Then You Throw Away Data Points

In logistic regression, throwing away data points is a form of regularization. It's implicitly reducing your capacity for formal fruladisation but it's not reducing how much you gt about e my noled sampas so is osential a. The ones we have are too easy or just don't make sense. Scale is a mad thing until you're able to solve the very simple problem with small amounts of data point. And transoltia networks would not be few in this way if there were no scale at all.