Intelligent decision system with predictive model in smart environments

Abstract

In this Ph.D. thesis a novel approach of digital education and hybrid education that presents effective individualized learning, adapting all knowledge domains to the learner’s needs and path through a predictive model supported by a decision binary tree, is presented. The thesis is based on papers of other researchers that were published in international journals, as well as on the author's papers, that were published in valid journals with a plethora of citations. The presented approach models the learning path in m-learning environments or how the student’s knowledge can be increased. In particular, it performs user modeling by dynamically identifying and updating the student’s knowledge level for all the concepts of the domain knowledge. Its operation is based on AI and a predictive model of learners' assessment(Fig. 1).

The presented novel approach was fully implemented and evaluated. Notably, an original system for digitalized m-learning was developed, based on the programming language JAVA, and the resulting app is called D-Quest. The specific knowledge domain that was chosen, was Mathematics since it has a connection with computer science. Supplementally, informatics and Mathematics have the best knowledge background and embedded implementation combination. Therefore, it is suitable for the implementation and evaluation of the thesis issue. D-Quest incorporates the presented student assessment modeling approach. Thereby, the predictive model of the decision tree visualizes dynamically the path for a new domain concept that is completely unknown to the learner. Furthermore, it recognizes when a previously known domain concept has been completely or partly forgotten by the learner.

The application of this approach is not limited to m-learning environments, but it can also be used in other systems with changeable user states, such as e-shops and smart cities, where consumers’ preferences change over time and affect one another. Consequently, the approach constitutes a novel predictive model for the evaluation of big data sets, which offers a dynamic adaptation to users’ needs and preferences for intelligent systems.