Claim Missing Document
Check
Articles

KLASIFIKASI PENYAKIT TANAMAN NILAM BERDASARKAN CITRA DAUN MENGGUNAKAN GLCM DAN SVM Sarina; Rizki Yusliana Bakti; Muhammad Faisal; Muhammad Syafaat; Andi Makbul Syamsuri; Muhyiddin AM Hayat; Andi Lukman Anas
PROGRESS Vol 17 No 2 (2025): September
Publisher : P3M STMIK Profesional Makassar

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

Abstract

This study presents a classification model for detecting diseases in patchouli (Pogostemon cablin Benth) leaves using image processing techniques. The method combines Grey Level Co-occurrence Matrix (GLCM) for texture feature extraction and Support Vector Machine (SVM) for classification, optimised using the Particle Swarm Optimisation (PSO) algorithm. A total of 2,080 leaf images were collected and categorized into four classes: healthy, leaf spot, yellowing, and mosaic. Each image was augmented and converted to grayscale to enhance the dataset and reduce computational complexity. Four GLCM features—contrast, correlation, energy, and homogeneity—were extracted to represent leaf textures. The classification model achieved an accuracy of 89.74% using SVM alone, and improved to 97.12% when optimized with PSO. The results indicate that the integration of GLCM, SVM, and PSO provides an effective and accurate solution for early detection of patchouli leaf diseases, potentially supporting farmers in decision-making and improving crop productivity and quality.
IMPLEMENTASI DEEP LEARNING MENGGUNAKAN HYBRID SENTENCE-TRANSFORMERS DAN K-MEANS UNTUK PERBANDINGAN JURNAL Muhammad Asygar Faeruddin; Muhammad Faisal; Rizki Yusliana Bakti; Muhammad Syafaat; Muhyiddin AM Hayat; Andi Makbul Syamsuri; Andi Lukman Anas
PROGRESS Vol 17 No 2 (2025): September
Publisher : P3M STMIK Profesional Makassar

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

Abstract

This study addresses the challenge of identifying semantic relatedness between scientific journal articles by developing a classification system based on deep learning. The system applies an unsupervised learning approach using the Sentence-Transformers model and K-Means clustering to generate semantic similarity scores and categorical labels. Abstracts from journal PDFs are extracted and processed to determine similarity levels across four predefined categories. The optimal number of clusters was determined using Elbow Method, Silhouette Score, and Davies-Bouldin Index, resulting in k = 4. The system is implemented as a web-based application that allows users to upload two PDF files, compare them semantically, and receive both a similarity score and an AI-generated narrative explanation. Functional testing showed that all core features performed as expected. This system significantly reduces the time required to assess relatedness between journal articles, offering an efficient tool for academic research navigation.
IMPLEMENTASI K-MEANS DAN ANALISIS SENTIMEN KRITIK SARAN BERBASIS NLP PADA DATA MONEV BBPSDMP KOMINFO MAKASSAR Syahril Akbar; Muhammad Faisal; Rizki Yusliana Bakti; Muhammad Syafaat; Andi Makbul Syamsuri; Muhyiddin AM Hayat; Andi Lukman Anas
PROGRESS Vol 17 No 2 (2025): September
Publisher : P3M STMIK Profesional Makassar

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

Abstract

Manual analysis of large-scale and unstructured textual feedback data is often inefficient and subjective, thereby hindering data-driven decision-making. This study aims to design and implement an integrated analytical workflow to automatically filter, cluster, and classify feedback data consisting of criticisms and suggestions. The research employs a hybrid approach that begins with TF-IDF-based data filtering, followed by dimensionality reduction using Latent Semantic Analysis (LSA), and topic clustering through K-Means clustering optimized with the Silhouette Score. The resulting cluster labels are then used as training data to build a Multinomial Naive Bayes classification model. The results show that this workflow successfully identified two main thematic clusters, namely "Criticism and Expectations" and "Suggestions and Compliments", and the classification model achieved an overall accuracy of 91%. Although class imbalance affected the recall of the minority class (47%), the model demonstrated high precision (95%) for that class. It is concluded that this hybrid approach effectively transforms raw data into structured insights, and utilizing clustering results as training data is an efficient strategy for automating feedback categorization, providing a reliable tool for institutional analysis.
IMPLEMENTASI HYBRID LEXICON-BASED DAN SVM UNTUK KLASIFIKASI ANALISIS SENTIMEN TERHADAP PELATIHAN BBPSDMP KOMINFO MAKASSAR Nur Alam; Muhammad Faisal; Rizki Yusliana Bakti; Muhammad Syafaat; Andi Makbul Syamsuri; Muhyiddin AM Hayat; Andi Lukman Anas
PROGRESS Vol 17 No 2 (2025): September
Publisher : P3M STMIK Profesional Makassar

