Claim Missing Document
Check
Articles

Indonesian Sentence Boundary Detection using Deep Learning Approaches Santoso, Joan; Setiawan, Esther Irawati; Purwanto, Christian Nathaniel; Kurniawan, Fachrul
Knowledge Engineering and Data Science
Publisher : citeus

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Detecting the sentence boundary is one of the crucial pre-processing steps in natural language processing. It can define the boundary of a sentence since the border between a sentence, and another sentence might be ambiguous. Because there are multiple separators and dynamic sentence patterns, using a full stop at the end of a sentence is sometimes inappropriate. This research uses a deep learning approach to split each sentence from an Indonesian news document. Hence, there is no need to define any handcrafted features or rules. In Part of Speech Tagging and Named Entity Recognition, we use sequence labeling to determine sentence boundaries. Two labels will be used, namely O as a non-boundary token and E as the last token marker in the sentence. To do this, we used the Bi-LSTM approach, which has been widely used in sequence labeling. We have proved that our approach works for Indonesian text using pre-trained embedding in Indonesian, as in previous studies. This study achieved an F1-Score value of 98.49 percent. When compared to previous studies, the achieved performance represents a significant increase in outcomes.
Indonesian Language Term Extraction using Multi-Task Neural Network Santoso, Joan; Setiawan, Esther Irawati; Ferdinandus, Fransiskus Xaverius; Gunawan, Gunawan; Collantes, Leonel Hernandez
Knowledge Engineering and Data Science
Publisher : citeus

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

The rapidly expanding size of data makes it difficult to extricate information and store it as computerized knowledge. Relation extraction and term extraction play a crucial role in resolving this issue. Automatically finding a concealed relationship between terms that appear in the text can help people build computer-based knowledge more quickly. Term extraction is required as one of the components because identifying terms that play a significant role in the text is the essential step before determining their relationship. We propose an end-to-end system capable of extracting terms from text to address this Indonesian language issue. Our method combines two multilayer perceptron neural networks to perform Part-of-Speech (PoS) labeling and Noun Phrase Chunking. Our models were trained as a joint model to solve this problem. Our proposed method, with an f-score of 86.80%, can be considered a state-of-the-art algorithm for performing term extraction in the Indonesian Language using noun phrase chunking.
Maximum Marginal Relevance and Vector Space Model for Summarizing Students' Final Project Abstracts Gunawan, Gunawan; Fitria, Fitria; Setiawan, Esther Irawati; Fujisawa, Kimiya
Knowledge Engineering and Data Science
Publisher : citeus

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Automatic summarization is reducing a text document with a computer program to create a summary that retains the essential parts of the original document. Automatic summarization is necessary to deal with information overload, and the amount of data is increasing. A summary is needed to get the contents of the article briefly. A summary is an effective way to present extended information in a concise form of the main contents of an article, and the aim is to tell the reader the essence of a central idea. The simple concept of a summary is to take an essential part of the entire contents of the article. Which then presents it back in summary form. The steps in this research will start with the user selecting or searching for text documents that will be summarized with keywords in the abstract as a query. The proposed approach performs text preprocessing for documents: sentence breaking, case folding, word tokenizing, filtering, and stemming. The results of the preprocessed text are weighted by term frequency-inverse document frequency (tf-idf), then weighted for query relevance using the vector space model and sentence similarity using cosine similarity. The next stage is maximum marginal relevance for sentence extraction. The proposed approach provides comprehensive summarization compared with another approach. The test results are compared with manual summaries, which produce an average precision of 88%, recall of 61%, and f-measure of 70%.
Timbre Style Transfer for Musical Instruments Acoustic Guitar and Piano using the Generator-Discriminator Model Nagari, Widean; Santoso, Joan; Setiawan, Esther Irawati
Knowledge Engineering and Data Science
Publisher : citeus

