I am trying to map specific words in a list to another list of tokenized sentences and if the word is found in the sentence then I append a 1 to a list of its category and 0 to the rest of categories. For example:
category_a=["stain","sweat","wet","burn"]
category_b=["love","bad","favorite"]
category_c=["packaging","delivery"]
tokenized_sentences=['this deodorant does not stain my clothes','i love this product','i sweat all day']
for i in category_a:
for j in tokenized_sentences:
if(i in nltk.word_tokenize(j)):
list_a.append(j)
tag_a,tag_b,tag_c=([],)*3
tag_a.append(1)
tag_b.append(0)
tag_c.append(0)
final=tag_a+tag_b+tag_c
Similarly for category_b and category_c
Expected output:this deodorant does not stain my clothes-->[1,0,0]
i love this product-->[0,1,0]
i sweat all day-->[1,0,0]
great fragrance-->[0,0,0]
I am getting duplicate outputs for each sentence like: i love this product-->[1,0,0] i love this product-->[1,0,0] and also like this:[i love this product,i sweat all day]-->[0,1,0]
Also, if a sentence has words from two different categories Ex: 'this product does not stain and i love it'
the expected output would be [1,1,0]
How do I get the output in the required format?
This should do the job:
category_b = ["love", "bad", "favorite"]
category_c = ["packaging", "delivery"]
sentences = ['this deodorant does not stain my clothes', 'i love this product', 'i sweat all day']
results = []
for sentence in sentances:
cat_a = 0
cat_b = 0
cat_c = 0
for word in sentance.split():
if cat_a == 0:
cat_a = 1 if word in category_a else 0
if cat_b == 0:
cat_b = 1 if word in category_b else 0
if cat_c == 0:
cat_c = 1 if word in category_c else 0
results.append((sentance, [cat_a, cat_b, cat_c]))
print(results)
This code will check if each sentence contains word of each of the given categories and save both of them (the sentence and result) in form of a tuple. All tuples will be appended to a list called results.
Output:
[('this deodorant does not stain my clothes', [1, 0, 0]), ('i love this product', [0, 1, 0]), ('i sweat all day', [1, 0, 0])]