NATURAL LANGUAGE PROCESSING IN THE STATE DECENTRALIZED SYSTEMS OF GEORGIA
As we know, the State Public Registry uses Blockchain Technology in Georgia. Naturally, the document turnover proceeds in the Georgian language. However, due to the peculiarities of this language and, also, the small number of Georgian-speaking people, there has not been made significant progress in the computer processing of the Georgian language, so far. The present paper investigates the implementation of well-known machine learning algorithms for the Georgian language in the public registry using the blockchain system (this, in turn, will facilitate the complete computer processing of the Georgian language (NLP) and use it for other purposes as well). However, it should be noted that the Georgian language differs from other languages a lot, which makes it necessary to modify the algorithms in order to successfully work, for example, for English or other already processed languages, to the needs of Georgian with its grammatical peculiarities. Moreover, it should be noted that the use of the approaches proposed by us in this paper further accelerates the work of this algorithm at the expense of reducing iterations and saving computing resources. A more general purpose of the research is to make it possible to classify texts in the blockchain system the existing documents in registry repositories, identify minimal editing distance between words, calculate the probability of probable sequences between given words, and so on.