Search

This study explores the performance of gradient boosting methods in bankruptcy prediction for a highly imbalanced dataset. We developed different heterogenous ensemble models based on three popular gradient boosting methods—XGBoost, LightGBM, and CatBoost. Our ensemble models were optimized using the cross‐validation method and the results of the hold‐out test sets showed that the optimized ensemble models not only outperform their base learners, but also improve the state‐of‐the‐art benchmark results on the same dataset. Interestingly, we observed that the data oversampling technique that is commonly used to address the class imbalance issue had an adverse impact on our ensemble models' performance. This indicates that our models are robust to the imbalanced dataset problem that typically degrades the classification performance of machine learning models. [ABSTRACT FROM AUTHOR]

The aim of this paper is to verify the applicability and accuracy of the Altman model in the case of Slovenian companies. The use of the Altman model is hugely popular and widespread among financiers, analysts and other stakeholders who want to determine the creditworthiness of a company’s operations and the likelihood of it running into financial difficulties in the coming years.

This paper extends the theory of fuzzy diseases predictions and bankruptcy likelihood in order to improve cause detection by introducing Fuzzy Entropy, as an alternative measure for the analysis of causes. In this case, the fuzzy entropy allows to assess a level of uncertainty in the expert's valuation of the firm and the visualization of certain problems in experts' valuation levels, which can cause higher levels of relative uncertainty in the analysis of causes in business diagnosis. Also, with this extension, the model can be useful to develop suitable computer systems for monitoring companies' problems, warning of failures and facilitating decision-making. [ABSTRACT FROM AUTHOR]

Bankruptcy prediction is considered one of the most important research topics in the field of finance and accounting. The rapid increase of data science, artificial intelligence, and machine learning has led researchers to build an accurate bankruptcy prediction model. Recent studies show that ensemble methods perform better than traditional machine learning models for predicting corporate failure, especially with highly imbalanced datasets. However, the black box property of these techniques remains challenging to interpret the result and generate corporate classes without any explanation. To this end, we propose to build an accurate and interpretable classification model that generates a set of prediction rules for output. Tri-eXtreme Gradient Boosting (Tri-XGBoost), a semi-supervised technique, is recommended in this paper. The proposed method combines Borderline-Smote (BLSmote) based on Edited Nearest Neighbor (ENN) sampling techniques with three different XGBoost methods as weak classifiers (gbtree, gblinear, and dart). First, the resampling techniques are used to produce more representative synthetic data and balance the distribution of the datasets. To this end, BLSmote is applied to increase the proportion of instances in the minority class (bankrupt data). Then, ENN is used to eliminate the noisy samples from both classes. In addition, the most crucial features that increase predictive accuracy are chosen using XGBoost. Finally, in order to make the model more understandable for both applicants and experts, our result is presented as "IF–THEN" rules. Our proposed model is validated using the imbalanced Polish and Taiwan bankruptcy datasets. Our obtained results demonstrate that our suggested model performs better than the existing models based on the area under the ROC curve (AUC), F1-score, and G-mean performance measures. Our proposed model significantly improves classification accuracy, which is greater than 95% for Polish datasets and more than 93% for Taiwanese dataset in terms of AUC, G-mean and F1-score. [ABSTRACT FROM AUTHOR]

The bankruptcy of companies is essential in financial literature, and the development of bankruptcy forecasting techniques and models is the priority of financial research goals. Many studies have been conducted on predicting the bankruptcy of companies. This study first used a combination of theoretical and expert analysis to determine the financial ratios and macroeconomic variables affecting bankruptcy. Thus, bankrupt companies were distinguished from non-bankrupt ones referring to Black and Scholes's asset pricing models based on the intrinsic value of liabilities and assets. Therefore, 144 companies were studied in the 12 years of 2010-2021 in the screening process. The analysis of multilayer artificial neural networks for evaluating the reliability of the results in identifying the factors affecting the prediction of bankruptcy and prioritizing these factors showed that the least important factor was the ratio of capital to the net profit of the company and the most critical factor was the ratio of profit before interest and taxes to the total assets of the company. [ABSTRACT FROM AUTHOR]

This study seeks to identify the impact of sustainable development goals (SDGs) in predicting corporate financial performance (CFP) in the information communications technology (ICT) industry. Data over the period of 2016–2020 that are relevant to financial reporting and corporate social responsibility (CSR) reporting have been extracted for 208 firms in the ICT industry. Important variables have been identified to help predict the financial performance in the following years upon the publication of CSR reports. Drawing on resource‐based view and stakeholder theory, the purpose of this study is to find the quintessential variables that influence the prediction accuracy of financial performance. To better forecast earnings per share (EPS), machine learning feature selection methods have been implemented. The findings suggest that certain variables such as return on total assets, SDGs adoption and whether the firm has established KPI for SDGs achievements can help enhance EPS prediction. With the various predictive models, the artificial neural network model is the most effective in predicting CFPs. Most importantly, the adoption of SDGs can be utilized to sharpen the forecast on financial performance as it enables firms to bolster stakeholder engagement and evaluate environmental, social, and corporate governance efforts. [ABSTRACT FROM AUTHOR]

