EKSTRAKSI INFORMASI BERBASIS KLASIFIKASI KOMPONEN ABSTRAK PADA JURNAL ILMIAH BERBAHASA INGGRIS
| Main Author: | Masyhudi, Alfian Edi |
|---|---|
| Format: | Thesis NonPeerReviewed Book Image |
| Bahasa: | eng |
| Terbitan: |
, 2019
|
| Subjects: | |
| Online Access: |
http://eprints.umm.ac.id/46613/1/PENDAHULUAN.pdf http://eprints.umm.ac.id/46613/2/BAB%20I.pdf http://eprints.umm.ac.id/46613/3/BAB%20II.pdf http://eprints.umm.ac.id/46613/4/BAB%20III.pdf http://eprints.umm.ac.id/46613/5/BAB%20IV.pdf http://eprints.umm.ac.id/46613/6/BAB%20V.pdf http://eprints.umm.ac.id/46613/7/POSTER.jpg http://eprints.umm.ac.id/46613/ |
Daftar Isi:
- Research published in the form of a journal every year is always new and more and more, to get a certain information that will be needed from a study will be very difficult and time consuming. It's like looking for certain information by reading the whole abstraction one by one which is not effective. Therefore, in the hope of developing a system that is able to retrieve certain information from a document correctly and automatically, this study uses a machine learning approach, a type of classification-based Information Extraction. In this study the authors used 25 features, as well as 3 classification Algorithms SMO, Random Forest (RF), and KNN, with 5 target classes Background Information, Hypothesis, Method, Result, and Implication / Future research from abstract components that became target information. Testing is carried out in two stages, namely testing the model and testing the results of classification. The SMO test results obtained an accuracy of 93.6725%, then KNN of 99.2635% and the highest result of RF was 99.6194%. and the results of classification testing with abstract test data 0.488. Future research is expected to improve the performance of the model, and utilize abstract extraction of this journal into other implementations.