Executive Summary

January 28, 2026

Machine learning in the prediction of reimbursements for electrical damages

Benjamim Abreu de Souza; Wilson Tarantin Junior

Summary prepared by the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege focused on synthesis and writing.

The supply of electrical energy, an essential service, is subject to interruptions that generate complaints and requests for compensation for electrical damage, often associated with environmental factors (ANEEL, 2024). This study employs machine learning techniques to model and predict the occurrence of these requests in municipalities of a utility company in Southeast Brazil. The research adopts an exploratory approach, combining quantitative and qualitative data from public sources, including meteorological, socioeconomic, and energy distribution technical information (Cipoli, Nogueira e Rios, 2003). The modeling was performed in the Spyder environment, using the Python language and libraries such as Sklearn and Statsmodels (Fávero e Belfiore, 2024). The hierarchical structure of the data was considered at two levels: municipalities within climatologically homogeneous regions (INCAPER, 2025).
The methodology involved collecting data from various sources. The first, named indgercomercial, accessed the Management Indicators of Distribution (INDGER) from ANEEL via API, covering the period from 2023 to 2024. The collected variables included the municipality’s IBGE code, the number of active consumer units, and the number of damage compensation requests (ANEEL, 2025). The second collection, interrupcoesaneel, also used Python code to access data on interruptions per energy distribution substation, also between 2023 and 2024. The information obtained detailed the substation, the interruption order number, the start and end dates, and the generating event (ANEEL, 2024). A third database, dados_climaticos, was obtained from the National Institute of Meteorology (INMET), containing meteorological information from 12 stations in Espírito Santo, organized by month and year (INMET, 2025). Additionally, the study incorporated the Socioeconomic Index (INSE) of the municipalities, made available by the state government (INEP, 2023).
To model the compensation requests, supervised machine learning techniques were applied: Multilevel regression modeling, count data regression model, and Random Forest (Géron, 2019). Multilevel modeling, suitable for data with a nested structure, allowed for the analysis of individual and between-group variations (Fávero e Belfiore, 2024). The count data regression model used the Negative Binomial distribution, appropriate for non-negative integers with overdispersion (Cameron e Trivedi, 1990). The Random Forest algorithm, in turn, aggregated a set of decision tree predictors (Géron, 2019). Cross-validation of the models used 5 folds and the R² coefficient of determination as a performance metric.
The results revealed that the Multilevel model outperformed the Negative Binomial and Random Forest models, although all provided relevant contributions. Municipal characteristics, such as the number of consumer units and the socioeconomic index, proved to be statistically significant in at least one of the models, as did the average wind speed, a parameter of climatologically homogeneous regions. Exploratory data analysis identified a high correlation between predictor variables such as consumer units, interruptions, wind speed, gusts, humidity, pressure, and temperature (Vasconcellos e Alves, 2000). Multilevel modeling found heterogeneity in compensation requests among regions, with the intercept variance referring to the regions being significantly different from zero. The final model, after Box Cox transformation, showed statistical significance for the variables units, wind speed, and wind gusts. The Negative Binomial model also considered the INSE variable as statistically significant. The Random Forest model, in turn, indicated the importance of electrical energy interruptions, although without statistical significance in the other models.
The joint analysis of the models indicated that the variables with statistical significance in the Multilevel and Negative Binomial models have the potential to influence compensation requests. The number of consumer units with active contracts in the municipality emerged as the most important variable, followed by wind speed and gusts, and the socioeconomic index of the municipalities. The application of different modeling techniques combined with public data proved sufficient to investigate the factors influencing requests for compensation for electrical damage. It is concluded that the objective was achieved: it was identified that the number of consumer units, wind speed, and socioeconomic index are determining factors for the occurrence of requests for compensation for electrical damage, with the Multilevel model presenting the best predictive performance.

