Keynote Speakers


...

Prof. DR. Tsukasa Hirashima

Hiroshima University

Tentative Title:

EXTERNALIZATION OF THINKING TASK AND LEARNING ANALYTICS WITH PROCESS EVIDENCE

Abstract:

To Be Announced (TBA)


...

Prof. DR. Abdurazzag Ali Aburas

University of KwaZulu-Natal

Tentative Title:

BIG DATA: REDUCTION NOT COMPRESSION

Abstract:

To Be Announced (TBA)


...

Prof. DR. Anton Satria Prabuwono

King Abdulaziz University

Tentative Title:

To Be Announced (TBA)

Abstract:

To Be Announced (TBA)


...

Prof. DR. Halimah Badioze Zaman

Universiti Kebangsaan Malaysia

Tentative Title:

To Be Announced (TBA)

Abstract:

To Be Announced (TBA)


...

Prof. Ir. Dwi Hendratmo W., M.Sc., Ph.D.

Institut Teknologi Bandung

Tentative Title:

CONVERSATIONAL RECOMMENDER SYSTEM BASED ON FUNCTIONAL REQUIREMENTS

Abstract:

Specifying requirements based on technical features of the product is often difficult for most customers in purchasing a multi-function and multi-feature products. A more natural way to identify customers’ needs is by asking what they really want to use with the product they are looking for. To address this issue, we propose a conversational recommender system (CRS) that is based on user functional requirement. The proposed system covers ontology structure and interaction mechanism that can explore user’s preference from their functional requirements. The ontology is composed of three different interacting hierarchical classes: functionality requirement class, component class and product class. Interaction in this system is performed through question-answering dialogue, product recommendation and explanation, similar to the interaction between customer and professional sales support. Based on our user studies, most of expert users (familiar with product technical features) and novice users (not familiar with product technical features) perceive that the CRS based on our proposed functional requirement and knowledge design is more useful than that of recommender system based on technical features. The evaluation results also show that our proposed model is able to narrow down the average result's size significantly within short interactions.