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

Abstract

The evaluation of government training programs is often hindered by manual analysis of unstructured qualitative feedback, making the process inefficient and subjective. This study aims to implement and evaluate a sentiment classification model using a hybrid Lexicon-Based and Support Vector Machine approach to analyze participants’ perceptions of the Vocational School Graduate Academy training organized by BBPSDMP Kominfo Makassar, as well as to compare the performance of a standard SVM model with a model optimized using Particle Swarm Optimization. This quantitative research employs 2,313 unstructured review data, which undergo text preprocessing, initial lexicon-based labeling, and TF-IDF feature extraction before being classified using an SVM with an RBF kernel. The results show that the SVM model optimized with PSO consistently outperforms the standard model across all four evaluation aspects, with the most significant accuracy improvement observed in the instructor category from 84.71% to 89.02% and in the assessor category reaching 91.46%. PSO optimization has proven effective in enhancing the model’s ability to identify negative sentiments, which represent the minority class. The hybrid approach with PSO optimization is capable of producing a more accurate and balanced classification system, with practical implications as an objective automated evaluation tool.
IMPLEMENTASI SISTEM DETEKSI PRODUK BOIKOT BERBASIS WEBSITE REAL-TIME MENGGUNAKAN METODE YOLOv10 Ahmad Nur Rahman; Emil Agusalim Habi Talib; Fahrim Irhamna Rachman; Rizki Yusliana Bakti; Muhammad Faisal; Muhammad Syafaat S.Kuba
PROGRESS Vol 18 No 1 (2026): April
Publisher : P3M STMIK Profesional Makassar

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

Abstract

Manual identification ofboycott products remains a challenge for the public due to limited access to information and the complexity of brand affiliations. This study aims to develop a real-time, website-based boycott product detection system using the You Only Look Once version 10 (YOLOv10) algorithm. The dataset consists of images of food and beverage product packaging collected from various online sources, annotated using the bounding box method, and classified into five categories. The model was trained and tested using separate test data, while performance evaluation was conducted using a confusion matrix with precision, recall, and f1-score metrics. In addition, functional testing of the system was performed using the Black Box Testing method. The result indicate that the YOLOv10 model is capable of detecting boycott product with good performance and can be effectively integrated into a real-time web-based system. The proposed system is expected to assist users in identifying boycott products more quickly and accurately.
PERBANDINGAN CNN DAN YOLO PADA SISTEM PENGENALAN WAJAH BERBASIS PRESENSI Nurfadillah; Ida; Darniati; Rizki Yusliana Bakti; Titin Wahyuni; Muhammad Faisal
PROGRESS Vol 18 No 1 (2026): April
Publisher : P3M STMIK Profesional Makassar

