Financial Crisis Prediction Research Articles

Internet of Things Enabled Financial Crisis Prediction in Enterprises Using Optimal Feature Subset Selection-Based Classification Model.

At present time, an effective tool becomes essential to forecast business failure as well as financial crisis on small- to medium-sized enterprises. This article presents a new optimal feature selection (FS)-based classification model for financial crisis prediction (FCP). The proposed FCP method involves data acquisition, preprocessing, FS, and classification. Initially, the financial data of the enterprises are collected by the use of the internet of things devices, such as smartphones and laptops. Then, the pigeon-inspired optimization (PIO)-based FS technique is applied to choose an optimal set of features. Afterward, the extreme gradient boosting (XGB)-based classification optimized by the Jaya optimization (JO) algorithm called JO-XGB is employed to classify the financial data. The application of the JO algorithm helps to tune the parameters of the XGB model. A detailed experimental validation process takes place to ensure the performance of the presented PIO-JO-XGBoost model. The obtained simulation results verified the effectiveness of the presented model over the compared methods.

Computational Intelligence-Based Financial Crisis Prediction Model Using Feature Subset Selection with Optimal Deep Belief Network.

At present times, financial decisions are mainly based on the classifier technique, which is utilized to allocate a collection of observations into fixed groups. A diverse set of data classifier approaches were presented for forecasting the financial crisis of an institution using the past data. An essential process toward the design of a precise financial crisis prediction (FCP) approach comprises the choice of proper variables (features) that are related to the issues at hand. This is termed as a feature selection (FS) issue that assists to improvise the classifier results. Besides, computational intelligence techniques can be used as a classification model to determine the financial crisis of an organization. In this view, this article introduces a new FS using elephant herd optimization (EHO) with modified water wave optimization (MWWO) algorithm-based deep belief network (DBN) for FCP. The EHO algorithm is applied as a feature selector, and MWWO-DBN is utilized for the classification process. The application of the MWWO algorithm helps to tune the parameters of the DBN model, and the choice of optimal feature subset from the EHO algorithm leads to enhanced classification performance. The experimental results of the proposed model are tested against three benchmark data sets, namely AnalcatData, German Credit, and Australian Credit. The obtained simulation results indicated the superior performance of the proposed model by attaining maximum classification performance.

Volatility spillover between economic sectors in financial crisis prediction: Evidence spanning the great financial crisis and Covid-19 pandemic

This paper measures volatility spillovers between sectors of economic activity using network connectivity measures. Volatility spillovers are an accurate proxy for the transmission of risk across sectors and are particularly informative during crisis periods. To do this, we apply the novel methodology proposed in Diebold and Yilmaz (2012) to seven economic sectors of U.S. economic activity and find that Banking&Insurance, Energy, Technology and Biotechnology are the main channels through which shocks propagate to the rest of the economy. Banking&Insurance is especially relevant during the 2007–2009 global financial crisis while the Energy sector and Technology are especially relevant during the COVID-19 crisis. We also show that volatility spillovers exhibit ability to predict high episodes of volatility for the S&P 500 index being useful as early financial crisis indicators.

Credit Growth, the Yield Curve and Financial Crisis Prediction: Evidence from a Machine Learning Approach

We develop early warning models for financial crisis prediction using machine learning techniques on macrofinancial data for 17 countries over 1870–2016. Machine learning models mostly outperform logistic regression in out-of-sample predictions and forecasting. We identify economic drivers of our machine learning models using a novel framework based on Shapley values, uncovering non-linear relationships between the predictors and crisis risk. Throughout, the most important predictors are credit growth and the slope of the yield curve, both domestically and globally. A flat or inverted yield curve is of most concern when nominal interest rates are low and credit growth is high.

