An Approach and Software Prototype for Translation of Natural Language Business Rules into Database Structure Andrii Kopp, Dmytro Orlovskyi and Sergey Orekhov National Technical University “Kharkiv Polytechnic Institute”, Kyrpychova str. 2, Kharkiv, 61002, Ukraine Abstract In the recent decades data has indeed become one of the most valuable assets for government institutions, private businesses, and individual persons. Nowadays almost any software, from social networks and dating mobile applications to large information systems and analytical services for enterprise management, accumulates, stores, and processes data to solve certain problems in their subject areas. Extremely large data volumes are organized in databases that are used as the baseline for almost all of modern software applications. As the most important components of software systems, databases should be carefully designed, since drawbacks at the stage of requirements elicitation may result in exponential growth of defects fixing costs at testing and maintenance phases. Therefore, this study proposes an approach and software tool to database schema generation from textual requirements also known in database design domain as business rules. This may help database designers to rapidly obtain usable database schemas in order to detect and fix defects as early as possible. Moreover, proposed solution may simplify the database design process, since database creation scripts are generated from business rules directly. Thus, instead of coding all the required statements, engineers are only need to check obtained schema and make certain adjustments to data types, unique attributes, or used naming style. This research considers relational model and relational databases, since they are most widely used nowadays. State-of-the-art analysis is made, proposed approach is described in details, software tool with its brief usage examples is described, conclusions are made, and further research directions are formulated. Keywords 1 Relational Database Design, Business Rule, Database Schema, Relational Model, Natural Language 1. Introduction Databases are essential components of almost all modern software systems despite their usage area or architectural complexity, or users demographic. Databases could be considered as computer-based structures that store collections of raw facts, valuable for database users (so called end-user data), and metadata (i.e. data about data) that describes how end-user data is managed. Database management systems (DBMS) are specialized software systems that manage database structures, make collections of data persistent and shareable in a secure way [1]. Originally databases and information systems (IS) that use databases were utilized by enterprises (except some small ones), which had data they needed to store in ways which will be easy to retrieve later [2]. In [2] authors made an example of a ledger of names and addresses of persons or other companies that deal with the enterprise. For small businesses such lists could have been kept on paper, in text files, or spreadsheets. However, tremendous growth of business process complexity, data volumes, and information technology adoption made impossible to store enterprise data without using databases. Nowadays, even simple mobile applications, such as to-do lists, address books, or budget managers, use databases to store users’ data in a persistent secure COLINS-2021: 5th International Conference on Computational Linguistics and Intelligent Systems, April 22–23, 2021, Kharkiv, Ukraine EMAIL: kopp93@gmail.com (A. Kopp); orlovskyi.dm@gmail.com (D. Orlovskyi); sergey.v.orekhov@gmail.com (S. Orekhov) ORCID: 0000-0002-3189-5623 (A. Kopp); 0000-0002-8261-2988 (D. Orlovskyi); 0000-0002-5040-5861 (S. Orekhov) ©฀ 2021 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CEUR Workshop Proceedings (CEUR-WS.org)