Markov models are the most commonly used approaches for this
type of web access prediction. Although the low order Markov models may provide higher coverage,
but they are complicated with ambiguous rules. In this paper, we introduced the use of default rule in
resolving ambiguous web access predictions in the Integrating Markov models and Fuzzy ART model
(IMFM). IMFM working pattern was of 2 parts. Firstly, IMFM divided web user sessions into groups
with similar web access patterns, and secondly, each of those groups was analyzed by Markov model.
This model was found to provide better predictions than individual Markov models. Input of the IMFM
was also examined to optimize model performance, and it was found that the sequence representation
by using the precedence-based sequence matrix provided better performance than using
transition-
based sequence matrix.
Markov models are the most commonly used approaches for this type of web access prediction. Although the low order Markov models may provide higher coverage, but they are complicated with ambiguous rules. In this paper, we introduced the use of default rule in resolving ambiguous web access predictions in the Integrating Markov models and Fuzzy ART model (IMFM). IMFM working pattern was of 2 parts. Firstly, IMFM divided web user sessions into groups with similar web access patterns, and secondly, each of those groups was analyzed by Markov model.This model was found to provide better predictions than individual Markov models. Input of the IMFM was also examined to optimize model performance, and it was found that the sequence representation by using the precedence-based sequence matrix provided better performance than usingtransition-based sequence matrix.
การแปล กรุณารอสักครู่..