Using artificial intelligence to predict students' STEM attitudes: an adaptive neural-network-based fuzzy logic model

Abstract

Numerous artificial intelligence methods have lately been applied in education. This study proposes an Adaptive Neural-network-based Fuzzy Logic (ANFIS) model combining fuzzy logic and artificial neural networks for predicting students' STEM attitudes. The inputs of the research were determined as grade levels and academic achievement scores, and the output was determined as STEM attitudes. The hybrid optimisation method trained the fuzzy inference system (FIS) to recognise the more effective ANFIS model. Afterward, the effects of the input variables on the output variable were examined with statistical techniques. Finally, the students' artificial and real STEM attitude scores were compared. 600 middle school students participated in the study, and data were gathered using a Personal Information Form and STEM Attitude Scale. The study revealed a significant positive correlation (r = 0.298; p = 0.000) between the generated ANFIS scores and real scores, the real and artificial scores didn't indicate a statistically significant difference. Therefore, the results obtained through ANFIS correctly predict students' STEM attitude scores. The grade level and academic achievement variables showed statistically significant differences in STEM attitudes. This study is a concrete example as it shows that it is possible to know some characteristics of students using artificial intelligence.