|
Article information
2014 , Volume 19, ¹ 2, p.76-93
Lysyak A.S., Ryabko B.Y.
Forecasting methods for time series with large alphabets based on universal measure and decision trees
We suggest and experimentally investigate methods to construct forecasting algorithms based on universal measure and decision trees. The classical method for forecasting is based on the universal coding along with methods for its optimization, and the method for forecasting on the basis of the universal measure. We propose a new approach to forecasting algorithms - the method of separation of the alphabet. This method can be implemented in an arbitrary prediction algorithm and it can significantly reduce the complexity and increase the accuracy of predictions. We show how to implement the method of separation of the alphabet into random prediction algorithm.
[full text]
Keywords: universal measure, decision trees, forecasting, prediction, R-method, time series
Author(s):
Lysyak Alexander Sergeevich
Position: Assistent
Office: Novosibirsk state university
Address: 630090, Russia, Novosibirsk
E-mail: accemt@gmail.com
Ryabko Boris Yakovlevich
Dr. , Professor
Position: Head of Laboratory
Office: Federal Research Center for Information and Computational Technologies, Novosibirsk State University
Address: 630090, Russia, Novosibirsk, Academician M.A. Lavrentiev avenue, 6
Phone Office: (383) 334-91-24
E-mail: boris@ryabko.net
SPIN-code: 5580-5794
Bibliography link:
Lysyak A.S., Ryabko B.Y. Forecasting methods for time series with large alphabets based on universal measure and decision trees // Computational technologies. 2014. V. 19. ¹ 2. P. 76-93
|
|
|
|
