Article Per Year (5 Year)
p-Index From 2021 - 2026
2.147
P-Index
This Author published in this journals
All Journal FORUM STATISTIKA DAN KOMPUTASI Jurnal Gaussian Jurnal Media Infotama Indonesian Journal of Applied Statistics BAREKENG: Jurnal Ilmu Matematika dan Terapan IJID (International Journal on Informatics for Development) Xplore: Journal of Statistics Journal of Mathematics, Computation and Statistics (JMATHCOS) Indonesian Journal of Statistics and Its Applications Diophantine Journal of Mathematics and Its Applications
Cici Suhaeni
Statistics And Data Science, School Of Data Science, Mathematics, And Informatics, IPB University, Indonesia
Agriculture, Biological Sciences & Forestry Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Earth & Planetary Sciences Economics, Econometrics & Finance Energy Engineering Environmental Science Industrial & Manufacturing Engineering Mathematics Mechanical Engineering Physics Transportation Other

Published : 21 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 21 Documents
Search

 1   2   3 
Identifikasi Tema Perbincangan Masyarakat Tentang Vaksinasi Covid-19 di Media Sosial Cici Suhaeni; , Bagus Sartono
Diophantine Journal of Mathematics and Its Applications Vol. 1 No. 1 (2022)
Publisher : UNIB Press

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33369/diophantine.v1i1.25216

Abstract

Informasi mengenai vaksin covid-19 dan program vaksinasi pemerintah merupakan isu yang mendapat perhatian besar masyarakat dan menjadi perbincangan utama di media sosial, termasuk twitter. Beragam tema dan sudut pandang telah disampaikan oleh masyarakat, dan penelitian ini berupaya mengidentifikasi opini apa saja yang berkembang. Pengetahuan ini dapat menjadi masukan bagi pemerintah dan pemangku kepentingan lain untuk secara bersama membantu proses pemulihan dampak pandemi. Identifikasi opini masyarakat mengenai vaksin covid-19 dilakukan menggunakan metode text clustering terhadap tweets hasil crawling data di Twitter dalam kurun waktu 1 s.d 7 Agustus 2021. Hasil analisis menunjukkan terdapat enam tema besar yang menjadi isu perbincangan yaitu: (1) Kepercayaan terhadap efek vaksin, (2) keikutsertaan dalam vaksinasi untuk mencegah terpapar covid, (3) vaksinasi sebagai upaya herd immunity, (4) keampuhan vaksin melawan virus, (5) jenis-jenis vaksin (6) riset medis tentang.
LDA Topic Modeling Analysis of Public Discourse on Indonesia’s Free Nutritious Meals Program (MBG) Cici Suhaeni; Mualifah, Laily Nissa Atul; Wijayanto, Hari
IJID (International Journal on Informatics for Development) Vol. 14 No. 1 (2025): IJID June
Publisher : Faculty of Science and Technology, UIN Sunan Kalijaga Yogyakarta

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

Abstract

This study investigates public discourse on Indonesia's Free Nutritious Meals (Makan Bergizi Gratis/MBG) program through Latent Dirichlet Allocation (LDA) topic modeling of YouTube comments. Filling a research gap on online public opinion regarding the MBG policy, this study identifies dominant themes and discursive patterns in public perception. A three-topic model, validated through coherence score evaluation and pyLDAvis visualization, reveals key topics: concerns over food prices and distribution, perceived benefits for children and society, and emotionally and politically driven reactions. The findings provide valuable insights into public opinion, while also highlighting challenges in processing Indonesian-language text, such as informal language and noisy data. This study contributes to understanding public perceptions of social policies in digital environments and recommends future research directions, including improved text preprocessing and alternative topic modeling approaches. By shedding light on online public discourse, this research informs policymakers and stakeholders about the effectiveness and potential areas for improvement in the MBG program.
EVALUATING ROBERTA AND GPT-BASED MODELS FOR SDG MULTICLASS TEXT CLASSIFICATION ACROSS DIFFERENT DOCUMENT LENGTHS Uswatun Hasanah; Agus Mohamad Soleh; Cici Suhaeni; Anwar Fitrianto
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 20 No 3 (2026): BAREKENG: Journal of Mathematics and Its Application
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30598/barekengvol20iss3pp2645-2664

