Araneum Uzbecicum: A Gigaword Web-Crawled Uzbek Corpus

We want to introduce a Project of creation of a web-crawled Uzbek corpus. The main steps related to data capture, pre-processing, tokenization, deduplication, PoS tagging and lemmatization are discussed, and some query examples are shown. This article analyzes the problems in the process of creating and lemmatization the corpus of Araneum Uzbecicum. We also show word embeddings created from the Uzbek data. The Word embeddings function serves as a useful tool for researching the semantic relations in the Uzbek language, word context and valence. Also, the paradigm created by the Word embeddings function helps to perfect dictionaries.