HELP NEEDED WITH TAACO

[HAN MIN]

Dear professor,
      I'm afraid that I don't understand what you mean "variables".
     Just for further explanation, the article I mentioned  computed lexical cohesion features using TACCO with LSA, LDA, and Word2vec. It is said that the "LDA output was computed as divergence; hence, LDA scores were reversed (i.e., by subtracting them from one)" and aggregated all three measures by computing the mean as the cohesion score. And its cohesion scores range from 0 to 1.
     What I want to know is that what is the range of LSA, LDA, and Word2Vec respectively, and if I obtained negative LSA, Word2Vec scores, is that normal?
                                                                                                                                          Your sincerely

[sacro...@gmail.com]

LSA is usually between -1 and 1. So negative weights are very much possible.