Héctor Raúl Chávez

Ms. Héctor Raúl Chávez

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 November 2025 in Energies
Verified
Journal Article
Open access
PDF file
TMThomas Monteiro
Thomas Monteiro
GCGlaucio José Bezerra Cavalcante Castor
Glaucio José Bezerra Cavalcante Castor
CCCarlos Castillo Correa
Carlos Castillo Correa
Héctor Raúl ChávezHéctor Raúl Chávez
Héctor Raúl Chávez
Dionicio Zocimo Ñaupari HuatucoDionicio Zocimo Ñaupari Huatuco
Dionicio Zocimo Ñaupari Huatuco
YRYuri Percy Molina Rodriguez
Yuri Percy Molina Rodriguez
This paper proposes a hybrid machine learning framework for detecting electricity fraud within the broader context of Non-Technical Losses (NTLs) in power-distribution systems. The framework combines unsupervised anomaly detection using Isolation Forest with supervised classification through XGBoost, exploiting the complementary strengths of both algorithms. Using real consumption data from a Peruvian utility, the approach integrates domain-informed feature engineering to capture behavioral, temporal, and contextual indicators of irregular usage. To address the extreme class imbalance inherent to fraud datasets, the SMOTETomek hybrid resampling technique was applied, enhancing minority-class representation and decision boundary clarity. Experimental results achieved high predictive performance on the test set (AUC-ROC = 0.999, F1-score = 0.77) using an optimized decision threshold of 0.6. Moreover, SHAP-based interpretability analysis identified extreme monthly variations, prolonged low-consumption periods, and tariff category as key behavioral predictors of fraudulent activity. The robustness of the proposed framework was further validated through a 5-fold cross-validation procedure during the training phase, ensuring consistent performance across different data partitions. Overall, the proposed framework demonstrates not only robust and explainable performance but also practical operational value, providing utilities with a scalable data-driven tool to optimize inspection strategies and maximize recovery of non-technical losses.

19 January 2025 in Energies
Verified
Journal Article
Open access
PDF file
Héctor Raúl ChávezHéctor Raúl Chávez
Héctor Raúl Chávez
YRYuri Percy Molina Rodriguez
Yuri Percy Molina Rodriguez
This paper proposes an innovative methodology for geospatial forecasting of electrical demand across various consumption segments and scales, integrating machine learning and discrete convolution within the framework of global system projections. The study was conducted in two phases: first, machine learning techniques were utilized to classify and determine the relative growth of segments with similar consumption patterns. In the second phase, convolution methods were employed to produce accurate spatial forecasts by incorporating the influence of neighboring areas through a “core matrix” and accounting for geographical constraints in regions with and without consumption. The proposed approach enhances the precision of spatial forecasts, making it suitable for large-scale distribution systems and implementable within short timeframes. The proposed method was validated using data from a Peruvian distribution system serving over one million users, employing 204 historical records and analyzing three georeferenced consumption segments at scales of 1:10,000, 1:1000, and 1:100. The results demonstrate its effectiveness in forecasting across different time horizons, thereby contributing to improved planning of electrical infrastructure.