Data level approach for imbalanced class handling on educational data mining multiclass classification

Pristyanto, Yoga; Pratama, Irfan; Nugraha, Anggit Ferdita

doi:10.1109/icoiact.2018.8350792

Cited by 37 publications

(25 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Traditional ML methods will not perform well for class-imbalanced datasets as these methods will be biased towards the majority class [47]. Several techniques have been proposed both at the data level and at the algorithmic level to solve class-imbalance problems [48,49,50,51].…”

Section: Handling Imbalanced Classesmentioning

confidence: 99%

Athlete Health Prediction Using Machine Learning Methods

Sharif¹

2023

Preprint

View full text Add to dashboard Cite

Due to an increase in sports activities, the prediction of athletes’ health (AH) has recently become an important research topic. However, it is a challenging task to predict AH because of the nature of the data and the limitations of predictive models. The main objective of this work is to develop appropriate models that can forecast AH using historical data. This work will enable sport organizations to monitor the well-being of their athletes. In this thesis, we explore the applicability of various machine learning (ML) methods for predicting AH. Traditional ML methods do not perform well for class-imbalanced data as these methods are biased towards the majority class. In this work, we propose to use ensemble-based methods which utilize downsampling, bootstrap sampling, and boosting techniques to improve the classification performance. Various metrics are used to evaluate and to compare the model performance. Our results show the superiority of ensemble-based methods over traditional approaches. The random forest and the RUSBoost classier models are in particular found to produce the best performance in handling imbalanced classes.

show abstract

Section: Handling Imbalanced Classesmentioning

confidence: 99%

Athlete Health Prediction Using Machine Learning Methods

Sharif¹

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…DM approaches are used more and more to examine educational information and to discover trends for improving educational practices. EDM is used to collect interesting, useful, and unique knowledge from educational data repositories, as is the case with other DM practices [18]. This research employed 2 well-established technology for DM: pattern discovery or association rules for student data.…”

Section: ) Data Miningmentioning

confidence: 99%

Efficient Pattern Generation using Declat Algorithm Based on Association Rule Mining in Educational Data Mining

Sharma,

Tamrakar

2020

IJEAT

View full text Add to dashboard Cite

A new application called DM Educational Data Mining (EDM) involves data extraction and analysis from the classroom or area of education. In order for educators to deliver quality education to students, EDM integrates various educational information into its review. The EDM works by translating raw data from education systems invaluable information which could have a major effect on the study of education. The output of each student is measured from the database and must be sufficiently accurate to withstand changes in the academic record. Then we have transformed the overall arrangement into a modified relation for the adequacy of the Declat algorithm. The purpose of this work is to examine how prior researchers, as well as recent data mining trends in educational research, have dealt with data mining. In this paper, collected data comprised of 200 students. We define academic performance & impact of additional issues on the basis of these course’s last grades, indications of attendance, class tests, and term last answer substance add up to marks and so on. Here, we compare the FP-Growth and Eclat with Declat algorithm on the bases of confidence and support value in a relation of execution time & no. of patterns generated. This paper uses a declat algorithm to create patterns or delete effective patterns. Such patterns help to illustrate a growing student's success.

show abstract

“…Sedangkan geometric mean (g-mean) adalah salah satu pengukuran paling komprehensif untuk mengevaluasi kinerja algoritme klasifikasi khususnya dalam permasalahn ketidakseimbangan kelas pada dataset. G-Mean dapat menunjukkan akurasi keseluruhan dari akurasi kelas minoritas dan akurasi kelas mayoritas [9], [23]. Berikut ini persamaan untuk menghitung akurasi, specificity, sensitivity, dan g-mean [2].…”

Section: Pra Pengolahan Dataunclassified

Penerapan Metode Ensemble Untuk Meningkatkan Kinerja Algoritme Klasifikasi Pada Imbalanced Dataset

Pristyanto

2019

JTI

Self Cite

View full text Add to dashboard Cite

Pada bidang data mining sering kali para peneliti tidak memperhatikan keseimbangan distribusi kelas pada dataset. Hal ini dapat menimbulkan kesulitan yang cukup serius pada algoritme klasifikasi. karena secara teori mayoritas classifier mengasumsikan distribusi yang relatif seimbang, sehingga menyebabkan kinerja suatu algoritme klasifikasi menjadi kurang maksimal. Oleh karena itu, pada penelitian ini diterapkan metode ensemble dengan penambahan adaptive boosting untuk menyelesaikan permasalahan tersebut. Dari hasil pengujian yang dilakukan pada penelitian ini, metode ensemble dengan penambahan adaptive boosting dapat meningkatkan nilai kinerja algoritme klasifikasi. Nilai kinerja algoritme Naive Bayes dengan Adaptive Boosting akurasi yang dihasilkan sebesar 91.98%, sensitifitas sebesar 91.98%, spesifisitas sebesar 96.49%, dan g-mean sebesar 94.21%. Nilai kinerja algoritme Support Vector Machine dengan Adaptive Boosting akurasi yang dihasilkan sebesar 91.52%, sensitifitas sebesar 91.52%, spesifisitas sebesar 96.29%, dan g-mean sebesar 93.88%. Sedangkan Nilai kinerja algoritme Decision Tree dengan Adaptive Boosting akurasi yang dihasilkan sebesar 94.37%, sensitifitas sebesar 94.37%, spesifisitas sebesar 97.73%, dan g-mean sebesar 96.03%. Hal ini menunjukkan bahwa metode ensemble dengan Adaptive Boosting dapat menjadi solusi untuk meningkatkan kinerja algoritme pada imbalanced dataset.Kata Kunci: adaptive boosting, data mining, ensemble, ketidakseimbangan kelas, klasifikasi.

show abstract

Data level approach for imbalanced class handling on educational data mining multiclass classification

Cited by 37 publications

References 11 publications

Athlete Health Prediction Using Machine Learning Methods

Athlete Health Prediction Using Machine Learning Methods

Efficient Pattern Generation using Declat Algorithm Based on Association Rule Mining in Educational Data Mining

Penerapan Metode Ensemble Untuk Meningkatkan Kinerja Algoritme Klasifikasi Pada Imbalanced Dataset

Contact Info

Product

Resources

About