Performance analysis of C5.0 and naïve bayes classification algorithm for pattern recognition of student graduates

Preparing preventive measures against students who may encounter difficulties in the learning process is one of the issues that need to be addressed to raise the standard of higher education at UIN Walisongo. The length of the study session, which might be short or long, is a possible barrier for students. Using data from 2015 and 2016 mathematics study programs graduates, we create a classification system for student graduate patterns using the C5.0 and Naive Bayes algorithm models to address this issue. Two categories—graduating on time and not—will be used to classify the pattern of student graduates. The confusion matrix testing and accuracy value are methods to measure the effectiveness of the two algorithms. The test results demonstrate that the C5.0 algorithm performs better than the Naive Bayes algorithm, as seen by the accuracy value of 94.12% for the C5.0 algorithm compared to 91.18% for the Naive Bayes algorithm. We use data mining with C5.0 and Nave Bayes algorithms to forecast the pattern of graduates from the mathematics study program at UIN Walisongo Semarang in the categories of graduating on time and not graduating on time. Overall, the variables that affect the predicted results of the graduate pattern are the 1st achievement semester index, the total number of credits taken, the number of credits taken in the 6th semester, the entry or selection path chosen when first registering, gender, and the 8th achievement semester index.