Abstract

Multiclass text classification remains a difficult task, primarily due to semantic ambiguity and differences in input length. This study evaluates RoBERTa and GPT-based models for multiclass text classification, focusing on how prompting strategies and document length affect accuracy and robustness. Experiments were conducted using the OSDG Community Dataset, which contains approximately 15,000 labeled samples. The dataset was partitioned into four subsets based on input length: short, medium, long, and all combined. Three GPT variants (zero-shot, few-shot, and fine-tuned) were compared against a RoBERTa baseline. Fine-tuning was implemented via OpenAI’s supervised API with prompt-response formatting. Performance was assessed through F1-score, precision, recall, and balanced accuracy. Fine-tuned GPT achieved the strongest results in all settings, with a macro F1-score of 0.9204 on the all-combined dataset, representing a 4.61% improvement over RoBERTa. Consistent gains were also observed across short (8.63%), medium (3.83%), and long (20.31%) texts. The largest improvement occurred on long documents, while medium-length inputs provided the most stable performance across models. These findings highlight the effectiveness of task-specific fine-tuning in enhancing GPT’s capability to classify SDG-related texts across diverse input lengths.
OPTIMASI XGBOOST DALAM PREDIKSI KECEPATAN KENDARAAN SECARA REAL-TIME : PERBANDINGAN METODE TUNING HYPERPARAMETER Panji Lokajaya Arifa; Kusman Sadik; Agus M Soleh; Cici Suhaeni
Jurnal Gaussian Vol 15, No 1 (2026): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.15.1.01-11

Abstract

Real-time vehicle speed prediction plays a vital role in the development of intelligent transportation systems aimed at improving traffic flow and safety. This study investigates the performance of the XGBoost algorithm enhanced with three hyperparameter tuning techniques: Grid Search, Bayesian Optimization, and Genetic Algorithm. A simulated dataset was constructed reflect diverse urban traffic scenarios, incorporating environmental variables such as weather, road conditions, and traffic density. The models were assessed using 5 and 10-fold cross-validation based on prediction metrics (MSE, RMSE, MAE and R²) as well as computational efficiency in terms of training and inference time. The findings reveal that Bayesian Optimization achieves the highest prediction accuracy, while Grid Search offers the fastest training time. Genetic Algorithm demonstrates a balanced trade-off between accuracy and computational efficiency, making it a competitive and practical choice. These results highlight the importance of selecting hyperparameter tuning strategies based on specific system needs in real-time traffic prediction using XGBoost. 
A Two-Stage Framework for Unsupervised Sentiment Analysis with Model Selection and Semantic Similarity Evaluation Cici Suhaeni; Fani Fahira; Hari Wijayanto; La Ode Abdul Rahman; Hwan-Seung Yong
Indonesian Journal of Statistics and Applications Vol 10 No 1 (2026): Vol 10 Issue 1 June 2026
Publisher : Statistics and Data Science Program Study, SSMI, IPB University, in collaboration with the Forum Pendidikan Tinggi Statistika Indonesia (FORSTAT) and the Ikatan Statistisi Indonesia (ISI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29244/ijsa.v10i1p01-22

Abstract

Sentiment analysis is widely used to extract user opinions from large-scale textual data. However, in practical settings, sentiment labels are often unavailable, making it difficult not only to perform sentiment classification but also to evaluate whether the predicted labels are reliable. This study proposes a two-stage framework for unsupervised sentiment analysis with model selection and semantic similarity evaluation. The dataset consists of Gemini app reviews, in which a labeled subset was used in the first stage to investigate the behavioral characteristics and predictive patterns of three sentiment analysis approaches: lexicon-based, transformer-based, and large language model (LLM)-based methods. The model with the most suitable performance was then selected and applied to predict sentiment labels for the remaining unlabeled data in the second stage. The predicted labels were further evaluated using embedding-based cosine similarity to assess semantic consistency within sentiment classes and separability between classes. The results show that the LLM-based method using Gemini 2.0 Flash achieved the best performance, with accuracy, balanced accuracy, and F1-score values above 0.91, followed by the transformer-based IndoBERT model, while the InSet lexicon-based method showed the weakest performance. In the second stage, semantic similarity evaluation revealed a high average intra-class similarity of 0.6699 and a low inter-class similarity of 0.1917, resulting in a similarity gap of 0.4783. These findings indicate that the proposed framework can support reliable sentiment prediction in largely unlabeled datasets by combining sample-based model selection with semantic validation.
KAJIAN SIMULASI PENDUGAAN SELANG KEPERCAYAAN BOOTSTRAP BAGI ARAH MEDIAN DATA SIRKULAR Cici Suhaeni; I Made Sumertajaya; Anik Djuraidah
Indonesian Journal of Statistics and Applications Vol 2 No 1 (2018)
Publisher : Statistics and Data Science Program Study, SSMI, IPB University, in collaboration with the Forum Pendidikan Tinggi Statistika Indonesia (FORSTAT) and the Ikatan Statistisi Indonesia (ISI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29244/ijsa.v2i1.64

