Calculation Model for Correlating Quantitative and Qualitative Peer Assessment using Fuzzy Logic and Generating Individual and Collective Scores with Measures of Central Tendency Adjustment

New teaching-learning paradigms promote the integration of various evaluations (qualitative/quantitative) of a task simultaneously, however, unifying them manually takes a long time. Fuzzy reasoning systems model the inherent uncertainty in the human reasoning process, translating knowledge and experience into a set of linguistic expressions that use words instead of numerical values to correlate multiple variables. The purpose of this study is to present a correlation model between numerical and sentiment scores generated by the predictive model for each criterion, and calculations of scores for each evaluator and the collective. An experimental methodology was applied. Fuzzy logic techniques and measures of central tendency adjustment were used. The results reveal that the assessment score for each criterion is equitable when no penalty is applied in the fuzzy rules, and that the median has the best fit for generating a reliable collective score in peer assessment processes. As future work, it is planned to expand the model with the calibration of the task evaluation score considering the evaluator’s profile.