REKOMENDASI ARTIKEL ILMIAH MENGGUNAKAN METODE CONTENT BASED FILTERING

Nisa, Amalia Cahya Uswatun (2023) REKOMENDASI ARTIKEL ILMIAH MENGGUNAKAN METODE CONTENT BASED FILTERING. Other thesis, UPN "Veteran" Yogyajarta.

[thumbnail of ABSTRAK_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
ABSTRAK_AMALIA CAHYA USWATUN NISA_123180111.pdf
Download (13kB)
[thumbnail of COVER_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
COVER_AMALIA CAHYA USWATUN NISA_123180111.pdf
Download (135kB)
[thumbnail of DAFTAR ISI_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
DAFTAR ISI_AMALIA CAHYA USWATUN NISA_123180111.pdf
Download (67kB)
[thumbnail of DAFTAR PUSTAKA_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
DAFTAR PUSTAKA_AMALIA CAHYA USWATUN NISA_123180111.pdf
Download (145kB)
[thumbnail of PENGESAHAN PEMBIMBING_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
PENGESAHAN PEMBIMBING_AMALIA CAHYA USWATUN NISA_123180111.pdf
Download (676kB)
[thumbnail of PENGESAHAN PENGUJI_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
PENGESAHAN PENGUJI_AMALIA CAHYA USWATUN NISA_123180111.pdf
Download (712kB)
[thumbnail of SKRIPSI FULLTEXT_AMALIA CAHYA USWATUN NISA_123180111.pdf] Text
SKRIPSI FULLTEXT_AMALIA CAHYA USWATUN NISA_123180111.pdf
Restricted to Repository staff only
Download (4MB)

Abstract

vi
ABSTRAK
Artikel ilmiah sangat bermanfaat untuk ilmu pengetahuan hingga penyelesaian masalah
dalam kehidupan. Untuk menemukan artikel ilmiah peneliti dan pembaca bisa mencari jurnal
elektronik yang ada di internet. Internet menyediakan kemudahan akses namun banyak sekali
artikel yang tersebar atau tsunami data akan menyulitkan pembaca untuk mencari artikel. Dan
diperlukan ketelitian untuk menghindari publikasi jurnal predator. Untuk itu, dibutuhkan
rekomendasi untuk mengelola informasi agar pembaca mendapat rekomendasi artikel yang
sesuai dengan kebutuhan.
Penelitian ini membantu pembaca mengetahui artikel yang dibutuhkan dengan
mengimplementasikan content based filtering dalam memberi rekomendasi. Dengan data
artikel ilmiah yang berasal dari data sekunder pada website Kaggle milik Ammarabbasi.
Conten based filtering adalah metode yang memperhatikan kemiripan konten atau informasi.
Pada penelitian ini dataset akan melalui proses preprocessing sebelum diolah. Kemudian data
akan melalui pembobotan kata dengan TF-IDF, lalu dilanjutkan perhitungan similaritas
menggunakan cosine similarity.
Rekomendasi yang dibangun akan diuji dengan menggunakan confusion matrix. Dari
tabel confusion matrix akan didapat nilai precision dan recall. Dengan threshold 0.10 hasil
yang diperoleh dari perhitungan precision@k yaitu k=5 88%, k=8 82%, dan k=10 83%.
Sedangkan hasil recall@k pada yaitu k=5 51%, k=8 73%, dan k=10 87%.
Kata kunci: artikel ilmiah, rekomendasi, content based filtering, TF-IDF, cosine similarity
vii
ABSTRACT
Scientific articles are very useful for science and solving problems in life. To find
scientific articles, researchers and readers can search for electronic journals on the internet.
The internet provides easy access, but there are so many articles scattered around or a tsunami
of data that makes it difficult for readers to find articles. And care is needed to avoid
publication in predatory journals. For this reason, recommendations are needed for managing
information so that readers get recommended articles that suit their needs.
This research helps readers find out what articles they need by implementing content-
based filtering in providing recommendations. With scientific article data originating from
secondary data on Ammarabbasi's Kaggle website. Content based filtering is a method that
pays attention to similarities in content or information. In this research, the dataset will go
through a preprocessing process before being processed. Then the data will go through word
weighting with TF-IDF, then continue with similarity calculations using cosine similarity.
The recommendations built will be tested using a confusion matrix. From the confusion
matrix table, precision and recall values will be obtained. With a threshold of 0.10, the results
obtained from the precision@k calculation are k=5 88%, k=8 82%, and k=10 83%.
Meanwhile, the results for recall@k are k=5 51%, k=8 73%, and k=10 87%.
Keywords: scientific articles, recommendations, content based filtering, TF-IDF, cosine
similarity

Item Type: Thesis (Other)
Uncontrolled Keywords: scientific articles, recommendations, content based filtering, TF-IDF, cosine similarity
Subjects: Z Bibliography. Library Science. Information Resources > ZA Information resources
Divisions: Faculty of Engineering, Science and Mathematics > School of Engineering Sciences
Depositing User: Eko Yuli
Date Deposited: 13 Dec 2023 03:42
Last Modified: 13 Dec 2023 03:42
URI: http://eprints.upnyk.ac.id/id/eprint/38329

Actions (login required)

View Item View Item