Below is a gensim's example, but whenever I execute it, it show different result, so I couldn't believe gensim works well.
from gensim import corpora, models, similarities
from collections import defaultdict
documents = ["Human machine interface for lab abc computer applications", # 0
"A survey of user opinion of computer system response time", # 1
"The EPS user interface management system", # 2
"System and human system engineering testing of EPS", # 3
"Relation of user perceived response time to error measurement", # 4
"The generation of random binary unordered trees", # 5
"The intersection graph of paths in trees", # 6
"Graph minors IV Widths of trees and well quasi ordering", # 7
"Graph minors A survey"] # 8
stoplist = set('for a of the and to in'.split())
texts = [[word for word in document.lower().split() if word not in stoplist]
for document in documents]
frequency = defaultdict(int)
for text in texts:
for token in text:
frequency[token] += 1
texts = [[token for token in text if frequency[token] > 1]
for text in texts]
dictionary = corpora.Dictionary(texts)
corpus = [dictionary.doc2bow(text) for text in texts]
lda = models.LdaModel(corpus, id2word=dictionary, num_topics=2)
index = similarities.MatrixSimilarity(lda[corpus])
doc = "Human computer interaction"
vec_bow = dictionary.doc2bow(doc.lower().split())
vec_lda = lda[vec_bow]
sims = index[vec_lda]
sims = sorted(enumerate(sims), key=lambda item: -item[1])
print(sims)
print(lda.get_document_topics(vec_bow))
result
[(0, 0.9986434), (4, 0.99792993), (2, 0.99722278), (3, 0.99651831), (1, 0.99158639), (5, 0.53059661), (6, 0.4146674), (8, 0.38019019), (7, 0.36143348)] [(0, 0.18366596), (1, 0.81633401)]
[(1, 0.999605), (4, 0.9991864), (0, 0.998689), (5, 0.62957084), (6, 0.48837978), (8, 0.48152202), (3, 0.4541581), (7, 0.41751832), (2, 0.40637407)] [(0, 0.80285221), (1, 0.19714773)]
[(7, 0.99957085), (8, 0.99660784), (0, 0.99202132), (5, 0.78449017), (6, 0.77530348), (2, 0.56972337), (3, 0.47117239), (4, 0.47092015), (1, 0.4172135)] [(0, 0.25292286), (1, 0.74707717)]
Result 7 doesn't look simiar with "Human computer interaction" at all. Thanks.
