Computational Estimate Visualisation and Evaluation of Agent Classified Rules Learning System

📝 Abstract

Student modelling and agent classified rules learning as applied in the development of the intelligent Preassessment System has been presented in [10],[11]. In this paper, we now demystify the theory behind the development of the pre-assessment system followed by some computational experimentation and graph visualisation of the agent classified rules learning algorithm in the estimation and prediction of classified rules. In addition, we present some preliminary results of the pre-assessment system evaluation. From the results, it is gathered that the system has performed according to its design specification.

💡 Analysis

Student modelling and agent classified rules learning as applied in the development of the intelligent Preassessment System has been presented in [10],[11]. In this paper, we now demystify the theory behind the development of the pre-assessment system followed by some computational experimentation and graph visualisation of the agent classified rules learning algorithm in the estimation and prediction of classified rules. In addition, we present some preliminary results of the pre-assessment system evaluation. From the results, it is gathered that the system has performed according to its design specification.

📄 Content

PAPER COMPUTATIONAL ESTIMATE VISUALISATION AND EVALUATION OF AGENT CLASSIFIED RULES LEARNING SYSTEM Computational Estimate Visualisation and Evaluation of Agent Classified Rules Learning System http://dx.doi.org/10.3991/ijet.v11i1.5001 Kennedy E. Ehimwenma, Martin Beer and Paul Crowther Sheffield Hallam University, United Kingdom

Abstract!Student modelling and agent classified rules learning as applied in the development of the intelligent Pre- assessment System has been presented in [10],[11]. In this paper, we now demystify the theory behind the development of the pre-assessment system followed by some computa- tional experimentation and graph visualisation of the agent classified rules learning algorithm in the estimation and prediction of classified rules. In addition, we present some preliminary results of the pre-assessment system evaluation. From the results, it is gathered that the system has per- formed according to its design specification. Index Terms—agent learning, speech acts, ontology, classi- fication, pre-assessment, student evaluation, visualisation, prediction, artificial intelligence I. INTRODUCTION Learning is change in the mental state of humans or machines after a sequence of some acquired experiences. Whether these experiences has caused any changes in the “knower” is left to be determined by some form of as- sessment. Learning can be permanent or temporary — meaning that a concept or process can be learned or un- learned. One way to determine the occurrence of learning is through some form of assessment in order to ascertain whether a concept is learned or has been unlearned.
Like humans, machines have the ability to learn. But these abilities are inherent in the chosen type of learning technique. For machines to learn, models—mathematical or symbolic—are chosen or developed suitably to match or solve a learning problem. In this work we have used classification learning in a multiagent system (MAS) for pre-assessing and predicting students’ true state of cogni- tion for appropriate leaning materials based on some measurable modelled parameters. The act of using exist- ing knowledge, features or trained examples to make decision is classification learning. Aside having pre- defined knowledge (or beliefs) for decision making, an agent acquires new knowledge either from self-perception of activities in its environment or through peer-to-peer communication by speech act performatives [1], [22] within a multiagent system. Both predefined knowledge and acquired knowledge amounts to a rise in agent knowledge base (KB) or belief base (BB). In this paper, we now present in details the theory be- hind the Pre-assessment System design, the principles applied in the development of the classified rules as well as some computational experimentation and graph visuali- sation of the agent classified rule learning algorithm and how they make accurate prediction for the required num- ber of classified rules. Also we present the preliminary results of the pre-assessment system evaluation in which the results showed that the system has performed accord- ing to its design specification. As revealed from this ex- perimentation, the learning algorithms only holds for a regular ontology i.e. an ontology with equal number of leave-nodes across all parent class nodes [11].
The hallmark of this work is the use of description logic tool – Jason AgentSpeak – in the development of an intel- ligent tutoring system (ITS) in which agents communicate interoperable knowledge in the format of triples, thus causing changes in their mental state as they carry out the overall system’s objective—which is to identify gaps in human learning.
This paper continues with related works in Section I. Section II is BDI: Belief, Desire and Intention in agents, and agent environment. In Section III we present the Pre- assessment agents, and multiple classifications learning in Section IV. Section V presents report on algorithmic ex- perimentation and the results obtained; and Section VI is conclusions and further work. A. Related Work: Learning Systems and Strategies of Development Works in literature has it that several systems has emerged to support learning, teaching, and assessment (LTA). How these systems operate is perhaps determined by the strategy employed in their development e.g. com- puter assisted assessment (CAA), computer based testing (CBT), intelligent learning system (ILS), computer assist- ed learning (CAL), computer adaptive testing (CAT), learning management system (LMS) and web-based learn- ing systems. To assess learning for instance, the CBT employs the strategy of presenting predefined sets of questions, while the CAT dynamically select and present questions depending on students’ performance [16]. Though varying needs has influenced the design of differ- ent systems, holistically, computers in LTA was borne on the need to use t

This content is AI-processed based on ArXiv data.