The following information was submitted:
Transactions: WSEAS TRANSACTIONS ON COMPUTERS
Transactions ID Number: 53-409
Full Name: DAI Da Meng
Position: Associate Professor
Age: ON
Sex: Female
Address: Northwestern Polytechnical University
Country: CHINA
Tel:
Tel prefix:
Fax:
E-mail address: jsj_ddm@126.com
Other E-mails:
Title of the Paper: prediction symbolic representation patterns in time series
Authors as they appear in the Paper: dai da-meng mu de-jun
Email addresses of all the authors: jsj_ddm@126.com mudejun@ nwpu.edu.cn
Number of paper pages: 10
Abstract: It is an interesting problem that predicts time series in time series analysis and time series data mining community. In this paper, we propose a novelty approach of predicting symbolic representation patterns of time series. Proposed method is based on symbolization representation of time series. Given a size of prediction window and a size of alphabet, at each crisp sample point, historical data of a time series is segmented and transformed into the Piecewise Aggregate Approximation (PAA) representation, and the PAA representation is then further symbolized into a symbol string. Naive Bayesian predictor is then employed to predict future trend from the symbol string. The experimental results demonstrate that proposed method predict trend patterns of stock price time series and stock index time series with high precision.
Keywords: Time series; Piecewise aggregate approximation; Predict trend patterns
EXTENSION of the file: .doc
Special (Invited) Session:
Organizer of the Session:
How Did you learn about congress:
IP ADDRESS: 220.190.59.175
