What is Avalingua?
Avalingua automatically evaluates the linguistic quality of texts, both in Galician and Spanish (the Portuguese version is in process). From a written text or a set of texts, Avalingua makes a detailed analysis using different natural language processing tools, including a syntactic dependency analyser, and various language resources. It has been developed by the Tecnoloxías da Linguaxe Natural program of CiTIUS, University of Santiago de Compostela, Galiza.
What advantages does this technology have?
It allows us to go beyond the usual grammar checkers, because it not only allows us to detect spelling and lexical errors, but also, thanks to an automatic syntactic analyzer, it identifies more complex grammatical problems between syntactically related words that are at a great distance from each other.
What research is Avalingua based on?
Avalingua is a set of PLN modules with hybrid technology, combining expert knowledge, heuristics, language resources and machine learning. The number of types of errors and linguistic problems that can be found in texts is so high that it is outside the scope of any classifier, as there are not enough training dates with sufficient diversity. For this purpose, we believe hybrid technology is needed that includes explicit expert knowledge. Syntactic knowledge is encoded with DepPattern technology, a dependency analyser compiled from grammars provided with an intuitive formalism for linguists.
What are the applications of Avalingua?
The most direct application is in the field of education, with two clear directions: it can be used as a monitoring and evaluation system for primary and secondary students, or/and as a second language self-learning tool for foreign students.