making queries to systems with natural language

splitting text up with tokenization + removing "textual noise" + reducing tokens with stemming and lemmatization