Rabbani, Muhammad Fikri (2022) Graph-Based Process Mining Untuk Pemodelan Model Proses Hirarki Berbasis Bpmn Dengan Mempertimbangkan Semantic Similarity Dan Kategorisasi. Other thesis, Institut Teknologi Sepuluh Nopember.
![]() |
Text
05111840000165-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (3MB) | Request a copy |
Abstract
Semua kegiatan untuk mengatur berjalannya suatu organisasi harus memiliki aturan dan dicatat dalam suatu laporan untuk mencapai tujuan organisasi yang dikenal dengan proses bisnis. Ada dua jenis model proses, yaitu model proses tunggal dan model proses hierarki. Sebuah model proses hirarki menggambarkan hubungan antara kegiatan, di mana setiap kegiatan mewakili sub-model. Model dari aktivitas tersebut dan hubungannya adalah model proses tingkat atas, sedangkan setiap sub-model adalah model proses tingkat bawah. Banyak metode penemuan proses telah muncul; namun, sebagian besar fokus pada membangun model proses tunggal. Penelitian ini mengusulkan algoritma Graph-based Procses Discovery in BPMN-Notation (GHPDBN) untuk memodelkan model proses hierarki dalam notasi Business Process Modeling Notation (BPMN) secara otomatis. GHPDBN memetakan semua aktivitas log peristiwa ke dalam kategori menggunakan kesamaan semantik dan kategorisasi, yaitu menggunakan metode GloVe Embedding. Kemudian, dengan mempertimbangkan kategori yang diperoleh, GHPDBN menetapkan aturan untuk membuat model proses top-level dan aturan untuk menggambarkan model proses bottom-level. Terakhir, GHPDBN membentuk relasi model proses sesuai notasi BPMN. Tugas Akhir ini mengevaluasi algoritma GHPDBN dengan membandingkan hasil model proses GHPDBN dan hasil model Fuzzy Mining berdasarkan kualitas model proses. Aspek-aspek Quality Measurement digunakan, yaitu Fitness, Precision, Generalization dan Simplicity untuk pengukuran kualitas model proses. Model proses secara keseluruhan (model proses global) dan model proses bottom-level GHPDBN memiliki nilai Fitness, Precision, Generalization dan Simplicity lebih tinggi dari model Fuzzy Mining. Model proses top-level GHPDBN memiliki nilai Precision dan Generalization lebih tinggi dari model Fuzzy Mining. Dari hasil kualitas model proses, metode GHPDBN membentuk model proses hirarki yang lebih fit, lebih presisi, lebih general, dan lebih sederhana dari metode Fuzzy Mining.
=================================================================================================================================
All activities to regulate the running of an organization must have rules and be recorded in a report to achieve organizational goals known as business processes. There are two types of process models: single process models and hierarchical process models. A hierarchical process model describes the relationships between activities, where each activity represents a sub-model. The model of these activities and their relationships is a top-level process model, while each sub-model is a bottom-level process model. Many process discovery methods have emerged; however, most focus on building single process models. This study proposes the Graph-based Procses Discovery in BPMN-Notation (GHPDBN) algorithm to automatically model the hierarchical process model in the form of Business Process Modeling Notation (BPMN). GHPDBN maps all event log activities into categories using Semantic Similarity and categorization, i.e., the GloVe Embedding method. Then, considering the obtained categories, GHPDBN establishes rules for creating a top-level process model and rules for discovering bottom-level process models. This final project evaluates the GHPDBN algorithm by comparing its discovered process models with process models produced by Fuzzy Mining based on the quality of the process model. Aspects of Quality Measurement, such as Fitness, Precision, Generalization and Simplicity, are used for measuring the quality of process models. A global process model and bottom-level process models produced by GHPDBN have a higher Fitness, Precision, Generalization and Simplicity than those depicted by Fuzzy Mining. The top-level process model produced by GHPDBN has a higher Precision and Generalization than that represented by Fuzzy Mining. Based on the quality results, the GHPDBN method can form a hierarchical process model that is more fit, precise, general, and simpler than Fuzzy Mining.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Business Process Management, Hierarchical Process Model, Process Discovery, Process Modelling, |
Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
Depositing User: | Mr. Marsudiyana - |
Date Deposited: | 13 Oct 2025 05:51 |
Last Modified: | 13 Oct 2025 05:51 |
URI: | http://repository.its.ac.id/id/eprint/128578 |
Actions (login required)
![]() |
View Item |