Statistical Methods in the Problem of Studying Apology Speech Formulas and Their Satellites in the English Language

摘要

This article describes the algorithm of the analysis of justification speech formulas in the English language. We also consider methods of preprocessing text data (normalization, part-of-speech tagging, functional tagging). Such methods as frequency analysis and Apriori (an algorithm for frequent item set mining and association rule learning) have been used for texts analyzing. Based on the statistical analysis results, 11 general justification formulas have been derived. These formulas have also been used to compose an attribute vector for the classifier. The support vector machine classifier (SVM) is used for texts classifying. After training the classifier, 12 000 texts of the language corpus have been analyzed.