Impact of Morphological Preprocessing of Training Corpus on Quality of Russian Distributional Semantic Models

Student: Tushkanov Vladislav

Supervisor: Ekaterina Artemova

Educational Programme: Language Theory and Computational Linguistics (Master)

Final Grade: 10

Year of Graduation: 2017

This paper examines the impact of morphological preprocessing on quality of distributional language models when applied to text classification tasks, covering both theoretical grounds of distributional language modelling and morphological preprocessing, and their practical applications. Experiments presented in this paper include three kinds of morphological preprocessing (stemming, lemmatization, part-of-speech-tagging) as well as three types of distributional modelling algorithms (word2vec skip-gram, FastText skip-gram and GloVe). Using four different classification approaches, including neural network-based, we show how preprocessing affects the results and describe the cases where the effect is positive, and the cases where preprocessing should be avoided.

Full text (added May 24, 2017)

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses