A NEURO-FUZZY MODEL FOR PREDICTING SUCCESSFUL PASSING OF ENTRANCE EXAMS OF APPLICANTS TO HIGHER EDUCATION

This scientific article describes the development of a neuro-fuzzy system for predicting the success of applicants in university entrance exams. In the study, the ANFIS model was used to predict the applicant's probability of success in the upcoming exam based on the average grade at school, the period of preparation for the exam, the applicant's experience, and other factors. Data from 256 applicants who applied for the 2022 entrance exam in Uzbekistan were collected as a database. 80% of this data was used for model training and 20% for model testing. The study highlights the possibilities of using neuro-fuzzy systems in the field of education to create opportunities for applicants to choose the right higher education institution for the upcoming exam. The ANFIS model's inclusion of quantitative and qualitative data, as well as its adaptability to changing conditions, make it a promising tool for predicting applicant success in different contexts. An experiment was conducted comparing the performance of the ANFIS model with prediction models such as LinearRegression, Random Forest, XGBoost, DecisionTreeRegressor, and K-nearest Neighbors. The results showed that the ANFIS model outperformed the other models and showed high accuracy in predicting the scores that the applicants could score. Overall, the paper provides valuable insights into the potential of neuro-fuzzy systems to predict academic success.

Ҳали таржима қилинмаган