Hi, I am trying your code. I wonder whether the number before the time in file alltimes is necessary? For instance, you have experimented the email data, what is the number in that case? Thank you in advance!

I have following two question regarding updating of parameter psi :

1. Why 1 is added to psi[i][0] and psi[i][1] in GetMethodOfMomentsEstimatesForPsi apart from other factor mentioned in the paper ?
   psi[i][0] = 1 + timestamp_mean_common_factor
   psi[i][1] = 1 + (1-timestamp_mean)_common_factor

2. Why time stamp is updated once for every document rather than updating for each word of every document in TopicsOverTimeGibbsSampling ?

   par['psi'] = self.GetMethodOfMomentsEstimatesForPsi(par)
   Thanks

Hello,
I run the code ,after a while, the Terminal tell me that there is an error:

     in TopicsOverTimeGibbsSampling
           new_topic = list(np.random.multinomial(1, topic_probabilities, size=1)[0]).index(1)
     ValueError: 1 is not in list

Can you tell me why the error will happen? Thanks very much!

hello,i got a problem like that
Traceback (most recent call last): File "main_pnas.py", line 44, in <module> main() File "main_pnas.py", line 33, in main documents, timestamps, dictionary = tot.GetPnasCorpusAndDictionary(documents_path, timestamps_path, stopwords_path) File "/home/mere/topics_over_time/src/tot.py", line 43, in GetPnasCorpusAndDictionary stopwords.update(Set(line.lower().strip().split())) NameError: name 'Set' is not defined

Pls, can we get an evaluation metrics for the model?

Hello, your project is really a good job. I wonder can I use the trained model(pans_tot) on a tweets dataset? If so, how should I preprocess my dataset?

Thanks!

I want to know if I can get the words change of a topic over time.

Hello,

I am noticing some duplication in the code for tot.py which doesn't seem right.

Lines 32-33 show:

		for line in fileinput.input(stopwords_path):
			stopwords.update(set(line.lower().strip().split()))

which seems to repeat in lines 42-43:

		for line in fileinput.input(stopwords_path):
			stopwords.update(Set(line.lower().strip().split()))

with the difference being the use of Set in line 43 vs the uset of set in line 33.
As far as I know there is no Set construct in Python, so that's likely an error.

I assume the latter code duplicate can be deleted as it's erroneous?