LF_TrainRegression Processing Resource

The regression training PR allows you to train a classifier suitable for learning a numerical target, for example learning to predict the star rating a reviewer assigns based on the text of their review. The PR has no init time parameters. Here are the runtime parameters:

Set your instance to the annotation type that you wish to classify, for example “Review”. These annotations must already be present in the input annotation set. For learning to predict the star rating assigned by a reviewer based on the text of their review, a good starting point for features would be the words in the review, as in this example feature specification:

<ML-CONFIG>

<NGRAM>
<NUMBER>1</NUMBER>
<TYPE>Token</TYPE>
<FEATURE>string</FEATURE>
</NGRAM>

</ML-CONFIG>