Corporate Default Prediction Model Using Deep Learning Time Series Algorithm, RNN and LSTM (딥러닝 시계열 알고리즘 적용한 기업부도예측모형 유용성 검증)
-
- Journal of Intelligence and Information Systems
- /
- v.24 no.4
- /
- pp.1-32
- /
- 2018
In addition to stakeholders including managers, employees, creditors, and investors of bankrupt companies, corporate defaults have a ripple effect on the local and national economy. Before the Asian financial crisis, the Korean government only analyzed SMEs and tried to improve the forecasting power of a default prediction model, rather than developing various corporate default models. As a result, even large corporations called 'chaebol enterprises' become bankrupt. Even after that, the analysis of past corporate defaults has been focused on specific variables, and when the government restructured immediately after the global financial crisis, they only focused on certain main variables such as 'debt ratio'. A multifaceted study of corporate default prediction models is essential to ensure diverse interests, to avoid situations like the 'Lehman Brothers Case' of the global financial crisis, to avoid total collapse in a single moment. The key variables used in corporate defaults vary over time. This is confirmed by Beaver (1967, 1968) and Altman's (1968) analysis that Deakins'(1972) study shows that the major factors affecting corporate failure have changed. In Grice's (2001) study, the importance of predictive variables was also found through Zmijewski's (1984) and Ohlson's (1980) models. However, the studies that have been carried out in the past use static models. Most of them do not consider the changes that occur in the course of time. Therefore, in order to construct consistent prediction models, it is necessary to compensate the time-dependent bias by means of a time series analysis algorithm reflecting dynamic change. Based on the global financial crisis, which has had a significant impact on Korea, this study is conducted using 10 years of annual corporate data from 2000 to 2009. Data are divided into training data, validation data, and test data respectively, and are divided into 7, 2, and 1 years respectively. In order to construct a consistent bankruptcy model in the flow of time change, we first train a time series deep learning algorithm model using the data before the financial crisis (2000~2006). The parameter tuning of the existing model and the deep learning time series algorithm is conducted with validation data including the financial crisis period (2007~2008). As a result, we construct a model that shows similar pattern to the results of the learning data and shows excellent prediction power. After that, each bankruptcy prediction model is restructured by integrating the learning data and validation data again (2000 ~ 2008), applying the optimal parameters as in the previous validation. Finally, each corporate default prediction model is evaluated and compared using test data (2009) based on the trained models over nine years. Then, the usefulness of the corporate default prediction model based on the deep learning time series algorithm is proved. In addition, by adding the Lasso regression analysis to the existing methods (multiple discriminant analysis, logit model) which select the variables, it is proved that the deep learning time series algorithm model based on the three bundles of variables is useful for robust corporate default prediction. The definition of bankruptcy used is the same as that of Lee (2015). Independent variables include financial information such as financial ratios used in previous studies. Multivariate discriminant analysis, logit model, and Lasso regression model are used to select the optimal variable group. The influence of the Multivariate discriminant analysis model proposed by Altman (1968), the Logit model proposed by Ohlson (1980), the non-time series machine learning algorithms, and the deep learning time series algorithms are compared. In the case of corporate data, there are limitations of 'nonlinear variables', 'multi-collinearity' of variables, and 'lack of data'. While the logit model is nonlinear, the Lasso regression model solves the multi-collinearity problem, and the deep learning time series algorithm using the variable data generation method complements the lack of data. Big Data Technology, a leading technology in the future, is moving from simple human analysis, to automated AI analysis, and finally towards future intertwined AI applications. Although the study of the corporate default prediction model using the time series algorithm is still in its early stages, deep learning algorithm is much faster than regression analysis at corporate default prediction modeling. Also, it is more effective on prediction power. Through the Fourth Industrial Revolution, the current government and other overseas governments are working hard to integrate the system in everyday life of their nation and society. Yet the field of deep learning time series research for the financial industry is still insufficient. This is an initial study on deep learning time series algorithm analysis of corporate defaults. Therefore it is hoped that it will be used as a comparative analysis data for non-specialists who start a study combining financial data and deep learning time series algorithm.