Financial Management Early‐Warning Mechanism Construction and Decision Analysis Research Based on Wireless Sensor Network and Data Mining

With the gradual and complete establishment of the current socialist legal market economy management system with Chinese characteristics and the continuous investment in deepening system reform and continuous improvement in the later period, the social financial industry and corporate financial management have gradually increased their risk awareness of corporate financial management. This paper deeply analyzes and studies the statistical methods of financial data‐related legal rule interactive mining and proposes a new improved statistical algorithm of financial‐related legal rules, which greatly improves the work efficiency of financial data interactive mining. At the same time, a multilevel analysis model based on the concept of corporate financial crisis risk assessment and a corporate financial crisis risk early‐warning analysis model for decision‐making risk evaluation are proposed. Finally, it can be determined how to choose more internationally representative corporate financial management risk analysis indicators, which have more objectivity and practical application significance than traditional analysis methods. Finally, it is concluded that the accuracy of this model is better than that of other models. The accuracy rate of financial crisis prediction reached 62.35%.

Predicting a corporate financial crisis using letters to shareholders

This study utilizes the textual financial information—letters to shareholders to propose a scheme for corporate financial crisis prediction instead of traditional numerical financial ratios. In the scheme, the letters to shareholders were first parsed and analyzed to establish a library of financial crisis feature terms. Based on the financial crisis feature term library, queen genetic algorithm and support vector machine were then used to classify letters to shareholders (i.e., financial crisis and non-financial crises). This scheme can effectively enhance the accuracy of corporate financial crisis detection and reduce the resulting capital damage to enterprises and investors. To achieve the above objective, the following tasks were performed: (1) a process for predicting corporate financial crises by using letters to shareholders was designed, (2) techniques involved in the process of financial crisis prediction were developed, and (3) the use of the proposed approach was demonstrated and evaluated.

Analysis of Financial Crisis Early Warning Model of Listed Enterprises in China

In recent years, many studies show that corporate financial crisis usually has a long hidden period. If this hypothesis can predict the financial crisis of enterprises, it is expected to control the destructive power brought about by the financial crisis to a minimum. Therefore, this paper takes 108 listed financial crisis companies (ST companies) and 435 non-ST listed companies in China’s A-share annual report in 2019 as reference samples. Then, on the basis of these indicators, the corresponding models are established. The results show that through the analysis of various indicators, we can find the potential financial crisis prediction information and predict the company’s financial distress. From the results of the model analysis, we can see that in the year before the financial crisis, the variables of the financial model are the most. This shows that the closer the year of ST occurs, the higher the accuracy of the model.

Financial crisis prediction in Indonesia using combined of volatility and Markov switching models based on real interest rate on deposit and nominal exchange rate indicators

The financial crisis has happened Indonesia 1997 and 2008. The crisis had brought the disadvantages impact on the economy of Indonesia. To anticipate the impact of the crisis, an early detection system that can detect some signs of a financial crisis is needed so that the government as a decision maker in maintain the stability of the economy. The crisis occurred due to hight volatility and structural changes in macroeconomic indicators. One of model that can be used to captured signals crisis is a combination of volatility and Markov switching. Fluctuations can be explained using volatility model, while some changes in conditions are explained through Markov switching. This study uses indicators of the real interest rate on deposit and nominal exchange rate to detect signals crisis in Indonesia. The results showed that the MS-ARCH(2,1) models of the real interest rate on deposit could only explain the crisis that occurred in 1997 and the MS-GARCH(3,1,1) models of nominal exchange rate could explain the crisis that occurred in 1997 and 2008. Based on the predicted value of smoothed probability, it shows that from July 2019 to June 2020, there are no signals of financial crisis in Indonesia.

