Cristián Bravo

Elena Tiukhova, Emiliano Penaloza, María Óskarsdóttir et al.

Leveraging network information for prediction tasks has become a common practice in many domains. Being an important part of targeted marketing, influencer detection can potentially benefit from incorporating dynamic network representation. In this work, we investigate different dynamic Graph Neural Networks (GNNs) configurations for influencer detection and evaluate their prediction performance using a unique corporate data set. We show that using deep multi-head attention in GNN and encoding temporal attributes significantly improves performance. Furthermore, our empirical evaluation illustrates that capturing neighborhood representation is more beneficial that using network centrality measures.

Elena Tiukhova, Emiliano Penaloza, María Óskarsdóttir et al.

Leveraging network information for predictive modeling has become widespread in many domains. Within the realm of referral and targeted marketing, influencer detection stands out as an area that could greatly benefit from the incorporation of dynamic network representation due to the continuous evolution of customer-brand relationships. In this paper, we present INFLECT-DGNN, a new method for profit-driven INFLuencer prEdiCTion with Dynamic Graph Neural Networks that innovatively combines Graph Neural Networks (GNNs) and Recurrent Neural Networks (RNNs) with weighted loss functions, synthetic minority oversampling adapted to graph data, and a carefully crafted rolling-window strategy. We introduce a novel profit-driven framework that supports decision-making based on model predictions. To test the framework, we use a unique corporate dataset with diverse networks, capturing the customer interactions across three cities with different socioeconomic and demographic characteristics. Our results show how using RNNs to encode temporal attributes alongside GNNs significantly improves predictive performance, while the profit-driven framework determines the optimal classification threshold for profit maximization. We compare the results of different models to demonstrate the importance of capturing network representation, temporal dependencies, and using a profit-driven evaluation. Our research has significant implications for the fields of referral and targeted marketing, expanding the technical use of deep graph learning within corporate environments.

Ricardo Muñoz-Cancino, Cristián Bravo, Sebastián A. Ríos et al.

For more than a half-century, credit risk management has used credit scoring models in each of its well-defined stages to manage credit risk. Application scoring is used to decide whether to grant a credit or not, while behavioral scoring is used mainly for portfolio management and to take preventive actions in case of default signals. In both cases, network data has recently been shown to be valuable to increase the predictive power of these models, especially when the borrower's historical data is scarce or not available. This study aims to understand the creditworthiness assessment performance dynamics and how it is influenced by the credit history, repayment behavior, and social network features. To accomplish this, we introduced a machine learning classification framework to analyze 97.000 individuals and companies from the moment they obtained their first loan to 12 months afterward. Our novel and massive dataset allow us to characterize each borrower according to their credit behavior, and social and economic relationships. Our research shows that borrowers' history increases performance at a decreasing rate during the first six months and then stabilizes. The most notable effect on perfomance of social networks features occurs at loan application; in personal scoring, this effect prevails a few months, while in business scoring adds value throughout the study period. These findings are of great value to improve credit risk management and optimize the use of traditional information and alternative data sources.

Mahsa Tavakoli, Rohitash Chandra, Fengrui Tian et al.

Knowing which factors are significant in credit rating assignment leads to better decision-making. However, the focus of the literature thus far has been mostly on structured data, and fewer studies have addressed unstructured or multi-modal datasets. In this paper, we present an analysis of the most effective architectures for the fusion of deep learning models for the prediction of company credit rating classes, by using structured and unstructured datasets of different types. In these models, we tested different combinations of fusion strategies with different deep learning models, including CNN, LSTM, GRU, and BERT. We studied data fusion strategies in terms of level (including early and intermediate fusion) and techniques (including concatenation and cross-attention). Our results show that a CNN-based multi-modal model with two fusion strategies outperformed other multi-modal techniques. In addition, by comparing simple architectures with more complex ones, we found that more sophisticated deep learning models do not necessarily produce the highest performance; however, if attention-based models are producing the best results, cross-attention is necessary as a fusion strategy. Finally, our comparison of rating agencies on short-, medium-, and long-term performance shows that Moody's credit ratings outperform those of other agencies like Standard & Poor's and Fitch Ratings.

Sherly Alfonso-Sánchez, Jesús Solano, Alejandro Correa-Bahnsen et al.