Business failure prediction has become crucially important for today's firms, enabling them to reduce financial risks and make informed decisions. This study uses a dataset of 6819 companies and 96 financial and macroeconomic variables to present a comparative analysis of machine learning (ML) models for predicting corporate bankruptcies. Behind this research is to improve the accuracy of bankruptcy prediction, which can help companies make more informed decisions and reduce financial risks. This study aims to evaluate the effectiveness of 16 ML algorithms in terms of accuracy, sensitivity, and other relevant metrics. The work uses methodologies that include data collection and cleaning, exploratory data analysis, model preprocessing and training, and model performance evaluation. Data preprocessing and hyperparameter optimization techniques were used to improve model performance. The evaluated algorithms include Classifiers such as Stacking Classifier (SCC), Randomized Search Classifier (RCV), Historical Gradient Boosting Classifier (HGBC), MLP Classifier (MLPC), K-Neighbors Classifier (KNC), Decision Tree Classifier (DTC), XGBRF Classifier (XGBRFC), Support Vector Classifier (SVC), Logistic Regression Classifier (LR), Linear SVC Classifier (LSVC). With an accuracy of 97.63 %, recall of 97.63 %, and F1-score of 97.63 %, the results show that the SCC algorithm was the best. Other models, such as RCV and DTC, also showed good results, with accuracies above 97 %. However, models such as PAC and BNB had lower performance and accuracy below 90 %. Finally, this study compares the results of ML models in predicting business failures and highlights their effectiveness. The SCC algorithm is considered the most suitable model for this task, as it suggests that it can help economic actors make more informed decisions and reduce financial risks in the context of firms.

The present paper aims to compare the predictive performance of five models namely the Linear Discriminant Analysis (LDA), Logistic Regression (LR), Decision Trees (DT), Support Vector Machine (SVM) and Random Forest (RF) to forecast the bankruptcy of Tunisian companies. A Deep Neural Network (DNN) model is also applied to conduct a prediction performance comparison with other statistical and machine learning algorithms. The data used for this empirical investigation covers 25 financial ratios for a large sample of 732 Tunisian companies from 2011–2017. To interpret the prediction results, three performance measures have been employed; the accuracy percentage, the F1 score, and the Area Under Curve (AUC). In conclusion, DNN shows higher accuracy in predicting bankruptcy compared to other conventional models, whereas the random forest performs better than other machine learning and statistical methods. [ABSTRACT FROM AUTHOR]

Bankruptcy analysis is needed to anticipate bankruptcy. Errors in predicting bankruptcy often cause bankruptcy. Machine learning with high accuracy to analyze reversal must continuously improve its accuracy. Many machine learning models have been applied to predict bankruptcy. However, model improvisation is still needed to improve prediction accuracy. We propose a combination model to improve the accuracy of bankruptcy prediction based on a genetic algorithm-support vector machine (GA-SVM) and stacking ensemble method. This study uses the Taiwanese Bankruptcy dataset from the Taiwan Economic Journal. Then, we implement a synthetic minority oversampling technique for handling imbalanced datasets. We select the best feature using GA-SVM, adopt a new strategy by stacking the classifier, and use extreme gradient boosting as a meta-learner. The results show superior accuracy obtained by the stacking model-based GA-SVM with an accuracy of 99.58%. The accuracy obtained is higher than just applying a single classifier. Thus, this study shows that the proposed method can predict bankruptcy with superior accuracy. [ABSTRACT FROM AUTHOR]

Financial distress identification remains an essential topic in the scientific literature due to its importance for society and the economy. The advancements in information technology and the escalating volume of stored data have led to the emergence of financial distress that transcends the realm of financial statements and its' indicators (ratios). The feature space could be expanded by incorporating new perspectives on feature data categories such as macroeconomics, sectors, social, board, management, judicial incident, etc. However, the increased dimensionality results in sparse data and overfitted models. This study proposes a new approach for efficient financial distress classification assessment by combining dimensionality reduction and machine learning techniques. The proposed framework aims to identify a subset of features leading to the minimization of the loss function describing the financial distress in an enterprise. During the study, 15 dimensionality reduction techniques with different numbers of features and 17 machine-learning models were compared. Overall, 1,432 experiments were performed using Lithuanian enterprise data covering the period from 2015 to 2022. Results revealed that the artificial neural network (ANN) model with 30 ranked features identified using the Random Forest mean decreasing Gini (RF_MDG) feature selection technique provided the highest AUC score. Moreover, this study has introduced a novel approach for feature extraction, which could improve financial distress classification models. [ABSTRACT FROM AUTHOR]