Abstract

The median direction is one of central tendency of circular data. The estimation process usually requires information about sampling distribution of statistic that want to be used as a parameter estimate. Theoretically, sampling distribution derived from population distribution. But, it is not easy to get sampling distribution of median although the population distribution is known. When the sampling distribution cannot be derived easily from population distribution, the bootstrap method can be an alternative to handle it. This study wants to evaluate the effect of increasing concentration parameter to the performance of bootstrap confidence interval estimation for median direction through simulation study. Three methods were used to estimate the interval which are equal-tailed arc (ETA), symmetric arc (SYMA), and likelihood-based arc (LBA). The most important criterion to evaluate them were true coverage and interval width. The simulation results that in general, the increasing of concentration parameter followed by more narrow interval. For small concentration parameter (k<1), all methods give unstable true coverage and interval width. The authors also identify that those three methods produce intervals with identical width when the parameter concentration is 20 or more. In terms of coverage and interval width, the best method was ETA.
Exploring a Large Language Model on the ChatGPT Platform for Indonesian Text Preprocessing Tasks Cici Suhaeni; Sabrina Adnin Kamila; Fani Fahira; Muhammad Yusran; Gerry Alfa Dito
Indonesian Journal of Statistics and Applications Vol 9 No 1 (2025)
Publisher : Statistics and Data Science Program Study, SSMI, IPB University, in collaboration with the Forum Pendidikan Tinggi Statistika Indonesia (FORSTAT) and the Ikatan Statistisi Indonesia (ISI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29244/ijsa.v9i1p100-116

Abstract

Preprocessing is a crucial step in Natural Language Processing, especially for informal languages like Indonesian, which contain complex morphology, slang, abbreviations, and non-standard expressions. Traditional rule-based tools such as regex, IndoNLP, and Sastrawi are commonly used but often fall short in handling noisy, user-generated text. This study explores the capability of Large Language Model, particularly ChatGPT-o3, in performing Indonesian text preprocessing tasks, namely text cleaning, normalization, stopword removal, and stemming/lemmatization, and compares it to conventional rule-based approaches. Using two types of datasets, consisting of a small example dataset of five manually constructed sentences and a real-world dataset of 100 tweets about the Indonesian “Makan Bergizi Gratis” program, both preprocessing methods were applied and evaluated. Results show that ChatGPT-o3 performs equally well in text cleaning and significantly better in normalization. However, rule-based methods like IndoNLP and Sastrawi still outperform ChatGPT-o3 in stopword removal and stemming. These findings indicate that while ChatGPT-o3 demonstrates strong contextual understanding and linguistic flexibility, they may underperform in rigid, token-based operations without fine-tuning. This study provides initial insights into using Large Language Models as an alternative preprocessing engine for Indonesian text and highlights the need for hybrid approaches or improved prompt design in future applications.
Deteksi Polycystic Ovary Syndrome (PCOS) Berbasis Machine Learning: Kombinasi SMOTE, Random Forest, Gradient Boosting, dan Bayesian Optimization Naufalia Alfiryal; Kusman Sadik; Cici Suhaeni; Agus Mohamad Soleh
Indonesian Journal of Applied Statistics Vol 8, No 2 (2025)
Publisher : Universitas Sebelas Maret

