Please document the max range of each value

[benedictchen]

It is not totally clear regarding the resulting values what the max value could be.

Please add to the documentation.

[trakout]

It used to be +/- 5 (score, not the comparative score), however it appears that this is changed as I have gotten 6 or even 7 as a value.

[thisandagain]

The range for each token as well as the comparative score is +/- 5 as per the underlying AFINN dataset, but certainly we can make it more clear in the docs. Thanks for filing!

@trakout If you are seeing cases where that isn't true can you please file a bug report with details / example strings? "Score" has always simply been the sum of all tokens, so that can have an arbitrary range, but the comparative score should always be stable between +/- 5.

[thisandagain]

Here is a thread where I explain in detail:
#107

[pdw207]

Can this be closed or does this issue indicate an update to the README? @trakout were you able to reproduce the bug?

[trakout]

@pdw207
Sure, just re-tested and was able to reproduce. I have a demo here, some setup required for the Twitter Auth. https://github.com/trakout/simple-sentiment

If you uncomment this line, you can observe sentiment scores greater than 5 or less than -5.

If you don't want to go through the setup process, I snagged some example copy from the output.
The following statement consistently outputs score: -8, comparative: -0.17391304347826086.

(the following words are from twitter and do not necessarily reflect my personal opinion or sentiment)

console.log(sentiment.analyze('So to recap-Trump wants kids held hostage. Sen. Schumer in no rush to end the kid hostage crisis. Paul Ryan in no rush to end the hostage crisis unless the Hastert rule is met. But Mitch McConnell is the only leader urging an immediate clean-bill fix.'))

[pdw207]

@thisandagain I can take this on. Sounds like a good first issue to pick up.

[pdw207]

As @thisandagain mentioned:

the score is the sum of all tokens, so that can have an arbitrary range

In the example you provided there are 4 negative tokens: [ 'crisis', 'no', 'crisis', 'no' ] and no positive tokens. Each token has a value between -5 and 5. crisis has a AFINN score of -3 and no AFINN score of -1, so -3 + -1 + -3 + -1 = -8.

[pdw207]

I think this issue can be closed.

[ralyodio]

I'm also wondering why I always get around .02 to -.02 for comparitve score. Everything I run this on is "barely" positive or "barely" negative. I'm trying to create an indicator to reflect the negativity and positivity of an article but the marker is pretty much always in the middle making it rather useless.