Tint, the Swiss-Army Tool for Natural Language Processing in Italian

In this we paper present the last version of Tint, an opensource, fast and extendable Natural Language Processing suite for Italian based on Stanford CoreNLP. The new release includes a set of text processing components for fine-grained linguistic analysis, from tokenization to relation extraction, including part-of-speech tagging, morphological analysis, lemmatization, multi-word expression recognition, dependency parsing, named-entity recognition, keyword extraction, and much more. Tint is written in Java freely distributed under the GPL license. Although some modules do not perform at a state-of-the-art level, Tint reaches very good accuracy in all modules, and can be easily used out-of-the-box.