Analisis Judul Artikel dan Peneliti Berdasarkan Perilaku dari Rekam Jejak Publikasi

Fadhlurrahman, Muhammad Afif and Pangestu, Muhammad Farras (2022) Analisis Judul Artikel dan Peneliti Berdasarkan Perilaku dari Rekam Jejak Publikasi. Project Report. [s.n.], [s.l.]. (Unpublished)

[img] Text
05111840000093_05111840000134-Project_Report.pdf - Accepted Version
Restricted to Repository staff only

Download (1MB) | Request a copy
[img] Text
05111840000093_05111840000134-Project_Report.pdf - Accepted Version
Restricted to Repository staff only

Download (1MB) | Request a copy
[img] Text
05111840000093_05111840000134-Project_Report.pdf - Accepted Version
Restricted to Repository staff only

Download (1MB) | Request a copy

Abstract

Perkembangan teknologi dan informasi yang semakin pesat menuntut masyarakat untuk beradaptasi atau berkontribusi didalamnya. Berkembangnya teknologi yang ada juga diserta dengan pengembangan ide berbagai bidang di kehidupan sehari-hari. Penelitian mengenai suatu topik terus dilakukan agar dapat ditemukan inovasi-inovasi baru. Implementasi yang dilakukan adalah pembentukan node dan relationship pada Neo4j, pembobotan relationship antar peneliti, visualisasi persebaran data peneliti berdasarkan komunitas menggunakan algoritma t-SNE, pembuatan Linked Open Data (LOD) untuk studi kasus peneliti Indonesia, ekstraksi kata kunci pada tiap topik artikel penelitian, filtrasi judul pada tiap topik artikel penelitian, pelabelan subtopik artikel penelitian menggunakan Latent Dirichlet Allocation (LDA). Hasil dari penelitian mengenai analisis judul artikel dan peneliti berdasarkan perilaku dari rekam jejak , yaitu : Kedekatan antar peneliti menggunakan skenario 3 yaitu skenario dengan menggunakan rumus jaccard yang dihitung dari ‘value_similarity’ dengan menentukan threshold lebih besar dari 0.5 menghasilkan 16.116.059 peneliti yang memiliki nilai kedekatan satu sama lain. Hasil yang didapatkan untuk t-SNE peneliti bahwa sebanyak 1188 data author, sebagian besar author termasuk pada kelompok pallet warna nomor 1 (komunitas 8, 16, 13) atau pallet warna nomor 5 (komunitas 192, 427, 1008, 62, 129, 149). Hasil pembuatan LOD terdiri dari 4015 instance Researcher, 11 instance University, 179197 instance Article, 18 instance Topic, 360 instance Subtopic. Dalam pembuatan LOD ini menggunakan format file berupa file turtle. Proses filtrasi judul artikel dilakukan dengan mencari kata kunci masing-masing topik terlebih dahulu kemudian dilakukan filtrasi judul yang relevan berdasarkan kata kunci yang telah didapat. Persebaran dari setiap judul artikel penelitian yang ada dikelompokkan ke dalam 20 subtopik pada masing-masing topik berdasarkan hasil dari metode LDA. ===================================================================================================== The rapid development of technology and information requires people to adapt or contribute to it. The development of existing technology is also accompanied by the development of ideas in various fields in everyday life. Research on a topic continues to be carried out so that new innovations can be found. The implementation carried out is the establishment of nodes and relationships in Neo4j, weighting relationships between researchers, visualizing the distribution of research data based on the community using the t-SNE algorithm, making Linked Open Data (LOD) for case studies of Indonesian researchers, extracting keywords on each topic of research articles, title filtration on each research article topic, labeling research article subtopics using Latent Dirichlet Allocation (LDA). The results of the research regarding the analysis of article titles and researchers based on the behavior of track records, namely: The closeness between researchers using scenario 3, namely scenarios using the Jaccard formula which is calculated from 'value_similarity' by determining a threshold greater than 0.5 resulting in 16,116,059 researchers who have a value closeness to each other. The results obtained for the t-SNE researcher are 1188 author data, most of the authors belong to the color palette group number 1 (community 8, 16, 13) or color palette number 5 (community 192, 427, 1008, 62, 129, 149 ). The results of making LOD consist of 4015 Researcher instances, 11 University instances, 179197 Article instances, 18 Topic instances, 360 Subtopic instances. In making this LOD using a file format in the form of a turtle file. The article title filtration process is carried out by first searching for keywords for each topic and then filtering the relevant titles based on the keywords that have been obtained. The distribution of each existing research article title is grouped into 20 subtopics on each topic based on the results of the LDA method.

Item Type: Monograph (Project Report)
Uncontrolled Keywords: Neo4j, t-SNE, Linked Open Data, Filtration, LDA, Filtrasi
Subjects: Q Science > QA Mathematics > QA76 Computer software
Q Science > QA Mathematics > QA76.F56 Data structures (Computer science)
T Technology > T Technology (General) > T57.5 Data Processing
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User: Muhammad Afif Fadhlurrahman
Date Deposited: 19 Jan 2022 06:12
Last Modified: 19 Jan 2022 06:12
URI: https://repository.its.ac.id/id/eprint/92379

Actions (login required)

View Item View Item