Greetings,

I practice with TKE for episode mining, There is a parameter Max. Time duration which makes the number of sequences of the episode to be generated not greater than the value entered ( eg if 2 entered, it means the episode of 1 sequence, and 2 only will be generated).

Kindly need for the feature to work with my dataset "Minimum Time duration" (we can make it as an option with above "Max. Time duration" ) which will restrict the number of sequences in a generated episodes to be greater than the value entered only.

Any help/suggestions from the above request

Thank you

Dear Irfan,

Ok, thanks for your comment and also for the other one about RPGrowth (I have read it but did not answer you yet).

I will have a look at how to add this parameter to TKE. I think it should also be easy to do. I will try to do it today or tommorow, and then inform you about it. I think it is a good idea to have a minimum duration.

If you have some other suggestions for features, it is always welcome. Especially when the features are not hard to implement like this, I can do it quickly. But if it would be some more complicated features, maybe I cannot do it.

For RPGrowth, I will maybe rename the parameter minraresupport, but the original code was not by me. Thus before modifying the parameter name, I will check with those who implemented the code.

Best,

Philippe

Dear prof,

Thank you very much for the quick reply.

Dear Irfan,

I have checked the code of TKE for adding the minimum duration to episode mining. I spent 2-3 hours yesterday and today to try to implement but it is not as easy as I thought. It would require to change how the time intervals are calculated and a bit how the algorithm works, and there is some risk that it cause some other bug or problem, that may make the result wrong. Because it looks complicated to do, I have to postpone this feature for now. But I like the idea. So maybe I would do it later... but it would require to really think about it... it is not trivial to do.

Best regards,

Philippe

Dear prof,

Thank you for your time and feedback about the requested feature.

Maybe I have to wait for this. Do you have any plan for how long or when should I

expect to be included?

Thank you for your usual cooperation.

Hi Irfan,

I am not sure for now. At first, I wanted to implement that feature because I thought it was easy but after spending a few hours I realized that it would require changing a good part of how the algorithm is working because of some assumptions made by the algorithm. So it is actually much more work and it is not trivial to do I think.

Currently, I am very busy with some urgent projects and also we have the corona virus problem in China. Since implementing this feature may require a few days, I am not sure if I could do it in the coming weeks. And as I said, I am not sure that this feature will be easy to do. If it is too complicated i may not do it or postpone to later. We will see.

Best regards,

Philippe

Thank you for feedback Professor,

If there will be any other update on that I will inform you.