Optimization of Multiclass Support Vector Machine using Genetic Algorithm: Application to the Prediction of Corporate Credit Rating (유전자 알고리즘을 이용한 다분류 SVM의 최적화: 기업신용등급 예측에의 응용)
-
- Information Systems Review
- /
- v.16 no.3
- /
- pp.161-177
- /
- 2014
Corporate credit rating assessment consists of complicated processes in which various factors describing a company are taken into consideration. Such assessment is known to be very expensive since domain experts should be employed to assess the ratings. As a result, the data-driven corporate credit rating prediction using statistical and artificial intelligence (AI) techniques has received considerable attention from researchers and practitioners. In particular, statistical methods such as multiple discriminant analysis (MDA) and multinomial logistic regression analysis (MLOGIT), and AI methods including case-based reasoning (CBR), artificial neural network (ANN), and multiclass support vector machine (MSVM) have been applied to corporate credit rating.2) Among them, MSVM has recently become popular because of its robustness and high prediction accuracy. In this study, we propose a novel optimized MSVM model, and appy it to corporate credit rating prediction in order to enhance the accuracy. Our model, named 'GAMSVM (Genetic Algorithm-optimized Multiclass Support Vector Machine),' is designed to simultaneously optimize the kernel parameters and the feature subset selection. Prior studies like Lorena and de Carvalho (2008), and Chatterjee (2013) show that proper kernel parameters may improve the performance of MSVMs. Also, the results from the studies such as Shieh and Yang (2008) and Chatterjee (2013) imply that appropriate feature selection may lead to higher prediction accuracy. Based on these prior studies, we propose to apply GAMSVM to corporate credit rating prediction. As a tool for optimizing the kernel parameters and the feature subset selection, we suggest genetic algorithm (GA). GA is known as an efficient and effective search method that attempts to simulate the biological evolution phenomenon. By applying genetic operations such as selection, crossover, and mutation, it is designed to gradually improve the search results. Especially, mutation operator prevents GA from falling into the local optima, thus we can find the globally optimal or near-optimal solution using it. GA has popularly been applied to search optimal parameters or feature subset selections of AI techniques including MSVM. With these reasons, we also adopt GA as an optimization tool. To empirically validate the usefulness of GAMSVM, we applied it to a real-world case of credit rating in Korea. Our application is in bond rating, which is the most frequently studied area of credit rating for specific debt issues or other financial obligations. The experimental dataset was collected from a large credit rating company in South Korea. It contained 39 financial ratios of 1,295 companies in the manufacturing industry, and their credit ratings. Using various statistical methods including the one-way ANOVA and the stepwise MDA, we selected 14 financial ratios as the candidate independent variables. The dependent variable, i.e. credit rating, was labeled as four classes: 1(A1); 2(A2); 3(A3); 4(B and C). 80 percent of total data for each class was used for training, and remaining 20 percent was used for validation. And, to overcome small sample size, we applied five-fold cross validation to our dataset. In order to examine the competitiveness of the proposed model, we also experimented several comparative models including MDA, MLOGIT, CBR, ANN and MSVM. In case of MSVM, we adopted One-Against-One (OAO) and DAGSVM (Directed Acyclic Graph SVM) approaches because they are known to be the most accurate approaches among various MSVM approaches. GAMSVM was implemented using LIBSVM-an open-source software, and Evolver 5.5-a commercial software enables GA. Other comparative models were experimented using various statistical and AI packages such as SPSS for Windows, Neuroshell, and Microsoft Excel VBA (Visual Basic for Applications). Experimental results showed that the proposed model-GAMSVM-outperformed all the competitive models. In addition, the model was found to use less independent variables, but to show higher accuracy. In our experiments, five variables such as X7 (total debt), X9 (sales per employee), X13 (years after founded), X15 (accumulated earning to total asset), and X39 (the index related to the cash flows from operating activity) were found to be the most important factors in predicting the corporate credit ratings. However, the values of the finally selected kernel parameters were found to be almost same among the data subsets. To examine whether the predictive performance of GAMSVM was significantly greater than those of other models, we used the McNemar test. As a result, we found that GAMSVM was better than MDA, MLOGIT, CBR, and ANN at the 1% significance level, and better than OAO and DAGSVM at the 5% significance level.
