NLP MODULE FOR BULGARIAN TEXT PROCESSING

Abstract

The wide use of web-based information systems and a lack of highly skilled developers are the primary motivation to search for methods and approaches to optimize the building of such systems. This paper describes a model for creating web-based information systems by using a core of reusable, independent, and installable base modules. Such a system is easily adapted to a client’s needs and is extendable by adding specific modules that interact with the remainder of the system by following certain rules. The approach allows flexible and rapid development of applications for small to extremely large web-based systems, simply by adding modules with adequate functionality. The growing demand of Bulgarian customers for such systems is the reason for building a base module for automatic processing of Bulgarian text. This paper presents a module that performs automatic morphological analysis and synthesis, verifies syntactic agreement, automatically places stress, and processes complex verb forms, among other functions. The described functionality can be integrated with other modules using a suitable interface