How to group column fields in different tables in data labelling for a dataset of custom ML package?

Help AI Center
,
1

Hello,

I have tried searching in the forum as well as the documentation, but couldn’t find any answer.
I want extract data from a document by developing custom ML skill. I use data labelling to label the data to e extracted. My document has 3 different tables to be extracted.
In order to achieve my goal, I have defined 3 different table respective in the taxonomy.
Table1
col1 col2
Table2
col1 col2 col3
Then, I start the document labelling where I define 5 column fields.
However, in data labelling I can’t group them in different tables. All the column field get accumulated under 1 table. I then train the model and deploy the ML skill.
But, in data extraction activity, under configure ML field to taxonomy fields can’t be done for both the tables. Can map these ML column fields to either Table1
or Table2 columns defined in the taxonomy.
Any help is appreciated.
Thanks
AV

1 Like
2

HI @Aditi.av

I understood most of your question except the following…

Can you show me what exactly happens here?

3

Hi Lahiru,
Will try to explain as much possible
In the data labelling, the column fields are defined as below

image
image1770×135 18.2 KB

However in the Taxonomy, they appear in 2 different tables. Hence while mapping the ML fields in configure extractors, first I map the column for one table

image
image939×184 33.4 KB

When I try to map the other column field to second table, I can’t
image
image922×174 33.8 KB

As all the column fields in ML are in a single table items. Hence I can map either equipment table or origequipment table not both of them. How can I have them in separate tables while labeling?

Thanks in advance
AV