1. You are given a dataset with predictor variables X and labels Y. The dataset has...

Transcribed Image Text:

1. You are given a dataset with predictor variables X and labels Y. The dataset has the following scatter plot: Y iii) ŷ 55 = 50 45 40 35 30 25 Bo + B₁x + ₂x² 0 0.5 1 1.5 X 2 k 2.5 3 Consider the following three machine learning models, where x is the input variable, ŷ is the predicted label value, and Bo, B₁, B₂,... are the trainable parameters of the model. (here x is just a number, not a vector) i) ŷ = Bo + ₁x ii) ŷ Bo + B₁x + 3₂x² + 3x³ + B4 log(x + B5) + Be + 37 tan(x + 38) = 3.5 4 Suppose you train each of these three models using the above dataset. a) Which of these three models will have the lowest training error? Please explain. b) Which model is most likely to have the lowest test error? Please explain. c) Which of the three models is the best model to use for this dataset? Why? 1. You are given a dataset with predictor variables X and labels Y. The dataset has the following scatter plot: Y iii) ŷ 55 = 50 45 40 35 30 25 Bo + B₁x + ₂x² 0 0.5 1 1.5 X 2 k 2.5 3 Consider the following three machine learning models, where x is the input variable, ŷ is the predicted label value, and Bo, B₁, B₂,... are the trainable parameters of the model. (here x is just a number, not a vector) i) ŷ = Bo + ₁x ii) ŷ Bo + B₁x + 3₂x² + 3x³ + B4 log(x + B5) + Be + 37 tan(x + 38) = 3.5 4 Suppose you train each of these three models using the above dataset. a) Which of these three models will have the lowest training error? Please explain. b) Which model is most likely to have the lowest test error? Please explain. c) Which of the three models is the best model to use for this dataset? Why?

Answer rating: 100% (QA)

a Model ii will have the lowest training errorThis is because it is the most complex modeland it has ... View the full answer