Question about use of prior lexicon

[chachijun]

Hi victor7246,

I found the results of the model with lexicon of 2000 words and the one with lexicon of 1 word are exactly the same.
And I looked at the code,

s = sampleFromCategorical(sentimentDistribution)
prior_sentiment = lexicon_dict.get(w,1)
self.n_vts[w, t, s*prior_sentiment] += 1

Suppose the number of sentiment labels is 2, and s will produce 0 or 1 and prior sentiment will be -1 or 1;
if s is 0, then s*prior_sentiment will always produce 0;
if s is 1, then s*prior_sentiment will be -1 or 1 but n_vts[w,t,-1] is the same as n_vts[w,t,1] right?

I am so confused about this issue.
I will really appreciate if you can help me :)
Thank you!

Kind regards,
Jun