TEKNOLOGI MACHINE LEARNING PADA SISTEM PENDUKUNG PENELITIAN RSYS (RESEARCH SUPPORT SYSTEM)

• Main Author: Firdamdam Sasmita, - (Author)
• Format: Book
• Published: 2021-05-10.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upi.edu/60768/1/T_PTK_1906671_Title.pdf; http://repository.upi.edu/60768/2/T_PTK_1906671_Chapter1.pdf; http://repository.upi.edu/60768/3/T_PTK_1906671_Chapter2.pdf; http://repository.upi.edu/60768/4/T_PTK_1906671_Chapter3.pdf; http://repository.upi.edu/60768/5/T_PTK_1906671_Chapter4.pdf; http://repository.upi.edu/60768/6/T_PTK_1906671_Chapter5.pdf; http://repository.upi.edu/60768/7/T_PTK_1906671_Appendix.pdf
• Keywords: L Education (General); TJ Mechanical engineering and machinery

Students who are still unfamiliar with the revising process focus on the local level, such as grammar, spelling, punctuation, and sentence level. Meanwhile, students who are experts focus on the global level, such as focusing on improving writing goals, ideas, and meanings. When making revisions, students become too focused on the local level rather than the global level. Comments for the local level are ineffective as a guide in the revision process. Therefore, this study aims to build machine learning-based software with the ANN method to classify global or local comments. This study uses a design research methodology with the SDLC model of prototyping. The results show that the RSYS software was successfully built with machine learning accuracy in classifying 19 comments, obtained from 2 documents, with a 95: 0.5 ratio, 94.74%. Whereas in alpha testing, it was stated that the functionality of the RSYS system and machine learning was considered to be functioning correctly. For beta testing, the largest percentage was 85% for ease of operation and convenience in using the application, 75% for website display and navigation availability.