Comparative Analysis of Credit Risk Models and the Impact of Their Variables on Their Predictions

doi:10.22167/2675-6528-202602838

Article

September 30, 2026

Comparative Analysis of Credit Risk Models and the Impact of Their Variables on Their Predictions

Comparative Analysis of Credit Risk Models and the Impact of Their Variables on Their Predictions

Domingos Soares do Carmo Neto; Felipe Pinto da Silva

DOI: 10.22167/2675-6528-202602838

Article derived from a Final Course Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.

Summary

The growing non-payment in the financial system and the limitations of traditional credit assessment methods have highlighted the need for more robust and interpretable predictive models. The study compared the performance of three supervised machine learning algorithms — Logistic Regression, Decision Tree, and Artificial Neural Network (MLP) — applied to three public credit risk datasets: Statlog German Credit Dataset, Default of Credit Card Clients, and Give Me Some Credit. Data preprocessing was performed, including imputation by median, binary encoding, standardization, and strategies for imbalanced classes. The models were evaluated by accuracy, precision, recall, F1-score, and AUC-ROC, and subjected to hyperparameter optimization with five-fold cross-validation. Prediction interpretability was conducted using the SHAP method. The results revealed that no algorithm demonstrated consistent superiority in all scenarios. Logistic Regression achieved the best AUC-ROC on the German Credit dataset (0.803), the Decision Tree led on the Give Me Some Credit dataset (0.849), and the MLP stood out on the Taiwan Credit Card dataset (0.765). The SHAP analysis indicated that recent financial behavior, revolving credit utilization, and checking account availability were consistently the most predictive variables. It was concluded that the multi-model approach, combined with SHAP interpretability analysis, offers a more complete and reliable framework for credit risk classification than the isolated use of any single algorithm.

Keywords: Machine learning; Default classification; Predictive models; Credit risk; SHAP.

1. Introduction

The growing complexity of the financial system and the increase in default rates have made credit risk management one of the main challenges for financial institutions. In this context, assessing the payment capacity of credit takers assumes a strategic role, as mistaken decisions can expose lenders to significant losses and compromise the stability of the financial system as a whole (Sanches, Vander Lúcio et al., 2018).

Credit analysis plays a central role in this process, allowing for efficient resource allocation and delinquency mitigation. According to Joseph and Ciby (2013), it serves three fundamental purposes: revealing the borrower’s financial risks, guiding the operational and commercial analysis of the business, and determining the extent of the necessary financial support, especially for intermediaries such as banks.

Credit analysis is based on the evaluation of several factors, such as financial history, income, and customers’ payment behavior, with the objective of predicting the probability of non-payment. The importance of this study intensified after the 2008 financial crisis in the United States, when the collapse of high-risk mortgage-backed securities revealed flaws in traditional credit assessment models (Reinhart & Rogoff, 2009). The crisis highlighted the need for improvement in risk measurement methodologies, driving research into predictive models based on statistical techniques and machine learning.

Since then, several studies have explored the potential of machine learning techniques in credit risk assessment. Selau (2017) compared three distinct predictive modeling approaches, analyzing their effectiveness in classifying defaulting customers and highlighting the importance of choosing the appropriate model to optimize prediction accuracy. Additionally, Marques, Garcia, and Sanchez (2012) evaluated the performance of seven different forecasting models, including 1-Nearest Neighbor, Naive Bayes Classifier, Logistic Regression, Multilayer Perceptron, Radial Basis Function Neural Networks, Support Vector Machine with linear kernel, and Decision Tree, and also showed that the decision tree model constitutes an effective solution for predictive classification methods.

Beyond the model itself, several factors can influence the classification of default, with variable selection and engineering being one of the main ones. Khandani et al. (2010) demonstrated that incorporating a diverse set of attributes, such as credit history, payment behavior, and demographic data, into predictive algorithms can reveal complex interactions and improve the accuracy of credit scoring models. A widely used model is the “credit score”, based on discriminant analysis or logistic regression. Vicente (2003) states that it is widely used by financial institutions for credit granting and relates the score obtained by a customer to the probability of loss, highlighting variables such as marital status, number of loan installments, and income commitment.

Given this scenario, the comparative analysis of credit risk models has become essential to identify more effective and robust approaches in predicting default. The problem focused on the need to improve the accuracy of these predictions and ensure that the methods used were interpretable, avoiding biases that could compromise financial decisions. Thus, the present study sought to investigate the accuracy and applicability of different machine learning models, evaluating their effectiveness in identifying good and bad payers, and comparing the performance of three supervised machine learning algorithms — Logistic Regression, Decision Tree, and Artificial Neural Network of the Multilayer Perceptron (MLP) type — applied to three public credit risk databases: Statlog German Credit Dataset, Default of Credit Card Clients, and Give Me Some Credit, totaling nine model-base combinations, contributing to the advancement of credit analysis techniques and assisting financial institutions in decision-making.

2. Material and Methods

The present study adopted an explanatory applied research nature, according to Gil’s (2002) classification. The objective was the implementation of machine learning algorithms for the search of knowledge applicable in real contexts, aiming to understand the characteristics of customers who most contribute to default. The methodological approach was predominantly quantitative, centered on the analysis of predictive credit risk models and the mitigation of default through statistical techniques.

