Did you check what you have in column
If you have negative values then you have wrong values. It can’t sell negative number of items - it would means it buys them. You have to calculate it again.
it happens when i run on linear regression models, on tree based models like decision tree,random forest no negative values in columns
Hi ,you can take absolute value which makes negative values as positive by no.abs(value).
Better way to deal is by taking log transformation of target np.log(target) and you can get original shape after prediction by np.exp(prediction)