I have dataset of 5000 jobs descriptions out of which only 200 jobs are labelled with required English level score range between 0 to 9 and I want to predict remaining 4800 jobs required English level score? how to use clustering or multi classification in this scenario? Thanks
$\begingroup$
$\endgroup$
1
-
1$\begingroup$ How would you do it if you only had $10$ predictions to make instead of $4800?$ $\endgroup$– DaveApr 20, 2022 at 2:59
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
You can do it the usual way: train on 200 labelled instances, test on the remaining 4800 instances. But actually you should probably keep a labelled test set for evaluating performance first, or use cross-validation on the 200 instances.
However you might have an even more serious problem: it's not clear to me that the English required level can be predicted this way. Either it's explicitly mentioned in the description and then it's just a matter of extracting it, or it's not and there's no way to know it.
