I will add to that... Even if your data schema matches up to what's used in the accelerator, at the very least you will need to do some data exploration and data cleaning and redo the model training. The patterns found in the sample data will not be the same ones found in yours. More likely you will have different data and schema from the data used in the accelerator. Then you will just need to follow through the accelerator to use the assets as a template to do it yourself with your data. It will very much speed up the learning curve with the examples and code that you will need. You should be able to walk through it, using your data and following the same steps to go from data to deployed model
------------------------------
Tim Bohn
Data Scientist / Sr. Solution Architect
IBM Data Science and AI Elite
San Jose CA
------------------------------
Original Message:
Sent: Wed June 10, 2020 04:48 PM
From: Walaa Wageh
Subject: Importing the accelerator
thanks, that is exactly what iam asking about, appreciated..thanks again
------------------------------
Walaa Wageh
Original Message:
Sent: Wed June 10, 2020 12:42 PM
From: David Belton
Subject: Importing the accelerator
Hi,
The industry accelerators help the customer to kick-start their own data science projects, from data preparation to building machine learning models and deploying the models with WML. As you say, the customer will need to provide their own data. Depending on the data quality and schema, and the specific accelerator, they may need to modify the '1-model_training' notebook to handle their own custom data preparation steps. From there, the customer can run the notebooks to generate and deploy machine learning models and functions.
------------------------------
David Belton
Original Message:
Sent: Wed June 10, 2020 11:31 AM
From: Walaa Wageh
Subject: Importing the accelerator
Thanks , I know it is not connected with real data, but just was asking to make sure that when we need to make the same use case outcome , we will duplicate what is done in the accelerator and it is efficient.
thanks your interest is appreciated
Original Message:
Sent: 6/10/2020 11:26:00 AM
From: TOMASZ HANUSIAK
Subject: RE: Importing the accelerator
Hi,
Each accelerator clearly explains what's included and what outcome you will have:
https://community.ibm.com/community/user/cloudpakfordata/viewdocument/insurance-claims?CommunityKey=c0c16ff2-10ef-4b50-ae4c-57d769937235&tab=librarydocuments
as stated, they dont usually connect to live data, just use samples.
Really depends on what you mean by end to end.
Thanks
------------------------------
TOMASZ HANUSIAK
Original Message:
Sent: Wed June 10, 2020 11:18 AM
From: Walaa Wageh
Subject: Importing the accelerator
Thanks
as it is my first real project .. I asked to make sure that if I already follows the accelerator , I don't miss any other thing...I mean if the accelerator really end to end ?
Original Message:
Sent: 6/10/2020 11:15:00 AM
From: TOMASZ HANUSIAK
Subject: RE: Importing the accelerator
Hi,
To run an Industry Accelerator, you need to deploy all the required services first.
I'm not sure I understand your question.
Thanks
------------------------------
TOMASZ HANUSIAK
Original Message:
Sent: Wed June 10, 2020 11:05 AM
From: Walaa Wageh
Subject: Importing the accelerator
Yes thanks but I mean the needed add ons for the solutions , and the steps needed ..is there any thing else to do for a real one? For sure the dataset will be my customer's one and any different the schema need to be modified
and thanks again
appreciated
Original Message:
Sent: 6/10/2020 9:54:00 AM
From: TOMASZ HANUSIAK
Subject: RE: Importing the accelerator
Hi,
Please note that each industry accelerator comes with 'sample data sets and schemas' which may not be relevant to your customer/situation.
https://www.ibm.com/support/knowledgecenter/en/SSQNUZ_2.5.0/cpd/svc/industry-accel-svc.html
Thanks
------------------------------
TOMASZ HANUSIAK
Original Message:
Sent: Sun June 07, 2020 10:10 AM
From: Walaa Wageh
Subject: Importing the accelerator
Kindly clarify is the accelerator assets enough to reuse in a real project with a customer? or there are any other effort needed to be learned and prepared to be included in a real project to develop real accurate results to decision makers.
thank in advance
your help is really appreciated
------------------------------
Walaa Wageh
------------------------------
#CloudPakforDataGroup