Complementarily, an interpretative qualitative analysis was performed to understand the influence and weight of the most relevant variables in the classification of defaulters, according to Pinto (2021). For the analysis, three public databases were used: Statlog German Credit Dataset (Hofmann, 1994), with 1,000 observations; Default of Credit Card Clients (Yeh & Lien, 2009), with 30,000 records; and Give Me Some Credit (Kaggle, 2011), with 150,000 observations. All databases included a binary target variable for default.

The data underwent a structured four-step preprocessing. First, missing values in the Give Me Some Credit dataset were imputed, where the variables MonthlyIncome (19.8% missing) and NumberOfDependents (2.6% missing) were filled using the median, a method robust to extreme values (Chapman, 2005). Subsequently, the categorical variables from the German Credit dataset were converted using binary encoding (One-Hot Encoding).

Subsequently, all numerical variables were standardized by the StandardScaler method, which establishes zero mean and unit standard deviation, eliminating scale differences. Finally, each dataset was divided into training (80%) and testing (20%) sets, with stratification by the target variable to maintain proportionality between classes, as recommended by Kuhn and Johnson (2013). Class imbalance, accentuated in the Give Me Some Credit dataset, was treated with the `class_weight=’balanced’` parameter in the models and sample weights in the neural network.

Three supervised machine learning models were implemented: Logistic Regression, Decision Trees, and Multilayer Perceptron (MLP) Artificial Neural Networks. Logistic Regression was implemented by the `LogisticRegression` class from Scikit-learn (Pedregosa et al., 2011), with L2 regularization and the `lbfgs` solver. Decision Trees used the CART algorithm, implemented by the `DecisionTreeClassifier` class from Scikit-learn. The MLP was implemented by the `MLPClassifier` class from Scikit-learn.

The models were submitted to a hyperparameter optimization process with 5-fold cross-validation, adopting AUC-ROC as the optimization metric, as it is more suitable for imbalanced data. For Logistic Regression and Decision Tree, `GridSearchCV` was used. For the MLP neural network, `RandomizedSearchCV` was employed, aiming for greater computational efficiency in the larger configuration space.

The interpretability of the predictions was conducted in a unified manner by the SHAP (SHapley Additive exPlanations) method, applied to all nine models (Lundberg & Lee, 2017). The `LinearExplainer` was used for Logistic Regression, the `TreeExplainer` for the Decision Tree, and the `KernelExplainer` for the MLP. The Scikit-learn (Pedregosa et al., 2011), Seaborn (Waskom, 2021), and SHAP (Lundberg & Lee, 2017) libraries were employed for data development and analysis.

The comparison between the models was performed using five metrics established in the machine learning literature (Fawcett, 2006), derived from the confusion matrix: true positives (TP), true negatives (TN), false positives (FP), and false negatives (FN). The metrics included accuracy, precision, recall, F1-score, and AUC-ROC, allowing for a comprehensive evaluation of predictive performance in class imbalance scenarios.

3. Results and Discussion

The implementation of the three machine learning models on the three selected credit bases revealed a performance landscape that, in several aspects, challenged the initial expectations based on the literature. It was observed that no algorithm demonstrated consistent superiority in all evaluated scenarios, indicating that the choice of the most suitable model is intrinsically dependent on the specific characteristics of each database, including its volume, the distribution of classes, and the nature of the variables, in addition to the metric priorities of the business problem in question. This variability underscores the complexity of credit risk prediction and the need for a flexible approach in model selection.

The analysis of evaluation metrics, before hyperparameter optimization, demonstrated that Logistic Regression achieved the best AUC-ROC on the German Credit dataset, with a value of 0.803, and a recall of 0.800. On the Give Me Some Credit dataset, the Decision Tree stood out with the highest AUC-ROC, reaching 0.849, and a recall of 0.801. In turn, the Artificial Neural Network (MLP) obtained the best performance on the Taiwan Credit Card dataset, with an AUC-ROC of 0.765. These initial results already indicated the absence of a universally superior model, reinforcing the relevance of a detailed comparative analysis.

Regarding recall, a crucial metric for reducing losses to the creditor, the Decision Tree presented the best results on the Give Me Some Credit dataset, with 0.801, and on the German Credit dataset, with 0.750. Logistic Regression also demonstrated competitive recall on the German Credit dataset, with 0.800, and on the Give Me Some Credit dataset, with 0.670. The MLP Neural Network, although competitive in terms of AUC-ROC on some datasets, did not lead in recall in any of the scenarios, suggesting that high general predictive capacity does not always translate into greater sensitivity for identifying the minority class of defaulters.

These findings partially diverge from those of Marques, Garcia, and Sanchez (2012), who identified the Decision Tree as the most effective solution for predictive classification in credit scoring. In the present study, the Decision Tree excelled only on the largest dataset, Give Me Some Credit, which has 150,000 observations, being outperformed by other models in different contexts. This observation is in line with Selau (2017), who noted the variation in algorithm performance depending on the size and nature of the database, a result that aligns with the patterns identified in this research.