With the recent advances in machine learning (ML), several models have been successfully applied to financial and accounting data to predict the likelihood of companies' bankruptcy. However, time series have received little attention in the literature, with a lack of studies on the application of deep learning sequence models such as Recurrent Neural Networks (RNNs) and the recent Attention-based models in general. In this research work, we investigated the application of Long Short-Term Memory (LSTM) networks to exploit time series of accounting data for bankruptcy prediction. The main contributions of our work are the following: (a) We proposed a multi-head LSTM that models each financial variable in a time window independently and compared it with a single-input LSTM and other traditional ML models. The multi-head LSTM outperformed all the other models. (b) We identified the optimal time series length for bankruptcy prediction to be equal to 4 years of accounting data. (c) We made public the dataset we used for the experiments which includes data from 8262 different public companies in the American stock market generated in the period between 1999 and 2018. Furthermore, we proved the efficacy of the multi-head LSTM model in terms of fewer false positives and the better division of the two classes. [ABSTRACT FROM AUTHOR]

Corporations getting bankrupt has been a severe issue for investors, businesses as well as ordinary individuals. Several research works have been conducted over the years to accurately predict bankruptcy, with the earliest works depending on the financial metrics of the company taken under consideration and the latest ones trying to predict bankruptcy from various kinds of data of organizations. However, the present works do not capture the dynamic nature of the business world and the possibility of a turnaround scenario. Hence, predictive models that are spatiotemporally aware when they predict a firm's financial distress are needed. Considering this imminent problem, our work focuses on building a unique spatiotemporal context-aware bankruptcy prediction model that can predict bankruptcy with the help of daily news articles of a company or its related companies and key financial metrics to predict bankruptcy. Knowledge graphs were used to represent the vast amount of textual data. Their embeddings, along with the financial metrics, were used in the classification process. In the first stage, various machine learning algorithms were used for the financial metrics, while for the textual data or the embeddings, attention-based LSTM was used. Next, both were assembled together in the second stage to form the final predictive model, which has given an accuracy of 0.97 on the test set and an F1 score of 0.95. We hope our novel approach to this problem helps those who are uncertain about the future of any organization in predicting its bankruptcy beforehand and thereby timely decision making. [ABSTRACT FROM AUTHOR]

Machine and Deep Learning methods are widely adopted to predict corporate bankruptcy events for their effectiveness. Bankruptcy prediction is commonly modeled as a binary classification task over accounting data where the positive label is associated with companies with a high likelihood of bankruptcy and the negative label with a low risk of failure. Most of the models mainly focus on exploiting accounting, stock market data, and data augmentation to deal with the intrinsic unbalance of this task. More recently, financial reports such as the US SEC annual reports have been investigated for feature engineering to boost the accuracy of the classification task. However, these approaches only marginally leverage Natural Language Processing advanced techniques to improve the prediction, by usually only leveraging dictionary-based approaches and word frequencies for feature engineering. These fixed features suffer from concept drift over time leading to weaker predictive models by missing a data-driven architecture to extract text disclosures from financial reports to improve the task. This paper aims to fill the gap between the bankruptcy prediction domain and the recent advances in Natural Language Processing by proposing a Transformer-based architecture that combines: a) a text summarization module that extracts text disclosures from financial reports, over time, by leveraging the self-attention mechanism and learning which contents are more valuable for the prediction in a text communication; b) a multivariate time series modeling for accounting data that is aligned and optimized along with the text module. In this way, the architecture benefits from both data sources for the prediction and ensures continual model adaptation over time. We focused on public companies listed in the American stock market with a dataset including 6190 companies from 1999 to 2018. We have deeply analyzed the contribution of the two proposed modules, the accounting time series module (Accuracy 78%) and the text disclosures module (Accuracy 81%) to finally prove that a unique model that can leverage both data sources at the same time achieves better performance (Accuracy 87.5%). The architecture also outperforms the other baselines for Recall of default events (0.84) and for type II error (16.12).