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Music style transfer is a technique for creating new music by combining the input song's content and the target song's style to have a sound that humans can enjoy. This research is related to timbre style transfer, a branch of music style transfer that focuses on using the generator-discriminator model. This exciting method has been used in various studies in the music style transfer domain to train a machine learning model to change the sound of instruments in a song with the sound of instruments from other songs. This work focuses on finding the best layer configuration in the generator- discriminator model for the timbre style transfer task. The dataset used for this research is the MAESTRO dataset. The metrics used in the testing phase are Contrastive Loss, Mean Squared Error, and Perceptual Evaluation of Speech Quality. Based on the results of the trials, it was concluded that the best model in this research was the model trained using column vectors from the mel-spectrogram. Some hyperparameters suitable in the training process are a learning rate 0.0005, batch size greater than or equal to 64, and dropout with a value of 0.1. The results of the ablation study show that the best layer configuration consists of 2 Bi-LSTM layers, 1 Attention layer, and 2 Dense layers.
Cross Platform Waste Reuse, Reduce And Recycle Management Application With Prototyping Methodology Esther Irawati Setiawan; Patrick Hartono; Tong Nam Tuan Vu; Kevin Jonathan Halim; F. X. Ferdinandus; Joan Santoso
Applied Information System and Management (AISM) Vol. 7 No. 1 (2024): Applied Information System and Management (AISM)
Publisher : Depart. of Information Systems, FST, UIN Syarif Hidayatullah Jakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15408/aism.v7i1.37230

Abstract

The use of plastic bags by the public as containers for shopping goods is very high. People often throw away plastic waste carelessly, causing pollution everywhere. The biggest problem with the lack of recycling action lies in the lack of public awareness of the importance of implementing 3R (Reduce, Reuse, Recycle). People are less motivated to do 3R, one of which is because there are no rewards after doing 3R. This research develops an application provides rewards to the community for using Eco-Green Bags as an alternative to plastic bags. This research application has a system that makes it easier for people to carry out 3R. This proposed framework application uses the React Native framework for creating mobile apps and Next.JS for creating admin websites and APIs. With the React Native framework, application performance will be faster and smoother for users. The Next.JS framework allows developers to have a very clear project structure, because Next.JS uses file-based routing. With this research, it is hoped that people can be even more motivated to carry out 3R. The community can participate easily in implementing 3R without any coercion but through the community's own initiative. In this way, the application can be a real supporter of society in implementing 3R and avoiding the use of plastic bags which are also contributors to environmental pollution.
Kuntilanak as a Runtime Entity: Technical Integration of Javanese Folklore Using Manga Matrix in a 2D Horror Game Herman Thuan To Saurik; Harits Ar Rosyid; Aji Prasetya Wibawa; Esther Irawati Setiawan
International Journal of Engineering, Science and Information Technology Vol 5, No 3 (2025)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v5i3.961

Abstract

In this work, Kuntilanak, a mythological creature from Javanese mythology, is used as a dynamic element in a 2D horror game to provide a technical framework for integrating culturally infused folklore into interactive gaming. The design process breaks down the character's appearance, attire, and personality into workable technical specifications using the Manga Matrix framework as a guide. With C# scripted behaviours like unexpected appearances, animation state changes (controlled by Unity's Animator Controller), audio triggers (laughing, crying), and interactive reactions to in-game objects like yellow Bamboo (for hiding) and scissors (for repelling), Kuntilanak was created as a sprite-based runtime entity inside the Unity game engine. The character can be dynamically instantiated thanks to this technical approach, which supports procedural horror encounters and is consistent with traditional narratives. The effectiveness of the suggested technological integration was validated by a quantitative assessment using a Likert scale (N=50), which showed 82.2% agreement on cultural authenticity and 79.5% on emotional impact. The findings support the methodology's capacity to turn folklore characters into functional game entities and offer a replicable model for serious games that consider cultural sensitivity. The findings support the methodology's capacity to turn folklore characters into functional game entities and provide a replicable model for serious games that consider cultural sensitivity, with direct implications for designing engaging educational experiences that promote cultural heritage preservation.
Optimization of LPG Distribution for a Multiplatform-Based LPG Marketplace Herman Budianto; Farhan Faisal Zainul Mustaqin; Esther Irawati Setiawan; Joan Santoso
International Journal of Engineering, Science and Information Technology Vol 5, No 3 (2025)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v5i3.994

