• Proceedings of the Korea Database Society Conference
• /
• 1999.06a
• /
• pp.415-421
• /
• 1999

Altman의 연구(1965, 1977)나 Beaver의 연구(1986)와 같은 전통적 예측모형은 분석자의 판단에 따른 예측도가 높은 재무비율을 선정하여 다변량판별분석(MDA： multiple discriminant analysis), 로지스틱회귀분석 등과 같은 통계기법을 주로 이용해 왔으나 1980년 후반부터 인공지능 기법인 귀납적 학습방법, 인공신경망모형, 유전모형 둥이 부실기업예측에 응용되기 시작했다. 최근 연구에서는 인공신경망을 활용한 변수 및 모형개발에 관한 보고가 있다. 그러나 지금까지의 연구가 주로 기업의 재무적 비율지표를 고려한 모형에 치중되었으며 정성적 자료인 비재무지표에 대한 검증과 선정이 자의적으로 이루어져온 경향이었다. 또한 너무 많은 입력변수를 사용할 경우 다중공선성 문제를 유발시킬 위험을 내포하고 있다. 본 연구에서는 부실기업예측모형을 수립하기 위하여 정량적 요인인 재무적 지표변수와 정성적요인인 비재무적 지표변수를 모두 고려하였다. 재무적 지표변수는 상관분석 및 요인분석들을 통하여 유의한 변수들을 도출하였으며 비재무적 지표변수는 조직생태학내에서의 조직군내 조직사멸과 관련된 생태적 과정에 대한 요인들 중 조직군 내적요인으로 조직의 연령, 조직의 규모, 조직의 산업밀도를 도출하여 4개의 실험집단으로 분류하여 비재무적 지표변수를 보완하였다. 인공신경망은 다층퍼셉트론(multi-layer perceptrons)과 역방향 학습(back-propagation )알고리듬으로 입력변수와 출력변수, 그리고 하나의 은닉층을 가지는 3층 퍼셉트론(three layer perceptron)을 사용하였으며 은닉충의 노드(node)수는 3개를 사용하였다. 입력변수로 안정성, 활동성, 수익성, 성장성을 나타내는 재무적 지표변수와 조직규모, 조직연령, 그 조직이 속한 산업의 밀도를 비재무적 지표변수로 산정하여 로지스틱회귀 분석과 인공신경망 기법으로 검증하였다. 로지스틱회귀분석 결과에서는 재무적 지표변수 모형의 전체적 예측적중률이 87.50%인 반면에 재무/비재무적 지표모형은 90.18%로서 비재무적 지표변수 사용에 대한 개선의 효과가 나타났다. 표본기업들을 훈련과 시험용으로 구분하여 분석한 결과는 전체적으로 재무/비재무적 지표를 고려한 인공신경망기법의 예측적중률이 높은 것으로 나타났다. 즉, 로지스틱회귀분석의 재무적 지표모형은 훈련, 시험용이 84.45%, 85.10%인 반면, 재무/비재무적 지표모형은 84.45%, 85.08%로서 거의 동일한 예측적중률을 가졌으나 인공신경망기법 분석에서는 재무적 지표모형이 92.23%, 85.10%인 반면, 재무/비재무적 지표모형에서는 91.12%, 88.06%로서 향상된 예측적 중률을 나타내었다.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 1999.03a
• /
• pp.415-421
• /
• 1999