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

Abstract

Face recognition based on image data has been widely applied in automated attendance systems; however, it still faces challenges related to accuracy and efficiency under varying lighting conditions and facial pose variations. This study aims to compare the performance of Convolutional Neural Network (CNN) and You Only Look Once (YOLO) methods for face detection and recognition in a deep learning–based attendance system. The dataset consists of facial images collected from students in a limited campus environment with several variations in viewpoint and illumination. The research stages include image preprocessing, training of CNN and YOLO models, and performance evaluation using accuracy, precision, recall, and computation time metrics. The experimental results indicate that YOLO outperforms CNN in terms of detection speed and performance stability, while CNN demonstrates competitive classification performance on limited datasets. This study provides empirical insights into the characteristics of both methods in attendance system scenarios and can serve as a reference for selecting appropriate models for real-world implementation. The main limitations of this study are the dataset size and the restricted data acquisition scope.
Analisis Kontribusi Deteksi Tangan dan Ekstraksi Landmark untuk Pengenalan Alfabet BISINDO Wa Nanda Sulystrian; Muhammad Faisal; Fahrim Irhamna Rachman; Abd Rakhim Nanda; Rizki Yusliana Bakti; Muhammad Syafaat S. Kuba; Andi Makbul Syamsuri
Buletin Sistem Informasi dan Teknologi Islam (BUSITI) Vol 7, No 3 (2026)
Publisher : Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/busiti.v7i3.3506

Abstract

Komunitas tunarungu menghadapi hambatan dalam aspek bahasa dan komunikasi sehingga membutuhkan dukungan teknologi yang sesuai dengan karakteristik visual mereka. Kondisi tersebut menunjukkan pentingnya pengembangan teknologi untuk mendukung akses komunikasi yang lebih inklusif melalui pengenalan Bahasa Isyarat Indonesia (BISINDO). Penelitian ini menganalisis kontribusi deteksi tangan dan ekstraksi landmark terhadap performa pengenalan alfabet BISINDO menggunakan pendekatan berbasis computer vision. Sistem yang diusulkan mengintegrasikan YOLOv8 untuk deteksi tangan, MediaPipe Hands untuk ekstraksi dua puluh satu landmark, normalisasi landmark, serta Multi-Layer Perceptron sebagai model klasifikasi. Evaluasi dilakukan menggunakan dataset berisi 1066 citra alfabet BISINDO dari 26 kelas melalui tiga skenario eksperimen. Hasil penelitian menunjukkan bahwa kombinasi deteksi tangan dan normalisasi landmark menghasilkan performa terbaik dengan nilai accuracy 0.915, precision 0.913, recall 0.915, dan F1-score 0.896, serta meningkatkan accuracy 14.8% dibandingkan pendekatan tanpa deteksi tangan. Temuan ini menunjukkan bahwa deteksi tangan dan representasi landmark berkontribusi penting terhadap peningkatan akurasi sistem. Pendekatan yang diusulkan berpotensi diterapkan pada aplikasi penerjemah alfabet BISINDO berbasis kamera secara real-time untuk mendukung komunikasi yang lebih inklusif bagi komunitas tunarungu di Indonesia.
PENERAPAN ALGORITMA GAUSSIAN NAIVE BAYES UNTUK KLASIFIKASI TINGKAT RISIKO PELANGGAN POTENSIAL PENUNGGAK TAGIHAN LISTRIK PADA DATA PLN Ari Ahmad Dahril; Rizki Yusliana Bakti; Muhammad Faisal
Journal of Computer Science and Information Technology Vol. 3 No. 4 (2026): September
Publisher : Yayasan Nuraini Ibrahim Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70248/jcsit.v3i4.4434

