0

I am trying to test the performance of the Lexicon-Analyzer from the nltk module to compare it with different models.

I have the following csv which I put in a Pandas Dataframe, it looks as follows:

Sentiment   Text
0   1   "When will the suffering end?"\r\nSubaru enjoy...
1   1   ~This is a preliminary review. As of writing t...
2   1   As of the time I'm writing this review I've on...
3   0   Fine then. I'll be the first one to say it. Th...
4   1   here are two responses I get, without fail, ev

Pandas Code:

   anime = pd.read_csv("Models/Benchmarks/AnimeReviews.csv", encoding="utf-8")
    anime.shape
    anime.dtypes
    anime.convert_dtypes().dtypes
    anime.dtypes
    x_anime = anime.iloc[:, 1]
    y_anime = anime.iloc[:,0]

sentiment is returned as int64 and the sentences as an object

Using this, I want to create an array to save all the results to then use sklearn.metrics.accuracy_score to compare the results and get the accuracy.

I did the following:

results = np.array(object)
for sentence in x_anime:
    result = demo_liu_hu_lexicon(sentence)
    np.append(results, result)

If I do this, the array is shown as empty. I tried it with just one sentence (no for loop) and it saved the result in the array. That means the sentence doesn't get passed through to the analyzer to analyze it.

How can I solve this?

EmreYe
  • 53
  • 1
  • 5
  • Does this answer your question? [must capture output of a function that has no return statement](https://stackoverflow.com/questions/47337547/must-capture-output-of-a-function-that-has-no-return-statement) – Let's try Aug 25 '20 at 08:08

1 Answers1

0

Try the following. Numpy append is not inplace:

results = np.append(results, result)
Let's try
  • 1,044
  • 9
  • 20
  • The problem is that the sentence is not passed to the analyzer but thanks! – EmreYe Aug 25 '20 at 07:52
  • Have a look [here](https://stackoverflow.com/questions/47337547/must-capture-output-of-a-function-that-has-no-return-statement) then – Let's try Aug 25 '20 at 08:08