The prediction of financial crisis in Indonesia based on the smoothed probability value from the combination of volatility and Markov switching models

Indonesia is a country that applies an open economic system. The application of an open economic system can lead to the transmission of crises from other countries, for example the major financial crisis that experienced by Indonesia in 1997 and 2008. The financial crisis caused disruption to the economy stability of the country itself, so it is necessary to predict the financial crisis with the aim to protect economy stability. The indicators used in this research are the Indonesia Composite Index (ICI) and foreign exchange reserves. The method used is a combination of volatility and Markov switching models, namely the MS-ARCH(2,1) for ICI indicator and MS-GARCH(2,1,1) for foreign exchange reserves indicator. The result showed that based on the smoothed probability value of the combination of those two models, the ICI indicator could explain the crisis that occurred in 1997 and 2008, while the foreign exchange reserves indicator could only explain the crisis that occurred in 1997. In addition, the results of financial crisis prediction in Indonesia show that Indonesia will not experience a financial crisis for the next year. However, if the ICI and foreign exchange reserves indicators decrease by at least 32%, a crisis can occur.

Enhancement in Performance of Financial Crisis Prediction using Hybridization of Machine Learning Classifiers

Financial Crisis has been the stern problem experienced by various organizations or even common people when interested in investing in any Financial institutions like banks, Funds development institutions etc. Hence it is mandatory that a reliable prediction system should be applied in early prediction of Financial Crisis Prediction thereby preventing investment in weak financial institutions that might lead to bankruptcy. The Paper focuses on designing a Hybrid Optimized Algorithm called Hybrid Unified Machine Classifier (HUMC) based on Machine Learning Technique that would be capable of identifying categorized and continuous variables in a financial crisis dataset and determine the confusion matrix that can be instilled in performance analysis tool comprising of analytics and prediction related to Accuracy, F-Score, Sensitivity, Specificity, False Positive Rate (FPR) and False Negative Rate (FNR) respectively. Early testing with the training set of Australian credit dataset were tested with machine learning classifiers like Decision Tree, PART, Naive Bayesian, RBF Network and Multilayer Perceptron algorithms with accuracies 85.50%, 83.62%, 77.24%, 82.75% and 84.93% respectively. The Algorithm HUMC was developed based on combining classification features from decision tree, identifying hidden nodes and model with boosting technique that could enhance the performance levels of the Financial Crisis Prediction. The design of algorithm comprised of best characteristics of both classification and neural networks that are capable to find categorization criteria in the dataset at the first level and also to find the hidden continuous data during the second stage respectively. The design of HUMC was implemented and tested with MATLAB. The Result showed that HUMC algorithm showed greater accuracy (86.25%) in comparison to other classifier models along with other performance measures. Thus, this algorithm enhances the prediction of Financial Crisis predictions with good performance.

Credit Growth, the Yield Curve and Financial Crisis Prediction: Evidence from a Machine Learning Approach

We develop early warning models for financial crisis prediction using machine learning techniques on macrofinancial data for 17 countries over 1870–2016. Machine learning models mostly outperform logistic regression in out-of-sample predictions and forecasting. We identify economic drivers of our machine learning models using a novel framework based on Shapley values, uncovering non-linear relationships between the predictors and crisis risk. Throughout, the most important predictors are credit growth and the slope of the yield curve, both domestically and globally. A flat or inverted yield curve is of most concern when nominal interest rates are low and credit growth is high.

Back to the future: Backtesting systemic risk measures during historical bank runs and the great depression

Abstract We evaluate the performance of two popular systemic risk measures, CoVaR and SRISK, during eight financial panics in the era before FDIC insurance. Bank stock price and balance sheet data were not readily available for this period. We rectify this shortcoming by constructing a novel dataset for the New York banking system before 1933. Our evaluation exercise focuses on two challenges: ranking systemically important financial institutions (SIFIs) and financial crisis prediction. We find that CoVaR and SRISK meet the SIFI ranking challenge. That is, they help identify systemic institutions in periods of distress beyond what is explained by standard risk measures up to six months before panics. In contrast, aggregate CoVaR and SRISK are only somewhat effective at predicting financial crises.