Bibliographic References:

Agência Nacional de Energia Elétrica [ANEEL]. 2021a. Módulo 9 do PRODIST: Procedimentos de Distribuição de Energia Elétrica no Sistema Elétrico Nacional. Disponível em: <https://www2.aneel.gov.br/cedoc/aren202195628.pdf>. Acesso em: 16 mar. 2025.
Agência Nacional de Energia Elétrica [ANEEL]. 2021b. Resolução Normativa nº 1000, de 2021. Diário Oficial da União, Brasília, 23 dez. 2021. Disponível em: <https://www2.aneel.gov.br/cedoc/ren20211000.html>. Acesso em: 16 mar. 2025.
Agência Nacional de Energia Elétrica [ANEEL]. 2024. Qualidade do Fornecimento de Energia Elétrica. Disponível em: <https://www.gov.br/aneel/pt-br/assuntos/distribuicao/qualidade-do-fornecimento-de-energia-eletrica>. Acesso em: 16 mar. 2025.
Agência Nacional de Energia Elétrica [ANEEL]. 2025. Indicadores Gerenciais da Distribuição. Disponível em: <https://dadosabertos.aneel.gov.br/dataset/indger-indicadores-gerenciais-da-distribuicao>. Acesso em: 16 mar. 2025.
Brasil. Instituto Nacional de Estudos e Pesquisas Educacionais Anísio Teixeira [INEP]. 2023. Saeb 2021: Indicador de Nível Socioeconômico do Saeb 2021: nota técnica. Inep, Brasília, DF, Brasil.
Cameron, A. C.; Trivedi, P. K. 1990. Regression-based test for over dispersion in the Poisson model.Journal of Econometrics, v.46, n. 3, p. 347 – 364.
Cipoli, J.A.; Nogueira, E.D.B.; Rios, A.J.S. 2003. Ressarcimento de Danos Elétricos. Disponível em: <http://www.mfap.com.br/pesquisa/arquivos/20090428131506-Ressarcimento%20de%20Danos%20Eletricos.pdf>. Acesso em: 28 fev. 2025.
Fávero, L.; Belfiore, P. 2024. Manual de análise de dados: estatística e machine learning com Excel, SPSS, Stata, R e Python. 2ed. LTC, Rio de Janeiro, RJ, Brasil.
Géron, A. 2019. Mãos à obra: aprendizado de máquina com Scikit-Learn e TensorFlow. Alta Books, Rio de Janeiro, RJ, Brasil.
Instituto Capixaba de Pesquisa, Assistência Técnica e Extensão Rural [INCAPER]. 2025. Legenda de Regiões Meteorológicas. Disponível em: <https://meteorologia.incaper.es.gov.br/legenda_regioes>. Acesso em: 25 set. 2025.
Instituto Nacional de Meteorologia [INMET]. 2025. Glossário. Disponível em: <https://portal.inmet.gov.br/glossario/glossario>. Acesso em: 29 set. 2025.
Vasconcellos, M.; Alves, D. 2000. Manual de econometria. Atlas, São Paulo, Brasil.

Executive summary from the Final Project of the Specialization in Data Science and Analytics from the MBA USP/Esalq

Learn more about the course; click here:

Who edited this article

Most recent

You may also like

October 27, 2026

Effectiveness of the Brazilian Central Bank’s monetary policy and its possible social costs (1999-2024)

The conduct of monetary policy in Brazil, particularly under the inflation targeting regime, has been widely debated due to its possible impacts on inflation and economic activity. The study aimed to evaluate the effectiveness of Brazilian monetary policy, guided by the Selic rate, in controlling inflation, and to investigate its possible social costs during the period from 1999 to 2024. For this purpose, quarterly data on Gross Domestic Product (GDP), Broad Consumer Price Index (IPCA), Selic rate, and exchange rate, extracted from the Central Bank of Brazil, were used. The methodology employed a Vector Error Correction Model (VECM), preceded by stationarity and cointegration tests, and complemented by Impulse Response Function and Variance Decomposition analyses. The results indicated that inflation showed a strong inertial component, being predominantly explained by its own dynamics, while shocks in interest rates and exchange rates demonstrated a statistically limited impact on the IPCA. Additionally, it was found that GDP did not respond significantly to variations in the Selic rate, suggesting low sensitivity of economic activity to monetary policy during the analyzed period. In summary, it was concluded that monetary policy demonstrated limited effectiveness in controlling inflation, without, however, generating direct and significant social costs on Gross Domestic Product, which acts as the main driver of job creation.