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.13057/ijas.v8i2.109931

Abstract

Polycystic ovary syndrome (PCOS) merupakan gangguan endokrin yang umum terjadi pada wanita usia reproduktif. Kondisi ini dapat menyebabkan gangguan ovulasi, ketidakseimbangan hormon, resistensi insulin, serta meningkatkan risiko penyakit kardiovaskular, obesitas, dan gangguan psikologis. Meskipun prevalensinya cukup tinggi, sekitar 75% kasus PCOS masih belum terdiagnosis dalam praktik klinis akibat kompleksitas gejala dan keterbatasan metode diagnosis yang digunakan saat ini. Untuk mengatasi permasalahan tersebut, penelitian ini mengusulkan pendekatan berbasis machine learning guna meningkatkan akurasi dan efisiensi deteksi PCOS. Penelitian ini membandingkan performa dua algoritma pembelajaran terawasi, yaitu random forest dan gradient boosting, dalam melakukan prediksi PCOS. Dataset yang digunakan diperoleh dari repositori publik dan memuat berbagai fitur klinis yang berkaitan dengan PCOS. Untuk menangani permasalahan ketidakseimbangan kelas, metode synthetic minority over-sampling technique (SMOTE) diterapkan pada data pelatihan. Selain itu, bayesian optimization digunakan untuk melakukan penyetelan hiperparameter pada masing-masing model agar diperoleh performa yang optimal. Evaluasi performa model dilakukan menggunakan beberapa metrik, dengan area under the curve–receiver operating characteristic (AUC-ROC) sebagai metrik utama. Hasil penelitian menunjukkan bahwa model Gradient Boosting memberikan performa terbaik dengan nilai AUC sebesar 0,8983 dan nilai recall sebesar 0,95, yang mengindikasikan sensitivitas tinggi dalam mengidentifikasi kasus PCOS. Temuan ini menunjukkan bahwa kombinasi SMOTE dan bayesian optimization efektif dalam meningkatkan akurasi prediksi, khususnya pada dataset medis yang tidak seimbang. Pendekatan yang diusulkan memiliki potensi untuk diintegrasikan ke dalam sistem pendukung keputusan klinis guna mendukung proses skrining PCOS yang lebih dini dan andal.Polycystic ovary syndrome (PCOS) is a common endocrine disorder among reproductive-aged women. This condition can lead to ovulatory dysfunction, hormonal imbalance, insulin resistance, and an increased risk of cardiovascular disease, obesity, and psychological disorders. Despite its high prevalence, approximately 75% of PCOS cases remain undiagnosed in clinical settings due to the complexity of symptoms and limitations of current diagnostic methods. To address this issue, a machine learning-based approach is proposed to improve the accuracy and efficiency of PCOS detection. This study compares the performance of two supervised learning algorithms random forest and gradient boosting for PCOS prediction. The dataset used was obtained from a public repository and contains various clinical features associated with PCOS. To address the class imbalance problem, the synthetic minority over-sampling technique (SMOTE) was applied to the training data. Additionally, bayesian optimization was employed to fine-tune the hyperparameters of each model for optimal performance. Model performance was evaluated using several metrics, with the area under the curve–receiver operating characteristic (AUC-ROC) as the primary measure. The Gradient Boosting model achieved the best results, with an AUC of 0.8983 and a recall of 0.95, indicating high sensitivity in identifying positive PCOS cases. These findings demonstrate that the combination of SMOTE and Bayesian Optimization is effective in enhancing predictive accuracy, especially in imbalanced medical datasets. The proposed approach shows promise for integration into clinical decision-support systems to facilitate earlier and more reliable PCOS screening.Kata Kunci: Bayesian optimization; gradient boosting; PCOS; random forest; SMOTE.Keywords : Bayesian optimization; gradient boosting; PCOS; random forest; SMOTE.
Bayesian Neural Network untuk Prediksi Diabetes: Uncertainty Quantification dalam Machine Learning Sabrina Adnin Kamila; Kusman Sadik; Cici Suhaeni; Agus Mohamad Soleh
Indonesian Journal of Applied Statistics Vol 9, No 1 (2026)
Publisher : Universitas Sebelas Maret