Deepening the analysis by database, in German Credit, which contains 1,000 observations and 30% of defaulters, Logistic Regression obtained the best AUC-ROC of 0.803 and the highest recall of 0.800. In contrast, the MLP Neural Network registered the worst initial performance, with an AUC of 0.703. The smaller data volume in this database proved unfavorable for neural networks, which typically require larger samples to converge to optimal configurations. Logistic Regression, by its linear nature, proved less sensitive to data volume and more capable of generalizing on smaller databases with variables of more direct relationships.

On the Taiwan Credit Card base, with 30,000 observations and 22.1% of defaulters, the MLP Neural Network achieved the best AUC-ROC of 0.765, followed by the Decision Tree with 0.759 and Logistic Regression with 0.708. The richness of behavioral history variables, such as the payment statuses from PAY_0 to PAY_6, with their nonlinear interactions, favored models with higher expressive capacity, such as MLP and Decision Tree. However, the high accuracy of the Decision Tree, 0.772, should be interpreted with caution, as in a scenario with 22% of defaulters, accuracies at this level may reflect a bias towards the majority class, rather than an actual effective discrimination between good and bad payers.

The Give Me Some Credit dataset, with 150,000 observations and only 6.7% of defaulters, presented a sharp class imbalance, which represented a significant methodological challenge. In this context, the Decision Tree achieved the highest AUC-ROC of 0.849 and the highest recall of 0.801, benefiting from the volume of data and its ability to capture non-linear relationships, especially between the frequency of past delays and future default. In scenarios of such pronounced imbalance, a trivial model that classified all customers as compliant would achieve an Accuracy of 93.3% without identifying any defaulters, evidencing that, under these conditions, AUC-ROC and recall are the truly informative and most robust metrics for evaluating performance.

In the context of credit granting, the cost of a false negative, which occurs when classifying a defaulter as a good payer, is considerably higher than that of a false positive, as it represents a concrete financial loss for the lender. For this reason, recall is the most critical metric for institutional decision-making, as it measures the proportion of actual defaulters correctly identified by the model. However, unrestricted maximization of recall can lead to degenerate behavior, where a model that classified all customers as defaulters would achieve a perfect recall of 1.0, but without any practical utility for the institution.

The AUC-ROC, in turn, by evaluating the model’s performance across all possible decision thresholds, offers a balanced measure of true discriminative capacity, being particularly suitable for imbalanced data, as highlighted by Fawcett (2006). The F1-score complements this analysis by harmonizing precision and recall into a single indicator, being especially recommended in scenarios with imbalanced classes. The joint use of these metrics allows for a more robust and complete evaluation than accuracy alone, which proves inadequate in contexts of strong class imbalance.

Hyperparameter optimization was conducted using AUC-ROC as the primary metric, as it is independent of the decision threshold and measures the intrinsic discriminative quality of the model across the entire spectrum of sensitivity and specificity. This two-step process allows, firstly, for the selection of the model with the greatest capacity to separate classes and, subsequently, for the adjustment of the decision threshold separately to favor operational recall, according to the institution’s risk tolerance. This approach ensures that the selected model is the most robust and that its production application can be calibrated to meet specific business needs.

The results of hyperparameter optimization revealed important findings. The MLP Neural Network achieved the greatest absolute gain, with an increase of 9.11 percentage points in the German Credit dataset (from 0.703 to 0.794), demonstrating the effectiveness of optimization in the architecture and regularization parameters and learning rate for this model. In contrast, the Decision Tree showed a reduction in AUC-ROC on two datasets, German Credit (-0.12 pp) and Taiwan Credit Card (-1.20 pp), indicating that cross-validation selected hyperparameters that overfitted the training set, not generalizing well to the test set. Logistic Regression, in turn, exhibited variations close to zero across all datasets (+0.35, -0.01, and +0.01 pp), suggesting that the model was already operating essentially at its maximum potential with the initial parameters.

The importance of the variables was analyzed in a unified way through the SHAP (SHapley Additive exPlanations) method, applied to all nine models, using the most suitable explainer for each structure. This approach allowed for a direct comparison of how each algorithm weighted the same variables in different contexts. In the German Credit dataset, the three models converged in identifying the absence of a checking account (`checking_account_A14`) as the variable with the greatest impact. For Logistic Regression, critical credit history (`credit_history_A34`) and single male status (`personal_status_A93`) emerged as secondary predictors, indicating that the combination of demographic profile and negative history substantially increases the risk of default.

In the Decision Tree for the German Credit dataset, loan duration (`duration`) occupied the second most prominent position, suggesting that longer contracts non-linearly amplify risk. The MLP, in turn, corroborated the absence of a checking account as a dominant factor and highlighted the history of timely paid credits (`credit_history_A32`) as the second predictor with a protective effect, interpreted by SHAP as a risk reduction. These differences in the hierarchy of secondary variables between the models reflect their distinct capabilities in capturing linear and non-linear relationships in the data.