Keywords: Cointegration; Econometrics; Inflationary inertia; Inflation targets; Gross Domestic Product.

October 27, 2026

Financial efficiency and risk mitigation in legal departments

The research investigated how governance mechanisms and internal controls, applied to “Legal Operations” practices, can strengthen the financial management and budgetary predictability of a legal department in a multinational company. The study aimed to diagnose critical points of the legal budget cycle, from forecasting to payment, and identify process improvements, indicators, and responsibilities between the legal and financial departments. To this end, an applied qualitative research was conducted, configured as a case study. Data collection included interviews with professionals from the legal and financial areas, documentary analysis of cost spreadsheets (2022-2024) and compliance policies, as well as mapping of the legal expenses flow. The results revealed fragilities in interdepartmental coordination, notably in the definition of cost centers and the use of budgetary control instruments. Gaps were observed in process monitoring and in the integration of information from external firms, as well as opportunities for routine automation and improvement of management indicators. Budgetary predictability was found to be impacted by exogenous factors, such as the dynamics of the Judiciary, and endogenous factors, such as the quality of information and the governance of the budget cycle. It was concluded that budgetary predictability depends on the maturity of governance applied to processes, data, and operational routines. A governance and control model for the legal budget was proposed, focusing on process standardization, strengthening monitoring, and improving decision-making, including process mapping, risk and control matrix, and performance indicators.

Keywords: Automation; Internal controls; Corporate governance; Legal Operations; Budgetary predictability.

October 27, 2026

Comparison of prepayment risk models in fixed-rate credit portfolios

The risk of early settlement in pre-fixed credit portfolios constituted a relevant challenge for Asset and Liability Management (ALM) in financial institutions, as it generated cash flow asymmetries and inefficiencies in hedging strategies. The work aimed to comparatively evaluate the predictive performance of traditional econometric and statistical models versus machine learning algorithms in estimating the Monthly Mortality Rate (SMM). To this end, a real database of credit operations originated between 2016 and 2025 was used, totaling approximately 500 million observations. Models such as Historical CPR, Beta Regression, Quantile Regression, Logistic Regression, Cox Proportional Hazards, and LightGBM were evaluated, with performance measured by Mean Absolute Error (MAE) and validated through temporal cross-validation with an expanded window. The results demonstrated the superiority of non-linear approaches, with the LightGBM model showing the best performance in the test period (Jan/2024 to Jul/2025), reducing MAE by 34.21% compared to the historical benchmark. Additionally, interpretability techniques based on SHAP values and survival analysis were employed to identify the contract rate spread, market rate, and bond age as the main economic determinants of prepayment. The application of the proposed model improved the accuracy of cash flow estimates and provided favorable evidence for refining the calculation of effective duration and the sizing of the institutional liquidity buffer, suggesting that more flexible models can optimize ALM management.

Keywords: Machine learning; ALM Management; Hedge; Financial modeling; Interest rate risk.

