Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

BUAN 348/448 Fall 2022

Exam Practice Problem Set

1. (48 pts) Multiple Choices (Only 1 correct answer) 

1) (8 pts) Which of the following data mining technique is NOT supervised learning?  

A) Simple Linear Regression

B) Multiple Regressions

C) Logistics Regressions

D) Association Rules

2) (8 pts) Given the following target variable, which one is suitable for classification methods?  

A) Annual sales

B) Lot size

C) Income

D) Ownership or not

3) (8 pts) What is the estimation method for linear regression model?

A) Maximum Likelihood Method

B) Minimum Description Length

C) Akaike Information Criteria

D) Least Squared Error 

4) (8 pts) Which of the following is NOT a layer in neural networks?

A) Input layer

B) Hidden layer

C) Dependent layer

D) Output layer

5) (8 pts) If probability = 0.65, the odds = ?

A) 1.86

B) 5

C) 15

D) 7

6) (8 pts) From the output of a multiple regression model, the coefficient of one independent variable has a p value of 0.8. This means:

A) This independent variable is significant.

B) This independent variable is not significant.

C) Not sure if this independent variable is significant.

D) The model is overfitted.

2. (20 points) Fit a multiple linear regression model and neural network (using 2 hidden layer and 3 nodes for each layer) for "Spending" using all variables in the given data set (Tayko.csv) and conduct the prediction performance evaluation for each model. Please answer which model is better? Copy and paste your r-code here (using set.seed(12345)).

3. (32 points) Suppose we have a Neural Network Model (see below):

 

(a) (10 pts) If the transfer function g is defined as g(s) = 1/(1+exp(-s)), show the calculation of the output from node 6. 

������������������6 =  = 0.506 

(b) (10 pts) Compute the error for node 6:

  0.506 1−0.506  1−0.506 =0.123 

(c) (12 pts) Assuming the learning rate = 0.5, how can we update ���6 and ���3,6?

���6 =−0.015+0.5  0.123 =0.047    ���3,6 =0.01+0.5  0.123 =0.072