Author:
Abdulrahman Al-Haimi
Affiliation:
University of Waterloo, Canada
Keyword(s):
Text Mining, Matching, Knowledge Users, Knowledge Creators, Clustering, Classification, Management, Experimentation, Exploration, Performance.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Biomedical Engineering
;
Business Analytics
;
Data Analytics
;
Data Engineering
;
Data Management and Quality
;
Data Mining
;
Databases and Information Systems Integration
;
Datamining
;
Enterprise Information Systems
;
Health Information Systems
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Predictive Modeling
;
Semi-Structured and Unstructured Data
;
Sensor Networks
;
Signal Processing
;
Soft Computing
;
Symbolic Systems
;
Text Analytics
;
Web Analytics
Abstract:
Matching knowledge users with knowledge creators from multiple data sources that share very little similarity in content and data structure is a key problem. Solving this problem is expected to noticeably improve research commercialization rate. In this paper, we discuss and evaluate the effectiveness of a comprehensive methodology that automates classic text mining techniques to match knowledge users with knowledge creators. We also present a prototype application that is considered one of the first attempts to match knowledge users with knowledge creators by analyzing records from Linkedin.com and BASE-search.net. The matching procedure is performed using supervised and unsupervised models. Surprisingly, experimental results show that K-NN classifier shows a slight performance improvement compared to its competition when evaluated in a similar context. After identifying the best-suited methodology, system architecture is designed. One of the main contributions of this research is t
he introduction and analysis of a novel prototype application that attempts to bridge the gap between research performed in industry and academia.
(More)