October 27, 2026

Imperative vs. reactive architecture: impact on debugging and error correction in iOS mobile development

The development of iOS mobile applications can be conducted under imperative and reactive architectures, directly impacting debugging and maintenance practices. The objective was to establish the impact of architectural choice on debugging and error correction in iOS applications, evaluating the automatic detectability of failures, the traceability of reports, and the utility of logs. An experimental study was conducted, in which two functionally equivalent native applications were developed with UIKit (imperative) and SwiftUI/Combine (reactive). Four controlled error scenarios were inserted, classified as explicit crashes and silent errors. Data were collected with Firebase Crashlytics and Xcode Debugger in forty sessions, and these were analyzed by descriptive statistics and content analysis. The results revealed that, for state inconsistencies, the UIKit architecture generated immediate crashes and automatic reports, while SwiftUI suppressed the crash and produced a silent warning invisible to automatic tools. For forced unwrapping, the architecture proved irrelevant. In silent scenarios, neither architecture offered automatic observability, which highlighted the limitations of the tools. It was concluded that the reactive architecture, by abstracting state control mechanisms, reduced explicit errors but introduced silent failures with lower traceability. The architectural decision must consider the implications for production monitoring, requiring manual log instrumentation in SwiftUI/Combine projects for errors that would be automatically captured in UIKit.

Keywords: Combine; Maintainability; Observability; SwiftUI; UIKit.

October 27, 2026

CRM IN DATA RETENTION FOR CUSTOMER POTENTIAL ANALYSIS IN THE B2B HEALTH SECTOR

Commercial management in the B2B healthcare sector faced constant challenges, requiring strategic vision and sharp attention to delivering results. The study investigated how the implementation of a CRM (Customer Relationship Management) system contributed to mitigating the loss of historical information and enhancing customer retention in the face of selling costs. A qualitative-quantitative approach of an exploratory-descriptive nature was adopted, and data were collected through a questionnaire answered by 42 professionals in the commercial area, working in the industry or in healthcare segment companies. The results revealed that, although 71.4% of respondents used computerized CRM, organizational culture imposed strong resistance, confirmed by 57.1% of professionals who relied on manual methods, such as diaries and isolated spreadsheets, for data recording and retention. A gap in training was also evident, as 24.4% of respondents rarely or never received specific training. The research inferred that the underreporting of data was not solely due to a lack of technology or technological failure, but to a deficiency in the strategic management of people and processes. The vision of the need for integrated management to generate competitive intelligence and sales predictability was reinforced, and attention was drawn to the need for more frequent training and development. Mitigating the cost of selling and enhancing customer retention depend on consistent integration between human capital and technology, generating managerial assets to increase commercial production and strategic data retention.

Keywords: Training; Sales strategy; Customer management; Strategic management; Commercial intelligence.

October 27, 2026

Evaluation of the application of Scrum in improving efficiency and adaptability in technology projects

The growing complexity of technology projects demanded approaches capable of balancing control, efficiency, and adaptability to constant changes, as traditional management models showed limitations in dynamic environments. Technology project management was evaluated before the implementation of an agile method, improvements based on iterative practices were proposed, and post-implementation performance was reassessed, seeking to identify gains in operational efficiency and adaptability throughout the project lifecycle. The research employed a descriptive case study, with a qualitative approach, in two technology sector companies. Data were collected through structured questionnaires applied before and after the intervention, complemented by documentary analysis and systematic observation over ninety days of solution application. The results showed consistent improvement in planning flexibility, greater alignment with client needs, increased delivery predictability, and enhanced team productivity. A productivity gain was the most expressive indicator, with 85.7% total agreement on the improvement of team performance after adopting the *framework*. It was concluded that the structured implementation of the method improved the management of the analyzed projects, reinforcing the capacity to respond to changes and promoting greater business value generation.

Keywords: Organizational adaptation; Incremental delivery; Agile management; Iterative planning; Team productivity.

