Internetworking and Media Communications Research Laboratories
Department of Computer Science,
Discovering Hidden Cognitive Skill Dependencies between Knowledge Units using Markov Cognitive Knowledge State Network
Prof. Javed I. Khan
Date Submitted: January 2019
Cognitive psychology models using two mechanisms of the mental process: knowledge structure, and the process of using this knowledge. Knowledge structure refers to the interrelationships among knowledge-units in learning materials. In this dissertation, the interconnections among knowledge-units are represented as a well-known cognitive theory called Bloom’s taxonomy (BT).
This dissertation proposed a model termed Markov Cognitive Knowledge State Network (MCKSN) to infer the cognitive skill dependencies (CSD) among concepts in the knowledge-units. The proposed model contains some key ingredients, these key ingredients can be coupled together to tackle different angles of the model. The three key ingredients of the presented model are: mapping the semantic knowledge graph to Markov Cognitive Knowledge State Network (MCKSN), using human knowledge to describe the skill inference rules (SIR) among the cognitive skill dependencies via first order logic (FOL), and using the Probability Graphical Inference to infer cognitive skill dependencies.
An experiment was conducted on Introduction to Algorithms, a textbook used in Computer Science classes at many universities. To evaluate the MCKSN model human judgment, the most widely accepted form of judgment, was used. The results of the experiment verify that the MCKSN model is suitable to solve the problem and makes well behavior to discover the cognitive skill dependencies among concepts compared with a human result.
Last Modified: May 2020