TY Software Ltd. is a software development company established in 2001 in METU Technopolis. It’s focused on Natural Language Processing (NLP) and computational linguistics technologies. TY took part in the development of first industry-level Machine Translation (MT) systems for several language-pairs. TY provides a suite of linguistic tools including POS tagger, text preprocessor, proper-noun recognizer, morphological analyzer, syntactic Parser, all with API and SDK integration tools. These tools compose a robust and powerful language independent MT development framework.

• Text preprocessor

• Rule based and language independent.

• Generic rule engine can be used for different purposes with different rules (e.g. punctuator or sentence chunk-ing).

• Rule syntax supports complex boolean expressions and special linguistic functions.

• Rule engine is fast enough to process large texts using thousands of rules.

• Morphological Analyzer

• Rule based and language independent.

• Derives uninflected forms of words for analysis purpose.

• Builds inflected forms of stems for generation purpose.

• Syntactic Parser

• Rule based and language independent.

• Robust unification chart parser.

• Powerful rule syntax.

• POS Tagger

• HMM based and language independent.

• Can be trained for different languages by a tagged corpus.

• Proper-Noun Recognizer

• Rule based.

• Human names, geographical names, etc. in Turkish.