ISYE 6501 Midterm
1. What does SVM stand for?: Support Vector Machine
2. Is written text structured or unstructured?: Unstructured
3. When we increase the sum of the square of the coefficients we...:
Decrease the distance between the lines
4. In SVM soft
classifier we tradeoff between maximizing and minimizing
: margin and errors
5. If lambda gets small what gets emphasized, large margin or
minimizing training error?,: Minimizing errors.
6. What is a support vector?: A point that holds up a shape.
7. Does ...[T(a-1)+1/3(a+1)] move an SVM classifier up or down?: Up
8. How do you make errors more costly in a soft SVM classifier?: include
a multiplier for the point-error term.
9. If an SVM coefficient is very close to zero...: that term is not very
important to the classification.
10.What is the difference between standardization and scaling?: Scaling is
bounded in range. Standardization is scaling to a normal distribution.
,Standardiza- tion is the (value - factor mean) / (factor standard
deviation)
11.What is the 2-norm?: Euclidean distance
12.What is the 1-norm?: The rectilinear (Manhattan) distance
13.What is the infinity norm?: The value of the largest dimension
14.Measuring the quality of a model is called?: Validation
15.What does a confusion matrix show?: The performance of a
classification model.
16.A time series outlier that seems "off the curve" is called a...:
contextual outlier.
17.A data element that is different from all other data in a set is
called a...: point outlier.
18.When something is missing in a range of points: it is called a...,
collective outlier.
19.The whiskers on a box plot extend to...: the 10th and 90th percentiles
(or 5th and 95th)
20. Why are hypothesis tests generally not sufficient for change detection?
: They are slow to detect changes.
,21.In CUSUM, T is and C is .,: Threshold and a "bring down
factor"
22. In a CUSUM model, you adjust T and C to manage the tradeoff between..
: early detection and false-alarms
23.In exponential smoothing, if the data is less random, then you want
to pick an alpha that is...,: Close to 1.
24.What is the initial condition for T in exponential smoothing with
trend- ing?: T_i=0
25.In cyclic exponential smoothing, L represents...,: The length of the
cycle or season
26.In cyclic exponential smoothing, C_1 ... C_L = ?,: 1. In other
words, initialize it to no initial cycle.
27.Exponential, trending and cyclic smoothing are also referred to as:
single double and triple.
28.Triple smoothing is also known as?: Winter's or Holt-Winter's
29. What is the optimization formula for Exponential Smoothing?: -
min(F_t-Xt)^2 where alpha and beta are between 0 and 1.
30.ARIMA stands for?: Autoregressive Integrated Moving Average
31.Exponential smoothing is an order autoregressive model.: Infinity.
It uses data going all the way back.
32.For ARIMA, the D parameter is used to specify .: , The order, or
the differences of the differences of the differences (d-times.)
, 33.For ARIMA, the P parameters is used to specify .,: The order of
periods (autoregression).
34.For ARIMA, the Q parameter is used to specify .,: The order of
the moving average.
35.ARIMA(0,1,1) is ?,: Exponential smoothing.
36.What is the order of the ARIMA parameters?: p d q
37.GARCH estimates what?: Variance.
38.Variance can be a proxy for or .: volatility or Risk
39.What parameter does GARCH not have the ARIMA has?: d
because GARCH doesn't deal with differences.
40.What is a simple linear regression?: Linear regression with one
predictor.
1. What does SVM stand for?: Support Vector Machine
2. Is written text structured or unstructured?: Unstructured
3. When we increase the sum of the square of the coefficients we...:
Decrease the distance between the lines
4. In SVM soft
classifier we tradeoff between maximizing and minimizing
: margin and errors
5. If lambda gets small what gets emphasized, large margin or
minimizing training error?,: Minimizing errors.
6. What is a support vector?: A point that holds up a shape.
7. Does ...[T(a-1)+1/3(a+1)] move an SVM classifier up or down?: Up
8. How do you make errors more costly in a soft SVM classifier?: include
a multiplier for the point-error term.
9. If an SVM coefficient is very close to zero...: that term is not very
important to the classification.
10.What is the difference between standardization and scaling?: Scaling is
bounded in range. Standardization is scaling to a normal distribution.
,Standardiza- tion is the (value - factor mean) / (factor standard
deviation)
11.What is the 2-norm?: Euclidean distance
12.What is the 1-norm?: The rectilinear (Manhattan) distance
13.What is the infinity norm?: The value of the largest dimension
14.Measuring the quality of a model is called?: Validation
15.What does a confusion matrix show?: The performance of a
classification model.
16.A time series outlier that seems "off the curve" is called a...:
contextual outlier.
17.A data element that is different from all other data in a set is
called a...: point outlier.
18.When something is missing in a range of points: it is called a...,
collective outlier.
19.The whiskers on a box plot extend to...: the 10th and 90th percentiles
(or 5th and 95th)
20. Why are hypothesis tests generally not sufficient for change detection?
: They are slow to detect changes.
,21.In CUSUM, T is and C is .,: Threshold and a "bring down
factor"
22. In a CUSUM model, you adjust T and C to manage the tradeoff between..
: early detection and false-alarms
23.In exponential smoothing, if the data is less random, then you want
to pick an alpha that is...,: Close to 1.
24.What is the initial condition for T in exponential smoothing with
trend- ing?: T_i=0
25.In cyclic exponential smoothing, L represents...,: The length of the
cycle or season
26.In cyclic exponential smoothing, C_1 ... C_L = ?,: 1. In other
words, initialize it to no initial cycle.
27.Exponential, trending and cyclic smoothing are also referred to as:
single double and triple.
28.Triple smoothing is also known as?: Winter's or Holt-Winter's
29. What is the optimization formula for Exponential Smoothing?: -
min(F_t-Xt)^2 where alpha and beta are between 0 and 1.
30.ARIMA stands for?: Autoregressive Integrated Moving Average
31.Exponential smoothing is an order autoregressive model.: Infinity.
It uses data going all the way back.
32.For ARIMA, the D parameter is used to specify .: , The order, or
the differences of the differences of the differences (d-times.)
, 33.For ARIMA, the P parameters is used to specify .,: The order of
periods (autoregression).
34.For ARIMA, the Q parameter is used to specify .,: The order of
the moving average.
35.ARIMA(0,1,1) is ?,: Exponential smoothing.
36.What is the order of the ARIMA parameters?: p d q
37.GARCH estimates what?: Variance.
38.Variance can be a proxy for or .: volatility or Risk
39.What parameter does GARCH not have the ARIMA has?: d
because GARCH doesn't deal with differences.
40.What is a simple linear regression?: Linear regression with one
predictor.
