Date of Award


Document Type

Thesis campus only


Computer Science

First Advisor

Sheng Tan

Second Advisor

Mark Lewis


Nowadays, the visible success of online platforms certainly put a spotlight on the importance of recommendation systems (RS) in multiple domains. Typically, the use of RS has been proved to lead to considerable improvements for e-commerce business by bringing in various types of positive effects. However, recommendation systems are also known to be controversial because of the concerns they introduce: lack of transparency, reduction of diversity, little to no user control, etc. On the other hand, with the significant progress in Natural Language Processing and gradual acceptance of Artificial Intelligence by end-users, one cannot ignore the fact that conversational systems, especially, virtual personal assistants, are drawing more attention in many industries.

In order to alleviate the issues introduced by RS in return for a better user-friendly experience, many researchers are seeking to combine recommendation systems with conversational systems through different means. Aiming towards that same goal, we designed and implemented a simple chatbot, Sophia, for demonstrating the potential of chatbots in improving the user experience for e-commerce platforms in terms of user control. In particular, this work, serving as the groundwork for a series of the proposed research, will focus on the design, current progress, and future plan of both the chatbot and an associated e-commerce website. These two components, along with a simple product recommendation system, were built and integrated altogether into one project: E-Commerce with Sophia (EWS). Aside from achieving user control for e-commerce’s RS via a creative conversation-based approach, unexpectedly, we discovered that EWS might have the prospect of becoming a general solution to implementing, presenting, and comparing different user control approaches in RS for e-commerce.