Altman의 연구(1965, 1977)나 Beaver의 연구(1986)와 같은 전통적 예측모형은 분석자의 판단에 따른 예측도가 높은 재무비율을 선정하여 다변량판별분석(MDA:multiple discriminant analysis), 로지스틱회귀분석 등과 같은 통계기법을 주로 이용해 왔으나 1980년 후반부터 인공지능 기법인 귀납적 학습방법, 인공신경망모형, 유전모형 등이 부실기업예측에 응용되기 시작했다. 최근 연구에서는 인공신경망을 활용한 변수 및 모형개발에 관한 보고가 있다. 그러나 지금까지의 연구가 주로 기업의 재무적 비율지표를 고려한 모형에 치중되었으며 정성적 자료인 비재무지표에 대한 검증과 선정이 자의적으로 이루어져온 경향이었다. 또한 너무 많은 입력변수를 사용할 경우 다중공선성 문제를 유발시킬 위험을 내포하고 있다. 본 연구에서는 부실기업예측모형을 수립하기 위하여 정량적 요인인 재무적 지표변수와 정성적 요인인 비재무적 지표변수를 모두 고려하였다. 재무적 지표변수는 상관분석 및 요인분석들을 통하여 유의한 변수들을 도출하였으며 비재무적 지표변수는 조직생태학내에서의 조직군내 조직사멸과 관련된 생태적 과정에 대한 요인들 중 조직군 내적요인으로 조직의 연령, 조직의 규모, 조직의 산업밀도를 도출하여 4개의 실험집단으로 분류하여 비재무적 지표변수를 보완하였다. 인공신경망은 다층퍼셉트론(multi-layer perceptrons)과 역방향 학습(back-propagation)알고리듬으로 입력변수와 출력변수, 그리고 하나의 은닉층을 가지는 3층 퍼셉트론(three layer perceptron)을 사용하였으며 은닉층의 노드(node)수는 3개를 사용하였다. 입력변수로 안정성, 활동성, 수익성, 성장성을 나타내는 재무적 지표변수와 조직규모, 조직연령, 그 조직이 속한 산업의 밀도를 비재무적 지표변수로 산정하여 로지스틱회귀 분석과 인공신경망 기법으로 검증하였다. 로지스틱회귀분석 결과에서는 재무적 지표변수 모형의 전체적 예측적중률이 87.50%인 반면에 재무/비재무적 지표모형은 90.18%로서 비재무적 지표변수 사용에 대한 개선의 효과가 나타났다. 표본기업들을 훈련과 시험용으로 구분하여 분석한 결과는 전체적으로 재무/비재무적 지표를 고려한 인공신경망기법의 예측적중률이 높은 것으로 나타났다. 즉, 로지스틱회귀 분석의 재무적 지표모형은 훈련, 시험용이 84.45%, 85.10%인 반면, 재무/비재무적 지표모형은 84.45%, 85.08%로서 거의 동일한 예측적중률을 가졌으나 인공신경망기법 분석에서는 재무적 지표모형이 92.23%, 85.10%인 반면, 재무/비재무적 지표모형에서는 91.12%, 88.06%로서 향상된 예측적중률을 나타내었다.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.33-56
• /
• 2016

Many researchers have focused on developing bankruptcy prediction models using modeling techniques, such as statistical methods including multiple discriminant analysis (MDA) and logit analysis or artificial intelligence techniques containing artificial neural networks (ANN), decision trees, and support vector machines (SVM), to secure enhanced performance. Most of the bankruptcy prediction models in academic studies have used financial ratios as main input variables. The bankruptcy of firms is associated with firm's financial states and the external economic situation. However, the inclusion of qualitative information, such as the economic atmosphere, has not been actively discussed despite the fact that exploiting only financial ratios has some drawbacks. Accounting information, such as financial ratios, is based on past data, and it is usually determined one year before bankruptcy. Thus, a time lag exists between the point of closing financial statements and the point of credit evaluation. In addition, financial ratios do not contain environmental factors, such as external economic situations. Therefore, using only financial ratios may be insufficient in constructing a bankruptcy prediction model, because they essentially reflect past corporate internal accounting information while neglecting recent information. Thus, qualitative information must be added to the conventional bankruptcy prediction model to supplement accounting information. Due to the lack of an analytic mechanism for obtaining and processing qualitative information from various information sources, previous studies have only used qualitative information. However, recently, big data analytics, such as text mining techniques, have been drawing much attention in academia and industry, with an increasing amount of unstructured text data available on the web. A few previous studies have sought to adopt big data analytics in business prediction modeling. Nevertheless, the use of qualitative information on the web for business prediction modeling is still deemed to be in the primary stage, restricted to limited applications, such as stock prediction and movie revenue prediction applications. Thus, it is necessary to apply big data analytics techniques, such as text mining, to various business prediction problems, including credit risk evaluation. Analytic methods are required for processing qualitative information represented in unstructured text form due to the complexity of managing and processing unstructured text data. This study proposes a bankruptcy prediction model for Korean small- and medium-sized construction firms using both quantitative information, such as financial ratios, and qualitative information acquired from economic news articles. The performance of the proposed method depends on how well information types are transformed from qualitative into quantitative information that is suitable for incorporating into the bankruptcy prediction model. We employ big data analytics techniques, especially text mining, as a mechanism for processing qualitative information. The sentiment index is provided at the industry level by extracting from a large amount of text data to quantify the external economic atmosphere represented in the media. The proposed method involves keyword-based sentiment analysis using a domain-specific sentiment lexicon to extract sentiment from economic news articles. The generated sentiment lexicon is designed to represent sentiment for the construction business by considering the relationship between the occurring term and the actual situation with respect to the economic condition of the industry rather than the inherent semantics of the term. The experimental results proved that incorporating qualitative information based on big data analytics into the traditional bankruptcy prediction model based on accounting information is effective for enhancing the predictive performance. The sentiment variable extracted from economic news articles had an impact on corporate bankruptcy. In particular, a negative sentiment variable improved the accuracy of corporate bankruptcy prediction because the corporate bankruptcy of construction firms is sensitive to poor economic conditions. The bankruptcy prediction model using qualitative information based on big data analytics contributes to the field, in that it reflects not only relatively recent information but also environmental factors, such as external economic conditions.

