UNIT-V

Q.N Level Question A B C D ANSWER

o
1 L1 A model's bias, bias, bias, None of A
generalizatio variance variance and standard the above
n error can and standard deviation
expressed as irreducibl deviation and RMSE
a sum of e error
three different
errors:
2 L3 The most widely Confusion Cost-sensitive Area under the All of the D
used metrics and matrix accuracy ROC curve above
tools to assess a
classification
model are:
3 L2 What is the To assess To judge how Both A and B None of C
purpose of the the trained the above
performing predictive model performs
cross-validation? performanc outside the
e of the sample on test
models data
4 L4 The most widely Confusion Cost-sensitive Area under the All of the A
used metrics and matrix accuracy ROC curve above
tools to assess a
classification
model are:

5 L3 Which of the Selecting Comparing Selecting All of the D

following is variables to predictors parameters in mentioned
correct use of include in a prediction
cross validation? model function
6 L3 Point out the True False False All of the C
wrong negative=c negative=correc positive=corre mentioned
combination. orrectly tly rejected ctly identified
rejected
7 L1 Which of the Sensitivity Median Specificity All of the D
following is a absolute mentioned
common error deviation
measure?
8 L2 Which of the gplot quickplot plot all of the A
following can be mention
used to create the ed
most common
graph types?
9 L2 In case of Larger Lesser Can’t Say None Of B
overfitting ,the The Above
training error is
_________ than
test error
10 L5 Given the larger Overfit Underfit Can’t Say None Of A
hypothesis space, The Above
there is a higher
tendency for the
model to
________ the
training dataset
11 L4 Given the 2 10 20 15 A
following models
trained using K-
 NN, the model
which could
result in
overfitting will
most likely have
the value of K as
___________
12 L4 Given the 30 5 1 10 A
following models
trained using K-
NN, the model
which could
result in
underfitting will
most likely have
the value of K as
___________
13 L2 A model High bias High variance Can’t Say None Of A
suffering from The Above
underfitting will
most likely be
having
_____________
14 L1 A model High bias High variance Can’t Say None Of B
suffering from The Above
overfitting will
most likely
having
15 L3 In Ridge True False Can’t Say None Of A
regression, as the The Above
regularization
parameter
increases, do the
regression
coefficients
decrease?
16 L4 Which L1 L2 Both None of A
regularization is the above
used to reduce
the over fit
problem?
17 L5 What type of L0 L1 L2 None of C
penalty is used the above
on regression
weights in Ridge
regression?
18 L3 Ridge Regression L1 L2 L1 & L2 both. None of B
uses which the above
norm?
19 L2 In Ridge Alpha. Gamma. Lambda. None of C
regression, A the above
hyper parameter
is used called
“_____________
” that controls
the weighting of
the penalty to the
loss function.
20 L1 Ridge regression Squared Absolute value Cube value of Root value A
takes value of of variables. variables of
______________ variables. variables.
__ value of
variables.
21 L2 The effect of True False alpha value is None of A
alpha value on fix for ridge the above
both ridge and regression.
 lasso regression
is same in terms
of value increase
and decrease.
22 L4 Grid search is Linear in Polynomial in D Exponential in None of C
D. D the above
23 L2 The advantage of It can be It can be It is easy to All of the D
Grid search is applied to applied to non- implement. above
(are), non- continuous
differentiab functions
le
functions.
24 L2 Machine learning Training Transfer data Data training None of A
algorithms build data these
a model based on
sample data,
known as
25 L3 A machine Clustering Classification Anomaly All of the C
learning detection above
technique that
helps in detecting
the outliers in the
data
26 L3 Which of the Can help Cannot help Cannot help All of the A
following is with bias with model with variance mentioned
correct about trade-off selection trade-off
regularized
regression?
27 L1 Point out the Model Model based Model based All of the C
wrong statement. based approach use approach are mentioned
approach Bayes theorem reasonably
may be inaccurate on
computatio real problems
nally
convenient
28 L2 Which of the ridge lasso relaxo all of the D
following mentioned
methods are
present in caret
for regularized
regression?
29 L2 Which of the Model Model Model None of A
following stacking combining structuring the
method can be mentioned
used to combine
different
classifiers?
30 L1 Point out the Combining Combining Combining All of the C
correct statement. classifiers classifiers classifiers mentioned
improves reduces improves
interpretabi accuracy accuracy
lity
31 L3 Which of the cl_forecast cl_nowcast cl_precast None of D
following the
function provides mentioned
unsupervised
prediction?
32 L4 Model based True False Can’t say None of B
prediction the above
considers
relatively easy
version for
covariance
matrix.
33 L5 Which of the quantmod quantile quantity mboost A
following is used
 to assist the
quantitative
trader in the
development?
34 L3 Which of the predict forecast ets All of the B
following mentioned
function can be
used for
forecasting?
35 L2 Predictive True False Can’t say None of B
analytics is same the above
as forecasting.
36 L1 Which of the Total Total Variation Total All of the B
following is the Variation = = Residual Variation = mentioned
correct formula Residual Variation + Residual
for total Variation – Regression Variation *
variation? Regression Variation Regression
Variation Variation
37 L2 Residual ______ RR PP QQ None of C
plots investigate the above
normality of the
errors.
38 L1 Which of the rstudent cooks.distance rstandard all of the C
following show mentioned
residuals divided
by their standard
deviations?
39 L3 The least squares True False Can’t say None of B
estimate for the the above
coefficient of a
multivariate
regression model
is exactly
regression
through the
origin with the
linear
relationships.
40 L2 Which of the Flexibly fit Uncover Build accurate All of the D
following things complicate complex prediction mentioned
can be d functions multivariate models
accomplished relationships
with linear
model?