Credit scoring or predicting bankruptcy is among the most crucial techniques for identifying high-risk and low-risk credit situations. Accordingly, enhancing the accuracy of bankruptcy prediction methods decreases the risk of inappropriate financial decisions. Also, increasing the accuracy of credit scoring models brings significant benefits such as improved turnover, credit market growth, proper and efficient allocation of financial resources, and sustained improvement of the profits of banks, investors, funds, and governments. Various statistical classification methods have been developed in the literature with different features and characteristics for more accurate bankruptcy prediction. However, despite all appearance differences in statistical classification approaches, they all adhere to a common idea and concept in their training procedures. The basic operation logic in whole-developed statistical classification methods focuses on maximizing a continuous distance-based cost function to yield the highest performance. Despite it being a common and frequently used procedure for classification purposes, it is an unreasonable and inefficient manner to achieve maximum accuracy in a discrete classification field. In this paper, a new discrete direction-based Logistic Regression that is a common statistical classifier method for bankruptcy forecasting is proposed. In the proposed Logistic Regression, in contrast to all traditionally developed statistical classifiers, the compatibility of the cost function and the training procedure is considered. While it can be shown overall that the performance of the presented discrete direction-based classifier will not be inferior to its continuous counterpart, an evaluation of the suggested classifier is conducted to ascertain its superiority. For this purpose, three credit scoring datasets are considered to assess the classification rate of the presented classifier. Empirical outcomes demonstrate that, as pre-expected, in all cases, the model put forward can attain a superior performance compared to conventional alternatives. These findings clearly demonstrated the significant influence of the consistency between the cost function and the training process on the classification capability, a consideration absent in any of the traditional statistical classification procedures. Consequently, the presented Logistic Regression can be considered an efficient alternative for credit scoring purposes to achieve more accurate results. [ABSTRACT FROM AUTHOR]

This study addresses the challenge of applying ant colony optimisation algorithms to imbalanced datasets, focusing on a bankruptcy dataset. The application of ant colony optimization (ACO) algorithms has been limited by their performance on imbalanced datasets, particularly within bankruptcy prediction where the some of bankruptcy cases leads to skewed data distributions. Traditional ACO algorithms, including the original Ant-Miner, often fail to accurately classify minority classes, which is a critical shortcoming in the context of financial distress analysis. Hence, this study proposes an improved algorithm, the Hellinger Distance Ant-Miner (HD-AntMiner), which employs Hellinger distance as the heuristic for ants to gauge the similarity or dissimilarity between probability distributions. The effectiveness of HD-AntMiner is benchmarked against established classifiers--PART and J48--as well as the conventional Ant-Miner, using public datasets and a specialized dataset of 759 Shariahcompliant securities companies in Malaysia. Utilising the Friedman test and F-score for validation, HD-AntMiner demonstrates superior performance in handling imbalanced datasets compared to other algorithms, as affirmed by the Friedman test. The F-score analysis highlights HD-AntMiner's excellence, achieving the highest F-score for Breast-cancer and Credit-g datasets. When applied to the Shariahcompliant dataset, HD-AntMiner is compared with Ant-Miner and validated through a t-test and F-score. The t-test results confirm HD-AntMiner's higher accuracy than Ant-Miner, while the F-score indicates superior performance across multiple years in the Shariahcompliant dataset. Although the number of rules and conditions is not statistically significant, HD-AntMiner emerges as a robust algorithm for enhancing classification accuracy in imbalanced datasets, particularly in the context of Shariah-compliant securities prediction. [ABSTRACT FROM AUTHOR]

Bankruptcy prediction is the process of measuring the possibility of a company facing bankruptcy or financial distress in the future. An accurate bankruptcy prediction model is valuable for creditors, investors, and financial institutions to assess credit risk, make informed investment decisions, and take appropriate risk management measures. Various methods have been built to contemplate bankruptcy, involving more advanced machine learning (ML) methods and traditional statistical techniques. Typically, this method utilizes financial ratios, accounting data, market performance indicators, and other related variables as input features for predicting the probability of bankruptcy. There has been a growing interest in leveraging the power of neural networks for anticipating the bankruptcy with the emergence of deep learning (DL) methods. With this motivation, this article introduces a new white shark optimizer with deep learning-based bankruptcy prediction for financial risk assessment (WSODL-BPFCA) technique. The presented WSODL-BPFCA technique utilizes a hyperparameter-tuned DL model to predict the existence of bankruptcy. To obtain this, the WSODL-BPFCA technique utilizes min-max normalization to transform the input data into a uniform format. For bankruptcy prediction, the WSODL-BPFCA technique introduces an attention-based long short-term memory (ALSTM) approach. Lastly, the hyperparameter tuning of the ALSTM model was carried out by employing of WSO approach. To exhibit the enhanced performance of the WSODL-BPFCA technique, a widespread set of simulations were performed. The comprehensive comparison study highlighted the improved results of the WSODL-BPFCA technique as 97.61% in terms of different metrics. [ABSTRACT FROM AUTHOR]