Abstract

Penunggakan pembayaran tagihan listrik pascabayar merupakan salah satu permasalahan yang berdampak pada arus kas PT Perusahaan Listrik Negara (Persero). Mekanisme penanganan tunggakan yang masih bersifat reaktif menyebabkan biaya operasional yang tinggi sehingga diperlukan pendekatan prediktif berbasis data untuk mengidentifikasi pelanggan berisiko sejak dini. Penelitian ini menerapkan algoritma Gaussian Naive Bayes untuk mengklasifikasikan tingkat risiko pelanggan listrik pascabayar PLN UP3 Makassar Selatan ke dalam tiga kategori, yaitu Rendah, Sedang, dan Tinggi, berdasarkan pola historis pembayaran. Data riwayat pembayaran diagregasi menjadi 670 data pelanggan dengan delapan atribut prediktor. Tahapan penelitian meliputi preprocessing data, standardisasi menggunakan StandardScaler, seleksi fitur dengan SelectKBest (ANOVA F-test), pembagian data 80:20 menggunakan Stratified Split, optimasi hyperparameter var_smoothing melalui Bayesian Optimization, serta evaluasi akhir menggunakan 5-Fold Stratified Cross Validation. Hasil pengujian menunjukkan model memperoleh akurasi sebesar 85,07%, dengan precision, recall, dan F1-score tertinggi pada kelas Rendah (0,90; 0,84; 0,87) dan performa lebih rendah pada kelas Tinggi (0,74; 0,57; 0,64) akibat jumlah data yang tidak seimbang. Atribut jumlah pembayaran tepat waktu, jumlah keterlambatan, dan status pembayaran bulan terakhir merupakan fitur yang paling berpengaruh dalam proses klasifikasi. Hasil ini menunjukkan bahwa Gaussian Naive Bayes efektif digunakan sebagai model pendukung Early Warning System bagi PLN dalam mengidentifikasi pelanggan yang berpotensi menunggak. Kata Kunci: Tunggakan Listrik, Gaussian Naive Bayes, Klasifikasi, Risiko Pelanggan, Data Mining.
Integrasi Principal Component Analysis dan DBSCAN untuk Klasterisasi Tingkat Keparahan Stunting pada Balita Azzah Aulia Syarif; Rizki Yusliana Bakti; Muhammad Faisal; Nini Apriani Rumata; Titik Khawa Abd Rahman; Andi Makbul Syamsuri; M. Agusalim
Buletin Sistem Informasi dan Teknologi Islam (BUSITI) Vol 7, No 3 (2026)
Publisher : Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/busiti.v7i3.3512

Abstract

Stunting masih menjadi masalah kesehatan masyarakat yang serius. Variasi karakteristik antropometri pada balita stunting membuat satu kategori klasterisasi tidak cukup merepresentasikan kondisi pertumbuhan secara menyeluruh. Penelitian ini mengusulkan integrasi Principal Component Analysis (PCA) dan DBSCAN untuk mengidentifikasi subkelompok keparahan stunting pada balita di Puskesmas Turikale (Agustus 2025; n = 406). Setelah pembersihan dan standardisasi usia, PCA diterapkan pada HAZ, WAZ, WHZ, dan usia saat  pengukuran, menghasilkan tiga komponen utama (PC1–PC3) yang menjelaskan 99.35% variasi data. DBSCAN pada ruang PCA (ε = 0.60; MinPts = 5; Euclidean) menghasilkan tiga klaster dengan 12.56% noise yang direalokasi menggunakan nearest-centroid agar seluruh sampel berlabel. Hasil menunjukkan tiga subkelompok: stunting sedang dengan underweight pada usia lebih tua (n = 357), stunting sedang disertai risiko gizi lebih pada usia lebih muda (n = 24), dan stunting berat pada usia dini (n = 25). Dibandingkan K-medoids dan Fuzzy C-Means, PCA–DBSCAN memberikan validasi internal terbaik (Silhouette = 0.277; DBI = 1.062; Dunn = 0.0765). Pendekatan ini menghasilkan profil antropometri yang dapat digunakan sebagai alat bantu penentuan prioritas pemantauan dan penilaian gizi di layanan kesehatan primer, tanpa menggantikan klasifikasi WHO.
PERBANDINGAN KINERJA SUPPORT VECTOR MACHINE DAN LONG SHORT-TERM MEMORY DALAM DETEKSI CYBERBULLYING Muhammad Aditya Yudhistira; Rizki Yusliana Bakti; Muhammad Faisal
Nusantara Hasana Journal Vol. 6 No. 3 (2026): Nusantara Hasana Journal, August 2026
Publisher : Yayasan Nusantara Hasana Berdikari

