machine learning engineer trained a non$-$linear logistic regression classifier with threshold $0\mathrm{.}5.100\%$ accuracy was observed on training set. Then the threshold was changed $0\mathrm{.}2$ and then $0\mathrm{.}7.$ still $100\%$ accuracy was observed for both cases on training set. Which of the following statements could be true? Give brief reasons Statement $1$: Dataset is linearly separable Statement $2$: Model is overfitting