Porter stemmer not stemming properly

[wooorm]

Out of 23532 input/outputs provided with the original algorithm, Natural fails on 558 cases (Natural@0.1.28).

Failures include:

• ace to ac;
• acknowledg to acknowledg;
• adoptedly to adoptedli;
• aedile to aedil;
• …

Source: See http://tartarus.org/martin/PorterStemmer/, and search for "To test the programs out".

[kkoch986]

Correct me if i'm wrong but acknowledg to acknowledg are the same

[wooorm]

You seem to be right 👍. Weird why that error’d in my tests. Let me check out.

[wooorm]

EDIT: Hold on. My above examples do not include the actual output Natural gave, but rather what it should give. Here are the actual return values:

> n.PorterStemmer.stem('ace')
'ace'
> n.PorterStemmer.stem('acknowledg')
'acknowlg'
> n.PorterStemmer.stem('adoptedly')
'adoptli'
> n.PorterStemmer.stem('aedile')
'ail'

[kkoch986]

I see, ill try to go over the porter stemmer in more detail ASAP. Thanks!

[chrisumbel]

Didn't mean to close. I've handled the cases mentioned initially, but more remain. Will be addressing.

156 cases remain.

[pranavatgithub]

Still bugs with PorterStemmer.Anyone working on it??
please check the examples

1. natural.PorterStemmer.stem("recharged") ---->> recharg
   2.natural.PorterStemmer.stem("subscribed") ----->> subscrib
   3.natural.PorterStemmer.stem("restored") ------->>restor

[pranavatgithub]

i think below algorithm is working better.
URL : - - http://takafumir.github.io/javascript-lemmatizer/html/lemmatizer_sample.html
git repo: -- https://github.com/takafumir/javascript-lemmatizer

[kkoch986]

assuming whats at the end of those arrows is the output from natural i think those are correct. i verified with another unrelated implementation.