Raharjana, Indra Kharisma (2023) Ekstraksi User Stories Dari Berita Daring Untuk Membantu Proses Elisitasi Kebutuhan Perangkat Lunak. Doctoral thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 05111860010002-Dissertation.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
05111860010002-Dissertation.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (4MB) | Request a copy |
Abstract
User Story adalah salah satu artefak populer dalam Agile Software Development. User Story merupakan salah satu bentuk spesifikasi dari kebutuhan perangkat lunak yang menunjukkan elemen penting dari kebutuhan, yaitu aspect of who,aspect of what, dan aspect of why. Dalam proses elisitasi kebutuhan perangkat lunak beberapa teknik tradisional seperti wawancara pengguna, kuesioner, analisis dokumen, dan brainstorming biasa digunakan oleh sistem analis. Beberapa tahun terakhir, peneliti mengembangkan beberapa teknik baru, seperti pengambilan kembali kebutuhan dari repositori perangkat lunak yang telah ada dan ekstraksi kebutuhan dari review pengguna, stack-overflow, dan media sosial. Pendekatan baru ini dapat meningkatkan efisiensi dalam mengidentifikasi kebutuhan pengguna. Namun pendekatan ini memiliki celah dimana keterlibatan aktif dari pengguna akhir sangatlah penting. Untuk kasus tidak terprediksi, seperti bencana dan kecelakaan, keterlibatan pengguna akhir dalam proses elisitasi di awal proses pengembangan perangkat lunak sangatlah sulit.
Penelitian ini mengusulkan metode rule-based untuk mengekstraksi aspect of who, what, dan why sebagai penyusun user story dari berita daring. Pendekatan yang digunakan untuk mengekstraksi user story adalah pola part-of-speech (POS) tagging, syntactic dependency parsing, lexical names dari WordNet, named entity recognition (NER), BloomSoft taxonomy, dan domain specificity. Tahapan dalam penelitian ini terdiri atas (i) praproses; (ii) melakukan POS chunking; (iii) identifikasi aspect of who; (iv) identifikasi aspect of what; (v) identifikasi aspect of why; (vi) komposisi user story; dan (vii) evaluasi. Evaluasi dilakukan dengan menghitung nilai precision dan recall dari hasil ekstraksi.
Pengunaan berita daring sebagai salah satu sumber diharapkan dapat membantu analis sistem untuk mempelajari domain sistem yang akan dibangun, terutama untuk mengidentifikasi keinginan pemangku kepentingan (stakeholder) terhadap suatu domain masalah. Keuntungan yang bisa didapatkan dengan penerapan pendekatan ini diantaranya adalah cepat, tidak perlu mengumpulkan stakeholder yang banyak.
===============================================================================================================================
User Story is one of the popular artifacts in Agile Software Development. User Story is one form of specification of software requirements that shows important elements of requirements, namely aspects of who, aspects of what, and aspects of why. To determine good software requirements, system analysts use several requirements elicitation techniques. There are several traditional techniques in the software elicitation, such as user interviews, questionnaires, document analysis, and brainstorming. In recent years, researchers have developed several new techniques, such as retrieving requirements from existing software repositories and extracting requirements from user reviews, stack overflows, and social media. This new approach can improve efficiency in identifying user needs. But this approach has a gap where the active involvement of end-users is very important. For unpredictable cases, such as disasters and accidents, the involvement of end-users in the elicitation process at the beginning of the software development process is very difficult. This study proposes a rule-based method to extract aspects of who, what, and why in user stories by combining multiple approaches such as the part-of-speech (POS) tagging pattern, syntactic dependency parsing, lexical names from WordNet, named entity recognition (NER), BloomSoft taxonomy, and domain specificity. This method's output is a list of user stories sourced from online news, which are expected to help the needs elicitation process. The stages in this research consisted of (i) preprocessing; (ii) do POS chunking; (iii) identify the aspect of who; (iv) identify the aspect of what; (v) identify the aspect of why; (vi) user story composition; and (vii) evaluation. Evaluation is done by calculating the precision and recall of the extraction results. The use of online news as a source is expected to help the system analyst identify stakeholder desires for a problem domain in requirements elicitation phrases. This technique offers fast, does not need to gather many stakeholders, and is cost-effective.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Uncontrolled Keywords: | agile software development, berita daring, natural language processing, software requirements, requirements elicitation, user stories, agile software development, user stories, online news, natural language processing, software requirements, requirements elicitation. |
| Subjects: | Q Science > QA Mathematics > QA76.758 Software engineering |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55001-(S3) PhD Thesis (Comp Science) |
| Depositing User: | Indra Kharisma Raharjana |
| Date Deposited: | 12 Feb 2023 14:31 |
| Last Modified: | 12 Feb 2023 14:31 |
| URI: | http://repository.its.ac.id/id/eprint/96853 |
Actions (login required)
 |
View Item |