Rozi, Nanang Fakhrur (2016) Perbaikan metode Rekomendasi Diskusi Pemrograman Dengan Normalisasi Identifier Menggunakan Lingua::Idsplitter. Masters thesis, Institut Teknologi Sepuluh Nopember Surabaya.
Preview |
Text
5112201045-Master_Thesis.pdf - Published Version Download (989kB) | Preview |
Abstract
Situs tanya-jawab Stack Overflow telah sering digunakan sebagai acuan
oleh programmer. Informasi atau solusi dalam proses pengembangan perangkat
lunak dapat dicari dengan bantuan mesin pencari pada situs. Namun, perbedaan
dalam gaya penulisan, terutama pada penulisan identifier program, sering
menyebabkan rekomendasi (pencarian) menjadi tidak sesuai dengan kebutuhan
programmer. Beberapa programmer menulis identifier dalam bentuk singkatan
sementara yang lain tidak sehingga menurunkankinerja rekomendasi. Penelitian
ini mengadopsi Lingua::IdSplitter untuk menormalkanidentifier pada data diskusi
Stack Overflow. Proses normalisasi dilakukan dengan memisahkan identifier yang
terdiri atas komposisi term serta memperluas singkatan yang ada padaidentifierke
bentuk penuh. Hasil penelitian menunjukkan bahwa normalisasi identifier
menggunakan Lingua::IdSplitter hanya mampumeningkatkan performa sistem
rekomendasi ketikaidentifier dengan unsur singkatanmendominasi pada data
diskusi
=============================================================================================
Stack Overflow as a question and answer (Q&A) site has often been used
as reference by programmers. Information or solutions in the software
development process can be searched with the help of search engine on the site.
However, differences in writing style, especially on the program identifier
writing, often causing the recommendation (search) become incompatible with the
programmers need. Some programmers write identifier in abbreviated form while
another are not so that it lowers the recommendation performance. This research
adopted the Lingua::IdSplitter to normalize the identifier on Stack Overflow
discussion. The normalization process is done by separating the identifier
comprising the composition of the term as well as expand the existing
abbreviation on the identifier to the full form. The results showed that the
identifier normalization using Lingua::IdSplitter only able to improve the system
performance when the abbreviated term within identifierdominate in the dataset
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTIf 006.7 Roz P |
| Uncontrolled Keywords: | sistem rekomendasi, identifier, normalisasi, stack overflow |
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK105.8883 Web authoring software (include web server) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.888 Web sites--Design. Web site development. |
| Divisions: | Faculty of Information Technology > Informatics Engineering > 55101-(S2) Master Thesis |
| Depositing User: | EKO BUDI RAHARJO |
| Date Deposited: | 20 Jul 2020 03:05 |
| Last Modified: | 20 Jul 2020 03:05 |
| URI: | http://repository.its.ac.id/id/eprint/76416 |
Actions (login required)
 |
View Item |