Various methods are used when building bankruptcy prediction models. New sophisticated methods that are already used in other scientific fields can also be applied in this area. Graph theory provides a powerful framework for analyzing and visualizing complex systems, making it a valuable tool for assessing the sustainability and financial health of businesses. The motivation for the research was the interest in the application of this method rarely applied in predicting the bankruptcy of companies. The paper aims to propose an improved dynamic bankruptcy prediction model based on graph theoretical modelling. The dynamic model considering the causality relation between financial features was built for the period 2015–2021. Financial features entering the model were selected with the use of Domain knowledge approach. When building the model, the weights of partial permanents were proposed to determine their impact on the final permanent and the algorithm for the optimalisation of these weights was established to obtain the best performing model. The outcome of the paper is the improved dynamic graph theoretical model with a good classification accuracy. The developed model is applicable in the field of bankruptcy prediction and is an equivalent sophisticated alternative to already established models. [ABSTRACT FROM AUTHOR]

This study uses generalized additive models to identify and analyze nonlinear relationships between accounting-based and market-based independent variables and how these affect bankruptcy predictions. Specifically, it examines the independent variables that Altman (J Financ 23:589–609, 1968; Predicting financial distress of companies. Revisiting the Z-score and ZETA® models. Working paper, 2000) and Campbell et al. (J Financ 63:2899–2939, 2008) used and analyzes what specific form these nonlinear relationships take. Drawing on comprehensive data on listed U.S. companies, we show empirically that the bankruptcy prediction is influenced by statistically and economically relevant nonlinear relationships. Our results indicate that taking into account these nonlinear relationships improves significantly several statistical validity measures. We also use a validity measure that is based on the profitability of the bankruptcy prediction models in the context of credit scoring. The findings demonstrate that taking into account nonlinear relationships can substantially increase the discriminatory power of bankruptcy prediction models. [ABSTRACT FROM AUTHOR]

Financial distress identification remains an essential topic in the scientific literature due to its importance for society and the economy. The advancements in information technology and the escalating volume of stored data have led to the emergence of financial distress that transcends the realm of financial statements and its’ indicators (ratios). The feature space could be expanded by incorporating new perspectives on feature data categories such as macroeconomics, sectors, social, board, management, judicial incident, etc. However, the increased dimensionality results in sparse data and overfitted models. This study proposes a new approach for efficient financial distress classification assessment by combining dimensionality reduction and machine learning techniques. The proposed framework aims to identify a subset of features leading to the minimization of the loss function describing the financial distress in an enterprise. During the study, 15 dimensionality reduction techniques with different numbers of features and 17 machine-learning models were compared. Overall, 1,432 experiments were performed using Lithuanian enterprise data covering the period from 2015 to 2022. Results revealed that the artificial neural network (ANN) model with 30 ranked features identified using the Random Forest mean decreasing Gini (RF_MDG) feature selection technique provided the highest AUC score. Moreover, this study has introduced a novel approach for feature extraction, which could improve financial distress classification models.

Corporate bankruptcy and financial distress prediction is a topic of interest for a variety of stakeholders, including businesses, financial institutions, investors, regulatory bodies, auditors, and academics. Various statistical and artificial intelligence methodologies have been devised to produce more accurate predictions. As more researchers are now focusing on this growing field of interest, this paper provides an up-to-date comprehensive survey, classification, and critical analysis of the literature on corporate bankruptcy and financial distress predictions, including definitions of bankruptcy and financial distress, prediction methodologies and models, data pre-processing, feature selection, model implementation, performance criteria and their measures for assessing the performance of classifiers or prediction models, and methodologies for the performance evaluation of prediction models. Finally, a critical analysis of the surveyed literature is provided to inspire possible future research directions.

Companies are becoming aware of the fact that corporate social responsibility (CSR) is becoming the imperative of their sustainable business model despite the potential costs it could generate. Researchers are mostly focused on estimating the relationship between CSR and financial performance where most of the findings indicate their positive relationship. This paper expands existing research and focuses on the relationship between CSR and the risk of bankruptcy using the data from 102 midsize and large companies from non-financial sectors using the data for four years. Research expands existing studies on the EU level according to the fact that most of the existing studies are performed among US companies.

This study addresses the challenge of applying ant colony optimisation algorithms to imbalanced datasets, focusing on a bankruptcy dataset. The application of ant colony optimization (ACO) algorithms has been limited by their performance on imbalanced datasets, particularly within bankruptcy prediction where the some of bankruptcy cases lead to skewed data distributions. Traditional ACO algorithms, including the original Ant-Miner, often fail to accurately classify minority classes, which is a critical shortcoming in the context of financial distress analysis. Hence, this study proposes an improved algorithm, the Hellinger Distance Ant-Miner (HD-AntMiner), which employs Hellinger distance as the heuristic for ants to gauge the similarity or dissimilarity between probability distributions. The effectiveness of HD-AntMiner is benchmarked against established classifiers—PART and J48—as well as the conventional Ant-Miner, using public datasets and a specialized dataset of 759 Shariah-compliant securities companies in Malaysia. Utilising the Friedman test and F-score for validation, HD-AntMiner demonstrates superior performance in handling imbalanced datasets compared to other algorithms, as affirmed by the Friedman test. The F-score analysis highlights HD-AntMiner’s excellence, achieving the highest F-score for Breast-cancer and Credit-g datasets. When applied to the Shariah-compliant dataset, HD-AntMiner is compared with Ant-Miner and validated through a t-test and F-score. The t-test results confirm HD-AntMiner’s higher accuracy than Ant-Miner, while the F-score indicates superior performance across multiple years in the Shariah-compliant dataset. Although the number of rules and conditions is not statistically significant, HD-AntMiner emerges as a robust algorithm for enhancing classification accuracy in imbalanced datasets, particularly in the context of Shariah-compliant securities prediction.

