16. a) Of the 504 melt ponds, land ice had the proportion of 38.89%.
b) Form the table we can see that 88 of 504 melt ponds has first tear ice. Therefore,
the proportion of the ponds that had first year ice is 88/504*100%=17.46%, which is
close to the est
Anon 1
Anon
Math 424
11/28/16
HCH8-Anon-Math424
4)
a)
To calculate the regression residuals, we use R to first calculate our fitted regression model
and then subtract our predicted fitted model from our observed values ( y i ^y i ) to get:
Observation and
Anon
Anon
Math 424
09/09/16
HW 1
17,23,27,37,46,62,68,69,72,77
17)
a.
For the graphical method, we went ahead and made a bar graph to show the difference between
the two well class types. It shows that we have more of the public wells than the private wel
Anon 1
Anon
Math 424
10/19/16
HW CH4
10)
a) The least squares prediction equation for weight change in
y
is:
y=12.20.0256 x10.458 x 2
where
x 1= digestion efficiency and
x 2= acid-detergent fiber.
b) Taking a look at (b), we see that as our value of diges
Anon 1
Anon
Math-424
10/28/16
CH5 HW
8) In order to see the difference between how each variable relates to Heat Rate, we simply plot
each pair to see if they have some obvious relationship that relates to what degree we use.
Our independent variables are
Anon 1
Anon
Math 424
11/14/16
HCH7-Anon-Math424
6)
In order to talk about the prediction equation of our data, we take a look at each summary for
the three ice types and the general depth of each pond, then plot out a three boxplots for each
ice type to s
Anon 1
Anon
Math 424
12/07/16
HCH9-Anon-Math424
24)
a) Is the overall logit model statistically useful for predicting geese flight response?
Were given the fitted logistic regression model as:
=ln
( 1
)= + x + x
0
1
1
2
2
In order to test the overall mod
Anon
20)
a) Using R, we went ahead and used the anova function to take a look at its analysis of the
variance table. In in doing so, we then look for the Mean Squares of the Residuals to see that
2
s =10.138 . Thus the estimate of the model standard devia
Math 424
HW #2
3.8
(a) The straight line model
y=
0 +
1 x
where
x = appraised property value
y = sale price
(b) The scatter plot shows that when x increases y increases. The straight line model will
appear an appropriate fit to the data.
(c) sale price =
MATH 424
HW #1
1.16
(a) Of the 504 melt ponds, what proportion had landfast ice?
38.89%
(b) The University of Colorado researchers estimated that about 17% of melt ponds in the Canadian
arctic have first-year ice. Do you agree?
Yes, I agree. Because there
6.
When we do predictions, we need to make sure that the predictors are in the range of
prediction. For this problem, we can figure out the range for depth by each ice type
and if there are correlations among the depths for each ice type.
The SAS System
T
24.
a)
H 0 : 1= 2=0
Ha:
at least one of the beta is not 0.
=0.01
From the MINITAB printout, we can see that the p-value for
1 2
are 0.004 and
0.000, which are less than 0.01. Therefore, we can conclude that the overall logit
model is statistically usefu
1a . Forward Selection
First we need a dependent variable y and some independent variables, lets say 5,
x2 x3 x4 x5
.
x1
Step 1:
We will first construct 5 models:
Model 1: y= 0+ 1 x 1
Model 2: y= 0+ 1 x 2
Model 3: y= 0+ 1 x 3
Model 4: y= 0+ 1 x 4
Model 5:
4.
a)
The SAS System
The REG Procedure
Model: MODEL1
Dependent Variable: VOLUME
Output Statistics
O
bs
Depende
Predicted
nt
Value
Variable
Std Error
Mean
Predict
Residua
l
Std
Student
Error
Residua
Residual
l
-2-1 0 1 2
Cook
's
D
1 100.0000 98.6149
0.4037
16.
H 0 : 1=2=3= 4= 5
H a : at least one pair is not equal to each other.
=0.10
The SAS System
The GLM Procedure
Class Level Information
Class
Levels
BOREHOLE 5
Values
SD SWRA UMRB-1 UMRB-2 UMRB-3
Number of Observations Read
2
6
Number of Observations Us
6.
a)
y = 0+ 1 x 1+ 2 x2
b)
From the SAS output table we can see that the least squares regression equation is:
y = 20.35201+13.35045 x 1+243.71446 x 2
c)
Interpretation for
0 : When the age and the number of hours worked are both 0, the value for
the ve
8.
a) Let x be the appraised property value and let y be the sale price. I propose a straight line
model as:
y= 0+ 1 x
b)
From the scatterplot of the data we can see that when x increases, y also increases. So it appears
that a straight line model will be
Math 424
Anon HW 6
Due 11/2/2016
Name
1. Write-up, in detail, the steps taken to do the following regressions in model building. Make
sure you clearly state the CRITERIA for model selection.
a) Forward selection
b) Backward selection
c) Stepwise selection