8 Maret 2021 | Tim Media UISI

Model Tree with Modified L1 Loss Function for Predicting Missing Attendance Data of Faculties

The problem of missing attendance data in our university often arises due to the negligence of faculties. In this study, we address the problem by directly predicting the work duration of faculties. The nature of the problem require us not only to make a

Mohammad Arif Rasyidi
Department of Informatics
Universitas Internasional Semen Indonesia
Gresik, Indonesia
mohammad.rasyidi@uisi.ac.id

 

Rachmadita Andreswari
Information System Department
Telkom University
Bandung, Indonesia
andreswari@telkomuniversity.ac.id

 

The problem of missing attendance data in our university often arises due to the negligence of faculties. In this study, we address the problem by directly predicting the work duration of faculties. The nature of the problem require us not only to make accurate predictions, but also to minimize the rate of overestimation. To address the problem, we propose the implementation of model tree with modified L1 loss function and simple prediction result reduction. Experimental results show that our proposed method is able to lower the verestimation rate while maintaining accuracy within acceptable range.
Keywords—model tree; loss function; prediction; attendance

 

Full Paper: Download Full Paper

Plagiarism Check: Download Check Plagiarism

Peer Review: Download Peer Review

Artikel Terkait