Entity extraction auto ML

[Bejoy Pallithazhe]

Hi Team,

I am just trying out entity extraction in auto ML.

I got a decent level of accuracy with an entity extraction model.

I am now looking at ways to further increase the accuracy of the model.

Could you please let me know the method that the auto ml entity extraction model uses to classify new words (i.e. words that the model has not encountered in the training set). We are trying to analyse customer feedback in relation to a repair experience. For example, 

"The engineer was fabulous." Ideally i would be looking for the model to pick engineer and fabulous. Assuming that fabulous was not in the training data. Could you please let me know if the system would pick up fabulous. If yes, could you please let me know the basis on which it would allocate it to a label.

Thanks,

Bejoy

[Monica (Google Cloud Platform)]

Hello,

I suggest you read our guide that explains how the models are functioning. Please address this video or this tutorial that will show the whole process. After implementing and evaluating your model with your data you'll be able to see what it's picking and you can decide depending on your use cases to remove / modify labels to improve the quality of the model.