Summary: We present a novel session identification method based
on statistical language modeling. Unlike standard time-
out methods, which use fixed time thresholds for ses-
sion identification, we use an information theoretic ap-
proach that yields more robust results for identifying
session boundaries. We evaluate our new approach by
learning interesting association rules from the seg-
mented session files. We then compare the performance
of our approach to three standard session identification
methods--the standard timeout method, the reference
length method, and the maximal forward reference
method--and find that our statistical language modeling
approach generally yields superior results. However, as
with every method, the performance of our technique
varies with changing parameter settings. Therefore, we
also analyze the influence of the two key factors in our
language-modeling­based approach: the choice of
smoothing technique and the language model order. We
find that all standard smoothing techniques, save one,
perform well, and that performance is robust to language

Source: An, Aijun - Department of Computer Science, York University (Toronto)
Huang, Jimmy - School of Information Technology, York University (Toronto)

Collections: Computer Technologies and Information Sciences