Stemming and lemmatization in the clustering of finnish text documents