Claim Missing Document
Check
Articles

Found 33 Documents
Search

Optimasi Ambang Batas Sensitif Biaya untuk Pemeliharaan Prediktif Menggunakan CSTOLR, CSTORF, dan CSTOXGB: Cost-Sensitive Threshold Optimization for Predictive Maintenance Using CSTOLR, CSTORF, and CSTOXGB Rahman, Abi Rafdi Hasbur; Dhenabayu, Riska; Kautsar, Achmad; Safitri, Anita
MALCOM: Indonesian Journal of Machine Learning and Computer Science Vol. 6 No. 4 (2026): MALCOM October 2026
Publisher : Institut Riset dan Publikasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.57152/malcom.v6i4.2846

Abstract

Ketidakseimbangan kelas dalam Predictive Maintenance (PdM) dapat menyebabkan model Machine Learning bias terhadap kelas mayoritas sehingga kegagalan mesin tidak terdeteksi. Risiko meningkat karena False Negative (FN) menimbulkan kerugian lebih besar dibandingkan False Positive (FP). Penelitian ini mengevaluasi integrasi Cost-Sensitive Learning (CS) dan Threshold Optimization (TO) untuk meminimalkan Total Misclassification Cost (TMC) pada berbagai rasio biaya FN terhadap FP. CS menerapkan pembobotan biaya selama pelatihan, sedangkan TO menentukan ambang keputusan optimal berdasarkan data validasi. Logistic Regression (LR), Random Forest (RF), dan Extreme Gradient Boosting (XGBoost) dikembangkan menjadi CSTOLR, CSTORF, dan CSTOXGB melalui kombinasi kedua pendekatan tersebut. Eksperimen menggunakan dataset sintetis AI4I 2020 Predictive Maintenance dengan rasio biaya 5:1, 10:1, 20:1, dan 50:1. Kinerja dibandingkan pada model tanpa pembobotan biaya, dengan pembobotan biaya, serta kombinasi pembobotan dan optimasi ambang batas. Hasil menunjukkan integrasi CS dan TO konsisten menurunkan TMC pada ketiga model. CSTOXGB menghasilkan TMC terendah sebesar 41 pada rasio 10:1, dengan ambang optimal 0,38 dan satu FN, atau turun 51,2% dibandingkan XGBoost dasar. Model ini juga mencatat TMC terendah atau setara pada skenario. CSTOLR mengurangi FN secara signifikan, tetapi menghasilkan FP tinggi. Temuan ini menunjukkan efektivitas integrasi CS dan TO dalam menangani klasifikasi PdM dengan biaya kesalahan asimetris.
SHAP-Interpretable Cross-Platform Machine Learning for Classifying High-Performing Influencers Using Ratio-Based Engagement Metrics Hakim, Muchammad Luqmanul; Dhenabayu, Riska; Arifah, Ika Diyah Candra; Safitri, Anita
MALCOM: Indonesian Journal of Machine Learning and Computer Science Vol. 6 No. 4 (2026): MALCOM October 2026
Publisher : Institut Riset dan Publikasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.57152/malcom.v6i4.2887

Abstract

This study aims to develop and evaluate a Shapley Additive exPlanations (SHAP)-interpretable cross-platform machine learning framework for classifying high-performing influencer segments on TikTok and YouTube using ratio-based engagement metrics. The study addresses limitations of follower or subscriber-based evaluation by constructing a platform-stratified performance label and applying leakage-controlled feature selection to reduce circular prediction bias. A total of 2,000 influencer records, consisting of 1,000 TikTok and 1,000 YouTube influencers, were analyzed using Random Forest, Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and Support Vector Machine (SVM). Model performance was evaluated using accuracy, macro precision, macro recall, macro F1-score, and Receiver Operating Characteristic–Area Under the Curve (ROC-AUC), while SHAP was used to interpret feature contributions. The tuned CatBoost model produced the highest observed performance, with 0.9925 accuracy, 0.9904 macro F1-score, and 0.9995 ROC-AUC. Additional validation showed stable model ranking across alternative thresholds, while proxy leakage ablation and cross-platform transfer indicated that results should be interpreted as internal classification consistency rather than full platform-independent generalization. SHAP analysis showed that ratio-based engagement variables, particularly like ratio and engagement rate, contributed more strongly than raw audience size. The findings demonstrate that influencer performance can be represented as an interpretable
When Content Creates Anxiety: The Role of FOMO in Mediating Content Marketing and UGC on Netflix Purchase Decisions Ilyas, Marta Maura Olivia; Kharisma, Fresha; Arifah, Ika Diyah Candra; Dhenabayu, Riska
Jurnal Manajemen Pemasaran dan Perilaku Konsumen Vol. 5 No. 3 (2026)
Publisher : Fakultas Ekonomi dan Bisnis Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/jmppk.2026.05.3.25