The Impacts of Machine Learning in Financial Crisis Prediction

The most complicated and expected issue to be handled in corporate firms, small-scale businesses, and investors’ even governments are financial crisis prediction. To this effect, it was of interest to us to investigate the current impact of the newly employed technique that is machine learning (ML) to handle this menace in all spheres of business both private and public. The study uses systematic literature assessment to study the impact of ML in financial crisis prediction. From the selected works of literature, we have been able to establish the important role play by this method in the prediction of bankruptcy and creditworthiness that was not handled appropriately by others method. Also, machine learning helps in data handling, data privacy, and confidentiality. This study presents a leading approach to achieving financial growth and plasticity in corporate organizations. We, therefore, recommend a real-time study to investigate the impact of ML in FCP.

Mixed logistic model with two independent random coefficients for financial crisis prediction: Argentinean companies

The paper develops a mixed logistic financial distress prediction model with two independent random coefficients and validates it for public Argentinean companies. This study complements existing literature on bankruptcy prediction in emerging economies advancing the application of contemporary econometric methods (Caro et al., 2013). Anticipating bankruptcy risks increases portfolios' profitability. Emerging economies and frontier markets differ from developed economies in political, cultural, social and institutional terms. Given those differences, investors and lenders need specific bankruptcy and financial distress prediction models. The model developed achieves an excellent performance using financial statements from firms listed in the Buenos Aires Stock Exchange during 1993-2000 with ratios accepted in the literature (Altman, 1993; Jones and Hensher, 2004). Results show that profitability, assets turnover and cash flow from operations reduce the likelihood of financial distress while leverage increases it for companies operating in a frontier market such as Argentina.

Mixed logistic model with two independent random coefficients for financial crisis prediction: Argentinean companies

Mixed logistic model with two independent random coefficients for financial crisis prediction: Argentinean companies

Machine Learning Based Classification Models for Financial Crisis Prediction

Financial Crisis Prediction (FCP) being the most complicated and expected problem to be solved from the context of corporate organization, small scale to large scale industries, investors, bank organizations and government agencies, it is important to design a framework to determine a methodology that will reveal a solution for early prediction of the Financial Crisis Prediction (FCP). Earlier methods are reviewed through the various works in statistical techniques applied to solve the problem. However, it is not sufficient to predict the results with much more intelligence and automated manner. The major objective of this paper is to enhance the early prediction of Financial Crisis in any organization based on machine learning models like Multilayer Perceptron, Radial basis Function (RBF) Network, Logistic regression and Deep Learning methods and conduct a comparative analysis of them to determine the best methods for Financial Crisis Prediction (FDP). The testing is conducted with globalized benchmark datasets namely German dataset, Weislaw dataset and Polish Dataset. The testing is performed in both WEKA and Rapid Miner Framework design and obtained with accuracies and other performance measures like False Positive Rate (FPR), False Negative Rate (FNR), Precision, Recall, F-score and Kappa that would determine the best result from specific algorithm that will intelligently identify the financial crisis before it actually occurs in an organization. The results achieved the algorithms DL, MLP, LR and RBF Network with accuracies 96%, 72.10%, 75.20% and 74% on German Dataset, 91.25%, 85.83%, 83.75% and 73.75% on Weislaw dataset, 99.70%, 96.30%, 96.21% and 96.14 on Polish dataset respectively. It is evident from all the predictive results and the analytics in Rapid Miner that Deep Learning (DL) is the best classifier and performer among other machine learners and classifiers. This method will enhance the future predictions and would provide efficient solutions for financial crisis predictions.

Indonesian Financial Crisis Prediction Using Combined Volatility and Markov Regime Switching Model Based on Indicators of Real Interest Rate on Deposits and Lending Interest Rate/Deposit Interest Rate

The financial crisis that occurred in mid-1997 made it aware of the importance of building an early detection system because the crisis had a severe impact on the Indonesian economy. The crisis occurred due to several macroeconomic indicators experiencing very high fluctuations and the existence of changes in the structure (regime). The combined volatility and Markov regime switching models are very suitable to explain the crisis. Indicators of real interest rates on deposits and lending interest rate/deposit interest rates from 1990 to 2018 were used to build the combined model. The result showed that the MRS-ARCH (2,1) model for real interest rate on deposits indicator and the MRS-GARCH (3,1,1) model for lending interest rate/deposit interest rate indicator could explain the crisis. Markov regime switching predictions indicate that in 2019 there are no signs of a financial crisis in Indonesia.