Reinforcement learning has been explored for many problems, from video games with deterministic environments to portfolio and operations management in which scenarios are stochastic; however, there have been few attempts to test these methods in banking problems. In this study, we sought to find and automatize an optimal credit card limit adjustment policy by employing reinforcement learning techniques. Because of the historical data available, we considered two possible actions per customer, namely increasing or maintaining an individual's current credit limit. To find this policy, we first formulated this decision-making question as an optimization problem in which the expected profit was maximized; therefore, we balanced two adversarial goals: maximizing the portfolio's revenue and minimizing the portfolio's provisions. Second, given the particularities of our problem, we used an offline learning strategy to simulate the impact of the action based on historical data from a super-app in Latin America to train our reinforcement learning agent. Our results, based on the proposed methodology involving synthetic experimentation, show that a Double Q-learning agent with optimized hyperparameters can outperform other strategies and generate a non-trivial optimal policy not only reflecting the complex nature of this decision but offering an incentive to explore reinforcement learning in real-world banking scenarios. Our research establishes a conceptual structure for applying reinforcement learning framework to credit limit adjustment, presenting an objective technique to make these decisions primarily based on data-driven methods rather than relying only on expert-driven systems. We also study the use of alternative data for the problem of balance prediction, as the latter is a requirement of our proposed model. We find the use of such data does not always bring prediction gains.

Ricardo Muñoz-Cancino, Cristián Bravo, Sebastián A. Ríos et al.

Credit scoring models are the primary instrument used by financial institutions to manage credit risk. The scarcity of research on behavioral scoring is due to the difficult data access. Financial institutions have to maintain the privacy and security of borrowers' information refrain them from collaborating in research initiatives. In this work, we present a methodology that allows us to evaluate the performance of models trained with synthetic data when they are applied to real-world data. Our results show that synthetic data quality is increasingly poor when the number of attributes increases. However, creditworthiness assessment models trained with synthetic data show a reduction of 3\% of AUC and 6\% of KS when compared with models trained with real data. These results have a significant impact since they encourage credit risk investigation from synthetic data, making it possible to maintain borrowers' privacy and to address problems that until now have been hampered by the availability of information.

Bart Baesens, Andreas Goethals, Stefan Lessmann et al.

Predictive models play a pivotal role in credit risk management, guiding critical decisions through accurate estimation of default probabilities and losses. Extensive research has introduced new modeling techniques, complemented by large-scale benchmarking studies consolidating the state-of-the-art. Today, quasi-standards such as gradient-boosting models paired with SHAP explainers have emerged, yet continuous improvement of risk models remains a top priority. Concurrently, rapid advancements in AI, most notably large language models, have disrupted predictive modeling paradigms. Foundation models, pretrained on extensive datasets from diverse domains, have demonstrated remarkable performance by leveraging prior knowledge. While prevalent in natural language processing and computer vision, foundation models for tabular data have only recently emerged. We conjecture that pretraining on out-of-domain data is particularly beneficial in small-data settings, such as SME lending or specialized corporate portfolios, and may help address longstanding challenges including low default portfolios and class imbalance. This paper benchmarks recently proposed tabular foundation models against a broad set of competitors, including established and advanced machine learning techniques, across two core tasks: PD and LGD modeling. Our evaluation encompasses various datasets, performance indicators, and experimental conditions. We find that tabular foundation models generally perform best across datasets and tasks. Moreover, they offer significant improvement in predictive performance as dataset size shrinks. These results are remarkable given that the models are tested out-of-the-box, without hyperparameter tuning, ensuring ease of use and mitigating computational costs.

Matthew Stevenson, Christophe Mues, Cristián Bravo

LiDAR (short for "Light Detection And Ranging" or "Laser Imaging, Detection, And Ranging") technology can be used to provide detailed three-dimensional elevation maps of urban and rural landscapes. To date, airborne LiDAR imaging has been predominantly confined to the environmental and archaeological domains. However, the geographically granular and open-source nature of this data also lends itself to an array of societal, organizational and business applications where geo-demographic type data is utilised. Arguably, the complexity involved in processing this multi-dimensional data has thus far restricted its broader adoption. In this paper, we propose a series of convenient task-agnostic tile elevation embeddings to address this challenge, using recent advances from unsupervised Deep Learning. We test the potential of our embeddings by predicting seven English indices of deprivation (2019) for small geographies in the Greater London area. These indices cover a range of socio-economic outcomes and serve as a proxy for a wide variety of downstream tasks to which the embeddings can be applied. We consider the suitability of this data not just on its own but also as an auxiliary source of data in combination with demographic features, thus providing a realistic use case for the embeddings. Having trialled various model/embedding configurations, we find that our best performing embeddings lead to Root-Mean-Squared-Error (RMSE) improvements of up to 21% over using standard demographic features alone. We also demonstrate how our embedding pipeline, using Deep Learning combined with K-means clustering, produces coherent tile segments which allow the latent embedding features to be interpreted.

