Adrian Mitrea

Printed romanian modelling: a corpus linguistics based study with orthography and punctuation marks included

This paper is part of a larger study dedicated by the authors to the description of printed Romanian language as an information source. Here, the statistical investigation attempts to get an answer concerning the mathematical model of the language with orthography and punctuation marks included into the alphabet. To come out to an accurate result, the authors processed the information obtained out of multiple data sets sampled from a corpus linguistics, by using the following statistical inferences: probability estimation with multiple confidence intervals, test of the hypothesis that the probability belongs to an interval, and test of the equality between two probabilities. The second type statistical error probability involved in the tests was considered. The experimental results, which are new for printed Romanian, refer to the letter, digram and trigram statistical structure in a corpus linguistics of 93 books (about 50 millions characters).

A study on the statistical structure of words and of word digrams in a literary romanian corpus

By resuming and extending an original method for verifying natural language stationarity, the paper presents a study on statistical structures of words and word digrams (groups of two successive words) in printed Romanian. The paper also contains an evaluation of natural language redundancy based on word digrams. The experimental study was carried out on a literary linguistic corpus of novels and short stories summing up over 12.5 million words with orthography and punctuation marks.

A comparison between two literary printed Romanian corpora based on the statistical letter structure with orthography and punctuation marks

The problem under consideration is a comparison between a new literary linguistics corpus and an existing corpus for printed Romanian, with the purpose of strengthening the stationarity property of the language. By the applied statistical approach, the paper also includes a way of extending a corpus linguistics for mathematical purposes.

On the occurrences of two successive words in a literary Romanian corpus

The goal of the study was to investigate the statistical structure of groups of two successive words (digrams of words) in the literary field of printed Romanian. The paper brings into discussion the statistical effect that a preceding word has over the words following it. All the investigation presented here refers to natural language as a chain of words.