Can we have a Logit that punishes some missclassifications more harhsly than others ?

#1
Suppose I wish to estimate a logit model to predict Large losses when trading stocks, well I might say y_t = 1 if that trade was a large loss. But when I would then try to estimate a model to predict these large losses I would essentially just be modelling which stocks are most volatile. Ie. I would throw away both my best trades and my worst ones. Could I modify the logit model such that it avoided this ? - I was imagining adding a penalty to the log-likelihood contributions, such that a missclasification of a great trade was penalized more heavily than a mediocre one.

thoughts ?
 
#2
Re: Can we have a Logit that punishes some missclassifications more harhsly than othe

Why are you using a logit at all? That requires an arbitrary cutoff point that distinguishes "big" from "not big". Why not predict the profit of a trade?
 
#3
Re: Can we have a Logit that punishes some missclassifications more harhsly than othe

Thanks for the reply!

Samme issue unfurtunately, I have tried fitting a linear regression and using a cutoff point to categorize them but it also kills all the good trades.
 
#5
Re: Can we have a Logit that punishes some missclassifications more harhsly than othe

But the whole point was to try to predict the very worst trades, surely I need some cutoff point to categorize the forecasts ?