KLASIFIKASI PROPOSAL TUGAS AKHIR BERDASARKAN BIDANG KEAHLIAN DAN PEMINATAN DOSEN PEMBIMBING MENGGUNAKAN SUPPORT VECTOR MACHINE PADA JURUSAN TEKNIK INFORMATIKA DI STMIK INDONESIA BANJARMASIN

Abstrak

In STMIK Indonesia Banjarmasin each student who works on their final projects are given three lecturers determined by the outcome of the meeting of the concerned departments. Various obstacles may occur including the meeting time interfering with the course time, the delay in submitting students’ final project proposal draft to be discussed in the meeting and the results of the meeting not being suitable because the material covered in the final project is not mastered by the supervisor who has been specified. For that purpose, this research proposes to make a system that enables easy and appropriate classification of final assignment proposal draft of students to lecturers who are experienced and have the appropriate specialization and expertise.

Support Vector Machine (SVM) is one of the best ways to solve the problem of classification. Since SVM can only classify into two classes, multiclass SVM method is used. From the results of classification using SVM, sorting was done using Cosine Similarity for the final result.

In this research, 184 students' final project proposal drafts were used for the test. Seventy or as much as 129 documents were used for the training process, and the remainder, namely 30% or a total of 55 documents were used for the testing process. From the test results using SVM classification and sorting by Cosine Similarity, the success achieved was 85%.