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.13057/ijas.v9i1.103994

Abstract

Penelitian ini bertujuan mengevaluasi dan membandingkan kinerja tiga model machine learning, yaitu random forest (RF), feedforward neural network (FNN), dan bayesian neural network (BNN), dalam klasifikasi diabetes menggunakan Diabetes Health Indicators Dataset dari UCI Machine Learning Repository yang memiliki ketidakseimbangan kelas. Prapemrosesan data meliputi normalisasi fitur menggunakan StandardScaler dan penanganan ketidakseimbangan kelas dengan synthetic minority over-sampling technique (SMOTE). Evaluasi model dilakukan menggunakan metrik akurasi dan skor F1, yang didukung oleh classification report dan confusion matrix. Hasil evaluasi menunjukkan bahwa RF menghasilkan akurasi tinggi (0,8493) namun skor F1 yang rendah (0,3386), yang mengindikasikan rendahnya sensitivitas model terhadap kasus positif diabetes. FNN memberikan performa yang lebih seimbang dengan skor F1 sebesar 0,4490 setelah penyesuaian threshold optimal. Sementara itu, BNN mencapai akurasi 0,8498 dan skor F1 sebesar 0,4043, serta memiliki keunggulan tambahan berupa kemampuan mengukur ketidakpastian prediksi melalui pendekatan Monte Carlo Dropout. Dengan demikian, FNN lebih unggul dalam keseimbangan klasifikasi, sementara BNN lebih relevan untuk aplikasi medis yang membutuhkan informasi tingkat kepercayaan prediksi guna mendukung pengambilan keputusan klinis yang lebih andal.This study aims to evaluate and compare the performance of three machine learning models, namely random forest (RF), feedforward neural network (FNN), and bayesian neural network (BNN), for diabetes classification using the Diabetes Health Indicators Dataset from the UCI Machine Learning Repository, which exhibits significant class imbalance. Data preprocessing includes feature normalization using StandardScaler and class imbalance handling through synthetic minority over-sampling technique (SMOTE). Model performance is evaluated using accuracy and F1-score metrics, supported by classification report and confusion matrix analysis. The results show that RF achieves high accuracy (0.8493) but a low F1-score (0.3386), indicating poor sensitivity to positive diabetes cases. FNN provides more balanced performance with an F1-score of 0.4490 after optimal threshold adjustment. Meanwhile, BNN achieves an accuracy of 0.8498 and F1-score of 0.4043, while offering the additional advantage of uncertainty quantification through Monte Carlo Dropout. Therefore, FNN is more effective for balanced classification performance, while BNN is more suitable for medical applications that require prediction confidence information to support more reliable and informed clinical decision-making.Kata Kunci: Prediksi diabetes, kuantifikasi ketidakpastian, bayesian neural network, classification imbalance, machine learning.Keywords: Diabetes prediction, uncertainty quantification, bayesian neural network, classification imbalance, machine learning.
EVALUASI MODEL KLASIFIKASI DALAM DETEKSI PENIPUAN TRANSAKSI: STUDI KASUS PADA DATA TIDAK SEIMBANG Jefita Resti Sari; Kusman Sadik; Agus M. Soleh; Cici Suhaeni
Jurnal Gaussian Vol 14, No 2 (2025): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.14.2.565-576

Abstract

The rise in digital transactions increases the risk of credit card fraud, highlighting the need for a smart and accurate detection system. This study aims to develop a classification model that can effectively detect fraudulent transactions, despite data imbalance challenges. Data processing involves the use of the SMOTE technique to improve the representation of minority classes and Optuna for hyperparameter tuning. Three machine learning models are applied: Logistic Regression, Random Forest, and XGBoost. Model performance is evaluated using precision, recall, f1-score, and ROC-AUC. The results show that Random Forest achieves the best performance, with a precision of 0.91, recall of 0.74, and f1-score of 0.82. Logistic Regression achieves high recall but very low precision, while XGBoost produces a competitive AUC but a lower f1-score than Random Forest. This research highlights the importance of algorithm selection, data balancing with SMOTE, and parameter tuning to build an effective and adaptive fraud detection system for imbalanced data
Co-Authors Adib Roisilmi Abdullah Agus M Soleh Agus M. Soleh Agus Mohamad Soleh Agus Muhammad Saleh Anang Kurnia Anik Djuraidah Ayu Annisa Rahmah Bagus Sartono Bimandra Adiputra Djaafara Bimandra Adiputra Djaafara Budi Susetyo Eka Setiawaty Fani Fahira Fani Fahira Farit Mochamad Afendi Fitrianto, Anwar Gerry Alfa Dito Gita Lestari Hari Wijayanto Hwan-Seung Yong I Made Sumertajaya Itasia Dina Sulvianti Jefita Resti Sari Kesuma Millati Kusman Sadik La Ode Abdul Rahman Laily Nissa Atul Mualifah Muhammad Shalih Muhammad Yusran Naufalia Alfiryal Noer Endah Islami Panji Lokajaya Arifa Pika Silvianti Rahardiantoro, Septian Rere Kautsar Ristiyanti Ristiyanti Sabrina Adnin Kamila Sabrina Adnin Kamila Uswatun Hasanah Utami Dyah Syafitri Yuli Eka Putri