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59003/nhj.v6i3.2300

Abstract

The increasingly massive phenomenon of bullying spreading on social media drives the urgency of applying computational approaches to detect cyberbullying quickly and accurately. This study aims to compare the effectiveness of the Support Vector Machine (SVM) algorithm based on TF-IDF against the Long Short-Term Memory (LSTM) architecture based on Word Embedding in classifying bullying texts. The experiment was conducted utilizing a dataset of 6,520 Indonesian comments extracted from the social media platform X (Twitter). The entire data was then divided into 80% training data and 20% testing data. Performance measurements proved that the SVM model dominated the overall prediction accuracy, achieving an Accuracy of 0.8352 (83%), Precision of 0.8356 (83%), and F1-Score of 0.8363 (83%). Conversely, the LSTM model demonstrated superiority in the sensitivity of positive class identification, scoring a Recall of 0.8552 (85%) and ROC-AUC of 0.9088 (90%). In terms of computational efficiency, the SVM model proved to be far more resource-efficient, requiring only 42.90 seconds of training time and 74.87 MB of memory, in contrast to the LSTM which consumed up to 106.67 seconds and 126.80 MB of memory. This research concludes that SVM is the most optimal method overall because it is able to balance high detection accuracy with excellent operational load efficiency.
Co-Authors . Darniati A M Hayat, Muhyddin Abd Rakhim Nanda Abdul Rakhim Nanda Adrianingsih, Rizka Afil Anugrah Ahmad Nur Rahman Ahmad Risal ANDI AGUNG DWI ARYA BULU Andi Makbul Syamsuri ANDI MAWADDA TAIBA MAWADDA TAIBA Andi Yusri Ari Ahmad Dahril Azzah Aulia Syarif Burhanuddin, Fathurrahman Chatarina Umbul Wahyuni Danuputri, Chyquitha Darniati Desi Anggreani Dewi MJ, Wanda Tyrana Dewi, Syamrilla Emil Agus Salim Habi Talib Emil Agusalim Habi Talib Erika Yanti Fachrim Irhamna Rachman Fahrim I. Rahman Fahrim Irhamna Rachman Fahrim Irhamna Rachman Fahrim Irhmna Rachman Faturohman, Agung Fauzan Azhari Rahman Firdaus Fidaus Galbi Nadifah Hadawina Hadawina Haruna, Hanjas Hayat, Muhyddin A.M Ibnul Imamul Muttaqin Ida Ida Indriani, Lis Indriyanti Azis Iskandar, Aryansyah Ismail, La Ode Taufik Jihan Izzathul Mujidah Kamsurya, Rianita Kazman Riyadi La Ode Taufik Ismail La Ode Taufik Ismail Lukman Lukman Lukman Anas LUKMAN ANAS Lukman Lukman Lukman Lukman Lukman Lukman LUKMAN, LUKMAN M Agusalim M. Fikri Haikal Ayatullah Maharani, Afifah Makmur Jaya Nur Mar'atul Azizah Muh Nur Aqsal Aminullah Muh. Farrel Apta Indratama Muh.Dirham Rahim Muhammad Aditya Yudhistira Muhammad Asygar Faeruddin Muhammad Faisal Muhammad Faturrachman Iswan Muhammad Hasraddin Hasnan Muhammad Syafaat S. Kuba Muhyiddin A.M Hayat Muliana Muslimah, Nurul Aulia Muthalib, Ade Nirwani Abdurahman Mutiara yusuf Nabila Ismail Matta Nandy Rizaldy Najib Nini Apriani Rumata Nizirwan Anwar Nur Alam Nur Rahman, Ahmad Nurfadillah Nurfadillah Nurnawaty Prima Abdiguna, Aidhil Rahmania RAHMANIA Rahmania Rasyidi, Muhammad Fachri Reski Abbas Reski Awalia Rezki Asriani Ridwang Ridwang Ridwang Ridwang Ridwang, Ridwang Sa'ban Sa'ban Salam, Abd Sarina Sri Hastati Suandi Aritmawijaya Sulaeman Suriani Suriani Syahril Akbar Syahrul Ramadhan Syamsul Alam TANTRI INDRABULAN Thariq, Ahmad Titik Khawa Abd Rahman Titik Khawa Abdul Rahman Titin Wahyuni Ulil Amri Usman, Ansharullah Patiroi Utama, Prengki Putra Virgiawan, David Arian Wa Nanda Sulystrian Wibawa. Ar, Arya Yanti, Wilda Yumi