The Optimisation of Genetic Assessment Test Generation Based on Fuzzy Scoring

The most important aspect of an educational assessment is related to the closeness of the results to the actual knowledge level of the assessee. In this way, the fidelity of the assessment is ensured. In this matter, this paper presents the description and potential results of a model that determines a recognition of a detailed knowledge report related to the assessment topic, including the situation of partial knowledge. Thus, a model that details the usage of genetic algorithms (GAs) for establishing a method of partial scoring using fuzzy logic and weights given to specific parts of the assessment items is presented in this paper. Shortly, assessment items such as MCQ (multiple-choice) or cloze questions are given as examples of situations where this method of partial scoring can be applied. The genetic algorithm is used to generate the optimal scoring weights to the option of the assessment items related to a given value of entropy or scoring equilibrium between the scores chosen for the items. The purpose of this analysis is to obtain an optimal score configuration for an item which can offer chosen score equilibrium between the options score and which could give the opportunity of partial scoring, which has multiple benefits for the assessment process.