Abstract

Marketplace applications have become an essential digital solution supporting online transactions, including LPG distribution. The development of this application adopts a multiplatform approach, enabling the application to run on various devices, particularly Android platforms and websites. Using the React Native framework, developers can build applications with a single, efficient codebase for multiple platforms. This study aims to provide users with convenience in purchasing LPG without leaving their homes while offering a more practical and effective user experience. This research includes features for selling, buying, payment, and delivery via courier. The transaction feature facilitates sellers' recording of sales within the application. The results of alpha testing indicate that the Elpijiku marketplace app works well despite some significant errors or bugs. However, acceptance testing results were very positive, with 91% of respondents rating the application and user experience as good. These findings indicate that the Elpijiku application meets user needs in terms of convenience and efficiency and is suitable for use as a digital solution for LPG distribution.
Long short-term memory-based chatbot for vocational registration information services Yudo Sembodo Hastoro Langgeng; Esther Irawati Setiawan; Syaiful Imron; Joan Santoso
Journal of Applied Data Sciences Vol 4, No 4: DECEMBER 2023
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v4i4.128

Abstract

The development of chatbots can communicate fluently like humans thanks to the Natural Language Processing (NLP) technology. Using this technology, chatbots can provide more accurate and natural responses, providing an almost the same experience as human interaction. Therefore, chatbot technology is in great demand by companies and government agencies as a cost-effective solution for information and administrative services that require little human effort and can operate 24/7. The registration information service at BLK Surabaya still uses an operator who serves prospective trainees and answers questions via social media or chat. However, these operators have limitations in terms of time and effort. The purpose of this study is to examine how to use chatbots to answer questions about registration information training at BLK Surabaya using the Long Short Term Memory (LSTM) algorithm with a dataset of questions collected in the form of Frequently Asked Questions (FAQ) in Indonesian. The dataset consists of 2,636 labeled samples of questions, which were divided into three sets: 60% for training (1,581 pieces), 20% for validation (527 samples), and 20% for testing (528 samples) to evaluate the model's performance. Several steps were taken in implementing this research, including changing the list of questions and answers into the JSON data format, preprocessing, creating LSTM modeling, data training, and data testing. The test results show that Chatbot can provide accurate solutions related to training registration questions with Precision of 88.4%, Accuracy of 87.6%, and Recall of 87.3%.
Aspect-Based Sentiment Analysis of Healthcare Reviews from Indonesian Hospitals based on Weighted Average Ensemble Esther Irawati Setiawan; Patrick Tjendika; Joan Santoso; FX Ferdinandus; Gunawan Gunawan; Kimiya Fujisawa
Journal of Applied Data Sciences Vol 5, No 4: DECEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i4.328

Abstract

Public assessments are essential for evaluating hospital quality and meeting patient demand for superior medical treatment. This study offers a novel approach to aspect-based sentiment analysis (ABSA), which consists of aspect extraction, emotion categorization, and aspect classification. The goal is to examine patient reviews (6,711 reviews) from Google assessments of 20 Indonesian hospitals, broken down by categories including cost, doctor, nurse, and other categories. For example, there are 469 good, 66 negative, and 7 neutral ratings for cleanliness and 93 positive, 125 negative, and 19 neutral reviews for pricing in the sample, which covers a range of attitudes. Using the Conditional Random Field (CRF) approach, aspect phrase extraction was refined and word characteristics and positional tags were adjusted, resulting in an improvement in the F1-score from 0.9447 to 0.9578. The Support Vector Machine (SVM) model had the greatest F1-score of 0.8424 out of two strategies used for aspect categorization. With the addition of sentiment words, sentiment classification improved and led by SVM to an ideal F1-score of 0.7913. For aspect and sentiment classification, a Weighted Average Ensemble approach incorporating SVM, Naïve Bayes, and K-Nearest Neighbors was employed, yielding F1-scores of 0.7881 and 0.8413, respectively. The use of an ensemble technique for sentiment and aspect classification and the incorporation of hyperparameter optimization in CRF for aspect term extraction, which led to notable performance gains, are the innovative aspects of this work.
Retrieval Augmented Generation-Based Chatbot for Prospective and Current University Students Luluk Setiawati Hartono; Esther Irawati Setiawan; Vrijraj Singh
International Journal of Engineering, Science and Information Technology Vol 5, No 3 (2025)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v5i3.951

