A Conversational Recommender System for Exploring Pedagogical Design Patterns.

This work presents the preliminary results of developing a Conversational Recommender System (CRS) to recommend Pedagogical Design Patterns (PDPs) to educators. In this CRS, the user queries the system in the form of natural language. The dialogue manager unit, which is the core of this system, gets the user input query and extracts the most semantically relevant patterns from the knowledge-base by Natural Language Processing (NLP) and Machine Learning (ML) algorithms. Our findings on evaluation of this system show the recommended patterns are highly relevant and semantically similar to the user queries. This novel approach to the application of pedagogical design patterns can greatly benefit the educational community by helping them identify the best practices in the field without having to search through all the published repositories of patterns. Experts can contribute to the knowledge-base of this system by sharing their best practices with the community.