1. Select all the option(s) that are loss functions: (a) Cross entropy (b) Maximum likelihood (c) Squared error (d) 0-1

[answerhappygod]

1 Select All The Option S That Are Loss Functions A Cross Entropy B Maximum Likelihood C Squared Error D 0 1 1 (31.73 KiB) Viewed 57 times

1. Select all the option(s) that are loss functions: (a) Cross entropy (b) Maximum likelihood (c) Squared error (d) 0-1 loss (e) Hinge loss II. Select all the option(s) that are activation functions: (a) Sigmoid (b) ReLu (c) Hyperbolic tangent (d) Softmax (e) Entropy III. Select all the option(s) that are true when describing kernel methods: (a) Kernel methods are designed to reduce overfitting, (b) Kernel methods only work with Support Vector Machines (SVMs). (c) Common kernels include polynomial, Gaussian, and Lagrange functions. (d) Kernel methods are designed to map data into better representational space. (e) Linear SVMs always work better than kernel SVMs. IV. Select all the option(s) that are true about overfitting: (a) A bigger number of hidden nodes in a Multilayer Perceptron helps reduce overfitting with the same amount of data. (b) A smaller number of layers in a Multilayer Perceptron helps reduce overfitting with the same amount of data. (c) Pruning helps control overfitting in decision trees. (d) Random forest reduces overfitting compared with decision tree. (e) Data augmentation (e.g., creating copies of image data that are rotated or scaled versions of the original ones) helps reduce overfitting with the same amount of original data.