Determining WWW user's next access and its application to pre-fetching

Abstract

World-Wide Web (WWW) services have grown to levels where signi cant delays are expected to happen. Techniques like pre-fetching are likely to help users to personalize their needs, reducing their waiting times. However, prefetching is only e ective if the right documents are identi ed and if user's move is correctly predicted. Otherwise, pre-fetching will only waste bandwidth. Therefore, it is productive to determine whether a revisit will occur or not, before starting pre-fetching.In this paper we develop two user models that help determining user's next move. One model uses Random Walk approximation and the other is based on Digital Signal Processing techniques. We also give hints on how to use such models with a simple pre-fetching technique that we are developing.

A fast method for discovering critical edge sequences in e-commerce catalogs

“…The problem of finding the MPESs, the most difficult part of the CES discovery process, maps nicely to sequence mining discovery [5,11,8,22,24] which is a standard approach for finding sequential patterns in a dataset. Web usage mining (e.g., [8,22,26,18,23]) is a sub-area within the area of sequence mining particularly applicable to the MPES-discovery problem.…”

About the Heterogeneity of Web Prefetching Performance Key Metrics

“…Precision, defined in this way, just evaluates the algorithm without considering physical system restrictions; e.g., cache, network or time restrictions; therefore, it can be seen as a theoretical index. Other research studies refer to this index as Accuracy [4], [12], [13], [14], [15], [16], [17] while some others use a probabilistic notation; e.g., some Markov chains based models like [18] Pr(hit|match).…”

“…Precision [8], [9], [10], [11] Accuracy [4], [12], [13], [14], [15], [16], [17] Precision Pr(hit|match) [18] Recall [9], [10], [11] Usefulness [16], [17] Hit Ratio [19] Recall Predictability [6] 1. Prediction Applicability Applicability [8] Traffic increase [3], [4], [19] Wasted Bandwidth [2] Bandwidth ratio [21] Extra bytes [10] Data transfer [22] Traffic increase Latency [5], [23] Access time [3] 3.…”

On the Existence and Significance of Data Preprocessing Biases in Web-Usage Mining

“…Cunha and Jaccoud (1997) studied the problem of determining a user's next page accessed in a session for the purpose of determining how to pre-fetch pages and optimize a website's performance. The problem was modeled as a Markov process in which the next page accessed depends on the most recent sliding window.…”

“…Sliding window of fixed length w (Cunha and Jaccoud 1997. A single session of length n is broken down into n − w +1 sliding windows of length w. Hence, the implicit unit of analysis is a group of w consecutive clicks within a session.…”