ISYE 6501
PRACTICE EXAM QUESTIONS WITH
CORRECT DETAILED ANSWERS |
ALREADY GRADED A+<RECENT
VERSION>
1) Classification problems are commonly solved using what model(s)? -
answer support vector machine
2) clustering problems are commonly solved using what model(s)? -
answer k-means
3) response prediction questions are commonly solved using what model(s)?
- answer -arima
-cart
-exponential smoothing
-linear regression
-logistic regression
-random forest
4) validation questions are commonly solved using what model(s)? -
answer -cross validation
5) variance estimation questions are commonly solved using what model(s)?
- answer -garch
,6) examples of models that are designed for use with time series data -
answer -arima
-cusum
-exponential smoothing
-garch
7) in the soft classification svm model where we select coefficients a_0 ...
A_m to minimize sum(max(0, 1 - (sum(a_i * x_ij) + a_0 ) * y_j ) + c *
sum(a_i ^ 2) ). If we want to have a larger margin even though it means
possibly having more classification error, the value of c should get: -
answer larger
8) best way to split data - answer -70% for training
-15% for validation
-15% for test
9) purpose of a test set - answer estimate quality of selected model
10) purpose of a training set - answer fit parameters of all models
11) purpose of a validation set - answer compare all models and
select best
12) true or false: the most useful classification models are the ones that
correctly classify the highest fraction of data points. - answer false.
13) Explanation: sometimes the cost of a false positive is so high that
it's worth accepting more false negatives, or vice versa. Lesson 10.6
,14) a model is built to determine whether data points belong to a
category or not. A "true negative" result is: - answer a data point that
is not in the category, and the model correctly says so.
15) Explanation: true' and 'false' refer to whether the model is correct
or not, and 'positive' and 'negative' refer to whether the model says the
point is in the category. Lesson 10.5
16) a logistic regression model can be especially useful when the
response... - answer is binary (zero or one) or is a probability (a
number between zero and one). Lesson 10.4
17) true or false: when using a random forest model, it's easy to
interpret how its results are determined. - answer false
18) Explanation: unlike a model like regression where we can show the
result as a simple linear combination of each attribute times its regression
coefficient, in a random forest model there are so many different trees
used simultaneously that it's difficult to interpret exactly how any factor
or factors affect the result. Lesson 10.3
19) What does svm stand for? - answer support vector machine
20) Is written text structured or unstructured? - answer unstructured
21) When we increase the sum of the square of the coefficients we... -
answer decrease the distance between the lines
22) In svm soft classifier we tradeoff between maximizing ___ and
minimizing ___ - answer margin and errors
, 23) If lambda gets small what gets emphasized, large margin or
minimizing training error?, - answer minimizing errors.
24) What is a support vector? - answer a point that holds up a
shape.
25) Does ...[⅔(a-1)+1/3(a+1)] move an svm classifier up or down? -
answer up
26) How do you make errors more costly in a soft svm classifier? -
answer include a multiplier for the point-error term.
27) If an svm coefficient is very close to zero... - answer that term is
not very important to the classification.
28) What is the difference between standardization and scaling? -
answer scaling is bounded in range. Standardization is scaling to a
normal distribution. Standardization is the (value - factor mean) / (factor
standard deviation)
29) What is the 2-norm? - answer euclidean distance
30) What is the 1-norm? - answer the rectilinear (manhattan)
distance
31) What is the infinity norm? - answer the value of the largest
dimension
32) Measuring the quality of a model is called? - answer validation
PRACTICE EXAM QUESTIONS WITH
CORRECT DETAILED ANSWERS |
ALREADY GRADED A+<RECENT
VERSION>
1) Classification problems are commonly solved using what model(s)? -
answer support vector machine
2) clustering problems are commonly solved using what model(s)? -
answer k-means
3) response prediction questions are commonly solved using what model(s)?
- answer -arima
-cart
-exponential smoothing
-linear regression
-logistic regression
-random forest
4) validation questions are commonly solved using what model(s)? -
answer -cross validation
5) variance estimation questions are commonly solved using what model(s)?
- answer -garch
,6) examples of models that are designed for use with time series data -
answer -arima
-cusum
-exponential smoothing
-garch
7) in the soft classification svm model where we select coefficients a_0 ...
A_m to minimize sum(max(0, 1 - (sum(a_i * x_ij) + a_0 ) * y_j ) + c *
sum(a_i ^ 2) ). If we want to have a larger margin even though it means
possibly having more classification error, the value of c should get: -
answer larger
8) best way to split data - answer -70% for training
-15% for validation
-15% for test
9) purpose of a test set - answer estimate quality of selected model
10) purpose of a training set - answer fit parameters of all models
11) purpose of a validation set - answer compare all models and
select best
12) true or false: the most useful classification models are the ones that
correctly classify the highest fraction of data points. - answer false.
13) Explanation: sometimes the cost of a false positive is so high that
it's worth accepting more false negatives, or vice versa. Lesson 10.6
,14) a model is built to determine whether data points belong to a
category or not. A "true negative" result is: - answer a data point that
is not in the category, and the model correctly says so.
15) Explanation: true' and 'false' refer to whether the model is correct
or not, and 'positive' and 'negative' refer to whether the model says the
point is in the category. Lesson 10.5
16) a logistic regression model can be especially useful when the
response... - answer is binary (zero or one) or is a probability (a
number between zero and one). Lesson 10.4
17) true or false: when using a random forest model, it's easy to
interpret how its results are determined. - answer false
18) Explanation: unlike a model like regression where we can show the
result as a simple linear combination of each attribute times its regression
coefficient, in a random forest model there are so many different trees
used simultaneously that it's difficult to interpret exactly how any factor
or factors affect the result. Lesson 10.3
19) What does svm stand for? - answer support vector machine
20) Is written text structured or unstructured? - answer unstructured
21) When we increase the sum of the square of the coefficients we... -
answer decrease the distance between the lines
22) In svm soft classifier we tradeoff between maximizing ___ and
minimizing ___ - answer margin and errors
, 23) If lambda gets small what gets emphasized, large margin or
minimizing training error?, - answer minimizing errors.
24) What is a support vector? - answer a point that holds up a
shape.
25) Does ...[⅔(a-1)+1/3(a+1)] move an svm classifier up or down? -
answer up
26) How do you make errors more costly in a soft svm classifier? -
answer include a multiplier for the point-error term.
27) If an svm coefficient is very close to zero... - answer that term is
not very important to the classification.
28) What is the difference between standardization and scaling? -
answer scaling is bounded in range. Standardization is scaling to a
normal distribution. Standardization is the (value - factor mean) / (factor
standard deviation)
29) What is the 2-norm? - answer euclidean distance
30) What is the 1-norm? - answer the rectilinear (manhattan)
distance
31) What is the infinity norm? - answer the value of the largest
dimension
32) Measuring the quality of a model is called? - answer validation