Research Question: This research has been conducted with the aim of determining whether it is possible to generate a model that can reliably predict bankruptcy of Serbian small and medium-sized enterprises (SMEs) using Working Capital Management (WCM) and Asset Management (AM) efficiency ratios. Motivation: Motive for this research is the fact there are not many business failure prediction models related to SMEs. In addition, existing models are not focused on efficiency of the two above mentioned categories. Also, previously developed models, especially traditional and ground-breaking ones, are not necessarily aligned with accounting procedures and economic environment of all countries, which indicates the need to develop a model that is adapted for Serbian territory. Idea: The idea is to develop a model that has the ability to predict whether an entity has a tendency to initiate bankruptcy proceedings in the next year. This is useful both for external stakeholders and for SMEs’ owners themselves, as it allows them to better manage resources. Data: The research was conducted on a sample of 100 Serbian SMEs. Data for the calculation of ratio indicators is available on the Business Registers Agency webpage. Tools: The research was conducted as a combination of financial and statistical analysis instruments. Ratio indicators were used for financial analysis part, while statistical analysis was conducted in SPSS program v.26 and includes logistic (binary) regression. Findings: Results of the research indicate that AM efficiency indicators are suitable for SMEs bankruptcy prediction modelling, but also indicate that WCM ratios don’t have great contribution in bankruptcy prediction for Serbian SMEs. A model that has a classification accuracy of 79% has been developed. Contribution: This research empirically tests how selected ratio indicators support SMEs bankruptcy prediction, and therefore can be beneficial for all SMEs stakeholders, but also other researches since the research methodology is explained in details.

AbstractThis paper examines the distinct types of modified auditor opinions and the non-compliance with the legal certification of accounts, to assess whether they provide different relevant informational content on the risk of impending bankruptcies. The study also addresses the signalling effects when firms do not comply with disclosure obligations. Controlling for the a priori risk classification, we find that distinct opinion types have dissimilar marginal influences, with a disclaimer of opinion denoting the highest level of risk, followed by the non-compliance with the legal certification of accounts and the issuance of an auditor’s adverse opinion. The odds of a firm becoming failed are significantly greater when emphases and reserves are issued by a Big 4 auditor. These findings are based on the evidence of 36,509 firms in Portugal, a country characterized by a proportionately high number of small-sized audited firms and by a lack of independent oversight of auditors, which makes it a relevant setting to analyse.

We develop a framework to simultaneously compute the unobservable parameters underlying the structural-parametric models for bankruptcy prediction. More specifically, we compute the unobservable parameters such as, asset value and asset volatility, through learning by embedding in the structural models a neural network that maps the neural network's input space (e.g. companies' observable financial and market data) to the unobservable parameter space. Within such a 'neuro-structural' framework, the neural network and the structural model work together as a one unit during the learning phase by providing to each other forward and backward information, respectively, until the weights of the neural network are optimized according to a merit function. Empirical results show that structural models, like the Black-Scholes-Merton and the Down-and-Out option models, with parameters computed with our approach, perform better than alternative specifications of the structural models, out of sample, in terms of discriminatory power, information content and economic impact. Importantly, they also perform better than a standard neural network, suggesting that the co-joint dynamics between the neural network and the structural model are useful during the learning phase and can improve the prediction performance (and the training efficiency) of neural networks. Finally, our approach provides methodological (and empirical) enhancements over logit specifications such as, Campbell et al. [In search of distress risk. J Finance, 2008, 63, 2899–2939]. There, financial and market data are the inputs, and the output is the probability of bankruptcy whereas our approach includes an intermediary step to obtain the unobservable parameters and subsequently the probability of bankruptcy. [ABSTRACT FROM AUTHOR]

