History of the CORON System
-
The very first part of the system got ready on July 18, 2004. It can be considered as the birthday of Coron, though it was
just an implementation of the Close algorithm. The name of the project was "J-Close" at that time.
-
I (Laszlo) implemented the Titanic algorithm too as a separated project. But since the two algorithms shared lots of code, we
decided to put them together. Now the user had the possibility to choose which algorithm to execute on her dataset.
-
By January 30, 2005 the platform contained several algorithms, namely Apriori, Apriori-Close, Pascal, Pascal+, Zart, RMS
Carpathia, Close, and Titanic. The project gets the name Coron, which means "miner's village" in French. That is, Coron is a collection (village) of
data mining algorithms. (The name "coron" was actually proposed by my mother :)).
-
We had the possibility to generate association rules, but it was part of the core system. This part was moved in another
project called AssRuleX. The core part that extracts itemsets was also called Coron, just like the whole system. In order to distinguish them, we often
refer to the itemset extraction module as "Coron-base", and we say "Coron System" or "Coron Platform" when we refer to the entire system.
-
As the system got modularized, it was easy to add some more projects for pre-processing datasets, for post-processing and
visualising the results, etc.
-
On November 24, 2006 I (Laszlo) successfully defended my PhD thesis in Computer Science. Everything that I present in my
thesis is implemented in the Coron System.
-
The development of Coron did not stop. On the contrary. Since September, 2007 we have had an associated engineer who continues
the development. A new PhD student, Mehdi Kaytoue-Uberall, will also add some new features. And I (Laszlo) will not abandon "my child" either :).