| Popis: |
Skripsi atau tugas akhir menjadi suatu syarat untuk memperoleh gelar Sarjana Strata-1 di perguruan tinggi di Indonesia. Pada program studi Pendidikan Teknik Informatika dan Komputer Universitas Negeri Jakarta mahasiswa mengambil judul dan tema skripsi yang beragam, tidak semua tema skripsi yang dibuat sesuai dengan konsentrasi jurusan masing-masing. Dengan semakin bervariasinya dokumen skripsi maka dilakukan pemodelan topik skripsi dengan menggunakan metode Latent Dirichlet Allocation untuk mengetahui komposisi topik skripsi yang dapat digunakan sebagai referensi pada penelitian selanjutnya. Bahan penelitian yang digunakan berasal dari repository admin Prodi Pendidikan Teknik Informatika dan Komputer Universitas Negeri Jakarta berupa softcopy tahun lulus 2017-2022 sebanyak 329 dokumen skripsi. Dokumen skeipsi diproses melalui beberapa tahapan yaitu pre-processing yang terdiri dari casefolding, stopwords,normalisasi, stemming, setelah itu dilakukan pemodelan LDA dengan menggunakan library Gensim, pembobotan kata dengan TF-IDF, pengujian topik menggunakan coherence score dan perplexity. Sehingga didapatkan 10 topik yang sering dibahas yaitu topik ke-1 mengenai Pengembangan media pembelajaran sebanyak 50 dokumen, topik ke-2 Evaluasi pembelajaran sebanyak 42 dokumen, topik ke-3 Pengembangan sistem informasi berbasis website sebanyak 42 dokumen, topik ke4 Teknik komputer dan jaringan sebanyak 36 dokumen, topik ke-5 Pengembangan sistem informasi Universitas Negeri Jakarta berbasis website menggunakan user interface, user experience ,& backend sebanyak 36 dokumen, topik ke-6 Pengembangan dan evaluasi media pembelajaran e-learning menggunakan algoritma sebanyak 34 dokumen, topik ke-7 data mining sebanyak 28 dokumen, topik ke-8 Sistem requirement pengembangan perangkat lunak sebanyak 26 dokumen, topik ke-9 Standar pencapaian sumber daya Pendidikan sebanyak 14 dokumen, topik ke-10 Evaluasi pengembangan website, multimedia dan jaringan sebanyak 21 dokumen. Pengujian menggunakan coherence score menghasilkan 0.2614 dan perplexity -5.9895. ***** Thesis or final project is a requirement to obtain a bachelor's degree at a university in Indonesia. In the Informatics and Computer Engineering Education study program, State University of Jakarta, students take various thesis titles and themes, not all thesis themes are made according to the concentration of their respective majors. With the variety of thesis documents, the thesis topic modeling is carried out using the Latent Dirichlet Allocation method to determine the composition of the thesis topic which can be used as a reference in further research. The research material used comes from the admin repository of the Informatics and Computer Engineering Education Study Program, State University of Jakarta in the form of softcopy in the 2017-2022 graduation year as many as 329 thesis documents. The script document is processed through several stages, namely pre-processing which consists of casefolding, stopwords, normalization, stemming, after which LDA modeling is carried out using the Gensim library, word weighting with TF-IDF, topic testing using coherence scores and perplexity. So that there are 10 topics that are often discussed, namely the 1st topic regarding the development of learning media as many as 50 documents, the 2nd topic Evaluation of learning as many as 42 documents, the 3rd topic Development of website-based information systems as many as 42 documents, the 4th topic Computer engineering and a network of 36 documents, topic 5 Development of a website-based information system at the State University of Jakarta using a user interface, user experience, & backend as many as 36 documents, topic 6 Development and evaluation of e learning learning media using algorithms as many as 34 documents, topic 6 7 data mining 28 documents, the 8th topic System requirements software development 26 documents, the 9th topic Education resource achievement standards 14 documents, the 10th topic Evaluation of website, multimedia and network development 21 documents. Tests using a coherence score resulted in 0.2614 and a perplexity of -5.9895. |
