Being able to quantify the semantic similarity between two texts is important for many practical applications. SemantiKLUE combines unsupervised and supervised techniques into a robust system for measuring semantic similarity. At the core of the system is a word-to-word alignment of two texts using a maximum weight matching algorithm. The system participated in three SemEval-2014 shared tasks and the competitive results are evidence for its usability in that broad field of application.

SentiKLUE is an update of the KLUE polarity classifier -which achieved good and robust results in SemEval-2013 with a simple feature set -implemented in 48 hours.

For many applications in the field of natural language processing, a lexicon is needed. For the Albanian language a lexicon that can be used for these purposes is presented below. The lexicon contains around 75,000 entries, including proper names such as personal, geographical and other names. Each entry includes grammatical information such as parts of speech and other specific information, e.g. inflection classes for nouns, adjectives and verbs. The lexicon is part of a morphological tool, but can also be used as an independent resource for other tasks and applications or can be adapted for them. Sources for the creation and the extension of the presented lexicon include both information from traditional dictionaries, e.g. spelling dictionaries, and a balanced linguistic corpus using corpus-driven methods and tools. The lexicon is still work in progress, but aims to cover basic information for most frequent tasks of natural language processing.

No abstract