Improve the abstract and introduction

[fergunet]

After finishing issues #25 and #26, and because we are presenting the elements of our study (two fitness and two individual representation), a phrase indicating it should be added to abstract and introduction.

[unintendedbear]

I've been making changes these days that reference this, I'll leave it open until a whole revision has been made.

[unintendedbear]

Ok, I need to discuss the maint topic of the paper with you two @JJ @fergunet :

The aim is to propose a system which helps CSOs in the process of defining rules. Helping how? Eliminating, generalising, finding new rules... The aim is not comparing two fitness. The aim is not comparing two approaches Pittsburgh and Michigan. We, however, want to validate that the main approach we take (individual == set of rules instead of a single rule) by implementing the other approach and demonstrating by comparison that ours is indeed better. With regard to the fitness, we just use two as an additional feature to see if it converges faster or what. But to compare fitness is not our aim because in that case we should compare a way lot more than two.

Please let me know what is not clear about this or what do you think.

[fergunet]

So, we should emphasize in the paper how the chosen fitness and representations helps to achieve the objective. Is that ok? @JJ @unintendedbear

[JJ]

Well,

propose a system which helps CSOs in the process of defining rules

is what you want to get done in the long term. But is it what you have actually done in this particular paper?

[unintendedbear]

Of course! We have an initial working system. Can It be improved by taking into account the imbalance of the datasets? Sure, but the comparisons we make in the paper are the way to demostrate that the solution we propose is good. Comparing is not the aim, is the way to validate the achievement of our actual goal.

[JJ]

OK, your call. I would say that if you want to

propose a system which helps CSOs in the process of defining rules

you should have tried all, or at least a few, ways of doing so: Forest Tree, XGBoost, just to name two of the most popular and state-of-the-art algorithms, and then compare them with others, such as the one that you have actually tried in this paper. I agree with Pablo. You have explore one of the possible ways of achieving this objective, pursuing a sub-objective with its own challenges and conclusions, inside and outside the field of BYOD. But, once again, it's your call.

[unintendedbear]

Ok then! Right now I don't know what am I doing anymore, so anything you propose, I will write. Really, no hard feelings and I'm not being mean, I just want to have this done and sent.

[JJ]

Once again, I agree with Pablo.

[unintendedbear]

Ok, no prob, thanks

[unintendedbear]

Marked for review, please @JJ and @fergunet close when reviewed (or suggest changes)