Kamesh Korangi, Christophe Mues, Cristián Bravo

Apart from assessing individual asset performance, investors in financial markets also need to consider how a set of firms performs collectively as a portfolio. Whereas traditional Markowitz-based mean-variance portfolios are widespread, network-based optimisation techniques offer a more flexible tool to capture complex interdependencies between asset values. However, most of the existing studies do not contain firms at risk of default and remove any firms that drop off indices over a certain time. This is the first study to also incorporate such firms in portfolio optimisation on a large scale. We propose and empirically test a novel method that leverages Graph Attention networks (GATs), a subclass of Graph Neural Networks (GNNs). GNNs, as deep learning-based models, can exploit network data to uncover nonlinear relationships. Their ability to handle high-dimensional data and accommodate customised layers for specific purposes makes them appealing for large-scale problems such as mid- and small-cap portfolio optimisation. This study utilises 30 years of data on mid-cap firms, creating graphs of firms using distance correlation and the Triangulated Maximally Filtered Graph approach. These graphs are the inputs to a GAT model incorporating weight and allocation constraints and a loss function derived from the Sharpe ratio, thus focusing on maximising portfolio risk-adjusted returns. This new model is benchmarked against a network characteristic-based portfolio, a mean variance-based portfolio, and an equal-weighted portfolio. The results show that the portfolio produced by the GAT-based model outperforms all benchmarks and is consistently superior to other strategies over a long period, while also being informative of market dynamics.

Sahab Zandi, Kamesh Korangi, María Óskarsdóttir et al.

Whereas traditional credit scoring tends to employ only individual borrower- or loan-level predictors, it has been acknowledged for some time that connections between borrowers may result in default risk propagating over a network. In this paper, we present a model for credit risk assessment leveraging a dynamic multilayer network built from a Graph Neural Network and a Recurrent Neural Network, each layer reflecting a different source of network connection. We test our methodology in a behavioural credit scoring context using a dataset provided by U.S. mortgage financier Freddie Mac, in which different types of connections arise from the geographical location of the borrower and their choice of mortgage provider. The proposed model considers both types of connections and the evolution of these connections over time. We enhance the model by using a custom attention mechanism that weights the different time snapshots according to their importance. After testing multiple configurations, a model with GAT, LSTM, and the attention mechanism provides the best results. Empirical results demonstrate that, when it comes to predicting probability of default for the borrowers, our proposed model brings both better results and novel insights for the analysis of the importance of connections and timestamps, compared to traditional methods.

Sherly Alfonso-Sánchez, Cristián Bravo, Kristina G. Stankova

Geographic context is often consider relevant to motor insurance risk, yet public actuarial datasets provide limited location identifiers, constraining how this information can be incorporated and evaluated in claim-frequency models. This study examines how geographic information from alternative data sources can be incorporated into actuarial models for Motor Third Party Liability (MTPL) claim prediction under such constraints. Using the BeMTPL97 dataset, we adopt a zone-level modeling framework and evaluate predictive performance on unseen postcodes. Geographic information is introduced through two channels: environmental indicators from OpenStreetMap and CORINE Land Cover, and orthoimagery released by the Belgian National Geographic Institute for academic use. We evaluate the predictive contribution of coordinates, environmental features, and image embeddings across three baseline models: generalized linear models (GLMs), regularized GLMs, and gradient-boosted trees, while raw imagery is modeled using convolutional neural networks. Our results show that augmenting actuarial variables with constructed geographic information improves accuracy. Across experiments, both linear and tree-based models benefit most from combining coordinates with environmental features extracted at 5 km scale, while smaller neighborhoods also improve baseline specifications. Generally, image embeddings do not improve performance when environmental features are available; however, when such features are absent, pretrained vision-transformer embeddings enhance accuracy and stability for regularized GLMs. Our results show that the predictive value of geographic information in zone-level MTPL frequency models depends less on model complexity than on how geography is represented, and illustrate that geographic context can be incorporated despite limited individual-level spatial information.

Sahab Zandi, Kamesh Korangi, Juan C. Moreno-Paredes et al.

