RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

IBS standardized data in Severstal using machine learning

Customers: Severstal

Moscow; Metallurgical industry

Product: SAP Master Data Governance (SAP MDG)

Project date: 2019/02  - 2019/07

Content

2019

On July 1, 2019 the IBS company reported that Severstal automated data management of reference books of partners and material resources based on SAP MDG. According to the company, the project was implemented by forces of division of SAP Digital Business Services. Besides, within the project normalization of reference books, i.e. their reduction to a standard type, including using machine learning technology for fast data processing was carried out. As the technology partner of PJSC Severstal on normalization of reference books the IBS company acted.

Prerequisites

PJSC Severstal is steel-making and mining company for July, 2019 with 50 enterprises in Russia and abroad. Reference books of the material resources (MR) and partners allow to plan requirements and warehouse stocks, to control delivery performance accuracy.

Over time in reference books inaccuracies collect, and for improvement of quality of key business write processes should be normalized — are brought to the standardized type.

Project Tasks

The Severstal company faced such task. The situation was complicated by amount of data in the company (410 thousand materials and partners), besides, a part of records was conducted in the English and French languages.

Project Progress

According to the results of an open tender the IBS company became the partner of the project. To reduce the cost and terms of the project and to improve quality of data processing, the IBS command applied machine learning technologies.

File:Aquote1.png
Algorithms machine learning undertook a routine: processing of text names MTR, determination of characteristics, filling of templates. Specialists were responsible only for training of algorithms and check of results of machining. If one expert has a processing regulation — two thousand records a month, then the machine algorithm is capable to process them during the day.

told Bair Danilov, the head NSI in IBS
File:Aquote2.png

In the project was about 250 approving from 50 enterprises worldwide. Thanks to own instrument of the online approval "IBS Normalization" the customer could far off, in convenient time and to approve content with a minimum of costs and to make changes.

After normalization of data the IBS command executed their migration in the SAP MDG systems for further automation of maintaining reference materials of PJSC Severstal.

Project Results

File:Aquote1.png
Normalization of reference books allowed to unify and simplify identification of warehouse stocks and their involvement in other processes.

told Irina Povoroznik, the associate director on raw materials sales, purchases and logistics of Severstal company
File:Aquote2.png