Automatic readability classifier for European Portuguese

This paper describes a system that automatically classifies text readability for European Portuguese, while highlighting the key challenges on language features’ selection and text classification. To this goal, the system uses existing Natural Language Processing (NLP) tools to extract linguistic features from texts, which are then used by an automatic readability classifier. Currently, the system extracts 52 features grouped in 7 groups: parts-of-speech (POS), syllables, words, chunks and phrases, averages and frequencies, and some extra features. A classifier was created using these features and a corpus, previously annotated by readability level, using a five-level language classification official standard for Portuguese as Second Language. In a five-level (from A1 to C1) and three level (A, B and C) scenarios, the best-performing learning algorithm (LogitBoost) yields 79.25% and 86.32%, respectively.