Abstract

Universities utilize chatbots as assistants for users, especially prospective and current students, to access information and answer questions with relevant answers. This study introduces a new approach to an open-source model-based QA system using Gemma2-2b-it by combining Retrieval Augmented Generation (RAG) and Fine-tuning (FT) techniques. Previously, some studies have focused on only one approach, but this study will combine and compare both methods separately. Raw conversation data from WhatsApp, the main university website, and university PDF documents are used. The Retrieval Augmented Generation Assessment (RAGAS) framework will be used to evaluate the performance of the RAG model. In contrast, precision, recall, and similarity are used to assess the comparative performance of RAG and fine-tuning. The results of the RAGAS show that RAG using the base model is better than RAG using a fine-tuned model, which has 0.78 faithfulness, 0.64 answer relevancy, 0.81 context precision, and 0.68 context recall, so the overall RAGAS Score is 0.72. The comparison of precision and recall of fine-tuning are higher than those of using RAG, but the similarity score is not much different. Furthermore, the potential improvement for RAG of this study can be increased by adding a reranking process in the retrieved context, and fine-tuning of the embedding model can also be added to increase the retrieval process's performance. In addition, further experiments on various datasets and the challenge of overfitting in fine-tuning must be overcome so that the model can also perform better generalization.
Co-Authors Adi Sutanto Adriel Ferdianto Aidil Aji Prasetya Wibawa Albertus Josef Andika Arif Priyambodo Arya Tandy Hermawan Burhanudin Christian Nathaniel Purwanto Collantes, Leonel Hernandez Daniel Gamaliel Saputra Dhaniswara, Erwin Edwin Pramana Edwin Pramana Edwin Pramana Effendy Hadi Sutanto Eko Mulyanto Yuniarno Endang Setyati F. X. Ferdinandus F.X. Ferdinandus Fachrul Kurniawan Fachrul Kurniawan Farhan Faisal Zainul Mustaqin Ferdinandus, Fransiskus Xaverius Fitria Fitria Fitria Fitria Fujisawa, Kimiya FX Ferdinandus Gunawan Gunawan Gunawan Hans Juwiantho Hans Keven Budi Prakoso Harianto, Reddy Alexandro Harits Ar Rosyid Heppi Siswanto Herman Budianto Herman Thuan To Saurik Honoris Setiahadi I Ketut Eddy Purnama Ika Lestari Joan Santoso Johanes, Sugiharto Kevin Jonathan Halim Kimiya Fujisawa Kristian, Yosi Langgeng, Yudo Sembodo Hastoro Leonel Hernandez Lia Safitri Lim, Ernest Lukman Zaman Luluk Setiawati Hartono Mauridhi Hery Purnomo Mauridhi Hery Purnomo Mohamad Fahmi Yusuf Mohammad Farid Machfudin Muhammad Arief Rahman Nagari, Widean Natalie, Valerynta Patrick Hartono Patrick Tjendika Purwanto, Christian Nathaniel Putra, Adi Surya Rachmatullah, Sholeh Rahmatullah, Dewangga Samuel Budi Wardhana Kusuma Santoso , Joan Setya Ardhi Sugiono, Judi Prajetno Surya Sumpeno Syaiful Imron Titis Arwindarti Tjendika, Patrick Tjwanda Putera Gunawan Tong Nam Tuan Vu Vrijraj Singh Wardoyo, Nikko Riestian Putra Yosi Kristian Yudo Sembodo Hastoro Langgeng Yuliana Yuliana Yuni Yamasari Zakariyyah Hidayatulloh Zaman, Lukman