An Efficient Algorithm for Extracting High Utility Itemsets from Weblog Data

Abstract

High utility itemset refers to those set of items which has high utility such as profit in a database. High utility of itemset plays a crucial role in real life. In recent years, various algorithms have been proposed for finding high utility itemset but unfortunately they are not completely relevant at the time and space point of view. In the data mining field, high utility itemset can be found in different categories of data like time series, categorical, etc. Log data is useful for finding behaviour of the user in different aspects. In this paper, we have proposed an algorithm named HUIM (High Utility Itemsets Mining) and construct HUI-FP (High Utility Itemsets-Frequent Pattern) Tree for efficiently mining high utility itemsets from log database. The behaviour of the user can be predicted through the high utility of every visited page. We have also proposed pattern generation technique based on cosine similarities among itemsets. These techniques generate strong patterns, and customized users profile according to that pattern. The proposed algorithm is better than the previous state of the art algorithm for high utility itemset generation.

Jaccard Similarity based Mining for High Utility Webpage Sets from Weblog Database

“…It is pretty clear from the figure that the minimum threshold value decreases the runtime is increased but more number of web transactions is processed. Runtime decreases as the MTUV increases [26]. Here, HUWSM algorithm shows significant improvement over other algorithms.…”

“…High utility webpage set means, determining webpage set which satisfies the conditions of U(X) ≥MTUV [26].…”

“…This requires proper trade off among high utility item sets, memory and running time. [26] Proposed HUIM, an efficient algorithm for extracting the high utility item sets from weblog data sets. They applied the cosine similarity method and high utility item sets tree structure to find similar item sets from the web log data.…”

“…Further, the step-V builds the high utility webpage frequent pattern tree from the favourable webpages obtained in the previous step IV. To construct the HUW-FP [25] Tree, firstly, all the webpages are arranged in the descending order of their TWU values and the webpages of the lowest TWU [26] value is excluded from the computation work. The tree is helpful in finding the information about the webpages and their utilities.…”

High Fuzzy Utility Strategy Based Webpage Sets Mining from Weblog Database

“…Other algorithms [22 -24] of fuzzy data mining are still developed in progress. B. Bakariya et al [25] proposed HUIM, an efficient algorithm for extracting the high utility item sets from weblog data set. They applied the cosine similarity method and high utility item sets tree to find similar item sets from the web log data.…”