An Attribute-Value Pair (AVP) Model for Creating and Exploiting Annotations in Economic Intelligence.

Okunoye, O.B (2014)

A Thesis Submitted to the School of Postgraduate Studies, University of Lagos.

Thesis

This study is concerned with the design and development of an annotation model that gives users the power of expressivity as required in Economic Intelligence (EI) context. Most of the existing annotation platforms lack this power of expressivity. The study also discovered a need to develop an annotation model for creating and structuring annotations that will adequately capture the intention of the users (EI actors) in decision making process. In addition, there was a need for a technique that will allow actors to search for information based on the objective of the search. This study introduced an annotation model called Attribute-Value Pair (AVP) for creating and storing annotations. The study developed a mechanism for exploiting stored annotations based on the context of problem, and used the AVP model to develop a search algorithm that allowed actors to search for information based on the objective of the search. The study used Resource Description Framework (RDF) for the formalism of AVP annotation model. The exploitation phase was implemented using Explore, Query, Analyze and Annotate(EQuA2Te) architecture. A pattern-based algorithm called AVP search was developed that allowed actors to search for information based on the semantic of search objective. The study developed a prototype called Annotation Model and Tools for Economic Intelligence Actors (AMTEA) that used the AVP annotation model for creating, storing and exploiting annotations as well as performing search operations. Two search problems were used as scenarios to evaluate AMTEA system. Results obtained were compared with annotations made on the same set of documents by human agents. The performance evaluation shows that the new AMTEA system detected over 98% of manually made annotations by the human agents. In addition, AMTEA system was able to find new annotations that appear to be relevant. In essence, AMTEA assists human agents to discover new information that might be relevant in decision making process.

Collections: