Grožinės literatūros sakinių sudėtingumo analizė

LTŠio tyrimo tikslas – ištirti grožinės (vaikų) literatūros sakinių struktūrų sudėtingumą. Nagrinėjamas paprasčiausių sakinių sintaksinės struktūros gylis, išplečiančių sakinio dalių skaičius, atstumas tarp gretimo rango sakinio dalių. Taip pat gautas neblogas tiesinės regresijos modelis tarinį išplečiančioms sakinio dalims aprašyti. [Iš leidinio]

ENThe goal of this research is to explore the complexity of sentence structures of prose books for children. The following problems: 1) To see, or the algorithms based on trigrams and suitable, for example, for the English language are applicable to the Lithuanian language studies. 2) To make (if possible) the linear regression model, which described the parts of sentence in subject and predicate structures, and 3) to make (if possible) the model to describe the length of the sentence. The complexity of sentence structure can be conceived as depth and width of a sentence according to the words subordinate to the main parts of sentence (the subject and predicate). Also very important is the distance between the adjacent rank parts of the sentence. In this work were analyzed only the simple sentences the structure of which can be represented by a graphtree. So, it cannot be argued that the trigram statistics are suitable for Lithuanian language. Though, for example, even the 93,13% of the distances from the subject to directly related parts of the sentence are not more than 2, but for predicate – its 71,32%. Was made the linear regression model, which described the parts of sentence in predicate structures. This model correctly predicts 34,34% of data. was made the linear regression model, which described the length of the sentence, but only predicates and information about its go in this model. [From the publication]