A SENTENCE GENERATION METHOD FOR RULE-BASED LANGUAGES; AN EXAMPLE FROM TURKISH

Abstract: The most widely used language families today are Altaic, Indo-European, Hami-Semitic and Sino-Tietic. When these language families are evaluated in terms of phonology, morphology and syntax, it is clear that Altaic languages are the most advanced language family. The grammar of languages belonging to the Altaic language family are regular and logical. Due to these features, they are very suitable for artificial intelligence applications. It is difficult to say that Indo-European languages are rule-based like Altaic languages. There are many irregularities in these languages. Therefore, irregular situations must be memorized and sentences must be constructed according to certain patterns. In the first part of this article, the basic grammatical features of Altaic languages were introduced based on Turkish. In the second part of the article, first, it is shown how a sentence can be created using the grammatical features of Turkish language. A program was prepared to test and prove the introduced model and method. The proposed method was tested with this program and the results confirmed the suggestion. The developed method can be used in other Altaic languages with minor adaptations.