Small and Medium-sized Enterprises (SMEs) are known to play a vital role in economic growth, employment, and innovation. However, they tend to face significant challenges in accessing credit due to limited financial histories, collateral constraints, and exposure to macroeconomic shocks. These challenges make an accurate credit risk assessment by lenders crucial, particularly since SMEs frequently operate within interconnected firm networks through which default risk can propagate. This paper presents and tests a novel approach for modelling the risk of SME credit, using a unique large data set of SME loans provided by a prominent financial institution. Specifically, our approach employs Graph Neural Networks to predict SME default using multilayer network data derived from common ownership and financial transactions between firms. We show that combining this information with traditional structured data not only improves application scoring performance, but also explicitly models contagion risk between companies. Further analysis shows how the directionality and intensity of these connections influence financial risk contagion, offering a deeper understanding of the underlying processes. Our findings highlight the predictive power of network data, as well as the role of supply chain networks in exposing SMEs to correlated default risk.

Ricardo Muñoz-Cancino, Cristián Bravo, Sebastián A. Ríos et al.

The thin-file borrowers are customers for whom a creditworthiness assessment is uncertain due to their lack of credit history; many researchers have used borrowers' relationships and interactions networks in the form of graphs as an alternative data source to address this. Incorporating network data is traditionally made by hand-crafted feature engineering, and lately, the graph neural network has emerged as an alternative, but it still does not improve over the traditional method's performance. Here we introduce a framework to improve credit scoring models by blending several Graph Representation Learning methods: feature engineering, graph embeddings, and graph neural networks. We stacked their outputs to produce a single score in this approach. We validated this framework using a unique multi-source dataset that characterizes the relationships and credit history for the entire population of a Latin American country, applying it to credit risk models, application, and behavior, targeting both individuals and companies. Our results show that the graph representation learning methods should be used as complements, and these should not be seen as self-sufficient methods as is currently done. In terms of AUC and KS, we enhance the statistical performance, outperforming traditional methods. In Corporate lending, where the gain is much higher, it confirms that evaluating an unbanked company cannot solely consider its features. The business ecosystem where these firms interact with their owners, suppliers, customers, and other companies provides novel knowledge that enables financial institutions to enhance their creditworthiness assessment. Our results let us know when and which group to use graph data and what effects on performance to expect. They also show the enormous value of graph data on the unbanked credit scoring problem, principally to help companies' banking.

Kamesh Korangi, Christophe Mues, Cristián Bravo

In this paper, we study mid-cap companies, i.e. publicly traded companies with less than US $10 billion in market capitalisation. Using a large dataset of US mid-cap companies observed over 30 years, we look to predict the default probability term structure over the medium term and understand which data sources (i.e. fundamental, market or pricing data) contribute most to the default risk. Whereas existing methods typically require that data from different time periods are first aggregated and turned into cross-sectional features, we frame the problem as a multi-label time-series classification problem. We adapt transformer models, a state-of-the-art deep learning model emanating from the natural language processing domain, to the credit risk modelling setting. We also interpret the predictions of these models using attention heat maps. To optimise the model further, we present a custom loss function for multi-label classification and a novel multi-channel architecture with differential training that gives the model the ability to use all input data efficiently. Our results show the proposed deep learning architecture's superior performance, resulting in a 13% improvement in AUC (Area Under the receiver operating characteristic Curve) over traditional models. We also demonstrate how to produce an importance ranking for the different data sources and the temporal relationships using a Shapley approach specific to these models.

David Barrera Ferro, Sally Brailsford, Cristián Bravo et al.

Low attendance levels in medical appointments have been associated with poor health outcomes and efficiency problems for service providers. To address this problem, healthcare managers could aim at improving attendance levels or minimizing the operational impact of no-shows by adapting resource allocation policies. However, given the uncertainty of patient behaviour, generating relevant information regarding no-show probabilities could support the decision-making process for both approaches. In this context many researchers have used multiple regression models to identify patient and appointment characteristics than can be used as good predictors for no-show probabilities. This work develops a Decision Support System (DSS) to support the implementation of strategies to encourage attendance, for a preventive care program targeted at underserved communities in Bogotá, Colombia. Our contribution to literature is threefold. Firstly, we assess the effectiveness of different machine learning approaches to improve the accuracy of regression models. In particular, Random Forest and Neural Networks are used to model the problem accounting for non-linearity and variable interactions. Secondly, we propose a novel use of Layer-wise Relevance Propagation in order to improve the explainability of neural network predictions and obtain insights from the modelling step. Thirdly, we identify variables explaining no-show probabilities in a developing context and study its policy implications and potential for improving healthcare access. In addition to quantifying relationships reported in previous studies, we find that income and neighbourhood crime statistics affect no-show probabilities. Our results will support patient prioritization in a pilot behavioural intervention and will inform appointment planning decisions.

