The major objective of an intelligent tutoring system is to provide adaptivity of the course material according to the learners need and background knowledge. In Automatic Courseware Sequencing, the main idea is to generate the course material that is well suited for the learner according to the needs. The two main approaches have been identified for automatic courseware sequencing: Adaptive Courseware Generation and Dynamic Courseware Generation. This paper presents the dynamic courseware generation as incorporated in a web-based intelligent tutoring system (ITS). A brief review of related work in the area of dynamic courseware generation is presented. The proposed scheme for generating course material based on adaptive-network based fuzzy inference system (ANFIS) is described next. The paper concludes with discussion, advantages of such a scheme.
Keywords: Intelligent Tutoring System, Web-based Intelligent Tutoring System, Adaptive network-based inference system (ANFIS), Dynamic Courseware Generation.
[...] An Intelligent Tutoring System should be typically strongly adaptive, working in a wellstructured information space; gathering data about the user's movements and using this information to dynamically modify the presentation and functionality of the system in clearly defined ways. It is again to emphasize that adaptivity is not a technology, but a goal. Adaptivity is a common functional goal of intelligent systems. In summary, for a system to be called an adaptive system it must have the following characteristics [Eklund & Brusilovsky, 1999]: a. [...]
[...] “Neuro-Fuzzy and Soft Computing”, PTR Prentice Hall [Lesgold et al., 1987] Lesgold, A.M., Bona, J.G., Ivill, J.M., Bowen, A. (1987). “Toward a theory of curriculum for use in designing intelligent instructional systems”, In: Mandl, H., Lesgold, A.M. (eds.): Learning Issues for Intelligent Tutoring Systems. Springer-Verlag, New York. [Murray, 1996] Murray, T. (1996). “Special Purpose Ontologies and the Representation of Pedagogical Knowledge”, In Proceedings of International Conference for the Learning Sciences (ICLS-96), Evanston, IL AACE: Charlottesville, VA [PERSEUS, 2003] PERSEUS (2003). Webdocument available [...]
[...] learner according to his Table 4.1 ANFIS Rules for Lesson Generation Memory UnderConception Class standing Low Low Low 01 Low Low Medium 02 Low Low High 03 Low Medium Low 04 Low High Low 05 Medium Low Low 06 High Low Low 07 Low Medium Medium 08 Low Medium High 09 Low High Medium 10 Medium Low Medium 11 Medium Low High 12 Medium Medium Low 13 Medium High Low 14 High Low Medium 15 High Medium Low 16 Low High High 17 Medium Medium Medium 18 Medium Medium High 19 Medium High Medium 20 High Low High 21 High Medium Medium 22 High High Low 23 It is to be noted that the ANFIS is going to indicate the class of the content to be used in aggregation. The learning material is designed and organized through XML document type definition in such a manner that the material can be searched using its class with the unit, topic and the lesson number. [...]
[...] A lesson level adaptivity guided by the learners' knowledge level at each stage of lesson generation has been proposed and its implementation through computational intelligence techniques has been shown. The ANFIS directed lesson generation process alongwith class-oriented content structure provides real adaptivity. Following are the salient features and advantages of the proposed approach: i. An adaptive rule based approach provides the advantage of rule flexibilty. While we have shown a rule base designed on the basis of expertise gathered from several teachers, an individual teacher's expertise can also be formalised to reflect a personalized judgment. [...]
[...] Thus, using the class type taken from the ANFIS, the aggregation module simply retrieves the content pointers and submits them to the transformation module for subsequent processing ANFIS-BASED LESSON AGGREGATION MODEL As already described, adaptive network based fuzzy inference system utilizes a backpropagation algorithm or a hybrid algorithm to tune its parameters from the input-output data pairs. In case of lesson generation process, the rules are taken from the human experts. The criterion for rule framing is based on the learners' knowledge level. [...]
Online readingwith our online reader
Content validatedby our reading committee