Question

Problem 2. Linear Classifier (15 points) Let {X?, X?} denote two feature dimensions (e.g., age and income), and Y ? {0, 1} denote the class labels of the observations. In Figure 1, we have three positive observations ("+" for Y = 1) and one negative observation ("-" for Y = 0). Please answer the following questions. X? 2 | + + 1 | + - --|------ X? | 1 2 Figure 1: Four data points in a two-dimensional feature space. 1. (10 points) Is there any logistic regression classifier using X? and X? that can perfectly classify the examples in Figure 1? What if we change the label of point (1, 2) from "+" to "-"? 2. (5 points) For predicting the samples in Figure 1, which model is better: Logistic Regression or Linear Regression? Please explain the reason.

          Problem 2. Linear Classifier (15 points)
Let {X?, X?} denote two feature dimensions (e.g., age and income), and Y ? {0, 1} denote the class labels of the observations. In Figure 1, we have three positive observations ("+" for Y = 1) and one negative observation ("-" for Y = 0). Please answer the following questions.

X?
2 | + +
1 | + -
--|------ X?
  | 1 2

Figure 1: Four data points in a two-dimensional feature space.

1. (10 points) Is there any logistic regression classifier using X? and X? that can perfectly classify the examples in Figure 1? What if we change the label of point (1, 2) from "+" to "-"?

2. (5 points) For predicting the samples in Figure 1, which model is better: Logistic Regression or Linear Regression? Please explain the reason.

Problem 2. Linear Classifier (15 points)
Let X?, X? denote two feature dimensions (e.g., age and income), and Y ? 0, 1 denote the class labels of the observations. In Figure 1, we have three positive observations ("+" for Y = 1) and one negative observation ("-" for Y = 0). Please answer the following questions.

X?
2 | + +
1 | + -
–|—— X?
| 1 2

Figure 1: Four data points in a two-dimensional feature space.

1. (10 points) Is there any logistic regression classifier using X? and X? that can perfectly classify the examples in Figure 1? What if we change the label of point (1, 2) from "+" to "-"?

2. (5 points) For predicting the samples in Figure 1, which model is better: Logistic Regression or Linear Regression? Please explain the reason.

Added by Jessica H.

Question

Please give Ace some feedback