October 27, 2026

Challenges in the operationalization of rural credit: perception of employees of a financial institution in the Southeast of Brazil

Rural credit is a crucial instrument of Brazilian agricultural policy, essential for the viability of agricultural production cycles and the development of the sector. Despite its relevance, the operationalization of rural credit has faced significant challenges in financial institutions, stemming from regulatory complexity, documentary bureaucracy, and risks inherent to agricultural activity. In this context, the study aimed to analyze the perception of financial institution employees regarding the rural credit release process, identifying difficulties, operational challenges, and opportunities for improvement. A descriptive quantitative approach was adopted, with the application of a structured questionnaire to 37 employees who work directly or indirectly in the granting of rural credit. The sample was intentional, and the data were analyzed descriptively, with the support of tabulation and graphical representation. The results indicated that the main challenges faced by financial institutions were related to excessive bureaucracy, collateral requirements, risk analysis, limitation of agent autonomy, and the need for greater technical training. External factors, such as producer indebtedness, adverse climate events, and high production costs, also negatively impacted access to credit. It was concluded that the improvement of internal processes, combined with continuous training of professionals and coordination between regulatory agents and financial institutions, can contribute to making rural credit more efficient and accessible.

Keywords: Agribusiness; Bureaucracy; Agricultural policy.

October 27, 2026

Institutional Dissonance and Privacy Barriers in the Adoption of DREX in Brazil

The perception of different stakeholders on the implementation of DREX (Real Digital) in the Brazilian financial system was analyzed, focusing on the expected impacts on security, privacy, and adoption. The study aimed to investigate the perceptions of distinct groups regarding the risks and opportunities of DREX, offering subsidies for decision-making by financial institutions and regulators. To this end, three questionnaires were applied to 415 participants, divided among professionals in finance and technology (81), cryptoasset investors (149), and the general public (185). The collected data were tabulated and subjected to quantitative descriptive analysis, complemented by a comparative analysis between groups. A consistent institutional dissonance was identified, where technical trust in the Central Bank surpassed political trust in the Federal Government. Among cryptoasset investors, an ideological rejection of DREX was observed, driven by concerns about traceability and governmental control, with an inverse relationship between technical knowledge and adoption intention. The general public demonstrated pragmatic willingness, conditioned on the absence of explicit perception of financial surveillance, a pattern termed “Pix-Privacy Gap”. Financial privacy emerged as the main transversal obstacle to adoption across all groups, and a 1% cashback incentive proved insufficient to overcome this resistance. It was concluded that voluntary adoption of DREX tends to be fragmented and conditional, depending on structured communication initiatives, digital financial education, and auditable privacy architectures that reinforce the credibility and independence of the Central Bank before society.

Keywords: CBDC; Institutional trust; Risk perception; Traceability; Financial system.

October 27, 2026

Application of Six Sigma DMAIC and Gemba Walk for the analysis of mechanical availability of mobile equipment

The inefficient management of mobile equipment maintenance routines in infrastructure works had directly evidenced impacts on mechanical availability indicators and productive performance. The study aimed to analyze the mechanical availability of equipment in a highway superstructure work, employing the Six Sigma DMAIC cycle and the Gemba Walk practice. The adopted methodology was characterized as an applied case study, of a mixed nature. Data from operational reports were collected and processed, mechanical availability indicators were calculated, and qualitative analyses were conducted through the Ishikawa Diagram. In the Define phase, routines were mapped and historical data were collected. In Measure, mechanical loss records were standardized and mechanical availability was calculated, identifying critical equipment with performance below 85%. The Analyze stage used the Ishikawa Diagram to identify root causes, and in the Improve phase, improvement actions and an implementation schedule were structured. The results revealed that certain equipment families presented performance below the acceptable threshold, with the asphalt plant standing out for its unavailability of 79%, directly impacting production. The main identified causes included the absence of preventive planning, failures in maintenance management, and the unavailability of critical inputs. It was concluded that the integrated application of the methodologies allowed not only to identify the causes of low availability but also to structure a management model focused on operational reliability and continuous improvement, proposing the transition from a reactive model to proactive, field-based maintenance.

Keywords: Operational reliability; Asset management; Performance indicators; Maintenance; Continuous improvement.