Ayunin, Purina Qurota (2019) Rekomendasi Anotasi Otomatis Pada Konten Pembelajaran pada Konten Pembelajaran MOOC. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
05111540000008-Undergraduate_Thesis.pdf Download (1MB) | Preview |
Abstract
Saat ini proses belajar mengajar perkuliahan dapat dilakukan hanya dengan mengikuti kelas secara online melalui situs-situs yang menerapkan sistem Massive Open Online Course (MOOC). Namun pada praktiknya, ketika pengajar membuat kursus baru pada situs MOOC, pengajar kurang mendefinisikan secara detail perihal pemberian keterangan kursus. Khususnya pada MOOC yang menggunakan Moodle, pengajar cenderung hanya memberikan materi berupa attachment file atau tugas perminggunya.
Maka pada tugas akhir ini, akan dilakukan pelabelan dengan menganotasikan konten pembelajaran secara otomatis dengan metode pengekstrakan konten pembelajaran yang kemudian diklasifikasikan. Data yang diambil merupakan judul mata kuliah dan deksripsi, capaian pembelajaran, atau pokok bahasan mata kuliah. Kemudian dilakukan proses data mining berupa pengklasifikasian menggunakan metode tanpa Machine Learning dengan menerapkan beberapa rule dan dengan metode Machine Learning dengan Random Forest, Support Vector Machine, dan Naive Bayes.
Dari keempat metode yang diuji, metode dengan hasil terendah didapatkan pada metode klasifikasi tanpa Machine Learning dengan akurasi 71,7%. Sedangkan hasil terbaik diperoleh dari metode menggunakan Machine Learning menggunakan Random Forest Classifier dengan data training yang sudah di-over sampling dengan ADASYN mendapatkan nilai akurasi yaitu 93,3%. Model tersebut juga dikatakan terbaik karena terbukti menghasilkan keluaran label yang sesuai dari data uji baru.
===============================================================================================
Nowadays, teaching and learning process of lectures can be done only by taking classes online through sites that implement the Massive Open Online Course (MOOC) system. But in practice, when the teacher makes a new course on the MOOC website, the teacher does not define the course in detail in the about the course description. Especially in the MOOC that uses Moodle, teachers tend to only provide material in the form of file attachments or weekly assignments.
In this final project, labeling will be carried out by annotating learning content automatically by extracting the course content which is then classified. The data taken is the subject title and description, learning outcomes, or subject matter. Then the data mining process is carried out in the form of classifying using a method without Machine Learning by applying several rules and using Machine Learning methods with Random Forest, Support Vector Machine, and Naive Bayes.
From the four tested methods, the method with the lowest results was obtained from the classification method without Machine Learning with an accuracy of 71.7%. While the best results are obtained from the method of using Machine Learning, that is using a Random Forest Classifier with training data that has been over-sampled with ADASYN with an accuracy with of 93.3%. The model is also said to be the best because it is proven to produce the appropriate label output from the new test data.
Keywords: MOOC, Naïve Bayes, Random Forest, SVM
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSIf 005.12 Ayu r-1 2019 |
| Uncontrolled Keywords: | MOOC, Naïve Bayes, Random Forest, SVM |
| Subjects: | T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > T Technology (General) > T58.6 Management information systems |
| Divisions: | Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | Purina Qurota Ayunin |
| Date Deposited: | 11 Jun 2021 06:49 |
| Last Modified: | 11 Jun 2021 06:49 |
| URI: | http://repository.its.ac.id/id/eprint/60444 |
Actions (login required)
 |
View Item |