CODEWITHSUNDEEP
python
William Jungerman
William Jungerman

Reputation: 33

How do I check if the values I am getting are increasing/decreasing?

First, let me specify a few things:

cleaned_example = [['I', 'horrible', 'physics', 'pretty', 'good', 'chemistry', 'excellent', 'math'], ['First', 'concerned', 'worried', 'scared'], ['What', 'started', 'bar', 'fight', 'turned', 'confrontation', 'finally', 'gang', 'war'], ['Every', 'hill', 'mountain', 'shall', 'made', 'low', 'rough', 'places', 'made', 'plain,', 'crooked', 'places', 'made', 'straight'], ['This', 'blessed', 'plot', 'earth', 'realm', 'England']]

but in reality is composed of 20 groups of words.

I am using the sentiment function which is part of pattern.en and I want to see if the sentiment values of the words in cleaned_example[0], cleaned_example[1].... are increasing or decreasing. The sentiment function outputs two values of the form (a,b) but I am only interested in the first of these values.

Here is what I have done so far. I am encountering two problems. First, I am getting 40 outputs when I should only be getting 20. Second, all of these outputs are 'no' so it is pretty useless. 

for index in range(len(cleaned_example)):
for position in range(len(cleaned_example[index])-1):
    if sentiment(cleaned_example[index][position][0]) < sentiment(cleaned_example[index][position+1][1]):
        print('yes')
    else: 
        print('no')

Thanks in advance!

Upvotes: 2

Views: 2761

Answers (3)

tom10
tom10

Reputation: 69222

Since you have parameters (a,b) and you only care about a, then you always want [0] when choosing between these; and I suspect that you have your parentheses in the wrong place as well. That is, your difference line should read:

if sentiment(cleaned_example[index][position])[0] < sentiment(cleaned_example[index][position+1])[0]:

so the final index listed is 0 and not 1, and you use [0] on the value returned from sentiment (ie, the [0] is to the right of the function call).

A few side notes:

This would read more easily if you didn't index cleaned_example every time:

for word_list in cleaned_example:
    for position in range(len(word_list)-1):
        if sentiment(word_list[position])[0] < sentiment(word_list[position+1])[0]:
            print('yes')
        else: 
            print('no')

Finally, here you're calling sentiment twice for almost every word. If that's a problem then you should restructure your code a bit. With this in mind, it would probably be better to start with something like:

for word_list in cleaned_example:
    sentiments = [sentiment(word)[0] for word in word_list]
    for i in range(len(sentiments)):
        if sentiments[i] < sentiments[i+1]:
            print('yes')
        else: 
            print('no')

Upvotes: 1

Stefan Pochmann
Stefan Pochmann

Reputation: 28626

Going a step further than tom in avoiding indexing and using even more meaningful variable names (I googled that sentiment thing):

for sentence in cleaned_example:
    for word, next_word in zip(sentence, sentence[1:]):
        if sentiment(word)[0] < sentiment(next_word)[0]:
            print('yes')
        else: 
            print('no')

If for example sentence is ['a', 'b', 'c', 'd'], then sentence[1:] is ['b', 'c', 'd'] and zipping them gives you word pairs ('a', 'b'), ('b', 'c') and ('c', 'd').

I don't know why you think you should get only 20 outputs (I get 36, btw, not 40). I suspect you are working at the wrong level and should use sentiment on sentences, not words? Note how I named my variable names, good names can really help you understand your code. Triple indexes not so much.

Upvotes: 3

EFL
EFL

Reputation: 1008

If you are expecting to get only 20 outputs, is it correct to assume you only want either a "yes" or a "no" for each group of words? I am not certain how exactly you want to measure that because you can just compare the sentiment of the first word and the last word in a list of words for each set in cleaned_example. 

for words in cleaned_example:
    if sentiment(words[0])[0] < sentiment(words[-1])[0]:
        print "Increasing"
    else:
        print "Decreasing"

Another approach is to count the number of yes's and no's for each set of words. 

for words in cleaned_example:
    yes = 0
    no = 0
    for word1, word2 in zip(words, words[1:]):
        if sentiment(word1)[0] < sentiment(word2)[0]:
            yes = yes + 1
        else: 
            no = no + 1
    if yes > no:
        print "yes"
    elif yes < no:
        print "no"
    else:
        print "`yes` = `no`"

Another approach is to obtain the average sentiment of a set and compare it with the sentiment of the first word.

import numpy as np

for words in cleaned_example:
    a_values = []
    for word in words:
        a_values.append(sentiment(word)[0])
    if sentiment(words[0])[0] < np.mean(a_values):
        print "Increasing"

Upvotes: 0

Related Questions