf
DOI :
10.1109/WISA.2009.35
Filename :
5368078
Link To Document :
بازگشت