SYS 4021 Practice Test With
Complete Solutions
Why is evidence informed reasoning so important for systems engineering? -
ANSWER EISE applies the scientific method to the design of complex systems, it
uses evidence to inform engineering decisions, and it provides the best
assurance that the designed system will meet the user's needs
What steps do we take to implement evidence informed reasoning? - ANSWER
Problem Description: Situation, Goal, Metric, Hypotheses. Approach: Data,
Analysis, Evidence. Recommendation
List two sources of evidence - ANSWER Observations, Experiments
What is the difference between observations and experiments? - ANSWER
Experiment - apply treatments to experimental units and then observe effects
Observation - observe subjects and measure variables without assigning
treatments
Define a well-formed hypothesis. - ANSWER Actionable, testable and rejectable
What is the difference between cross sectional data and time series data? -
ANSWER Cross-sectional: many instances at one point in time
Time-series: one instance over many points in time
What is the most common problem in data visualization? - ANSWER Confirmation
bias - "seeing what we expect to see"
Compare and contrast a histogram and a density plot. - ANSWER Histogram -
graphical displays of data distributions by counts in intervals or bin of the
variable's values
Kernel Density Plot - graphical display of a data distribution that smooths the
display by summing over kernel functions at the data points.
What is the center line or point in the box on a box plot? - ANSWER Median
What are the observations above or below the whiskers in a box plot? - ANSWER
Outliers
What is a Q-Q plot and why is it useful? - ANSWER A graphical display of the
quantiles of a data distribution versus the quantiles of a reference distribution,
, typically the Gaussian distribution. It is used to check the validity of a
distribution
In R what plotting technique is particularly useful for visualizing a categorical
variable vs. a quantitative variables? - ANSWER Categorical box plots
What is the typical graphical display in two dimensions? - ANSWER Scatterplot
What display shows pairwise relations in more than two dimensions - ANSWER
Scatter Plot Matrix
What is the center line or point in the box on a box plot? - ANSWER Median
The first principal component is a projection that maximizes what at- tribute of
the data? - ANSWER Variance
How many principal components are there in a data set with only quantitative
variables? - ANSWER Number of components k = Number of variables p
What is the linear algebra solution for the first principal component? - ANSWER
The first eigenvector of S (variance-covariance matrix for X)
Why do we often use a correlation matrix and not the covariance matrix to
compute the principal components? - ANSWER To avoid issues of scale - you
use a covariance matrix when the variables' scales are similar and correlation
matrix when variables are on different scales
In the simple linear regression model Y = β0 + β1X1 + ε, name all of the terms in
the model. (Name Y , β0, β1, X1, and ε?) - ANSWER Y = predictand, B0 =
intercept, B1 = regression coefficient, X1 = predictor, ϵ = residual error
In the linear regression model Y = β0 + β1X1 + +β2X2 + ε, how many coefficients
and parameters are there? - ANSWER
What are the assumptions for the error term in a multiple regression? - ANSWER
expected value of error is 0. identical and independently distributed Gaussian
distribution. variance=sigma^2
What is the sum of squares decomposition of the total sum of squares used in
regression analysis? - ANSWER Total SS = Residual SS + Model SS
What does yiˆ mean in regression analysis? - ANSWER
What is the hypothesis for a t-test in multiple linear regression? - ANSWER null
hypothesis: Bi = 0
alternative hypothesis: Bi ≠ 0
Complete Solutions
Why is evidence informed reasoning so important for systems engineering? -
ANSWER EISE applies the scientific method to the design of complex systems, it
uses evidence to inform engineering decisions, and it provides the best
assurance that the designed system will meet the user's needs
What steps do we take to implement evidence informed reasoning? - ANSWER
Problem Description: Situation, Goal, Metric, Hypotheses. Approach: Data,
Analysis, Evidence. Recommendation
List two sources of evidence - ANSWER Observations, Experiments
What is the difference between observations and experiments? - ANSWER
Experiment - apply treatments to experimental units and then observe effects
Observation - observe subjects and measure variables without assigning
treatments
Define a well-formed hypothesis. - ANSWER Actionable, testable and rejectable
What is the difference between cross sectional data and time series data? -
ANSWER Cross-sectional: many instances at one point in time
Time-series: one instance over many points in time
What is the most common problem in data visualization? - ANSWER Confirmation
bias - "seeing what we expect to see"
Compare and contrast a histogram and a density plot. - ANSWER Histogram -
graphical displays of data distributions by counts in intervals or bin of the
variable's values
Kernel Density Plot - graphical display of a data distribution that smooths the
display by summing over kernel functions at the data points.
What is the center line or point in the box on a box plot? - ANSWER Median
What are the observations above or below the whiskers in a box plot? - ANSWER
Outliers
What is a Q-Q plot and why is it useful? - ANSWER A graphical display of the
quantiles of a data distribution versus the quantiles of a reference distribution,
, typically the Gaussian distribution. It is used to check the validity of a
distribution
In R what plotting technique is particularly useful for visualizing a categorical
variable vs. a quantitative variables? - ANSWER Categorical box plots
What is the typical graphical display in two dimensions? - ANSWER Scatterplot
What display shows pairwise relations in more than two dimensions - ANSWER
Scatter Plot Matrix
What is the center line or point in the box on a box plot? - ANSWER Median
The first principal component is a projection that maximizes what at- tribute of
the data? - ANSWER Variance
How many principal components are there in a data set with only quantitative
variables? - ANSWER Number of components k = Number of variables p
What is the linear algebra solution for the first principal component? - ANSWER
The first eigenvector of S (variance-covariance matrix for X)
Why do we often use a correlation matrix and not the covariance matrix to
compute the principal components? - ANSWER To avoid issues of scale - you
use a covariance matrix when the variables' scales are similar and correlation
matrix when variables are on different scales
In the simple linear regression model Y = β0 + β1X1 + ε, name all of the terms in
the model. (Name Y , β0, β1, X1, and ε?) - ANSWER Y = predictand, B0 =
intercept, B1 = regression coefficient, X1 = predictor, ϵ = residual error
In the linear regression model Y = β0 + β1X1 + +β2X2 + ε, how many coefficients
and parameters are there? - ANSWER
What are the assumptions for the error term in a multiple regression? - ANSWER
expected value of error is 0. identical and independently distributed Gaussian
distribution. variance=sigma^2
What is the sum of squares decomposition of the total sum of squares used in
regression analysis? - ANSWER Total SS = Residual SS + Model SS
What does yiˆ mean in regression analysis? - ANSWER
What is the hypothesis for a t-test in multiple linear regression? - ANSWER null
hypothesis: Bi = 0
alternative hypothesis: Bi ≠ 0