A clickstream-based collaborative filtering personalization model

Abstract

In recent years, clickstream-based Web personalization models for collaborative filtering recommendation have received much attention mainly due to their scalability [10,16,19]. The common personalization models are the Markov model, (sequential) association rule, and clustering. These models have shown strengths and weaknesses in their performance: for instance, the Markov model has higher precision and lower recall than (sequential) association rule and clustering, and vice versa [22]. In order to address the trade-off relationship of precision and recall, some study has combined two or more different models [22] or applied multi-order models [24,27]. The performance increases by these models, however, are at best marginal and still there is room for improving the performance because of their first order (one model type) application in making recommendation. We propose a new hybrid model for improving the performance, especially recall. The proposed hybrid model applies four prediction models -the Markov model, sequential association rule, association rule, and a default model [1,17] -in tandem in their precision order. We evaluated our model with Web usage data, and the result is promising.

Web Server Performance Optimization using Prediction Prefetching Engine

“…If association rules cannot cover the state, clustering algorithm is applied. Kim et al (2004) [19] work improved recall and it did not improve the Web page prediction accuracy.…”

“…Cadez et al [17] integrated clustering and first order Markov model to increase accuracy. Kim et al [18] combined association rules and Markov model for web access prediction. In this paper, Integrated data mining techniques are applied for extracting user access patterns from web access logs.…”

Flexible Query Answering Systems 2015

“…Montgomery et al (2004) [18] used a dynamic multinomial probit model to extract information from consumers' navigation path, which is helpful in predicting their future movements. Kim et al (2004) [11] used the clickstream data as implicit feedback to design a hybrid recommendation model, which resulted in better performance. Moe (2006) [17] proposed an empirical two-stage choice model based on clickstream data to capture observed choices for two stages: products viewed and products purchased.…”

“…The size of recommendation set K ranged from 5 to 15, that is, K ∈ {5, 6,7,8,9,10,11,12,13,14 where S is the collection of records in test set and AP (s) @K can be computed by:…”

Performance tuning and cost discovery of mobile web-based applications

“…Furthermore, there are many approaches for web log analysis aimed at classifying user paths (e.g., Spiliopoulou, 2000;Berkhin et al, 2001;Kim et al, 2004;Heer and Chi, 2002;Chi et al, 2000;Gillenson et al, 2000). Especially, the identification of long sequences described by Pitkow and Pirolli (1999) seems to be an important topic for the PETTICOAT concept.…”