dc.contributor.author | Momin, Afiz | |
dc.date.accessioned | 2016-12-19T19:25:26Z | |
dc.date.available | 2016-12-19T19:25:26Z | |
dc.date.issued | 2016-12-19T19:25:26Z | |
dc.identifier.uri | http://hdl.handle.net/10222/72603 | |
dc.description.abstract | We define expertise modeling as profiling an expert, a knowledgeable person in one or more domains, based on evidence from research articles into one or more research topics. The traditional text classification approach involves classifying a document into a class where classification hierarchy is limited to one level. However, the real-world problems are more complex and could be related to hierarchical structure and therefore, there has been numerous research in a hierarchical classification. Millions of enthusiastic researchers contribute in the form of research articles in conferences or journal publications and apply for research grants, and the task of assigning reviewers to research articles and correct research topic for the grant application is non-trivial.
For our research, we have trained a hierarchical classifier on titles and abstracts of research articles and it predicts one or more research topics for a given article of an expert. We have used traditional Bag-of-Words (BOW) representations of the text which is enriched using a semantic knowledge from Wikipedia's concepts (BOC) and categories (BOK). For each of these document representations, a hierarchical classifier is trained and their outputs are combined using consensus methods to predict a research topic. In reality, research articles can belong to multiple research topics and therefore two approaches to multi-label a research article are proposed.
We evaluate and compare the performance of the hierarchical model with a baseline, a flat classifier, and using different training set and different evaluation measures such as precision, recall, and f-measure. The combined outputs from hierarchical classifiers, BOW, BOC, and BOK, are compared with a flat classifier and a hierarchical classifier based on BOW. The results from various approaches, comparison of the performance of different hierarchical classifiers and current issues are also discussed. | en_US |
dc.language.iso | en | en_US |
dc.subject | hierarchical classification | en_US |
dc.subject | text analysis | en_US |
dc.subject | machine learning | en_US |
dc.title | Towards Expertise Modeling Using Hierarchical Classification and Wikipedia Knowledge | en_US |
dc.date.defence | 2016-12-12 | |
dc.contributor.department | Faculty of Computer Science | en_US |
dc.contributor.degree | Master of Computer Science | en_US |
dc.contributor.external-examiner | n/a | en_US |
dc.contributor.graduate-coordinator | Malcolm Heywood | en_US |
dc.contributor.thesis-reader | Abidalrahman Moh'd | en_US |
dc.contributor.thesis-reader | Qigang Gao | en_US |
dc.contributor.thesis-supervisor | Evangelos Milios | en_US |
dc.contributor.ethics-approval | Not Applicable | en_US |
dc.contributor.manuscripts | Not Applicable | en_US |
dc.contributor.copyright-release | Not Applicable | en_US |