Student graduation data is very important for universities because it is used in the accreditation process. Data continues to grow and is ignored because it is rarely used. Data of graduating students can provide useful information if processed optimally. This study processes data using data mining to obtain information in the form of a prediction of student graduation punctuality. The method used is the C4.5 algorithm. The criteria used are gender, regional origin, type of school origin, ranking and entry point. In its application, the C4.5 algorithm can be used in predicting student graduation times with a precision value of 70.70%, 60.4% recall, and 58.2% accuracy. In measuring the performance of the algorithm in pattern recognition or information retrieval it is recommended to use a minimum of two parameters namely precission and recall to detect bias, therefore in this study the F-Measure calculation is used. From the calculation of the F-Measure obtained a value of 71% which means that the C4.5 algorithm is considered good in classifying and predicting students who graduate on time

This study aims to determine the effect of the learning cycle 5e learning model assisted by PhET media on student learning outcomes in the subject matter of elasticity and Hooke's law. The research was conducted at SMA Negeri 1 Tapa in class XI during the odd semester of the 2022/2023 academic year, with a population of students in class XI IPA. The research subjects were class XI IPA 1 as the experimental class and XI IPA 2 as the control class, totaling 42 people. XI IPA 1 was taught using a learning cycle model assisted by PhET media, while XI IPA 2 was taught using a conventional learning model. The method used in this study was an experimental method with a research design using a pretest-posttest control group design. The variables in this study were independent and dependent variables. The instrument used was a written essay test to measure the level of student learning outcomes on elasticity and Hooke's law. Sampling in this study was done using a cluster random sampling technique. The data analysis techniques used in this research included data normality test, homogeneity test, hypothesis test, and n-gain test. The results showed that the average pretest value in the experimental class was 43.6, while in the control class it was 37.7. After treatment in each class, the posttest average values were 69.6 in the experimental class and 58.1 in the control class. Based on the results of testing the hypothesis, it was found that tcount = 6.26 ≥ ttable = 2.02, indicating that H0 was rejected and H1 was accepted. This demonstrates that learning using the learning cycle 5e learning model assisted by PhET media has a significant effect on student learning outcomes. The increase in student learning outcomes can be seen in the n-gain test analysis, with the experimental and control classes obtaining scores of 0.6 and 0.5 respectively, indicating moderate n-gain criteria. This means that the treatment given has an influence on student learning outcomes.