In the Taiwan Credit Card base, the payment status (`PAY_0`) was unanimously the dominant predictor in all three models, but the second most important variable differed among them. For Logistic Regression, the most recent month’s bill amount (`BILL_AMT1`) emerged as the second relevant predictor, indicating that higher bills, combined with delays, linearly increase risk. For the Decision Tree, the amounts paid two months ago (`PAY_AMT2`) and the credit limit (`LIMIT_BAL`) were the next, suggesting that low historical payments and reduced limits are risk signals captured by the model’s hierarchical structure.

The MLP, on the Taiwan Credit Card base, assigned a considerably high weight to the credit limit (`LIMIT_BAL`), almost comparable to the payment status (`PAY_0`), followed by the bill amounts (`BILL_AMT2`, `BILL_AMT5`). This pattern suggests that the neural network was able to capture complex nonlinear interactions between the credit limit and billing behavior, something that other models did not detect with the same clarity. The MLP’s ability to identify these complex interactions is an advantage in databases with rich behavioral characteristics.

The Give Me Some Credit base revealed the largest divergence between models regarding the importance of variables. Logistic Regression identified slight and recent delays (`NumberOfTime30-59DaysPastDueNotWorse`) as the main predictor, indicating they are a linear signal of future risk, followed by age (`age`), suggesting that younger clients tend to default, possibly due to a shorter credit history and higher income commitment. Notably, revolving credit utilization (`RevolvingUtilizationOfUnsecuredLines`) had near-zero importance in this linear model.

In contrast, the Decision Tree at the Give Me Some Credit base presented an opposite pattern, where the utilization of revolving credit (`RevolvingUtilizationOfUnsecuredLines`) was the absolutely dominant variable, followed by the delay indicators. The hierarchical structure of the tree captured that high limit utilization is the primary divisor between good and bad payers, evidencing its capacity to model non-linear relationships. The MLP, in turn, produced a third distinct pattern, with age (`age`) taking the lead, closely followed by the frequency of mild and recent delays (`NumberOfTime30-59DaysPastDueNotWorse`), while revolving credit utilization (`RevolvingUtilizationOfUnsecuredLines`) came in last. This result indicates that the neural network learned demographic and delay history interactions as combinatorially more predictive of limit utilization, highlighting the capacity of different algorithms to build fundamentally distinct risk representations from the same data.

The comparative analysis of the nine SHAPs revealed a clear pattern: the variables with the greatest impact were consistent across models in the three datasets, namely the absence of a checking account (`checking_account_A14`) in German Credit, the payment status (`PAY_0`) in Taiwan Credit Card, and the utilization of revolving credit (`RevolvingUtilizationOfUnsecuredLines`) in Give Me Some Credit (for Decision Tree and MLP). The divergence was concentrated in the following positions, where the linear nature of Logistic Regression led it to favor stable demographic variables, while the Decision Tree and MLP captured behavioral indicators of payment frequency and credit utilization with non-linear relationships. This finding justifies the joint use of multiple models in real-world credit scoring applications, especially when interpretability and predictive power need to be balanced.

In summary, the research results demonstrated that no machine learning model showed consistent superiority across all evaluated databases and metrics, reinforcing the need for a multi-model approach to credit risk classification. Logistic Regression proved effective on smaller and more linear datasets, while Decision Tree and MLP excelled on datasets with higher volume and complexity of non-linear interactions, respectively. SHAP analysis was fundamental in interpreting the predictions, revealing that recent financial behavior, revolving credit utilization, and checking account availability were consistently the most predictive variables, regardless of the algorithm, and that different models can build distinct risk representations from the same data, which validates the multi-model approach for a more complete and reliable analysis.

4. Conclusion

This study sought to investigate the accuracy and applicability of different machine learning models in credit risk classification, evaluating their effectiveness in identifying good and bad payers and improving the interpretability of predictions. It was found that none of the algorithms — Logistic Regression, Decision Tree, and Artificial Neural Network (MLP) — demonstrated consistent superiority across all evaluated datasets and metrics. Logistic Regression achieved the best performance on the German Credit dataset, while the Decision Tree stood out on the Give Me Some Credit dataset, and the MLP led on the Taiwan Credit Card dataset. Hyperparameter optimization revealed significant gains for the MLP, especially on the German Credit dataset, confirming its potential in scenarios with large data volumes and non-linear relationships. SHAP analysis was fundamental for interpreting the predictions, indicating that recent financial behavior, revolving credit utilization, and checking account availability were consistently the most predictive variables. This multi-model approach, combined with SHAP interpretability analysis, offers a more complete and reliable framework for credit risk classification, contributing to the advancement of credit analysis techniques and assisting financial institutions in decision-making.

However, the study used public databases that may not reflect the complexity and specificity of real financial institution portfolios, which represents a limitation. Additionally, the decision threshold was kept at the default value of 0.5 in all models, without calibration by risk policy, which may underestimate the actual operational performance in production environments. For future studies, it is suggested to explore the calibration of the threshold according to institutional risk tolerance, the inclusion of ensemble techniques to combine the strengths of different models, and the application of algorithms to Brazilian credit portfolios, aiming to improve risk analysis practices in the national context.

Bibliographic References

Chapman, A. D. 2005. Principles and Methods of Data Cleaning – Primary Species and Species-occurrence Data, version 1.0. Report for the Global Biodiversity Information Facility, Copenhagen.

