Russian scientists analyzed Tuva folklore using mathematical approaches
The formal concept analysis is one of the algebraic methods of data analysis where each object is described based on its central attributes. In the new work, the national corpus of Tuva language has been investigated with the texts being collected from the digitalized artistic and literally works, while storylines, introductions, main language clichés, and many other characteristics have been taken as attributes.
Next, each literature work has been correlated to a table recording the presence of attributes, whereas for the epos as a whole, the so-called «concept lattice», a special scheme showing global relations between various attributes, has been composed. Due to such formalized model, all epos works can be automatically classified at a semantic, i.e., at a qualitative level.
«A human perceives the world through the concepts. He defines the objects, extracts essential attributes of them and based on it classifies and systematizes fundamentals of surrounding world. The formal concept analysis represents the perception of the essence done by mathematics. However, to get reliable and stable knowledge, one should treat a large amount of data. Here, the mathematicians are faced to the „curse of dimensionality“: to analyze the needed amount of data, the whole life of a human is not enough,» - the Chief of the study, Professor at Chair of High and Applied Mathematics Valentin Bykov said.
Similar tasks are challenging even from the viewpoint of computer facilities. Exemplarily, according to the scientists, the complete treatment model with one hundred of attributes might take millions of computational time. Nevertheless, the Russian researchers have managed to optimize their algorithms, and now, an artificial brain can analyze the epos works much faster. The research article with the recent progress has been published in the «Journal of SFU. Mathematics and Physics».
The researchers point out that the «capability of unscrambling» folklore texts will not only assist in learning Tuva language but also will be a great support for translators and a notable help in learning and conservation of ethnic and cultural heritage of Tuva Republic. Now, the interpretation of developed mathematical model is mainly done by philologists and linguists from the Research and Educational Center «Tyurkology».
Analogous investigations of Russian scientists dealing with texts written in Turkic have provoked interested from mathematicians and linguists from M. Ulugbek National University of Uzbekistan.