Machine learning methods using for product suppliers evaluation
The paper is dedicated to the problem of the subjective factors influence on the choice of supplier. To make an objective decision on choosing a product supplier, machine learning models applying is suggested. Due to the use of machine learning models, the evaluation of suppliers is formed based on...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Peter the Great St. Petersburg Polytechnic University
2020-12-01
|
| Series: | Научно-технические ведомости СПбГПУ: Экономические науки |
| Subjects: | |
| Online Access: | https://economy.spbstu.ru/article/2020.86.07/ |
| _version_ | 1818718671165456384 |
|---|---|
| author | Potsulin Anton Sergeeva Irina Rudenko Vyacheslav |
| author_facet | Potsulin Anton Sergeeva Irina Rudenko Vyacheslav |
| author_sort | Potsulin Anton |
| collection | DOAJ |
| description | The paper is dedicated to the problem of the subjective factors influence on the choice of supplier. To make an objective decision on choosing a product supplier, machine learning models applying is suggested. Due to the use of machine learning models, the evaluation of suppliers is formed based on the analysis of the results of their activities, which minimizes the influence of subjective factors on the choice of supplier. The paper presents the results of the research based on a set of data, including the information obtained during the analysis of the annual report of the purchasing department of a meat processing enterprise, as well as open information published on the Rosselkhoznadzor (Federal Service for Veterinary and Phytosanitary Surveillance of Russian Federation) official site. A sample was formed for training the model for classifying suppliers into reliable and unreliable ones. Methods such as logistic regression and decision tree are used to solve the problem of supplier classification. Ordinal scales are proposed for evaluating suppliers, based on such criteria as the availability and correctness of the design of the product accompanying documentation, compliance with labeling, the presence of reactions to reviews, etc. This made it possible to design the structure of a database containing information about suppliers. In accordance with the specified structure, a sample was formed for training the model for classifying suppliers into reliable and unreliable ones. Methods such as logistic regression and decision tree are used to solve the problem of supplier classification. A comparative analysis of the selected methods was performed using the AUC metric. Modifications of the composition criteria will allow to use the proposed method not only for evaluating the suppliers of food products. Due to machine learning models using, the evaluation of suppliers is formed based on the analysis of their performance, which reduces the influence of subjective factors. The obtained results can simplify the process of selecting suppliers, promote competition in the commodity markets of the Russian Federation, allow enterprises to reduce management costs and save time on searching, evaluating and selecting bona fide suppliers. |
| first_indexed | 2024-12-17T19:54:45Z |
| format | Article |
| id | doaj.art-f13efc07f2de4d9ca2e9103fa0930c6f |
| institution | Directory Open Access Journal |
| issn | 2304-9774 2618-8678 |
| language | English |
| last_indexed | 2024-12-17T19:54:45Z |
| publishDate | 2020-12-01 |
| publisher | Peter the Great St. Petersburg Polytechnic University |
| record_format | Article |
| series | Научно-технические ведомости СПбГПУ: Экономические науки |
| spelling | doaj.art-f13efc07f2de4d9ca2e9103fa0930c6f2022-12-21T21:34:38ZengPeter the Great St. Petersburg Polytechnic UniversityНаучно-технические ведомости СПбГПУ: Экономические науки2304-97742618-86782020-12-0113610.18721/JE.1360720714726Machine learning methods using for product suppliers evaluationPotsulin Anton0https://orcid.org/0000-0003-1083-5442Sergeeva Irina1https://orcid.org/0000-0001-7314-7765Rudenko Vyacheslav2https://orcid.org/0000-0002-5762-3386St.Peterburg National Research University of Information Technologies, Mechanics and OpticsSt.Peterburg National Research University of Information Technologies, Mechanics and OpticsSt.Peterburg National Research University of Information Technologies, Mechanics and OpticsThe paper is dedicated to the problem of the subjective factors influence on the choice of supplier. To make an objective decision on choosing a product supplier, machine learning models applying is suggested. Due to the use of machine learning models, the evaluation of suppliers is formed based on the analysis of the results of their activities, which minimizes the influence of subjective factors on the choice of supplier. The paper presents the results of the research based on a set of data, including the information obtained during the analysis of the annual report of the purchasing department of a meat processing enterprise, as well as open information published on the Rosselkhoznadzor (Federal Service for Veterinary and Phytosanitary Surveillance of Russian Federation) official site. A sample was formed for training the model for classifying suppliers into reliable and unreliable ones. Methods such as logistic regression and decision tree are used to solve the problem of supplier classification. Ordinal scales are proposed for evaluating suppliers, based on such criteria as the availability and correctness of the design of the product accompanying documentation, compliance with labeling, the presence of reactions to reviews, etc. This made it possible to design the structure of a database containing information about suppliers. In accordance with the specified structure, a sample was formed for training the model for classifying suppliers into reliable and unreliable ones. Methods such as logistic regression and decision tree are used to solve the problem of supplier classification. A comparative analysis of the selected methods was performed using the AUC metric. Modifications of the composition criteria will allow to use the proposed method not only for evaluating the suppliers of food products. Due to machine learning models using, the evaluation of suppliers is formed based on the analysis of their performance, which reduces the influence of subjective factors. The obtained results can simplify the process of selecting suppliers, promote competition in the commodity markets of the Russian Federation, allow enterprises to reduce management costs and save time on searching, evaluating and selecting bona fide suppliers.https://economy.spbstu.ru/article/2020.86.07/supplier evaluationsupplier evaluation methodslogistic regressiondecision treeevaluation criteriamachine learning |
| spellingShingle | Potsulin Anton Sergeeva Irina Rudenko Vyacheslav Machine learning methods using for product suppliers evaluation Научно-технические ведомости СПбГПУ: Экономические науки supplier evaluation supplier evaluation methods logistic regression decision tree evaluation criteria machine learning |
| title | Machine learning methods using for product suppliers evaluation |
| title_full | Machine learning methods using for product suppliers evaluation |
| title_fullStr | Machine learning methods using for product suppliers evaluation |
| title_full_unstemmed | Machine learning methods using for product suppliers evaluation |
| title_short | Machine learning methods using for product suppliers evaluation |
| title_sort | machine learning methods using for product suppliers evaluation |
| topic | supplier evaluation supplier evaluation methods logistic regression decision tree evaluation criteria machine learning |
| url | https://economy.spbstu.ru/article/2020.86.07/ |
| work_keys_str_mv | AT potsulinanton machinelearningmethodsusingforproductsuppliersevaluation AT sergeevairina machinelearningmethodsusingforproductsuppliersevaluation AT rudenkovyacheslav machinelearningmethodsusingforproductsuppliersevaluation |