Deep Search of Experts Profile

Introduction

"Who is good at at Hidden Markov Model?". "I have a question about Naive Bayes". Deep experts search enables look up of information about experts at academic institute, especially about their research and teaching expertise.

Project I: Expert Search System

Expert Search is a web-based search engine system that is designed to search experts in institute like schools, faculties, research centres etc. based on their publications. The initial system searches the USM School of Computer Sciences academic staffs.

Phase I of expert search has three basic modules, The Claw, The Core and The Clue. The Claw is a tool that performs data extraction from data source, e.g. DBLP, which contains information of computer sciences bibliography. The Core is a tool that performs data indexing, text processing, classification algorithms etc. The Clue presents the search system with good user experience features like queries, visualizations, profiling etc.

In Phase II & III, we focus on extending the publications scope to alternate sources and formats like Google Scholar. The search system also explores on additional user experience feature, i.e. faceted search, infographics etc. to improve search results retrieval.

Expert Search I: The Claw, The Core and The Clue.
Expert Search II: The Source, The Search.
Expert Search III: The Classifier, Infographics.
Expert Search IV: The Transformer, The Insights and The Classifier New

Scope

People

Deliverables

System

Beta Version of Expert Search and Analytics IV released (1 Oct 2019) New.



Beta Version of Expert Search System II (with Infographics) released (9 Jan 2017).



Beta Version of Expert Search System II (with Faceted Search) released (27 May 2016).



Beta Version of Expert Search System I released (19 June 2015).

Presentation/Demo

Khoh Zhuo Yan, Goh Kau Yang and Chua San Thai: Enhancing Expert Bibliographical Profile Presentation via Infographics . CAT300/301 Undergraduate Project Demo, 19 Dec 2016, Sains@USM Penang, Malaysia.

Teh Chek Wei and Gan Keng Hoon: Faceted Search for Finding Expertise Bibliographies. AI3 & SOI Asia Joint Meeting Spring 2016, USM Penang, Malaysia.

Documentations

FYP System Analysis Slide - The Claw

FYP System Analysis Slide - The Core

FYP System Analysis Slide - The Clue

FYP Final Report - The Claw*

FYP Final Report - The Core*

FYP Final Report - The Clue*

FYP Final Report - The Search*

FYP Final Demo Slide - The Search

FYP Final Report - The Classifier*

FYP Final Demo Slide - The Classifier

* This file is password protected and reserved for internal viewing only. Please send a request to my email if you wish to use the file.

Project II: Expert Classification

This project explores techniques to map experts to their expertises using domain oriented classication systems.

Scope

People

Deliverables

Publications

Gan Keng Hoon, Gan Kian Min, Oscar Wong, Ooi Bong Pin, Chan Ying Sheng: Classifly: Classification of Experts by Their Expertise on the Fly. Web Intelligence 2015, 6-9 December, 2015, Singapore.

Project III: Faceted Search for Expert's Bibliographical Information

This project on faceted search is to explore on concept extraction technique to produce faceted keywords. The generated facets are used to optimize retrieval and navigation of experts. The uniqueness of this project is that the facets with values in the specific bibliographical profile which are extracted dynamically from the text. The facets we are looking is domain oriented and can be used in addition to current fixed facets like year, location, venue etc. Natural Language Processing analysis is used to extract facet's values from the unstructured text.

People

Deliverables

Publications

Gan Keng Hoon, Teh Chek Wei: Flexible Facets Generation for Faceted Search. The First EAI International Conference on Computer Science and Engineering, COMPSE 2016, 11-12 November, 2016, Penang, Malaysia.

Lim Heng Kuan, Gan Keng Hoon: Bibliographical-based Facets for Expertise Search. ACIS 2015, 15-17 October, 2015, Penang, Malaysia. slideshare