The study was designed to gain necessary basic data order to grasp health knowledge, attitude, practice level of middle and high school students and to analyse th problem and to point out the method of improvement in the field of school health education. The survery was carried out through this reporter's interview for 2,400 students who attend to ten schools in Young Nam area during the period of a month from 25 the June to 25th July 1989. The result of this study can be summaried as follows. 1. The total number of answers on the question was 2,346. As for general characteristics the percent of female middle school students was 60.6% and the percent of male students was 77.7%, 45.9% of high school students was evening school students. 52.9% of middle school students and 42.3% of high school students were borne in rural area. 2. The percentage of unknown and misunderstanding for Epidemic Hepatitis infection was 46.3% of middle school students and 29.6% of high school students. 3. The percentage of unknown and misunderstanding for Epidemic Hemorrhage fever infection was 85.6% of middle school students and 66.9% of high school students. 4. The percentage of right knowledge for AIDS infection was 66.0% of middle school students and 90.4% of high school students. 5. The percentage of right knowledge for Typhoid infection was 47.8% of middle school students and 69.4% of high school students. 6. The percentage of unknown and misunderstanding for Tuberculosis infection was 71.6% of middle school students and 62.2% of high school students. 7. As for personal hygiene, the percentage of toothbrushing after every meal was high level : 44.2% of middle school students and 42.0% of high school students. 8. 60.9% of middle school students take a bath twice a week, 49.2% oh high school students take a bath a week. Times of bath of middle school students was higher than that of high school students. 9.The percentage of washing hand after using toilet was 42.1% of middle school students and 35.1% of high school students. 49.0% of middle school students and 55.1% of high school students wash hand sometimes after using toilet. 10. The percentage of change of underwear twice a week was 57.6% of middle school students and 49.8% of high school students. 11. The percentage of habit of unbalanced diet was 30.% of middle school students and 27.6% of high school students. 50.8% of middle school students and 51.7% of high school students have balanced diet. 12. Index of health practice of personal hygiene can be summarized as follows. A. A case of middle school students. 1) The percentage of health practice index in male and female was 49.6% and 48.1% respectively. Index of female students was higher than that of male students. 2) As for parent's occupation, public servants and company emplyee was upper level. Farming was low level. 3) As for income level, middle, level with 56.5% was highest in high income level and low level with 27.4% was highest in low income level. B. A case of high school students. 1) Middle level of health practice index was 46.0% of male students, upper and low level was 32.4% and 28.0% of female students respectively. 2) Middle level of health practice index was high in farming and company employee and upper level was high in commerce and service, low level with 60.0% was high in unemployed. 3) Upper practice index 35.7% appears in the rich and low practice index 38.3% appears in the poor. 13. Average points of Health practice about personal hygiene were as follows. (Full marks at 4). A. A case of middle school. Female (1.87 point) was higher than male (1.26 point). Night time (2.03 point) was higher than day time (1.66 point) and middle or small cities (2.17 point) are high than any other places. As for parent's occupation, students whose parents are company clerk get high marks (2.32) and ten students whose parent's job are service get next high marks (2.20). B. A case of high school. Female (1.53 point) was higher than male (1.22 point), as parents educational level were higher the point were higher, and as income level was higher, the points of health practice (1.78) were higher, and as for parents occupation, service get highest point (1.93) and commerce get next high point (1.86) public servant get low point (1.66). 14. The percentage of experience in smoking was 11.9% of middle school students and 60.9% of high school students. 15. The percentage of experience in inhalation of bond and administrating LSD was 4.3% of male middle school students, 8.4% of female middle school students, 6.9% of male high school students and 4.2% of female high school students. The knowledge level of communicable disease infection are very low in middle and high school students and practice level of personal hygiene are also very low. As a whole we can evaluate that middle and high school students are low level of health knowledge and practice. In conclusion, we must consider preparation for school health education program through establishing of health subjects in the carriculum, and securing of health education teachers and using materials and media program of health education. It is very important to establish macroscopic policy and strategy for public health education and to get people have right knowledge and practice for health.