Skip to content
This repository has been archived by the owner on Mar 8, 2021. It is now read-only.

Is the test for compute_ARI() correct? #17

Open
vcidst opened this issue Jul 13, 2015 · 2 comments
Open

Is the test for compute_ARI() correct? #17

vcidst opened this issue Jul 13, 2015 · 2 comments
Labels

Comments

@vcidst
Copy link

vcidst commented Jul 13, 2015

In book branch, chapter 5 I believe the test for compute_ARI isn't correct. I have checked that the Automatic Readability Index for the sentence 'This was rather easy. Please give me something more challenging' is 6.033

The condition however is abs(compute_ARI(sentences) - 4.442) < 0.001 which will always be false

@fbkarsdorp
Copy link
Owner

Thanks for this. Chapter five still needs quite some polishing. This will be helpful in that.

@fbkarsdorp fbkarsdorp added the bug label Jul 13, 2015
@vcidst
Copy link
Author

vcidst commented Jul 13, 2015

Okay I figured out that the ARI of sentences when it is

sentences = [["this", "was", "rather", "easy"], 
             ["Please", "give", "me", "something", "more", "challenging"]]

is 6.033. However, if you sentences[0].append(".") or add a period there, the ARI of sentences become 4.442.

Not saying that the period should be there, it shouldn't. Only thought it'll be helpful to know where you got that magic number 4.442 in the condition from.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
Projects
None yet
Development

No branches or pull requests

2 participants