Brand switching data frequently used in market structure analysis is adequate to analyze non- durable goods, because it can capture competition between specific two brands. But brand switching data sometimes can not be used to analyze goods like automobiles having long term duration because one of main assumptions that consumer preference toward brand attributes is not changed against time can be violated. Therefore a new type of data which can precisely capture competition among durable goods is needed. Another problem of using brand switching data collected from actual purchase behavior is short of explanation why consumers consider different set of brands. Considering above problems, main purpose of this study is to analyze market structure for durable goods with consideration set. The author uses exploratory approach and latent class clustering to identify market structure based on heterogeneous consideration set among consumers. Then the relationship between some factors and consideration set formation is analyzed. Some benefits and two demographic variables - age and income - are selected as factors based on consumer behavior theory. The author analyzed USA automotive market with top 11 brands using exploratory approach and latent class clustering. 2,500 respondents are randomly selected from the total sample and used for analysis. Six models concerning market structure are established to test. Model 1 means non-structured market and model 6 means market structure composed of six sub-markets. It is exploratory approach because any hypothetical market structure is not defined. The result showed that model 1 is insufficient to fit data. It implies that USA automotive market is a structured market. Model 3 with three market structures is significant and identified as the optimal market structure in USA automotive market. Three sub markets are named as USA brands, Asian Brands, and European Brands. And it implies that country of origin effect may exist in USA automotive market. Comparison between modal classification by derived market structures and probabilistic classification by research model was conducted to test how model 3 can correctly classify respondents. The model classify 97% of respondents exactly. The result of this study is different from those of previous research. Previous research used confirmatory approach. Car type and price were chosen as criteria for market structuring and car type-price structure was revealed as the optimal structure for USA automotive market. But this research used exploratory approach without hypothetical market structures. It is not concluded yet which approach is superior. For confirmatory approach, hypothetical market structures should be established exhaustively, because the optimal market structure is selected among hypothetical structures. On the other hand, exploratory approach has a potential problem that validity for derived optimal market structure is somewhat difficult to verify. There also exist market boundary difference between this research and previous research. While previous research analyzed seven car brands, this research analyzed eleven car brands. Both researches seemed to represent entire car market, because cumulative market shares for analyzed brands exceeds 50%. But market boundary difference might affect the different results. Though both researches showed different results, it is obvious that country of origin effect among brands should be considered as important criteria to analyze USA automotive market structure. This research tried to explain heterogeneity of consideration sets among consumers using benefits and two demographic factors, sex and income. Benefit works as a key variable for consumer decision process, and also works as an important criterion in market segmentation. Three factors - trust/safety, image/fun to drive, and economy - are identified among nine benefit related measure. Then the relationship between market structures and independent variables is analyzed using multinomial regression. Independent variables are three benefit factors and two demographic factors. The result showed that all independent variables can be used to explain why there exist different market structures in USA automotive market. For example, a male consumer who perceives all benefits important and has lower income tends to consider domestic brands more than European brands. And the result also showed benefits, sex, and income have an effect to consideration set formation. Though it is generally perceived that a consumer who has higher income is likely to purchase a high priced car, it is notable that American consumers perceived benefits of domestic brands much positive regardless of income. Male consumers especially showed higher loyalty for domestic brands. Managerial implications of this research are as follow. Though implication may be confined to the USA automotive market, the effect of sex on automotive buying behavior should be analyzed. The automotive market is traditionally conceived as male consumers oriented market. But the proportion of female consumers has grown over the years in the automotive market. It is natural outcome that Volvo and Hyundai motors recently developed new cars which are targeted for women market. Secondly, the model used in this research can be applied easier than that of previous researches. Exploratory approach has many advantages except difficulty to apply for practice, because it tends to accompany with complicated model and to require various types of data. The data needed for the model in this research are a few items such as purchased brands, consideration set, some benefits, and some demographic factors and easy to collect from consumers.