Bankruptcy prediction is a crucial task in the determination of an organization's economic condition, that is, whether it can meet its financial obligations or not. It is extensively researched because it includes a crucial impact on staff, customers, management, stockholders, bank disposition assessments, and profitableness. In recent years, Artificial Intelligence and Machine Learning techniques have been widely studied for bankruptcy prediction and Decisionmaking problems. When it comes to Machine Learning, Artificial Neural Networks perform really well and are extensively used for bankruptcy prediction since they have proven to be a good predictor in financial applications. various machine learning models are integrated into one called the ensemble technique. It lessens the bias and variance of the ml model. This improves prediction power. The proposed model operated on quantitative and qualitative datasets. This ensemble model finds key ratios and factors of Bankruptcy prediction. LR, decision tree, and Naive Bayes models were compared with the proposed model's results. Model performance was evaluated on the validation set. Accuracy was taken as a metric for the model's performance evaluation purpose. Logistic Regression has given 100% accuracy on the Qualitative Bankruptcy Data Set dataset, resulting in the Ensemble model also performing well. [ABSTRACT FROM AUTHOR]

This article presents a comparative analysis of machine learning models for business failure prediction. Bankruptcy prediction is crucial in assessing financial risks and making informed decisions for investors and regulatory bodies. Since machine learning techniques have advanced, there has been much interest in predicting bankruptcy due to their capacity to handle complex data patterns and boost prediction accuracy. In this study, we evaluated the performance of various machine learning algorithms. We collect comprehensive data comprising financial indicators and company-specific attributes relevant to the Pakistani business landscape from 2016 through 2021. The analysis includes AdaBoost, decision trees, gradient boosting, logistic regressions, naive Bayes, random forests, and support vector machines. This comparative analysis provides insights into the most suitable model for accurate bankruptcy prediction in Pakistani companies. The results contribute to the financial literature by comparing machine learning models tailored to anticipate Pakistani stock market insolvency. These findings can assist financial institutions, regulatory bodies, and investors in making more informed decisions and effectively mitigating financial risks. [ABSTRACT FROM AUTHOR]

Corporate probability of default (PD) prediction is vitally important for risk management and asset pricing. In search of accurate PD prediction, we propose a flexible yet easy-to-interpret default-prediction single-index hazard model (DSI). By applying it to a comprehensive U.S. corporate bankruptcy database we constructed, we discover an interesting V-shaped relationship, indicating a violation of the common linear hazard specification. Most importantly, the single-index hazard model passes the Hosmer-Lemeshow goodness-of-fit calibration test while neither does a state-of-the-art linear hazard model in finance nor a parametric class of Box-Cox transformation survival models. In an economic value analysis, we find that this may translate to as much as three times of profit compared to the linear hazard model. In model estimation, we adopt a penalized-spline approximation for the unknown function and propose an efficient algorithm. With a diverging number of spline knots, we establish consistency and asymptotic theories for the penalized-spline likelihood estimators. Furthermore, we reexamine the distress risk anomaly, that is, higher financially distressed stocks deliver anomalously lower excess returns. Based on the PDs from the proposed single-index hazard model, we find that the distress risk anomaly has weakened or even disappeared during the extended period. [ABSTRACT FROM AUTHOR]

Financial companies are grappling with a burning issue about bankruptcy prediction. There are many methods for bankruptcy prediction, including statistical models and machine learning. Real-life datasets are often imbalanced with high dimensionality. Therefore, it is challenging to train a robust model to predict bankruptcy. Thus, we first applied an oversampling technique known as the Synthetic Minority Oversampling Technique (SMOTE) to reduce the skewness of the data. The balanced data was trained with the baseline models, the ensemble classifiers using different combination methods and the long short-term memory (LSTM) model. In addition, we employed an optimization technique called a genetic algorithm (GA) to optimize and determine the learning parameters of an LSTM network. We further determine the effects of using different training/testing ratios on the developed models. An autoencoder long short-term memory (LSTM) model was developed to extract the best feature representation of the input data. A comparative analysis was carried out between the LSTM-GA and autoencoder-LSTM. The results show that the improved LSTM-GA model with an accuracy of 98.11% performs better than other models. Overall, the research work concluded that all models and LSTM have good performances, while the optimized LSTM model via genetic algorithm outperforms the classical machine learning models. [ABSTRACT FROM AUTHOR]

