MDPI Analytics: Expert System
A key element of the MDPI is the use of an Expert System (ES). The ES interprets a student's performance to present a deeper insight into understanding of underlying concepts as well as those approaches that influence performance.
An ES is a computer application that emulates the decision-making ability of human experts. ESs are used extensively in the field of medicine. The MDPI extends their use to the assessment of mathematics and the generation of the Mathematics Learning Profile of a particular student. The ES incorporates the knowledge, research and experience that the authors have acquired over thirty years in clinical practice and decades of working in the classroom.
The MDPI expert system is a web-based application. This allows updates of the system as needed, e.g. when standards are adjusted or changed. The first step in using the system is the secure, online input of the performance and diagnostic test data by the professionals who administer the exam. Examiner communication with the system is provided by a web-based, natural language interface.
The internal structure of the MDPI expert system consists of three parts:
- the database, which includes the individual scores and observations from the MDPI diagnostic test of the student;
- the knowledge base, which is derived from the expertise of the MDPI author team;
- the inference engine that uses the knowledge base to help interpret the results of the diagnostic exam.
The most important part of the system is the knowledge base which holds the set of inference rules that are used in reasoning and diagnosis. Most expert systems use IF-THEN rules to represent the knowledge. This is the case with the MDPI approach.
Typically expert systems can have from a few hundred to a few thousand rules. The MDPI system uses rules that were established by the authors to interpret the results of the student's performance and recommend a range of options for an intervention program. These rules were developed by matching the inference engine output to the authors' interpretations, many of which have been validated in formal research efforts. Enough comparisons were made to have a high level of confidence in the expert system's output of the particular student's Math Learning Profile as representing the authors' best judgments.
It is important for users to realize that the MDPI expert system is for decision support, not for decision-making. Its function is to provide the professionals on the student's team with important issues to consider and alternatives to explore. The MDPI presumes that examiners will then use their own judgments to review the findings and add their own insights, emphases and recommendations where required.