Repair missing data to improve corporate credit risk prediction accuracy with multi-layer perceptron

Yang, Mei; Lim, Ming K.; Qu, Yingchi; Li, Xingzhi; Ni, Du

doi:10.1007/s00500-022-07277-4

Cited by 5 publications

(4 citation statements)

References 42 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lin et al conducted research on the effect of data discretization for continuous data, where MLP and DBN were significantly superior to the mentioned baseline imputation methods [ 40 ]. To repair missing data for credit risk prediction, Yang et al developed an ensemble MLP model with superior accuracy to the traditional machine learning model, which testified that repairing missing data can improve the model’s prediction ability [ 41 ]. However, more comprehensive consideration of missing mechanisms and missing rates is first required for wide application of the technique to imputing missing data.…”

Section: Materials and Methodologymentioning

confidence: 99%

Discrete Missing Data Imputation Using Multilayer Perceptron and Momentum Gradient Descent

Pan

et al. 2022

Sensors

View full text Add to dashboard Cite

Data are a strategic resource for industrial production, and an efficient data-mining process will increase productivity. However, there exist many missing values in data collected in real life due to various problems. Because the missing data may reduce productivity, missing value imputation is an important research topic in data mining. At present, most studies mainly focus on imputation methods for continuous missing data, while a few concentrate on discrete missing data. In this paper, a discrete missing value imputation method based on a multilayer perceptron (MLP) is proposed, which employs a momentum gradient descent algorithm, and some prefilling strategies are utilized to improve the convergence speed of the MLP. To verify the effectiveness of the method, experiments are conducted to compare the classification accuracy with eight common imputation methods, such as the mode, random, hot-deck, KNN, autoencoder, and MLP, under different missing mechanisms and missing proportions. Experimental results verify that the improved MLP model (IMLP) can effectively impute discrete missing values in most situations under three missing patterns.

show abstract

Section: Materials and Methodologymentioning

confidence: 99%

Discrete Missing Data Imputation Using Multilayer Perceptron and Momentum Gradient Descent

Pan

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Commonly used machine learning algorithms in CCR prediction include neural NN, SVM, LR, and ensemble algorithms (Bhatore et al, 2020;Lappas & Yannacopoulos, 2021;Ma & Lv, 2019;Yang et al, 2022). Take NN as an example.…”

Section: Corporate Credit Risk Predictionmentioning

confidence: 99%

Deep neural networks with L1 and L2 regularization for high dimensional corporate credit risk prediction

Yang

Lim

et al. 2023

Expert Systems with Applications

Self Cite

View full text Add to dashboard Cite

“…In 2022, M. Yang et al [35] proposed a multi-layer perceptron model that can predict risks in software development. The experimental findings show that the prediction accuracy is 83.11%, which is higher than that of typical machine learning models.…”

Section: Literature Reviewmentioning

confidence: 99%

“…The proposed Multihead-CNN obtains a high accuracy of 99.16% with a small error rate with 50 training epochs in the identification of risks. The proposed multi-head CNN network has been compared with existing techniques such as SVM [34], MLP [35], GLM [36], and CNN [37] using specific parameters such as accuracy, specificity, and sensitivity which are shown in Fig 6 . From the figure, it is clear that the proposed method achieves better accuracy than existing techniques. The proposed multi-head CNN network has been compared with existing techniques such as SVM [34], MLP [35], GLM [36], and CNN [37] using specific parameters such as accuracy, specificity, and sensitivity which are shown in Fig.…”

Section: Performance Analysismentioning

confidence: 99%

Multi-Head CNN-based Software Development Risk Classification

Ziana M.,

2023

Int. j. electr. comput. eng. syst. (Online)

View full text Add to dashboard Cite

Agile methodology for software development has been in vogue for a few decades, notably among small and medium enterprises. The omission of an explicit risk identification approach turns a blind eye to a range of perilous risks, thus dumping the management into strenuous situations and precipitating dreadful issues at the crucial stages of the project. To overcome this drawback a novel Agile Software Risk Identification using Deep learning (ASRI-DL) approach has been proposed that uses a deep learning technique along with the closed fishbowl strategy, thus assisting the team in finding the risks by molding them to think from diverse perspectives, enhancing wider areas of risk coverage. The proposed technique uses a multi-head Convolutional Neural Network (Multihead-CNN) method for classifying the risk into 11 classes such as over-doing, under-doing, mistakes, concept risks, changes, differences, difficulties, dependency, conflicts, issues, and challenges in terms of producing a higher number of risks concerning score, criticality, and uniqueness of the risk ideas. The descriptive statistics further demonstrate that the participation and risk coverage of the individuals in the proposed methodology exceeded the other two as a result of applying the closed fishbowl strategy and making use of the risk identification aid. The proposed method has been compared with existing techniques such as Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), Generalized Linear Models (GLM), and CNN using specific parameters such as accuracy, specificity, and sensitivity. Experimental findings show that the proposed ASRI-DL technique achieves a classification accuracy of 99.16% with a small error rate with 50 training epochs respectively.

show abstract

Repair missing data to improve corporate credit risk prediction accuracy with multi-layer perceptron

Cited by 5 publications

References 42 publications

Discrete Missing Data Imputation Using Multilayer Perceptron and Momentum Gradient Descent

Discrete Missing Data Imputation Using Multilayer Perceptron and Momentum Gradient Descent

Deep neural networks with L1 and L2 regularization for high dimensional corporate credit risk prediction

Multi-Head CNN-based Software Development Risk Classification

Contact Info

Product

Resources

About