Visualization of navigation patterns on a Web site using model-based clustering

By I. V. Cadez, David E. Heckerman, Christopher Meek, Padhraic Smyth, S. White. Knowledge Discovery and Data Mining, 2000

369 citations recorded by the source corpus.

We present a new methodology for visualizing navigation patterns on a Web site. In our approach, we rst partition site users into clusters such that only users with similar navigation paths through the site are placed into the same cluster. Then, for each cluster, we display these paths for users within that cluster. The clustering approach we employ is model based (as opposed to distance based) and partitions users according to the order in which they request Web pages. In particular, we cluster users by learning a mixture of rst-order Markov models using the ExpectationMaximization algorithm. Our algorithm scales linearly with both number of users and number of clusters, and our implementation easily handles millions of users and thousands of clusters in memory. In the paper, we describe the details of our technology and a tool based on it called WebCANVAS. We illustrate the use of our technology on user-traAEc data from msnbc.com.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/d38d30cf1fdfefb8277eea2802437483889ce38f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Signal Processing

Citations in the corpus

Introduction to the special topic section on mining Web resources for enhancing information retrieval (2007, methodology): “This evaluation method presents a similar idea to those proposed by Cadez et al. (2000) and Anderson Domingos, and Weld (2002).”
Outlier Analysis (2013, methodology): “Hidden Markov Models have also been used for point anomaly detection of small sequences [89, 413] by creating a mixture of hidden markov models in order to create the underlying clusters.”
From Visual Data Exploration to Visual Data Mining: A Survey (2003, background): “[16] describe an approach for exploratory analysis and visualization of the dynamic behavior of visitors of a particular Web site.”
Building a Recommender Agent for e-Learning Systems (2002, methodology): “There are specific tools for visualizing web data such as interconnections of web resources [13], tools for visualizing web navigation patterns [4], and tools to visualize the changes in web sites or web usage over time [6].”
Web Usage Mining as a Tool for Personalization: A Survey (2003, methodology): “However, there are cases where it has a very slow linear convergence and may therefore become computationally expensive, although in the results in Cadez et al. (2000), it is shown empirically that the algorithm scales linearly in all aspects of the problem.”
Anomaly Detection for Discrete Sequences: A Survey (2012, methodology): “Other probabilistic techniques such as mixture of HMMs [29], [30] or mixture of Maximum Entropy (maxent) models [23] can also be employed in the same manner for clustering-based anomaly detection.”
Anomaly detection: A survey (2009, methodology): “Similarly, Smyth [1997] and Cadez et al. [2000] use HMMs to cluster the set of sequences and detect any sequences which do not belong to any cluster as anomalies.”
Model-Based Clustering, Discriminant Analysis, and Density Estimation (2002, background): “Despite these apparent obstacles, model-based clustering seems to be emerging as an important component within schemes for the classi cation of large datasets (Meila 1999; Smyth 2000; Cadez et al. 2000; Posse 2001).”
Business process analysis in healthcare environments: A methodology based on process mining (2012, methodology): “Running the Sequence Clustering Algorithm The first step is to run the sequence clustering algorithm as described in [43, 37] in order to discover the behavioral patterns contained in the event log.”
Monitoring web browsing behavior with differential privacy (2014, background): “As web browsing sequences exhibit strong navigation patterns between adjacent web page requests, it has been shown that user navigation patterns can be well captured by ﬁrst-order Markov chain [4].”