Research Question: This research has been conducted with the aim of determining whether it is possible to generate a model that can reliably predict bankruptcy of Serbian small and medium-sized enterprises (SMEs) using Working Capital Management (WCM) and Asset Management (AM) efficiency ratios. Motivation: Motive for this research is the fact there are not many business failure prediction models related to SMEs. In addition, existing models are not focused on efficiency of the two above mentioned categories. Also, previously developed models, especially traditional and ground-breaking ones, are not necessarily aligned with accounting procedures and economic environment of all countries, which indicates the need to develop a model that is adapted for Serbian territory. Idea: The idea is to develop a model that has the ability to predict whether an entity has a tendency to initiate bankruptcy proceedings in the next year. This is useful both for external stakeholders and for SMEs' owners themselves, as it allows them to better manage resources. Data: The research was conducted on a sample of 100 Serbian SMEs. Data for the calculation of ratio indicators is available on the Business Registers Agency webpage. Tools: The research was conducted as a combination of financial and statistical analysis instruments. Ratio indicators were used for financial analysis part, while statistical analysis was conducted in SPSS program v.26 and includes logistic (binary) regression. Findings: Results of the research indicate that AM efficiency indicators are suitable for SMEs bankruptcy prediction modelling, but also indicate that WCM ratios don't have great contribution in bankruptcy prediction for Serbian SMEs. A model that has a classification accuracy of 79% has been developed. Contribution: This research empirically tests how selected ratio indicators support SMEs bankruptcy prediction, and therefore can be beneficial for all SMEs stakeholders, but also other researches since the research methodology is explained in details. [ABSTRACT FROM AUTHOR]

The primary tool for managing bank default risk is the credit rating of potential customers. The focus of the present study is on the logistic regression method used to construct 95% of the lender scorecards. The aim of the research is to determine how much the treatment of outliers improves the classification accuracy of the models when using a method that is highly sensitive to outliers, and which method of treating outliers results in the highest classification accuracy. Furthermore, what criteria should be used to determine the cut-off value of the models for a sample that does not contain solvent and insolvent businesses in equal proportions. The analysis was carried out on a sample of 1677 construction companies. The results show that the treatment of outliers significantly improves the predictive ability of the models, while the replacement of outliers with the closest non-outlier proved to be the most effective for treating outliers. When determining the cut-off, it is inappropriate to use the value that results in the highest classification accuracy, as this may lead to an increase in the proportion of first-order errors. The optimisation of this value may depend on the degree of credit risk taken by a given financial institution in its portfolio of loans. [ABSTRACT FROM AUTHOR]

With the recent advances in machine learning (ML), several models have been successfully applied to financial and accounting data to predict the likelihood of companies’ bankruptcy. However, time series have received little attention in the literature, with a lack of studies on the application of deep learning sequence models such as Recurrent Neural Networks (RNNs) and the recent Attention-based models in general. In this research work, we investigated the application of Long Short-Term Memory (LSTM) networks to exploit time series of accounting data for bankruptcy prediction. The main contributions of our work are the following: (a) We proposed a multi-head LSTM that models each financial variable in a time window independently and compared it with a single-input LSTM and other traditional ML models. The multi-head LSTM outperformed all the other models. (b) We identified the optimal time series length for bankruptcy prediction to be equal to 4 years of accounting data. (c) We made public the dataset we used for the experiments which includes data from 8262 different public companies in the American stock market generated in the period between 1999 and 2018. Furthermore, we proved the efficacy of the multi-head LSTM model in terms of fewer false positives and the better division of the two classes.

Copyright of Ruch Prawniczy, Ekonomiczny i Socjologiczny (0035-9629) is the property of Adam Mickiewicz University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Over the past two decades, convolutional neural networks (CNN) have been used to successfully perform classification tasks, such as image recognition, that rely on translation-invariant, locally and spatially correlated data. However, with certain classification problems, such as corporate bankruptcy forecasting, data are made up of unordered columns of numbers, and the distance between two columns is meaningless. Indeed, such data most often represent a firm's financial situation measured at a given point in time, and the variables that can embody this situation are not structured in a way that would make them suitable to be used with a CNN. Since CNN represent a class of classification techniques that is very efficient and as there are virtually no methods that make it possible to adapt accounting data to CNN properties, in this article we propose a technique that aims to transform data into a topological space—a space endowed with a spatial structure—in order to study the ability of CNN to forecast corporate bankruptcy and financial failure. The results of our study show that CNN-based models designed with this technique lead to models that are more accurate than traditional ones. [ABSTRACT FROM AUTHOR]

Over the past century, bankruptcy prediction has dominated economics headlines and continues to do so. An efficient way to deal with bankruptcy is to develop models that integrate multiple econometric measures and predicts a company's financial status. Different methods have been proposed in this domain ranging from statistical algorithms to artificial intelligence models. A novel approach to bankruptcy prediction was proposed in this paper. The novel method incorporates an improved particle swarm optimization (PSO) approach for determining key features and optimizing the long short-term memory system for bankruptcy prediction. This paper employs self-learning and multi-learning strategies to develop the cognitive and social learning parts of the PSO algorithm. The PSO algorithm then determines the optimal LSTM parameters. Our final bankruptcy prediction approach combines optimal feature selection with the LSTM model. Therefore, an improved method of feature selection was developed for the LSTM model. The data for evaluation is available in the machine learning repository at the University of California Irvine. We compare our improved model with the ordinary LSTM model, and the results show better performances in all the metrics for the improved LSTM model compared to the other models. [ABSTRACT FROM AUTHOR]