María Óskarsdóttir, Cristián Bravo

We present a multilayer network model for credit risk assessment. Our model accounts for multiple connections between borrowers (such as their geographic location and their economic activity) and allows for explicitly modelling the interaction between connected borrowers. We develop a multilayer personalized PageRank algorithm that allows quantifying the strength of the default exposure of any borrower in the network. We test our methodology in an agricultural lending framework, where it has been suspected for a long time default correlates between borrowers when they are subject to the same structural risks. Our results show there are significant predictive gains just by including centrality multilayer network information in the model, and these gains are increased by more complex information such as the multilayer PageRank variables. The results suggest default risk is highest when an individual is connected to many defaulters, but this risk is mitigated by the size of the neighbourhood of the individual, showing both default risk and financial stability propagate throughout the network.

Cristián Bravo, María Óskarsdóttir

In this paper we present a novel algorithm to study the evolution of credit risk across complex multilayer networks. Pagerank-like algorithms allow for the propagation of an influence variable across single networks, and allow quantifying the risk single entities (nodes) are subject to given the connection they have to other nodes in the network. Multilayer networks, on the other hand, are networks where subset of nodes can be associated to a unique set (layer), and where edges connect elements either intra or inter networks. Our personalized PageRank algorithm for multilayer networks allows for quantifying how credit risk evolves across time and propagates through these networks. By using bipartite networks in each layer, we can quantify the risk of various components, not only the loans. We test our method in an agricultural lending dataset, and our results show how default risk is a challenging phenomenon that propagates and evolves through the network across time.

Luisa Roa, Alejandro Correa-Bahnsen, Gabriel Suarez et al.

In this paper we present the impact of alternative data that originates from an app-based marketplace, in contrast to traditional bureau data, upon credit scoring models. These alternative data sources have shown themselves to be immensely powerful in predicting borrower behavior in segments traditionally underserved by banks and financial institutions. Our results, validated across two countries, show that these new sources of data are particularly useful for predicting financial behavior in low-wealth and young individuals, who are also the most likely to engage with alternative lenders. Furthermore, using the TreeSHAP method for Stochastic Gradient Boosting interpretation, our results also revealed interesting non-linear trends in the variables originating from the app, which would not normally be available to traditional banks. Our results represent an opportunity for technology companies to disrupt traditional banking by correctly identifying alternative data sources and handling this new information properly. At the same time alternative data must be carefully validated to overcome regulatory hurdles across diverse jurisdictions.

Matthew Stevenson, Christophe Mues, Cristián Bravo

Compared to consumer lending, Micro, Small and Medium Enterprise (mSME) credit risk modelling is particularly challenging, as, often, the same sources of information are not available. Therefore, it is standard policy for a loan officer to provide a textual loan assessment to mitigate limited data availability. In turn, this statement is analysed by a credit expert alongside any available standard credit data. In our paper, we exploit recent advances from the field of Deep Learning and Natural Language Processing (NLP), including the BERT (Bidirectional Encoder Representations from Transformers) model, to extract information from 60 000 textual assessments provided by a lender. We consider the performance in terms of the AUC (Area Under the receiver operating characteristic Curve) and Brier Score metrics and find that the text alone is surprisingly effective for predicting default. However, when combined with traditional data, it yields no additional predictive capability, with performance dependent on the text's length. Our proposed deep learning model does, however, appear to be robust to the quality of the text and therefore suitable for partly automating the mSME lending process. We also demonstrate how the content of loan assessments influences performance, leading us to a series of recommendations on a new strategy for collecting future mSME loan assessments.

María Óskarsdóttir, Cristián Bravo, Carlos Sarraute et al.

Credit scoring is without a doubt one of the oldest applications of analytics. In recent years, a multitude of sophisticated classification techniques have been developed to improve the statistical performance of credit scoring models. Instead of focusing on the techniques themselves, this paper leverages alternative data sources to enhance both statistical and economic model performance. The study demonstrates how including call networks, in the context of positive credit information, as a new Big Data source has added value in terms of profit by applying a profit measure and profit-based feature selection. A unique combination of datasets, including call-detail records, credit and debit account information of customers is used to create scorecards for credit card applicants. Call-detail records are used to build call networks and advanced social network analytics techniques are applied to propagate influence from prior defaulters throughout the network to produce influence scores. The results show that combining call-detail records with traditional data in credit scoring models significantly increases their performance when measured in AUC. In terms of profit, the best model is the one built with only calling behavior features. In addition, the calling behavior features are the most predictive in other models, both in terms of statistical and economic performance. The results have an impact in terms of ethical use of call-detail records, regulatory implications, financial inclusion, as well as data sharing and privacy.