QUERY EXPANSION WITH ENRICHED USER PROFILES FOR PERSONALIZED SEARCH UTILIZING FOLKSONOMY DATA
ABSTRACT
Query expansion has been widely adopted in Web search as a way of tackling the ambiguity of queries.Personalized search utilizing folksonomy data has demonstrated an extreme vocabulary mismatch problem that requires evenmore effective query expansion methods. Co-occurrence statistics, tag-tag relationships and semantic matching approaches area those favored by previous research. However, user profiles which only contain a user’s past annotation information may not be enough to support the selection of expansion terms, especially for users with limited previous activity with the system.We propose a novel model to construct enriched user profiles with the help of an external corpus for personalized queryexpansion. Our model integrates the current state-of-the-art text representation learning framework, known as wordembeddings, with topic models in two groups of pseudo-aligned documents. Based on user profiles, we build two novel queryexpansion techniques. These two techniques are based on topical weights-enhanced word embeddings, and the topicalrelevance between the query and the terms inside a user profile respectively. The results of an in-depth experimentalevaluation, performed on two real-world datasets using different external corpora, show that our approach outperformstraditional techniques, including existing non-personalized and personalized query expansion methods.
PROPOSED SYSTEM:
We adopt a different approach to personalizedQEutilizingfolksonomydata.Inourapproach,theexpansionprocess is based on an enriched user profile, 1041-4347 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for moreinformation. which contains tags and annotations together with documentsretrievedfromanexternalcorpus.Thiscorpuscanbeviewed as a knowledge base to enhance the informationstoredintheuserprofile.Thewholeprocedureofqueryadaptation is hidden to the user. It happens in animplicit way based on their choices of tags and the termsused on annotated web pages. We first propose a novelmodel to build the enriched user profiles. Our model integratesthe current state-of-the-art text representationlearning framework, known as word embeddings, withtopic models in two groups of pseudo-aligned documentsbetween user annotations and documents from the externalcorpus. We then present two novel QE techniques.The first technique approaches the problem by using topicalweights-enhanced WEs to select the best possible expansionterms.Thesecondmethodis based on the topicslearned. It calculates the topical relevance between thequery and the terms inside a user profile. This paper describes an in-depth experimental evaluationusing two different real-world folksonomy datasets,extracted from del.ico.us and Bibsonomy. We also exploretwo different external corpora for user profile enrichment.A comparative analysis of our findings with those obtainedby using well-known and state-of-the-art techniquessuch as those exploiting co-occurrence statistics,tag-tag relationships and semantic relatedness for personalizedQE, shows that our approach is able to achievesignificantly better retrieval results. 2The contribution of this paper can be summarized asfollows: i. We tackle the challenge of personalized QE utilizing folksonomydata in a novel way by integrating latent and deep semantics.ii.We propose a novel model that integrates word embeddingswithtopicmodelstoconstructenricheduserprofileswiththehelpofanexternalcorpus.iii.We suggest two novel personalized QE techniques basedon topical weights-enhanced word embeddings, and the topicalrelevance between the query and the terms inside a user profile.The techniques demonstrate significantly better results thanPreviously
EXISTING SYSTEM:
RELATED WORK Web users may not always be successful in using a representativevocabulary when locating objects in a system.Therefore, query expansion attempts to expand the termsof the user’s query with other terms, with the aim of retrievingmore relevant results. QE has a long standing
CONCLUSION
In this paper we study personalized search through enhanceduser profiles and personalized query expansionutilizing folksonomy data. We propose a novel model tobuild enriched user profiles. Our model integrates thecurrent state-of-the-art text representation learningframework, known as word embeddings, with topicmodels in two groups of pseudo-aligned documents betweenuser annotations and documents from the external corpus. Based on these enhanced user profiles, we thenpresent two novel QE techniques. The first technique approachesthe problem by using topical weights-enhancedword embeddings to select the best possible expansionterms. The second technique calculates the topical relevancebetweenthequeryandthetermsinsideauserprofile. The proposed models performed well on two real-world social tagging datasets produced by folksonomyapplications, delivering statistically significant improvementsover non-personalized and personalized representativebaselinesystems.Wealsoshowthatourmethodworkswell for users with small, moderate and richamounts of historical usage information. In future research,we aim to investigate incorporating more informationintothelatentsemanticmodelinordertocapturemoreaccurateuserprofiles.Futureworkwillalsoincludetheevaluation of different similarity models and weightingschemestobeusedinourmodels.
REFERENCES
[1] M. R. Bouadjenek, H. Hacid, and M. Bouzeghoub, Social networks and information retrieval, how are they converging?A survey, a taxonomy and an analysis of social informationretrieval approaches and platforms, Information Systems,2016, 56: 1-18.
[2] S. Xu, S. Bao, B. Fei, Z. Su, and Y. Yu, Exploringfolksonomy for personalized search, in Proceedings of the31st annual international ACM SIGIR conference onResearch and development in information retrieval, Singapore, Singapore, 155-162, 2008.
[3] G. Smith, Folksonomy: social classification, Blog article, August, 2004.
[4] M. Bertier, R. Guerraoui, V. Leroy, and A.-M. Kermarrec, Toward personalized query expansion, in Proceedings of theSecond ACM EuroSys Workshop on Social NetworkSystems, Nuremberg, Germany, 7-12, 2009.
[5] C. Biancalana, F. Gasparetti, A. Micarelli, and G. Sansonetti,Social semantic query expansion, ACM Trans. Intell. Syst.Technol., 2013, 4(4): 1-43.
[6] D. Zhou, S. Lawless, and V. Wade, Improving search via personalized query expansion using social media,Information Retrieval, 2012, 15(3-4): 218-242.
[7] C. Biancalana and A. Micarelli, Social Tagging in QueryExpansion: A New Way for Personalized Web Search, inProceedings of the 2009 International Conference onComputational Science and Engineering – Volume 04, 10601065,2009.
[8] M. Bender, T. Crecelius, M. Kacimi, S. Michel, T. Neumann