An E-Commerce Customer Service Robot Based on Intention Recognition Model

An E-Commerce Customer Service Robot Based on Intention Recognition Model

Minjing Peng (School of Economics and Management, Wuyi University, Guangdong, China), Yanwei Qin (Research Center of E-Commerce Augmented Reality of Guangdong Province, Wuyi University, Guangdong, China), Chenxin Tang (School of Economics and Management, Wuyi University, Guangdong, China) and Xiangming Deng (Research Center of E-Commerce Augmented Reality of Guangdong Province, Wuyi University, Guangdong, China)
Copyright: © 2016 |Pages: 11
DOI: 10.4018/JECO.2016010104
OnDemand PDF Download:
No Current Special Offers


There are three defects for providing human-labor customer services in e-commerce operations: high costs of human labors, staff turnover, and lack of service quality assurance. Breakthroughs made in artificial intelligence, natural language processing and related fields make it possible to replace human labors with online artificial intelligent robots to provide the e-commerce customer service, which indicates the online robots are the future of e-commerce customer services. However, most of the current robots are designed to reply with knowledge matching the key words in question sentences from the database, rarely involving in research on customer intentions that are key factors influencing user experience and online sales. In this research, an intention recognizing model was proposed to obtain intentions of e-commerce consumers by computing the strengths of candidate intention nodes in the intention graph, which was used to describe relations between different goods that could be the intentional targets of e-commerce consumers. The proposed robot was constructed based on the intention recognizing model to identify intentions of consumers and use the located knowledge combined with the AIML based sentence composition template to produce the response sentences for consumers. At last, the proposed robot was evaluated using F3 and ROUGE metrics by comparing with a keyword matching robot. And the evaluation results proved the validity of the proposed robot.
Article Preview

System Framework

As shown in Figure 1, the system consists of three layers: the interaction layer, the knowledge layer and the intention layer. Each layer plays its own role by executing functions of the modules.

Figure 1.

Three layers and the modules of the system framework


Interaction Layer

The interaction layer is responsible for interacting with consumers: getting questions from consumers and producing answers to consumers. There are four functional modules in the layer:

  • 1.

    AIML Files: AIML is an artificial intelligence markup language extended by XML. In this module, knowledge and question-answering patterns are stored using different tags. Some important tags are <template>, <pattern>, <topic>, <that>, <srai>, <set>, <get>, <random>, <star>, <condition> and <think>.

  • 2.

    Chinese Segmentation Module: Three parts in Figure 1 cooperate with this module. AIML memory tree and AIML reference engine are organized by the words which are the segmented results of user questions by the module under the support of AIML files.

  • 3.

    Sentence Composition Module: The AIML response engine is under supports of this module and the module of AIML files to composite answers according to the template logic.

  • 4.

    AIML Response Engine: This module receives intentions from the module of intention recognition and knowledge from the module of AIML memory tree, and then composites answers to the questions of e-commerce consumers under the support of the sentence composition module.

Complete Article List

Search this Journal:
Open Access Articles
Volume 19: 4 Issues (2021): 2 Released, 2 Forthcoming
Volume 18: 4 Issues (2020)
Volume 17: 4 Issues (2019)
Volume 16: 4 Issues (2018)
Volume 15: 4 Issues (2017)
Volume 14: 4 Issues (2016)
Volume 13: 4 Issues (2015)
Volume 12: 4 Issues (2014)
Volume 11: 4 Issues (2013)
Volume 10: 4 Issues (2012)
Volume 9: 4 Issues (2011)
Volume 8: 4 Issues (2010)
Volume 7: 4 Issues (2009)
Volume 6: 4 Issues (2008)
Volume 5: 4 Issues (2007)
Volume 4: 4 Issues (2006)
Volume 3: 4 Issues (2005)
Volume 2: 4 Issues (2004)
Volume 1: 4 Issues (2003)
View Complete Journal Contents Listing