ISYE 6501 MIDTERM 1 EXAM REVIEW
COMPLETE QUESTIONS AND SOLUTIONS
◉ Is written text structured or unstructured?
Answer: Unstructured
◉ When we increase the sum of the square of the coefficients we...
Answer: Decrease the distance between the lines
◉ In SVM soft classifier we tradeoff between maximizing ___ and
minimizing ___
Answer: margin and errors
◉ If lambda gets small what gets emphasized, large margin or
minimizing training error?,
Answer: Minimizing errors.
◉ What is a support vector?
Answer: A point that holds up a shape.
◉ Does ...[⅔(a-1)+1/3(a+1)] move an SVM classifier up or down?
Answer: Up
,◉ How do you make errors more costly in a soft SVM classifier?
Answer: include a multiplier for the point-error term.
◉ If an SVM coefficient is very close to zero...
Answer: that term is not very important to the classification.
◉ 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)
◉ What is the 2-norm?
Answer: Euclidean distance
◉ What is the 1-norm?
Answer: The rectilinear (Manhattan) distance
◉ What is the infinity norm?
Answer: The value of the largest dimension
◉ Measuring the quality of a model is called?
,Answer: Validation
◉ What does a confusion matrix show?
Answer: The performance of a classification model.
◉ A time series outlier that seems "off the curve" is called a...
Answer: contextual outlier.
◉ A data element that is different from all other data in a set is
called a...
Answer: point outlier.
◉ When something is missing in a range of points
Answer: it is called a..., collective outlier.
◉ The whiskers on a box plot extend to...
Answer: the 10th and 90th percentiles (or 5th and 95th)
◉ Why are hypothesis tests generally not sufficient for change
detection?
Answer: They are slow to detect changes.
, ◉ In CUSUM, T is _____ and C is _____.,
Answer: Threshold and a "bring down factor"
◉ In a CUSUM model, you adjust T and C to manage the tradeoff
between...,
Answer: early detection and false-alarms
◉ In exponential smoothing, if the data is less random, then you
want to pick an alpha that is...,
Answer: Close to 1.
◉ What is the initial condition for T in exponential smoothing with
trending?
Answer: T_i=0
◉ In cyclic exponential smoothing, L represents...,
Answer: The length of the cycle or season
◉ In cyclic exponential smoothing, C_1 ... C_L = ___?,
Answer: 1. In other words, initialize it to no initial cycle.
◉ Exponential, trending and cyclic smoothing are also referred to as
COMPLETE QUESTIONS AND SOLUTIONS
◉ Is written text structured or unstructured?
Answer: Unstructured
◉ When we increase the sum of the square of the coefficients we...
Answer: Decrease the distance between the lines
◉ In SVM soft classifier we tradeoff between maximizing ___ and
minimizing ___
Answer: margin and errors
◉ If lambda gets small what gets emphasized, large margin or
minimizing training error?,
Answer: Minimizing errors.
◉ What is a support vector?
Answer: A point that holds up a shape.
◉ Does ...[⅔(a-1)+1/3(a+1)] move an SVM classifier up or down?
Answer: Up
,◉ How do you make errors more costly in a soft SVM classifier?
Answer: include a multiplier for the point-error term.
◉ If an SVM coefficient is very close to zero...
Answer: that term is not very important to the classification.
◉ 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)
◉ What is the 2-norm?
Answer: Euclidean distance
◉ What is the 1-norm?
Answer: The rectilinear (Manhattan) distance
◉ What is the infinity norm?
Answer: The value of the largest dimension
◉ Measuring the quality of a model is called?
,Answer: Validation
◉ What does a confusion matrix show?
Answer: The performance of a classification model.
◉ A time series outlier that seems "off the curve" is called a...
Answer: contextual outlier.
◉ A data element that is different from all other data in a set is
called a...
Answer: point outlier.
◉ When something is missing in a range of points
Answer: it is called a..., collective outlier.
◉ The whiskers on a box plot extend to...
Answer: the 10th and 90th percentiles (or 5th and 95th)
◉ Why are hypothesis tests generally not sufficient for change
detection?
Answer: They are slow to detect changes.
, ◉ In CUSUM, T is _____ and C is _____.,
Answer: Threshold and a "bring down factor"
◉ In a CUSUM model, you adjust T and C to manage the tradeoff
between...,
Answer: early detection and false-alarms
◉ In exponential smoothing, if the data is less random, then you
want to pick an alpha that is...,
Answer: Close to 1.
◉ What is the initial condition for T in exponential smoothing with
trending?
Answer: T_i=0
◉ In cyclic exponential smoothing, L represents...,
Answer: The length of the cycle or season
◉ In cyclic exponential smoothing, C_1 ... C_L = ___?,
Answer: 1. In other words, initialize it to no initial cycle.
◉ Exponential, trending and cyclic smoothing are also referred to as
