Figure out the low confidence matches

[durgesh312]

Hi Sonal,

Thank you so much for your help with the previous query.

I am assigning the same label to all instances of the z_cluster "Supplier Name." However, some incorrect Supplier Names are also grouped under this label.

Could you please guide me on how to identify the low-confidence results?

I’ve attached a snapshot for your reference.

[sonalgoyal]

is this the only attribute you use for matching or are there more?

[durgesh312]

I only have this attribute

[sonalgoyal]

Ok. You can check the stopWords documentation and use that to improve accuracy in this case. If you have less training, try labelling a bit more - that can help too

[durgesh312]

I tried using the recommend stopwords logic but it did not work and also how do you configure the recommended stopwords to zingg arguments? Is there any example code I can refer to? The one at the following location is not very clear: https://github.com/zinggAI/zingg/blob/main/examples/databricks/FebrlExample.ipynb

[sonalgoyal]

This example is self explanatory, what do you find it is missing? Let us know and we will fix. Please refer to the docs as well. HTH