CLASSIFICATION OF QUERY RECOMMENDATION USING QUERY SEMANTIC FLOW GRAPH TECHNIQUE ON NOVEL AOL LOG METHOD
Abstract
vioft2nntf2t|tblJournal|Abstract_paper|0xf4ff5dd62b0000000f22060001000500
Web mining is one among the thrust area of research in the data mining domain. The classification of query recommendation can be divided into two major classes that are document-based approach and log-based approach. Log-based method can get relatively good query recommendation and find query inner relation. Query flow graph is one of log-based method and get relatively good recommendation. However, query flow graph cannot get query semantic information. And there are many isolated nodes because of data sparseness. Therefore, word2vec is used to define the query semantic and add query semantic to query flow graph. That can be able to modify query transfer probability which is calculated by query flow graph. At the same time, we can get connection between the isolated queries that are related but no connection due to the data sparseness and the inaccuracy of session split. Empirical tests are conducted in accordance with the AOL log. From the results the efficiency of the approach in suggesting queries and F1 value is about 20% higher than the traditional query flow graph.

Authors
R Balakrishnan
Rathnavel Subramaniam College of Arts and Science, India

Keywords
Query Recommendation, Word2vec, Semantics, Query Flow
Yearly Full Views
JanuaryFebruaryMarchAprilMayJuneJulyAugustSeptemberOctoberNovemberDecember
000000000000
Published By :
ICTACT
Published In :
ICTACT Journal on Soft Computing
( Volume: 11 , Issue: 3 , Pages: 2378-2383 )
Date of Publication :
April 2021
Page Views :
180
Full Text Views :
1

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.