how to delete trained data for specific Document type

[KrishnaElapavul]

Hi 

I am using Decipher Version: 1.120.6180
In my solution , A batch contains 4 Document types  , all are well trained.  unfortunately  document type - 1 didn't work correctly. I would like to re-train the specific document type-1  again freshly.
So...  I am planning as 

1. Delete  trained documents type-1  ( how to delete specific document type?) 
2. Train again the same document type-1 

Is it right approach ? if yes how to delete specific document type ? with out disturbing other trained documents ?
Please advise any one... 🙂


------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------

[joefarr]

Hi Krishna

If it is a specific document type causing you trouble, you can manage the document types in the Admin Panel. One thing to note though, machine learning models are applied to document types, however, they are also applied to the DFD associated with the document type, this means that any document types that use the same DFD will also use the same machine learning model. Best practice recommends that each document type has its own DFD and consequently its own machine learning model. This prevents issues where changes to the DFD resets the machine leaning model, or the machine learning isn't always relevant to the document type.

You can find more information in the user guide.

Hope this helps!

------------------------------
Joe Farr
Product Consultant
Blue Prism Limited
------------------------------

[KrishnaElapavul]

Thanks Joe,

I didn't enable ML as of now. 
My question is about deletion of specific document type is it possible in the specified version ? if yes how it can be carried ?

------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------

[joefarr]

Hi Krishna

You should still be able to delete the individual Document type that is causing you the problem, even if you're not using machine learning, unless I've misunderstood what you are looking to do? Could you share a screen grab of the issue?

------------------------------
Joe Farr
Product Consultant
Blue Prism Limited
------------------------------

[Tejaskumar_Darji]

So your Batch has 4 Document Types selected in it right.

Simply follow these steps and I believe it will work as per my experience.

1) Go to Document Types and delete the one that you no longer wanna use.
2) Come back to Batch Types and you should see now only 3 Document Types selected in the "Included document type" section
3) Now create your new DFD and Document type and train the new one individually. (Do this only for the document in which you want to reset training)
4) Once done go to Batch Types and tick mark this new Document Type.

Now run your entire batch with all Document Types included.

This way you'll not touch those DFD & Doc Types which are properly working for you.

Let us know how it goes for you so this thread can be helpful to others as well in the future.


------------------------------
Thanks & Regards,
Tejaskumar Darji
Sr. RPA Consultant-Automation Developer
------------------------------

[KrishnaElapavul]

Hi Tejas,
Thanks for  your suggestion. but here I want to deleted the trained data for respective Document type , not just a documents type
What i mean is ,
Already I have trained few files of Document type-1 and Decipher is able to map the fields for new files of same documents type-1. when i push. But it is not mapping  all fields correctly even after training around 30+ files .So  I want decipher should forget what it learned for the specific document type-1  may be by deleting the trained document type-1, 

I see one option in Admin panel to delete trained data , but look that is generic for all Document types not to specific document type. Hope I explained my  question clearly , 


------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------

[Ben.Lyons1]

Hi Krishna,

As you've correctly deduced, there is only a single option for deleting all training data when ML has not yet been enabled.

The reason for this is the training data is not linked to a DFD or document type. Decipher memorises each page it sees and its first step when seeing a new file is checking if the pages match something it's previously verified. It uses this to link the data to the locations and the other aspects specified in the DFD during the capture stage.

At this time it is not possible to delete the training data specific to a document type or DFD, however we are looking at ways this could be achieved in a future release.

Regards

------------------------------
Ben Lyons
Product Consultant
Blue Prism
UK
------------------------------