Prediction of Junior High School National Exam Results Based on Academic Report Using K-Nearest Neighbor
| Autor: | Utomo Pujianto, Mei Candra Kartikasari, Harits Ar Rosyid |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
0303 health sciences
03 medical and health sciences Computer science Performance comparison Resampling Classifier (linguistics) 0202 electrical engineering electronic engineering information engineering Mathematics education 020201 artificial intelligence & image processing 02 engineering and technology Report card 030304 developmental biology k-nearest neighbors algorithm |
| Zdroj: | 2020 6th International Conference on Science in Information Technology (ICSITech). |
| DOI: | 10.1109/icsitech49800.2020.9392052 |
| Popis: | The National Examination is a mechanism adopted by the Government of the Republic of Indonesia to evaluate the performance of the learning process at every level of education. This study sees the decline in national exam scores for junior high school students that occurred during 2016 to 2018 as a problem that needs to be resolved. The k-Nearest Neighbor method which is applied to the report card scores is used to predict the achievement of student performance in the national exam from the four subjects tested. The dataset containing 307 instances resulted from the acquisition of primary data from a state Junior High School in Malang, Indonesia. Two scenarios, one of which involved SMOTE resampling, were used in the performance comparison study. The results showed that the best performance was generated by a scenario involving k-Nearest Neighbor as the classifier, combined with SMOTE preprocessing. The best performance of national exam predictions can be seen in English, with an accuracy of 81.16%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|