Improved grey wolf optimization-based feature subset selection with fuzzy neural classifier for financial crisis prediction

In present days, prediction of financial crisis of a company is a hot research area. The use of data mining and machine learning algorithms assists to resolve the financial crisis prediction (FCP) problem. Since financial data contain more demographical and unwanted information, it might decrease the classification performance significantly. So, feature selection (FS) process is applied to choose useful data and remove the irrelevant repetitive data. This paper introduces a novel predictive framework for FCP model by the incorporation of improved grey wolf optimization (IGWO) and fuzzy neural classifier (FNC). An IGWO algorithm is derived by the integration of GWO algorithm and tumbling effect. The presented IGWO-based FS method is employed to discover the optimal features from the financial data. For classification purposes, FNC is employed. The proposed method is experimented on two benchmark data sets, namely Australian Credit and German data set under several of performance metrics. The experimental values verified the superior nature of the proposed FCP model over the compared methods.

Financial crisis prediction model using ant colony optimization

Financial decisions are often based on classification models which are used to assign a set of observations into predefined groups. Different data classification models were developed to foresee the financial crisis of an organization using their historical data. One important step towards the development of accurate financial crisis prediction (FCP) model involves the selection of appropriate variables (features) which are relevant for the problems at hand. This is termed as feature selection problem which helps to improve the classification performance. This paper proposes an Ant Colony Optimization (ACO) based financial crisis prediction (FCP) model which incorporates two phases: ACO based feature selection (ACO-FS) algorithm and ACO based data classification (ACO-DC) algorithm. The proposed ACO-FCP model is validated using a set of five benchmark dataset includes both qualitative and quantitative. For feature selection design, the developed ACO-FS method is compared with three existing feature selection algorithms namely genetic algorithm (GA), Particle Swarm Optimization (PSO) algorithm and Grey Wolf Optimization (GWO) algorithm. In addition, a comparison of classification results is also made between ACO-DC and state of art methods. Experimental analysis shows that the ACO-FCP ensemble model is superior and more robust than its counterparts. In consequence, this study strongly recommends that the proposed ACO-FCP model is highly competitive than traditional and other artificial intelligence techniques.

RETRACTED ARTICLE: Intelligent hybrid model for financial crisis prediction using machine learning techniques

Financial crisis prediction (FCP) plays a vital role in the economic phenomenon. The precise prediction of the number and possibility of failing firms acts as an index of the growth and strength of a nation’s economy. Traditionally, several methods have been presented for effective FCP. On the other hand, the classification performance and prediction accuracy and data legality is not good enough for practical applications. In addition, many of the developed methods perform well for some of the particular dataset but not adaptable to different dataset. Hence, there is a requirement to develop an efficient prediction model for better classification performance and adaptable to diverse dataset. This paper presents a cluster based classification model, comprises of two stages: improved K-means clustering and a fitness-scaling chaotic genetic ant colony algorithm (FSCGACA) based classification model. In the first stage, an improved K-means algorithm is devised to eliminate the wrongly clustered data. Then, a rule-based model is selected to design to fit the given dataset. At the end, FSCGACA is employed for seeking the optimal parameters of the rule-based model. The proposed algorithm is employed to a collection of three benchmark dataset which include qualitative bankruptcy dataset, Weislaw dataset and Polish dataset. A detailed statistical analysis of the dataset is also given. The results analysis ensured that the presented FCP model is superior to other classification model based on the different measures and also found to be more appropriate for diverse dataset.