The Data Mining Forum 
IMPORTANT: This is the old Data Mining forum.
I keep it online so that you can read the old messages.
Please post your new messages in the new forum: https://forum2.philippe-fournier-viger.com/index.php
Questions about Apriori Algorithm
Posted by:
Student
Date: November 15, 2014 03:19AM
Hi,
I want to become a researcher in the field of data mining,I would like specially improve the performance of the apriori algorithm in distributed envirennement to do this i want to integrate agent technology.
But before that i want to ask some questions:
i have no experience in java programming so to build apriori algorithm what are the things that i should learn to implement that kind of algorithm data mining using java language?
I want to build agent that execute apriori algorithm in autonomous manner so can We predect the best value of support and confidence without the intervention of user?
Thank you in advance.
I want to become a researcher in the field of data mining,I would like specially improve the performance of the apriori algorithm in distributed envirennement to do this i want to integrate agent technology.
But before that i want to ask some questions:
i have no experience in java programming so to build apriori algorithm what are the things that i should learn to implement that kind of algorithm data mining using java language?
I want to build agent that execute apriori algorithm in autonomous manner so can We predect the best value of support and confidence without the intervention of user?
Thank you in advance.
Re: Questions about Apriori Algorithm
Posted by:
webmasterphilfv
Date: November 15, 2014 05:13AM
Hi,
1) You should learn the basic Java language, how to read/write files, how to do a distributed Java program, and in particular if you want to implement Apriori efficiently, I would recommend to learn to use the data structures provided in Java such as ArrayList, HashMap, etc.
By the way Apriori is a slow algorithm. Faster algorithms have been proposed for the same task. But APriori is easy to implement.
If you want to see my Apriori implementation, you may look at the SPMF source code.
2) Interesting. But you should be aware that some work has been done on choosing automatically minsup and minconf thresholds. For example, you may look for Predictive Apriori, or algorithms for top-k pattern mining.
1) You should learn the basic Java language, how to read/write files, how to do a distributed Java program, and in particular if you want to implement Apriori efficiently, I would recommend to learn to use the data structures provided in Java such as ArrayList, HashMap, etc.
By the way Apriori is a slow algorithm. Faster algorithms have been proposed for the same task. But APriori is easy to implement.
If you want to see my Apriori implementation, you may look at the SPMF source code.
2) Interesting. But you should be aware that some work has been done on choosing automatically minsup and minconf thresholds. For example, you may look for Predictive Apriori, or algorithms for top-k pattern mining.
Re: Questions about Apriori Algorithm
Posted by:
Student
Date: November 15, 2014 02:49PM
Bonjour Monsieur Philippe Fournier-Viger
J'ai vu dans vos projets que vous travailler aussi sur les agents
Quelle plateforme utilisez vous pour l’implémentation de vos agents ,est ce que c'est JADE?
J'ai vu dans vos projets que vous travailler aussi sur les agents
Quelle plateforme utilisez vous pour l’implémentation de vos agents ,est ce que c'est JADE?
Re: Questions about Apriori Algorithm
Posted by:
webmasterphilfv
Date: November 18, 2014 09:45AM
Bonjour,
En fait, il y a longtemps que j'ai travaillé sur ce sujet. Pour ces projets, nous avions développé notre propre code pour les agents qui était spécifique à notre projet. J'ai déjà entendu parlé en bien de JADE, mais je ne l'ai jamais utilisé.
Bonne continuation,
En fait, il y a longtemps que j'ai travaillé sur ce sujet. Pour ces projets, nous avions développé notre propre code pour les agents qui était spécifique à notre projet. J'ai déjà entendu parlé en bien de JADE, mais je ne l'ai jamais utilisé.
Bonne continuation,