Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19
Because COVID-19 occurred in 2019, the behavioxr of humans has been changed and it will influence the business model of enterprise. Enterprise cannot predict its development according to past knowledge and experiment; so, it needs a new machine learning framework to predict enterprise performance. T...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-09-01
|
Series: | Mathematics |
Subjects: | |
Online Access: | https://www.mdpi.com/2227-7390/9/18/2215 |
_version_ | 1797518359058710528 |
---|---|
author | Jung-Kai Tsai Chih-Hsing Hung |
author_facet | Jung-Kai Tsai Chih-Hsing Hung |
author_sort | Jung-Kai Tsai |
collection | DOAJ |
description | Because COVID-19 occurred in 2019, the behavioxr of humans has been changed and it will influence the business model of enterprise. Enterprise cannot predict its development according to past knowledge and experiment; so, it needs a new machine learning framework to predict enterprise performance. The goal of this research is to modify AdaBoost to reasonably predict the enterprise performance. In order to justify the usefulness of the proposed model, enterprise data will be collected and the proposed model can be used to predict the enterprise performance after COVID-19. The test data correct rate of the proposed model will be compared with some of the traditional machine learning models. Compared with the traditional AdaBoost, back propagation neural network (BPNN), regression classifier, support vector machine (SVM) and support vector regression (SVR), the proposed method possesses the better classification ability (average correct rate of the proposed method is 88.04%) in handling two classification problems. Compared with traditional AdaBoost, one-against-all SVM, one-against-one SVM, one-against-all SVR and one-against-one SVR, the classification ability of the proposed method is also relatively better for coping with the multi-class classification problem. Finally, some conclusions and future research will be discussed at the end. |
first_indexed | 2024-03-10T07:27:48Z |
format | Article |
id | doaj.art-2c34604cd3754611946efe452e55e806 |
institution | Directory Open Access Journal |
issn | 2227-7390 |
language | English |
last_indexed | 2024-03-10T07:27:48Z |
publishDate | 2021-09-01 |
publisher | MDPI AG |
record_format | Article |
series | Mathematics |
spelling | doaj.art-2c34604cd3754611946efe452e55e8062023-11-22T14:05:00ZengMDPI AGMathematics2227-73902021-09-01918221510.3390/math9182215Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19Jung-Kai Tsai0Chih-Hsing Hung1Ph.D. Program in Finance and Banking, National Kaohsiung University of Science and Technology, Kaohsiung City 811532, TaiwanDepartment of Money and Banking, National Kaohsiung University of Science and Technology, Kaohsiung City 811532, TaiwanBecause COVID-19 occurred in 2019, the behavioxr of humans has been changed and it will influence the business model of enterprise. Enterprise cannot predict its development according to past knowledge and experiment; so, it needs a new machine learning framework to predict enterprise performance. The goal of this research is to modify AdaBoost to reasonably predict the enterprise performance. In order to justify the usefulness of the proposed model, enterprise data will be collected and the proposed model can be used to predict the enterprise performance after COVID-19. The test data correct rate of the proposed model will be compared with some of the traditional machine learning models. Compared with the traditional AdaBoost, back propagation neural network (BPNN), regression classifier, support vector machine (SVM) and support vector regression (SVR), the proposed method possesses the better classification ability (average correct rate of the proposed method is 88.04%) in handling two classification problems. Compared with traditional AdaBoost, one-against-all SVM, one-against-one SVM, one-against-all SVR and one-against-one SVR, the classification ability of the proposed method is also relatively better for coping with the multi-class classification problem. Finally, some conclusions and future research will be discussed at the end.https://www.mdpi.com/2227-7390/9/18/2215enterprise performancemachine learningAdaBoostCOVID-19 |
spellingShingle | Jung-Kai Tsai Chih-Hsing Hung Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19 Mathematics enterprise performance machine learning AdaBoost COVID-19 |
title | Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19 |
title_full | Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19 |
title_fullStr | Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19 |
title_full_unstemmed | Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19 |
title_short | Improving AdaBoost Classifier to Predict Enterprise Performance after COVID-19 |
title_sort | improving adaboost classifier to predict enterprise performance after covid 19 |
topic | enterprise performance machine learning AdaBoost COVID-19 |
url | https://www.mdpi.com/2227-7390/9/18/2215 |
work_keys_str_mv | AT jungkaitsai improvingadaboostclassifiertopredictenterpriseperformanceaftercovid19 AT chihhsinghung improvingadaboostclassifiertopredictenterpriseperformanceaftercovid19 |