Fawcett, T. 2006. An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861– 874.

Gil, A. C. 2002. Como elaborar projetos de pesquisa. 4. ed. São Paulo: Atlas.

Hofmann, H. 1994. Statlog (German Credit Data) [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5NC77.

Joseph, C. 2013. Advanced Credit Risk Analysis and Management. Chichester: John Wiley & Sons. (Wiley Finance Series).

Kaggle, 2011 [Referência completa não encontrada no documento original]

Khandani, Amir E., Kim, Andrew J., Lo, Andrew W. 2010. Consumer credit-risk models via machine-learning algorithms. Journal of Banking & Finance, [S.I.], v. 34, n. 11, p. 2767-2787.

Kuhn, M., & Johnson, K. 2013. Applied predictive modeling. New York, NY: Springer.

Lundberg, S. M., & Lee, S. I. 2017. A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30, 4765-4774.

Marqués, A. I.; García, V.; Sánchez, J. S. 2012. Exploring the behaviour of base classifiers in credit scoring ensembles. Expert Systems with Applications, v. 39, n. 11, p. 10244–10250, set. 2012. ISSN 0957-4174.

Pedregosa, F., et al. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12, 2825-2830.

Pinto, A. C. 2021. O poder preditivo dos modelos com aprendizado de máquina é superior aos modelos tradicionais para análise do risco de crédito? Revista Debates em Economia Aplicada – REDEA, v. 1, n. 1.

Reinhart, Carmen M.; Rogoff, Kenneth S. 2009. This time is different: eight centuries of financial folly. Princeton University Press.

Sanches, Vander Lúcio. et al. 2018. Análise de crédito: instituições financeiras minimizam os riscos de inadimplência por meio da análise de crédito. Revista Científica Multidisciplinar Núcleo do Conhecimento. Ano 03, Ed. 09, Vol. 09, pp. 127-151.

Selau, Lisiane Priscila Roldão. 2017. Construção de modelos de previsão de risco de crédito.. Dissertação (Mestrado em Modelagem Matemática) – Universidade Federal de Santa Maria, Santa Maria.

Vicente, E. 2003. Modelo para a avaliação do risco de crédito de municípios brasileiros. Tese (Doutorado em Administração) – Universidade de São Paulo.

Waskom, M. (2021). Seaborn: Statistical data visualization. Journal of Open Source Software, 6(60), 3021.

Yeh, I. C.; Lien, C. H. 2009. The comparisons of data mining techniques for the predictive accuracy of probability of default of credit card clients. Expert Systems with Applications, v. 36, n. 2, p. 2473-2480. https://doi.org/10.1016/j.eswa.2007.12.020.

Article originating from the Final Course Work of the Specialization in Data Science and Analytics of the MBA USP/Esalq

To learn more about the course, click here and access the MBX Academy platform

Most recent

Economy, Investments And Banking

October 23, 2026

Financial Education, Behaviors and Biases in Financial Decision Making

October 22, 2026

Challenges and Possibilities of the First School Year of Children with ASD for Teaching Work

October 22, 2026

Consumer Journey in the Decision to Purchase Infoproducts in the Digital Environment

October 22, 2026

Psychosocial Risks and Mental Health at Work in a Banking Institution

October 22, 2026

Digital Observatory of the Amazon: Data Science Applied to the Analysis of Social, Economic, Environmental, and Health Indicators

October 22, 2026

Continuous Learning as a Factor of Organizational Growth in the Perception of Collaborators

October 22, 2026

Statistical Regression in Microgrids for Operational Performance Estimation: Campusgrid Case

October 22, 2026

Forecast of Aggregate Default in the Brazilian Financial System by Sarimax Model with Macroeconomic Variables

October 22, 2026

Stakeholder Engagement Factors in International Technology Transfer to a Public Hospital

Personal financial management represents a growing challenge, influenced by cognitive biases that affect decision-making, even in individuals with high knowledge. The present study analyzed the relationship between the level of financial education and the occurrence of behavioral biases in the financial decision-making of adults residing in Brazil. An applied, quantitative, and descriptive-correlational field survey was conducted by administering a structured questionnaire to 163 respondents. The instrument included sociodemographic variables, financial knowledge and habits, and hypothetical situations to identify cognitive biases. The results indicated that higher levels of financial education were associated with better financial control, saving, and investment practices, with significant correlations with income and schooling. However, the presence of behavioral biases (anchoring, loss aversion, scarcity, optimism/negativity, present bias, status quo, overconfidence, and mental accounting) persisted even among individuals with higher schooling and income, with no statistically significant association with the analyzed sociodemographic variables. It was concluded that financial education, although essential for financial organization and control, was not sufficient to neutralize the influence of cognitive biases. The incorporation of behavioral approaches into financial education programs is suggested.

Keywords: Financial behavior; Behavioral economics; Personal finance; Cognitive heuristics; Cognitive biases.