Abstract

Purpose – This study examines how content marketing and user-generated content (UGC) influence Netflix subscription choices, with fear of missing out (FoMO) as a mediating psychological mechanism, addressing inconsistencies in previous findings regarding their direct effects on purchase decisions. Design/Methodology/Approach – A quantitative causal-associative design was applied to active Netflix subscribers in Indonesia. Data were collected through a structured online questionnaire using purposive sampling, resulting in 218 valid responses from 260 initial cases. Data were analyzed using PLS-SEM in SmartPLS 3, with significance assessed through 10,000-subsample bootstrapping using the BCa method. Findings – All seven hypotheses were supported. Content marketing and UGC significantly influenced both FoMO and purchase decisions, while FoMO significantly mediated the relationships between content marketing and purchase decisions and between UGC and purchase decisions. Originality/Value – The findings identify FoMO as a psychological mechanism that helps explain contradictory results in prior research on content marketing, UGC, and purchase decisions. The study also highlights the value of UGC-driven strategies in creating organic social urgency to encourage subscription decisions.
Co-Authors Achmad Fitro Adellia Nirma Azzahra Akbar, Muhammad Farhan Anita Safitri Arifah, Ika Diyah Candra Armawadin, Irsyad Daffa Audika Briantari Zaen Candra, Ika Diyah Danang Ary Dewangga Dendra Aptanata Nugraha Dhiyan Septa Wihara Dwi Wahyuni Erta Erwan Nicolas Fazlurrahman, Hujjatullah Febi Abdillah Achmad Febrinita, Filda Fitro, Achmad Fresha Kharisma Fresha Kharisma Hadi, Hafid Kholidi Hakim, Muchammad Luqmanul Hapsari Shinta Citra Puspita Dewi Hesty Puspita Sari Hujjatullah Fazlurrahman Husin, Nanang Ika Diyah Candra Arifah Ilyas, Marta Maura Olivia Irsyad Budhiraja Irsyad Daffa Armmawadin Izzati, Haidi Fildzah Kamila, Fanisa Ayu Cahya Kautsar, Achmad Kezia Anatasya Caroline Manalu Kurnia Paranita Kartika Lestanti, Sri M. Syamsul Hidayat Muhammad Ariq Fadhil Muhammad Ariqoh Bahrul Ulum Muhammad Fachmi Muhammad Fajar Wahyudi Rahman Muhammad Febrilian Dwi Syahputra Muhammad Jiwa Akmal Sani Arman Rachman Liga Musdholifah Musdholifah Najmi Khairunnisa Nanang Hoesen Hidroes Abbrori Naufal Rifqi Nasution Nurul Fajrin Ariyani Nurul Indawati Priambodo, Gumilang Purwohandoko Purwohandoko Putri Sabilla Qosimah, Rofiatul R.A. Sista Paramita Rahman, Abi Rafdi Hasbur Ramadhan Maruta Pradana Ramadhan, Gilang Bagus Ratih Amelia Ratih Amelia Renny Sari Dewi Rosmalia, Tria Rizky Ryan Ahmad Syah Putra Safitri, Anita Susi Handayani Suwarno, Arrahma Carine Syahda Hati Dendra Thaariq Ayub Arrafah Thwin Htoo Aung Tias Andarini Indarwati Tri Sudarwanto Trias Madanika Kusumaningrum Ulil Hartono Wahyu Suci Saputri Yoga Yuniadi Yusniarsi Primasari