APPLYING STATISTICAL METHODS FOR ANALYZING TEXTS OF THE PRESIDENTIAL ELECTION PROGRAMMES
Modern statistics is equipped with the methods of formalization (measurement) of the objects of different nature. This concerns in particular texts of the so called natural language. This article provides analysis (conducted by means of statistical methods) of the election programmes’ texts of the candidates for Ukraine’s Presidency in the 2019 election. With the method of multidimensional scaling, the data set was created that consists of two numerical characteristics that describe the peculiarities of the reviewed programmes’ texts. With the correlation analysis, the correlation was established between the texts of the candidates’ election programmes and the official results of the first round of the election, as well as the results of the nationwide exit poll. By applying the Ward’s method cluster analysis, the four groups of the candidates for Ukraine’s Presidency were outlined. Also, the peculiarities of the groups’ programmes texts were identified, as well as the key words clouds were created for quick apprehension of the most frequently used words and their distribution according to popularity. Data preparation and all statistical calculations were performed with the help of the statistical calculation environment R.