October 22, 2026

Challenges and Possibilities of the First School Year of Children with ASD for Teaching Work

The inclusion of children with Autism Spectrum Disorder (ASD) in Early Childhood Education presents significant challenges, given the difficulty of adapting to new environments and the intensity of stimuli. The objective was to analyze the teaching performance in Early Childhood Education in the initial schooling process of children with ASD. A case study was conducted, using narrative interview as a data production technique. The results revealed that the teaching performance was characterized by mediation and affection practices, with the flexibility of routine and the use of playful resources as essential strategies for the child’s development and sensory organization. The importance of the school support professional for the success of the pedagogical work was also highlighted. The child’s integral development was promoted through intentional teaching actions that prioritized environmental adaptation, strengthening of bonds, and attention to the specific needs of ASD. As a contribution, a guide of pedagogical strategies aimed at welcoming children with ASD in public daycare centers was developed.

Keywords: Early Childhood Education; School Inclusion; Teacher Narratives; Inclusive Pedagogical Practices; Autism Spectrum Disorder.

October 22, 2026

Consumer Journey in the Decision to Purchase Infoproducts in the Digital Environment

The expansion of e-commerce and technological transformation have driven the need to understand consumer behavior regarding infoproducts, whose perceived experience has become crucial for professional development. The study aimed to analyze the factors that influenced the purchase decision of infoproducts in the digital environment, considering the stages of the consumer journey. An applied, quantitative, and descriptive research was conducted through a field survey with 100 consumers from Serra Gaúcha. Data collection occurred via a structured questionnaire, using non-probabilistic sampling. The results revealed a qualified audience with strong “mobile-first” behavior. It was found that the availability of detailed information and social approval were the main attraction triggers in the pre-purchase phase. In the transactional and subsequent stages, the minimization of operational friction, the agility of immediate access, and system security proved crucial for conversion and satisfaction, although gaps in human support efficiency were identified. Retention was based on the trust generated by positive past experiences, overcoming the price factor. It was concluded that the infoproduct consumer journey operates as an integrated system, requiring alignment between technological infrastructure, content curation, and support humanization to ensure engagement.

Keywords: E-commerce; Consumer behavior; Consumer experience; Loyalty; Virtual business.

October 22, 2026

Psychosocial Risks and Mental Health at Work in a Banking Institution

Mental health in the workplace is a topic of growing relevance, impacting professional development and employee well-being. A study was conducted to identify aspects and situations that can influence absenteeism levels in a banking institution. The research involved 108 participants and used a structured questionnaire, based on the short version of COPSOQ II by Gonçalves (2021), which contained 45 closed Likert scale questions and two open-ended questions. Data were collected via an online platform and analyzed with a descriptive and quantitative approach. The results revealed a scenario that indicated the need for adjustments in the distribution of activities and in managers’ perception of behaviors and signs of disorders. Important evidence was identified in the participants’ open-ended responses. It was concluded that the study contributed to promoting the perception of a safe environment for mental health in the organization.

Keywords: Absences due to disorders; Professional development; Operational leadership; Medical leave; Mental health.

October 22, 2026

Digital Observatory of the Amazon: Data Science Applied to the Analysis of Social, Economic, Environmental, and Health Indicators

The state of Amazonas faced historical challenges in consolidating the Unified Health System (SUS), stemming from geographical barriers, socioeconomic inequalities, and structural limitations. This study described the development and implementation of a health information geoportal, designed to support monitoring, evaluation, and planning in the Amazonian context. A descriptive case study approach was adopted, based on documentary analysis and secondary data, covering the mapping of functionalities, technical infrastructure, and validation steps of the tool. The Portal was developed with open technologies and integrated multiple national databases, offering interactive dashboards, thematic maps, and a digital document repository. Among the results, highlighted were the georeferenced visualization of key indicators, the possibility of temporal comparative analyses, and the use of the platform in reviewing health plans and redirecting strategic actions. The portal can represent an advance in digital governance and evidence-based management, with potential for replication in other highly complex territories, provided it is accompanied by investments in technological infrastructure, professional training, and inter-institutional coordination.

Keywords: Amazon; Health Evaluation; Digital Health; Health Information Systems.

October 22, 2026

Continuous Learning as a Factor of Organizational Growth in the Perception of Collaborators

The impact of continuous learning and knowledge management on organizational growth was studied in a context of globalization and constant changes. The study’s objective was to analyze employees’ perception of learning culture and its importance for personal and professional development. Data were obtained through an anonymous survey, with a quantitative approach, using a form with “survey” type questions applied to 87 professionals from the technology, health, education, commerce, and services sectors. The results indicated that employees had a positive perception of the contribution of continuous learning to personal and organizational growth, recognizing its benefits for the development of new skills, innovation, and competitiveness. However, it was observed that the incentive for internal knowledge exchange and the offering of practical training still presented moderate levels of agreement. It was concluded that continuous learning significantly contributes to individual and organizational development, suggesting that companies should develop strategic policies for learning focused on new skills, aiming for greater competitive advantage and innovation.

Keywords: Continuous learning; Self-development; Organizational growth; Knowledge Management; Innovation.

