I want a Java Arabic stemmer

Question

I'm looking for a Java stemmer for Arabic. I found a lib called "AraMorph" , but its output is uncontrollable and it makes formation to words which is unwanted.

Is there any other stemmer for Arabic ?

Gymnasiarch

Answer 1

Here is new Arabic stemmer: Assem's Arabic light stemmer coded using Snowball framework and generated to many languages including Java. You can use it by downloading libstemmer for Java here.

Afford

Answer 2

You can find Kohja stemmer here:

http://zeus.cs.pacificu.edu/shereen/research.htm

Direct download:

http://zeus.cs.pacificu.edu/shereen/ArabicStemmerCode.zip

Antichrist

Thank you for your answer, @paradigmatic. I have asked my quesion because I did not know what stemming is. Following your answer I read about it a little bit. – Apomorphine

I want API or Lib so that I can use in my project Thanks any way :) – Gymnasiarch

@Kareem: It is an API or lib... Check the second link I've posted. – Antichrist

The code is under the GPL license though. Quite restrictive for a stemming library. Can't be used in a commercial product. I can't even wrap it in a TokenFilter and contribute it to Lucene, since Apache License and GPL are incompatible. – Possum

It can be used in a web-based commercial product, SaaS for example, This way no redistributables are involved, so GPL has no problems with that! – Wag

Answer 3

https://sourceforge.net/projects/arabicstemmer/

try this it is based on Shereen Khoja Algorithm.

Malefaction

Answer 4

after digging I found the best solution is to implement my own stemmer using porter Algorithm so that I can tune my stemmer

Gymnasiarch

What? It won't work! Arabic is typed in non-Latin letters and more importantly follows extremely different algorithmic approach than the other Latin languages. .... But I'm interested to know if it worked with you or not? – Univalent

Answer 5

You can use either Elkhoja stemmer or Lucene's light stemmer

Zymase