A comparative study of forecasting corporate credit ratings using neural networks, support vector machines, and decision trees

Golbayani, Parisa; Florescu, Ionuţ; Chatterjee, Rupak

doi:10.1016/j.najef.2020.101251

Cited by 64 publications

(48 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The output signal is determined by passing all collected signals through an activation function. As shown in Equation 2, 𝑋 is an input vector and y is a perceptron that produces a single output [18]. w is the value of the input weights, the b bias value and Φ a value corresponding to the nonlinear activation function.…”

Section: Artificial Neural Networkmentioning

confidence: 99%

Using Machine Learning Methods to Predict the Effect of Alternative Fuel Mixtures on Exhaust Emissions

BİLBAN¹,

Aydoğan²

2022

European Journal of Science and Technology

View full text Add to dashboard Cite

Machine learning is a science that deals with the design and development processes of algorithms that enable data-based learning. Machine learning methods try to find the most suitable model for new data prediction processes by using the past data. In this study, the data obtained from the engine trials with fuel mixtures of 5%, 10%, 15% by volume using 1-Propanol, 2-Propanol, AVGAS and gasoline fuel were used. Obtained data were compared with 100% gasoline values. In the study, a 4-cylinder engine with direct injection and turbocharging was used. With the obtained measurement results, a database was created to be used in machine learning. With the created database, estimation processes were carried out on ANN, GBA, SVM and AB machine learning models. At the end of the study, it was found that the most suitable model for the estimation of CO, CO2, HC, O2 values was ANN with an R 2 value of 0.9999. For the NO value, it was determined that the AB method was used with an R 2 value of 0.9996. In the estimation process of the CO value, GBA and AB methods are other machine learning methods that can be used as they have a higher value than 0.99 R 2 . CO2, HC and O2, and in the output value estimation process, GBA and AB are other methods that can be used instead of ANN as they have a higher value than 0.99 R 2 . It has been found that there is another machine learning method that can be used for NO value estimation, with an AB 0.99 R 2 value.

show abstract

Section: Artificial Neural Networkmentioning

confidence: 99%

Using Machine Learning Methods to Predict the Effect of Alternative Fuel Mixtures on Exhaust Emissions

BİLBAN¹,

Aydoğan²

2022

European Journal of Science and Technology

View full text Add to dashboard Cite

show abstract

“…The ability of the SVM method to produce accurate results with less data is important in choosing the method. According to Golbayani et al (2020), credit rating institutions tend to request long-term data. The long wait reduces the opportunity to work.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Successful machine learning programs can provide rapid analysis of credit scores. As a result of the research, it has been determined that neural networks and support vector machine methods provide superior performance compared to other methods [18].…”

Section: Literature Reviewmentioning

confidence: 99%

Data Mining Application for Financial Decision Optimization at Risk

KOÇOĞLU¹,

Ersöz

2021

International Journal of 3D Printing Technologies and Digital Industry

View full text Add to dashboard Cite

Financial decisions can add value to the existence of businesses or individuals, as well as a wrong financial decision can cause businesses to cease to exist. Hence, financial decision or financial assumptions are vital for businesses or individuals. In financial assumptions, risk refers to the probability of losing as a result of an investment made in an asset. Measures can be taken against possible risks in the future through financial assumptions. In this study, the Logistic Regression Analysis (LR), one of the traditional methods, and the machine learning algorithm, Support Vector Machines (SVM) technique, which is one of the new approaches, are compared in the loaning process. It is aimed to determine the importance of the compared methods, the accuracy of the model, the estimation power of the model, the estimation performance of the model, the determination of the importance of the independent variables that affect the non-repayment of the loan, and the superiority of the techniques. According to the analysis results, the SVM technique is superior to the LR technique in calculating accuracy rate and prediction rate, and the LR technique is superior to the SVM technique in assumption performance calculation. The most significant variable in the SVM technique is "Lending policy", the most significant variable in the LR technique is "Interest rate", the second significant variable is "Interest rate" in the SVM technique, and "Lending Policy" as the second important variable in the LR technique. It is seen that the third most crucial variable in the two techniques is the "Income" variable. The determination of the SVM technique as the more important variable of the loan policy is deemed more suitable to the opinion of the banking expert. Detecting more realistic results of the SVM technique compared to the LR technique has shown the superiority of the SVM technique.

show abstract

“…In previous studies, the related DT algorithms have typically been applied to functions of the relevant areas of specialization, such as churn prediction [ 51 ], forecasting corporate credit ratings [ 52 ], object classification in autonomous driving [ 53 ], and the identification of a route selection strategy in classification [ 49 ]. Thus, it is an effective method for the selection of medical research.…”

Section: Related Workmentioning

confidence: 99%

Identification of the Framingham Risk Score by an Entropy-Based Rule Model for Cardiovascular Disease

Chen

Cheng

Chen

et al. 2020

Entropy

View full text Add to dashboard Cite

Since 2001, cardiovascular disease (CVD) has had the second-highest mortality rate, about 15,700 people per year, in Taiwan. It has thus imposed a substantial burden on medical resources. This study was triggered by the following three factors. First, the CVD problem reflects an urgent issue. A high priority has been placed on long-term therapy and prevention to reduce the wastage of medical resources, particularly in developed countries. Second, from the perspective of preventive medicine, popular data-mining methods have been well learned and studied, with excellent performance in medical fields. Thus, identification of the risk factors of CVD using these popular techniques is a prime concern. Third, the Framingham risk score is a core indicator that can be used to establish an effective prediction model to accurately diagnose CVD. Thus, this study proposes an integrated predictive model to organize five notable classifiers: the rough set (RS), decision tree (DT), random forest (RF), multilayer perceptron (MLP), and support vector machine (SVM), with a novel use of the Framingham risk score for attribute selection (i.e., F-attributes first identified in this study) to determine the key features for identifying CVD. Verification experiments were conducted with three evaluation criteria—accuracy, sensitivity, and specificity—based on 1190 instances of a CVD dataset available from a Taiwan teaching hospital and 2019 examples from a public Framingham dataset. Given the empirical results, the SVM showed the best performance in terms of accuracy (99.67%), sensitivity (99.93%), and specificity (99.71%) in all F-attributes in the CVD dataset compared to the other listed classifiers. The RS showed the highest performance in terms of accuracy (85.11%), sensitivity (86.06%), and specificity (85.19%) in most of the F-attributes in the Framingham dataset. The above study results support novel evidence that no classifier or model is suitable for all practical datasets of medical applications. Thus, identifying an appropriate classifier to address specific medical data is important. Significantly, this study is novel in its calculation and identification of the use of key Framingham risk attributes integrated with the DT technique to produce entropy-based decision rules of knowledge sets, which has not been undertaken in previous research. This study conclusively yielded meaningful entropy-based knowledgeable rules in tree structures and contributed to the differentiation of classifiers from the two datasets with three useful research findings and three helpful management implications for subsequent medical research. In particular, these rules provide reasonable solutions to simplify processes of preventive medicine by standardizing the formats and codes used in medical data to address CVD problems. The specificity of these rules is thus significant compared to those of past research.

show abstract

A comparative study of forecasting corporate credit ratings using neural networks, support vector machines, and decision trees

Cited by 64 publications

References 63 publications

Using Machine Learning Methods to Predict the Effect of Alternative Fuel Mixtures on Exhaust Emissions

Using Machine Learning Methods to Predict the Effect of Alternative Fuel Mixtures on Exhaust Emissions

Data Mining Application for Financial Decision Optimization at Risk

Identification of the Framingham Risk Score by an Entropy-Based Rule Model for Cardiovascular Disease

Contact Info

Product

Resources

About