October 22, 2026

Statistical Regression in Microgrids for Operational Performance Estimation: Campusgrid Case

A multiple statistical regression model was proposed to estimate the operational performance of UNICAMP’s CampusGrid microgrid, based on internal electrical quantities. The microgrid integrates photovoltaic generation, batteries, and a gas generator. Energy management system data were collected at five-minute intervals over one month. The Operational Reliability Index, defined as the ratio between supplied and demanded energy, was established as the dependent variable. The model was adjusted for the islanded operation mode, evaluating the impact of demand, state of charge, solar generation, and supplied energy. The results revealed high explanatory power, indicating that operational reliability depends on the interaction between supply and demand. It was observed that the nighttime period concentrated the highest rate of partial supply, attributed to the reduction in solar generation. It was concluded that multiple regression offers an interpretable and useful framework for microgrid monitoring and control strategy support.

Keywords: Battery storage; Photovoltaic generation; Reliability Index; Islanded operation; Energy Management System.

October 22, 2026

Forecast of Aggregate Default in the Brazilian Financial System by Sarimax Model with Macroeconomic Variables

Aggregate non-payment in the Brazilian financial system is a crucial indicator of credit risk, the prediction of which is regulated by CMN Resolutions n.° 4.557/2017 and 4.966/2021 for the prospective measurement of expected losses. Aggregate non-payment was modeled and predicted using the SARIMAX model, projecting its trajectory until December 2026 under optimistic, base, and pessimistic scenarios. 164 monthly observations (July/2012 to February/2026) and macroeconomic variables such as unemployment, Selic, IPCA, IBC-Br, and household debt were used. The specification with contemporaneous variables was selected based on information criteria and coefficient stability, and residual diagnostics confirmed statistical adequacy, with a MAPE of 3.12%. Historical non-payment ranged between 4.33% and 7.64%. The estimation identified unemployment (β = 0.2534; p < 0.001) and Selic (β = 0.2128; p < 0.001) as the main positive determinants, with strong temporal persistence (AR(1) = 0.8965). Selic showed increasing correlation with non-payment (r = 0.62 at lag 12), while unemployment demonstrated a contemporaneous effect (r = 0.33). Projections indicated non-payment between 6.30% and 7.98% in December 2026, with a range of 1.68 percentage points, highlighting direct implications for provisioning in financial institutions.

Keywords: Unemployment; Predictive modeling; Monetary policy; Loss provisioning; Time series.

October 22, 2026

Stakeholder Engagement Factors in International Technology Transfer to a Public Hospital

This study described the critical factors perceived for stakeholder engagement in international technology transfer in a public hospital, with a focus on planning. The objective was to describe these conditioning factors, according to the perception of professionals. A single case study was conducted, applying a questionnaire to 25 professionals and using content analysis, complemented by descriptive tabulations. The main obstacles identified were resistance to change (72%), funding (64%), lack of communication (60%), and bureaucracy (40%). Only 28% of respondents recognized the existence of adequate tools for stakeholder management, while 48% denied it and 24% did not know. Exploratory cross-tabulations indicated variations in the assessment of influence according to time of practice, but did not establish causal relationships. Lack of engagement was associated with delays, low adherence, waste of resources, communication failures, compromised knowledge absorption, and risks to patient safety. The conclusions were limited to the perceptions of the researched group and supported preliminary recommendations for intersectoral communication and coordination.

Keywords: Case study; Project Management; Health innovation; Stakeholders; Public sector.

Economy

October 19, 2026

End of the 6×1 scale: what do the numbers allow us to conclude?

Following the promulgation of the end of the 6×1 shift system, the debate on the shorter workday moves to implementation and requires distinguishing social benefits from economic costs. The end of the 6×1 shift system transitioned from legislative discussion to a constitutional change in Brazil. On October 16, 2026, the National Congress promulgated Constitutional Amendment No. 141/2026, originating from Proposal […]

Artificial Intelligence

October 16, 2026

Why study if AI exists?

When answers become easy, learning comes to mean more than just arriving at them. Why do I need to learn this if I can ask an artificial intelligence? This type of question has accompanied education for decades. Why learn an operation if a calculator exists? Why memorize information if it can be found […]

Tax Management

October 15, 2026

Pink tax and the gender of price

How different marketing treatments for men and women are incorporated by consumption taxation Imagine a shelf on which two razors of the same brand and intended for the same purpose are displayed side by side: one has dark packaging and promises strength, precision, and efficiency; the other comes in shades of pink or […]

Health

October 13, 2026

The missing link in health transformation

Why is innovation not enough to drive changes in organizations and health systems Digital transformation has become a strategic priority for health systems worldwide. Electronic health records, artificial intelligence, telemedicine, remote monitoring, and data analysis are now part of the strategic agenda of governments and health organizations. Paradoxically, […]

Marketing

October 09, 2026

Bitcoin or pets: what Artificial Intelligence would not see

A real-life consumer intelligence case shows what gets left out of the brief A few years ago, an American insurance company hired a specialized consumer intelligence firm to answer three questions: The insurer was also considering launching paid insurance in bitcoin. At the time, the consultancy’s workflow was as follows: the analyst created a query […]

School Management

October 06, 2026

The school is not a casino: when trading cards become currency in the playground

Why, for certain students, does school manage to offer so few experiences of achievement? The World Cup is over. Brazil left the competition, the school holidays came, the final was played and the football calendar moved on. It would be reasonable to imagine that, with the end of the tournament and the return to classes, […] would also begin

Energy Transition

September 30, 2026

Renewable energies: the Brazilian advantage

Brazil needs to invest in the capacity to transform natural resources into innovation, productivity, and development. Talking about energy is talking about development. Behind industry, transportation, services, cities, and increasingly artificial intelligence, there is an energy infrastructure that sustains the economy. Understanding a country’s energy matrix is understanding […]

Data Analysis

September 30, 2026

When the data changes, has reality changed?

