如何运行时警告：LDA中的exp中遇到溢出

Question

ldamodel, fequency_list,vect = create_ldamodel(documents = chatTurn.case_content, num_topics = 6)

I am running the above LDA code and it appears the statement as below.

/Users/user/anaconda3/lib/python3.6/site-packages/gensim/models/ldamodel.py:497: RuntimeWarning: overflow encountered in exp expElogthetad = np.exp(Elogthetad)

The LDA model function that I used is as follow:

def create_ldamodel(documents, num_topics):
    vect = CountVectorizer(stop_words = 'english')
    X = vect.fit_transform(documents.apply(lambda x:x.lower()))
    corpus = gensim.matutils.Sparse2Corpus(X, documents_columns=False)
    id_map = dict((v,k) for k, v in vect.vocabulary_.items())
    ldamodel = gensim.models.ldamodel.LdaModel(corpus,      num_topics=num_topics,
                                          id2word = id_map)

fequency_list = dict()
for i in list(ldamodel[corpus]):
    for j_k, j_v in i:
        if j_k in fequency_list:
            fequency_list[j_k] += j_v
        else:
            fequency_list[j_k] = 0

return (ldamodel, fequency_list,vect)

May I know is there any solution to it?

Answer 1

Because your numpy is mismatch the gensim, maybe update numpy in your current environment or specify the numpy version to match your gensim version. Hope this is helpful to you. ^^

Answer 2

My runtime erros was the following: RuntimeWarning: overflow encountered in exp expElogthetad = np.exp(Elogthetad) After trying so many times and things, I came across with a reading recommending uninstalling numpy and reinstalling it back again:

python3.6 pip uninstall numpy and to install it back: python3.6 -m pip install -U numpy