Help for Creditworthy Scoring Model

ferdinand_papa

Hello Guys,

I hope you can assist me in my task. 

We need to develop a model to predict creditworthiness raiting within a given dataset. 
I have given you the training set its called "zahlungsausfall_train.txt" as well as the set to predict "zahlungsausfall_class.txt".

In the training dataset you will find the attribute "TARGET_BETRUG" it can either be YES or NO it means basically if someone is creditworthy or not creditworthy.  

The goal is to create a "CLASS" that Contains "H" for NO and "N" for YES" 

As you can see in my training data Set If you let it run through it converts "TARGET_BETRUG" into "Klasse" and therefore it puts them into the 2 categories N or H. 

The thing is I tried to get it done with a decision tree.. as you can see in the process but as soon as I let it run through it just gives me N in the dataset where we are supposed to predict correctly. 

My question is which parameters or operators do I have to add or adjust to get a similar outcome as in the training set.


greetings

varunm1

Hello @ferdinand_papa

I tried your model with Decision Tree and see that it is unable to predict 'H' label. I cannot say it is wrong as your training set does not have an output label to get the performance. I used a Gradient Boosted tree algorithm which is complex and powerful and observe that it is able to predict both 'H' and 'N' classes for your test data. But as there are no labels (outputs) for your test data I cannot comment whether the performance is good or not.

Do you by any chance have labels for your test data which you are applying for apply model operator? This will provide you with performance.

Thanks

ferdinand_papa

I think I finished the model.
If you want you can check for any mistakes

regards