When continuous variables are not binned, a negative weight might support intuition and not contradict it. E.g. if a cancer cell is benign and the weight for the cell's size is negative, it mean that a smaller cell would even more support the label "benign" for that cancer cell. However, plot_features plots the negative weight as contradictory. That makes no sense, does it?
When continuous variables are not binned, a negative weight might support intuition and not contradict it. E.g. if a cancer cell is benign and the weight for the cell's size is negative, it mean that a smaller cell would even more support the label "benign" for that cancer cell. However,
plot_features
plots the negative weight as contradictory. That makes no sense, does it?