An analysis of almost 20 million teaching positions-year between 2020 and 2025 shows how changes in population composition and record production can alter the interpretation of historical series Between 2020 and 2025, the number of formal teaching employment positions identified in RAIS (Annual Relation of Social Information), from the Ministry of Labor, […]

Artificial Intelligence

September 28, 2026

Artificial intelligence and the trends

Why do so many people join in on the fun with ChatGPT? I was never the trend person. I still am not. Some pass through my timeline, I see them, I find them interesting but not to the point of trying them out. This was the case, inclusive, with some recent waves of images made by artificial intelligence, like the 80s trend, the one that […]

Tax Management

September 01, 2026

Tax reform and estate planning in the countryside

Efficiency in managing taxes facilitates governance and allows for the organization of family succession. Rural wealth organization has become a strategic measure in light of the tax reform in Brazil, which came into effect this year. The planning of activities and rural wealth management, through tools such as the transition from an Individual to […]

Renata de Gaspari Valdejão Almeida

Artificial Intelligence

August 05, 2026

Inside the company and off the radar: the risks of Shadow AI

According to Microsoft, 75% of professionals already use artificial intelligence at work without company approval. The use of generative artificial intelligence is already part of the corporate routine. Tools like ChatGPT, Copilot, and other platforms help professionals produce texts, analyze data, create presentations, and accelerate tasks in practically all sectors. But this advance […]

Renata de Gaspari Valdejão Almeida

Strategy

July 02, 2026

The future as a competence

Academic interest reinforces the projection of scenarios as an essential tool for decision-making. While some companies are still trying to get the next step right, others are already working with multiple possible futures — and preparing for all of them. But after all, what does it mean to “think about the future”? Is it a trend, a strategy, or a new competence that […]

Renata de Gaspari Valdejão Almeida

Innovation

Technology

May 21, 2026

Agro SP Corridor: innovation, institutional articulation and future

Pecege commands the executive operation of the 400 km ecosystem activities that became known as the Brazilian agro Silicon Valley In 2025, in addition to taking over the management of the Piracicaba Technology Park, the Pecege Institute was also designated to command the executive operation of the State Agribusiness Innovation Corridor activities, the so-called […]

Education

April 27, 2026

How to turn a TCC into a business

New management of the Piracicaba Technology Park offers consulting to MBA graduates The new project director of the Piracicaba Technology Park (PTP), Alexandre Barreto, often says that one of the greatest rewards he has ever had in life was seeing his MBA USP/Esalq students structure companies from their final projects […]

Technology

March 20, 2026

The human repertoire in the age of AI

Critical thinking and creativity are irreplaceable in the work of a marketing professional Artificial Intelligence (AI) is capable of incredible feats, but it does not yet possess the judgment, empathy, intuition, and strategic creativity necessary to do the job of a good marketing professional. In this area, it is important to know how to use the tool to automate […]

Renata de Gaspari Valdejão Almeida

Agribusiness

Innovation

February 23, 2026

ESG as a growth strategy in a cooperative in RS

Santa Clara balances social and environmental responsibility with innovation Founded in 1911 by Italian immigrants in Carlos Barbosa, in the Serra Gaúcha region, Cooperativa Santa Clara is an example that it is possible to balance economic growth with social and environmental responsibility. From a small cheese production company, the organization has evolved into a […]

Renata de Gaspari Valdejão Almeida

Agribusiness

Innovation

January 22, 2026

Brazil free of Foot-and-Mouth Disease: what changes for Brazilian beef?

With a herd of over 200 million head of cattle, the country is currently the leader in exports. Brazil recently received, for the first time, the certification of country free of foot -and-mouth disease without vaccination, conferred by the World Organisation for Animal Health (WOAH). The seal strengthens the country’s position as a reliable supplier of meat for the […]

Renata de Gaspari Valdejão Almeida

Innovation

December 09, 2025

Innovation: the new cycle of the Piracicaba Technology Park

Instituto Pecege assumes management, betting on the creation of new businesses Starting in 2026, Instituto Pecege will assume the management of the Piracicaba Technology Park (PTP). Among its responsibilities will be the administration, operation, maintenance, and expansion of the park, in addition to fostering activities in innovation, entrepreneurship, and science and technology. The choice was made […]

Renata de Gaspari Valdejão Almeida

July 06, 2026

Large-Scale Architecture: DSR-SCRUM Change Management

Mônica Alessandra Guerios; Maria do Carmo Assis Todorov Abstract generated by ResumeAI, an artificial intelligence solution developed by Instituto Pecege focused on synthesis and writing. Architecture and urbanism are configured as essential bases in an increasingly complex and challenging urban space, where the intrinsic connection between social, environmental, and technological issues […]