Claim Missing Document
Check
Articles

Found 2 Documents
Search

PER CAPITA CONSUMPTION ESTIMATION IN SURABAYA USING ENSEMBLE MODEL APPROACH Sutikno, Sutikno; Purnomo, Jerry Dwi Trijoyo; Harfianto, Unggul; Irfandi, Yoga Prastya; Anisa, Kartika Nur; Cahyoko, Fajar Dwi
MEDIA STATISTIKA Vol 16, No 2 (2023): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.16.2.170-181

Abstract

The categorization of the Low-Income Community category is based on the poverty indicators in the Multidimensional Poverty Index, including the dimensions of health, education, and living standards. The Proxy Means Test (PMT) can estimate household income or consumption by taking into account household conditions that are readily observable and cannot be manipulated. This method offers the advantage of being capable of determining both the poverty level of a household and the household's characteristics based on asset ownership and socio-demographic conditions. This study aims to estimate per capita consumption using OLS, Robust, Quantile, LASSO, and Ensemble methods. The application of these methods is intended to address various issues, including the presence of outlier data, multicollinearity, and uncertainties. The results indicate that none of the four methods used achieved the highest accuracy based on the MSE, MAE, and sMAPE criteria. Consequently, employing an ensemble model becomes essential to accommodate the element of uncertainty present in these four models. The application of the ensemble method is not only as a comparison between the models, but also as a means to capture the uncertainty contained in each model
INTERPRETABLE MACHINE LEARNING DENGAN PENDEKATAN MODEL AGNOSTIK PADA PREDIKSI FUEL CONSUMPTION RATE MINING HAUL TRUCK Arbianto, Domy Guruh Dwi; Purnomo, Jerry Dwi Trijoyo
JOURNAL OF SCIENCE AND SOCIAL RESEARCH Vol 9, No 1 (2026): February 2026
Publisher : Smart Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54314/jssr.v9i1.5551

Abstract

Abstract: Machine learning (ML) models are frequently characterized as "black boxes" due to their complexity, which renders them difficult for humans to interpret. Model interpretability is crucial for understanding the underlying drivers of specific predictions. In the context of mining operations, explaining the fuel consumption rate (FCR) patterns of mining haul trucks through predictive modeling is essential; however, engineers often struggle to identify the most significant contributors quickly and easily. Because standard ML models do not disclose the logic behind their decisions, engineers face ambiguity when analyzing conditions and prioritizing necessary repairs. Such prioritization is vital, as maintenance costs, technical difficulty, and downtime directly impact productivity. Consequently, a model-agnostic approach is required to bridge this gap. This research aims to develop a predictive model to analyze FCR behavior and patterns, subsequently interpreting them through model-agnostic techniques. The study utilized Vehicle Health Monitoring System (VHMS) data from August 2024 to February 2025, incorporating outlier and multicollinearity management. The Random Forest Regressor (RFR) was employed as the primary machine learning algorithm. Global interpretations were conducted using Partial Dependence Plots (PDP), Feature Interaction, and Permutation Feature Importance, while local interpretations were performed using Local Interpretable Model-Agnostic Explanations (LIME) and Shapley Values (SHAP). The performance evaluation results demonstrate that the RFR predictive model maintains consistent performance regardless of the data treatment applied. The optimal configuration was the RFR model without normalization or outlier handling, achieving an RMSE of 3.7312, a SMAPE of 4.64%, and an R-squared of 0.7936. Both global and local interpretations identified engine speed, road angle, and boost pressure as the top three factors significantly contributing to FCR. Keywords: Fuel Consumption Rate; Interpretable Machine Learning; Agnostic Model; Random Forest Abstrak: Model machine learning (ML) sering disebut sebagai “Black-Box” karena kerumitannya sehingga sulit diinterpretasikan oleh manusia. Interpretabilitas model menjadi sangat penting untuk memahami penyebab sebuah prediksi tertentu dibuat. Salah satunya dalam memahami perilaku dan menjelaskan pola fuel consumption rate (FCR) dari mining haul truck menggunakan model prediksi. Seorang engineer akan kesulitan untuk menentukan kontributor paling signifikan secara mudah dan cepat. Pada sebuah prediksi, sebuah model ML tidak akan memberi tahu bagaimana sampai pada sebuah keputusan. Hal ini akan menimbulkan kebingungan engineer pada saat akan menganalisa kondisi dan menentukan prioritas perbaikan yang diperlukan. Prioritisasi perbaikan perlu dilakukan karena pertimbangan biaya, tingkat kesulitan, dan downtime yang sangat mempengaruhi produktivitas. Oleh karena itu, pendekatan model agnostik perlu dilakukan. Penelitian ini bertujuan untuk menghasilkan model prediksi untuk memahami perilaku dan pola FCR kemudian menginterpretasikannya dengan model agnostik. Penelitian ini menggunakan data Vehicle Health Monitoring System (VHMS) dari Agustus 2024 hingga Februari 2025 dengan penanganan outlier dan multikolinieritas. Algoritma ML yang digunakan adalah Random Forest Regressor (RFR). Model agnostik yang digunakan untuk interpretasi global adalah Partial Dependence Plot (PDP), Feature Interaction, dan Permutation Feature Importance. Sedangkan interpretasi lokal menggunakan Local Interpretable Model-Agnostic Explanations (LIME) dan Shapley Value (SHAP). Hasil evaluasi performa model menunjukkan bahwa model prediksi RFR memiliki performa yang konsisten bagaimanapun perlakuan data diterapkan. Model prediksi terbaik yang dipilih adalah model RFR Tanpa Normalisasi – Tanpa Penanganan Outlier dengan nilai RMSE 3,7312, SMAPE 4,64%, dan R-Squared 0,7936. Hasil interpretasi global dan lokal menunjukkan bahwa top three faktor yang berkontribusi signifikan terhadap FCR adalah engine speed, road angle, dan boost pressure. Kata Kunci: Fuel Consumption Rate; Interpretable Machine Learning; Model Agnostik; Random Forest