• Journal of Intelligence and Information Systems
• /
• v.18 no.2
• /
• pp.29-45
• /
• 2012

Bond rating is regarded as an important event for measuring financial risk of companies and for determining the investment returns of investors. As a result, it has been a popular research topic for researchers to predict companies' credit ratings by applying statistical and machine learning techniques. The statistical techniques, including multiple regression, multiple discriminant analysis (MDA), logistic models (LOGIT), and probit analysis, have been traditionally used in bond rating. However, one major drawback is that it should be based on strict assumptions. Such strict assumptions include linearity, normality, independence among predictor variables and pre-existing functional forms relating the criterion variablesand the predictor variables. Those strict assumptions of traditional statistics have limited their application to the real world. Machine learning techniques also used in bond rating prediction models include decision trees (DT), neural networks (NN), and Support Vector Machine (SVM). Especially, SVM is recognized as a new and promising classification and regression analysis method. SVM learns a separating hyperplane that can maximize the margin between two categories. SVM is simple enough to be analyzed mathematical, and leads to high performance in practical applications. SVM implements the structuralrisk minimization principle and searches to minimize an upper bound of the generalization error. In addition, the solution of SVM may be a global optimum and thus, overfitting is unlikely to occur with SVM. In addition, SVM does not require too many data sample for training since it builds prediction models by only using some representative sample near the boundaries called support vectors. A number of experimental researches have indicated that SVM has been successfully applied in a variety of pattern recognition fields. However, there are three major drawbacks that can be potential causes for degrading SVM's performance. First, SVM is originally proposed for solving binary-class classification problems. Methods for combining SVMs for multi-class classification such as One-Against-One, One-Against-All have been proposed, but they do not improve the performance in multi-class classification problem as much as SVM for binary-class classification. Second, approximation algorithms (e.g. decomposition methods, sequential minimal optimization algorithm) could be used for effective multi-class computation to reduce computation time, but it could deteriorate classification performance. Third, the difficulty in multi-class prediction problems is in data imbalance problem that can occur when the number of instances in one class greatly outnumbers the number of instances in the other class. Such data sets often cause a default classifier to be built due to skewed boundary and thus the reduction in the classification accuracy of such a classifier. SVM ensemble learning is one of machine learning methods to cope with the above drawbacks. Ensemble learning is a method for improving the performance of classification and prediction algorithms. AdaBoost is one of the widely used ensemble learning techniques. It constructs a composite classifier by sequentially training classifiers while increasing weight on the misclassified observations through iterations. The observations that are incorrectly predicted by previous classifiers are chosen more often than examples that are correctly predicted. Thus Boosting attempts to produce new classifiers that are better able to predict examples for which the current ensemble's performance is poor. In this way, it can reinforce the training of the misclassified observations of the minority class. This paper proposes a multiclass Geometric Mean-based Boosting (MGM-Boost) to resolve multiclass prediction problem. Since MGM-Boost introduces the notion of geometric mean into AdaBoost, it can perform learning process considering the geometric mean-based accuracy and errors of multiclass. This study applies MGM-Boost to the real-world bond rating case for Korean companies to examine the feasibility of MGM-Boost. 10-fold cross validations for threetimes with different random seeds are performed in order to ensure that the comparison among three different classifiers does not happen by chance. For each of 10-fold cross validation, the entire data set is first partitioned into tenequal-sized sets, and then each set is in turn used as the test set while the classifier trains on the other nine sets. That is, cross-validated folds have been tested independently of each algorithm. Through these steps, we have obtained the results for classifiers on each of the 30 experiments. In the comparison of arithmetic mean-based prediction accuracy between individual classifiers, MGM-Boost (52.95%) shows higher prediction accuracy than both AdaBoost (51.69%) and SVM (49.47%). MGM-Boost (28.12%) also shows the higher prediction accuracy than AdaBoost (24.65%) and SVM (15.42%)in terms of geometric mean-based prediction accuracy. T-test is used to examine whether the performance of each classifiers for 30 folds is significantly different. The results indicate that performance of MGM-Boost is significantly different from AdaBoost and SVM classifiers at 1% level. These results mean that MGM-Boost can provide robust and stable solutions to multi-classproblems such as bond rating.