Faculty Publications

Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736

Publications by NITK Faculty

Browse

Filters

2011 - 20132

Settings

Author: search.filters.author.Poornalatha, G.Subject: search.filters.subject.Web usage mining

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Alignment based similarity distance measure for better web sessions clustering
    (Elsevier B.V., 2011) Poornalatha, G.; Raghavendra, P.S.
    The evolution of the internet along with the popularity of the web has attracted a great attention among the researchers to web usage mining. Given that, there is an exponential growth in terms of amount of data available in the web that may not give the required information immediately; web usage mining extracts the useful information from the huge amount of data available in the web logs that contain information regarding web pages accessed. Due to this huge amount of data, it is better to handle small group of data at a time, instead of dealing with entire data together. In order to cluster the data, similarity measure is essential to obtain the distance between any two user sessions. The objective of this paper is to propose a technique, to measure the similarity between any two user sessions based on sequence alignment technique that uses the dynamic programming method. © 2011 Published by Elsevier Ltd.
  • No Thumbnail Available
    Item
    Web sessions clustering using hybrid sequence alignment measure (HSAM)
    (Springer-Verlag Wien michaela.bolli@springer.at, 2013) Poornalatha, G.; Raghavendra, S.R.
    Web usage mining inspects the navigation patterns in web access logs and extracts previously unknown and useful information. This may lead to strategies for various web-oriented applications like web site restructure, recommender system, web page prediction and so on. The current work demonstrates clustering of user sessions of uneven lengths to discover the access patterns by proposing a distance method to group user sessions. The proposed hybrid distance measure uses the access path information to find the distance between any two sessions without altering the order in which web pages are visited. R2 is used to make a decision regarding the number of clusters to be constructed. Jaccard Index and Davies–Bouldin validity index are employed to assess the clustering done. The results obtained by these two standard statistic measures are encouraging and illustrate the goodness of the clusters created. © 2012, Springer-Verlag.