Sort math scores from highest to lowest Score: 100%
Predict math score for each ENEM candidate Technique: Regression Model: Cubist Regression (script in R) - rule-base tree model. Reference: https://rulequest.com/cubist-info.html Score: 90,91%
Predict last 5 responses for each candidate Technique: Classification Models: My approach was testing a different model for each question and exam color. Those models were Hidden Markov Model, Random Forest Classifier, Naive Bayes and Support Vector Machine. Score: 32,10%
General ranking: 20º