Individualized recommending exercises and improving DINA algorithm by CLP-parameter adjustments

The traditional mode of education has been integrated with the technologies of internet to give rise to smart and or intelligent assistant education. A recommender “recommending items to customers” has been applied with great success in many commercial platforms, such as Amazon, Google, JD.com, and Tiktok, but there still has been no such an assisted learning system that effectively “recommending exercises (as the items) to learners (as the customers)”. This paper addresses the problems and challenges in applying recommender technology to the design of an intelligence-assisted learning system for individualized exercises recommendation. After providing an overview with some analysis on state-of-the-art researches, a model is proposed to improve DINA algorithm by four adjustments with changeable parameters of cognitive level perception (CLP), together with three primary steps of implementation in practice. It aims to improve recommendation accuracy of the mostly-concerned DINA algorithm based on widely-concerned cognitive diagnosis method (CDM). The proposed algorithm would have better performances on individualized exercises recommendation due to better CLP-capability. Another part of this study will provide further verification of the performances.