BADM 211 Test 2 – Questions With Ready
Solutions
Save
Terms in this set (92)
In a dataset, rows and records, variables(features)
columns correspond to
________ and _________
respectively.
continuous or integer
A numerical variable can
be defined as ___________.
Two types of categorical text or integer
variables are ____________
and ____________.
We need to dummy code nominal categorical variables
_____________ .
compute the z-score of the variables
Two ways of normalizing
rescale a uniform range (subtract the mean and
the data are ____________ .
divide by standard deviation)
The main reason for overfitting
doing data partitioning is
_____________ .
Scatter plots represent relationship between numerical variables
___________ .
We use _________ to box plot and histograms
visualize the entire
distribution of a variable.
, y=beta knot + beta one X one + beta n X n + error
used to fit a relationship between a numerical
What is linear regression?
outcome variable (Y) and predictors (x1, x2, xp)
supervised learning
What is predictive predictive: y-hat, focusing on outcomes
modeling and explanatory: beta,
explanatory modeling? 1st explanatory, then predictive
What are the take past data and predict beta
differences?
Beta coefficients weight
represent ______________ .
The higher the error , the less
______ model.
In explanatory modeling R^2
goodness of fit can be
assessed using ________.
of whether the predictions are on average, over or
.Mean error gives us under predicting the outcome variable
information ____________ . you need absolute mean error because they could
cancel out
1) follows normal distribution
2)the choice of predictors and their form is correct
Four assumptions in (linearity) (assuming what you are given is correct)
linear regression are 3) the records are independent of each other
_____________ . (avoidance of multicollinearity)
4) homoskedasticity (same direction no matter x or
error)
1) naive benchmark (average of outcomes in training
set)
.Two measures that can
2) Prediction Accuracy Measures (MAE (mean
be used to assess
absolute error), ME (mean error), RMSE)
predictive performance
are ___________ .
MAE and ME for training
RMSE for validation
Solutions
Save
Terms in this set (92)
In a dataset, rows and records, variables(features)
columns correspond to
________ and _________
respectively.
continuous or integer
A numerical variable can
be defined as ___________.
Two types of categorical text or integer
variables are ____________
and ____________.
We need to dummy code nominal categorical variables
_____________ .
compute the z-score of the variables
Two ways of normalizing
rescale a uniform range (subtract the mean and
the data are ____________ .
divide by standard deviation)
The main reason for overfitting
doing data partitioning is
_____________ .
Scatter plots represent relationship between numerical variables
___________ .
We use _________ to box plot and histograms
visualize the entire
distribution of a variable.
, y=beta knot + beta one X one + beta n X n + error
used to fit a relationship between a numerical
What is linear regression?
outcome variable (Y) and predictors (x1, x2, xp)
supervised learning
What is predictive predictive: y-hat, focusing on outcomes
modeling and explanatory: beta,
explanatory modeling? 1st explanatory, then predictive
What are the take past data and predict beta
differences?
Beta coefficients weight
represent ______________ .
The higher the error , the less
______ model.
In explanatory modeling R^2
goodness of fit can be
assessed using ________.
of whether the predictions are on average, over or
.Mean error gives us under predicting the outcome variable
information ____________ . you need absolute mean error because they could
cancel out
1) follows normal distribution
2)the choice of predictors and their form is correct
Four assumptions in (linearity) (assuming what you are given is correct)
linear regression are 3) the records are independent of each other
_____________ . (avoidance of multicollinearity)
4) homoskedasticity (same direction no matter x or
error)
1) naive benchmark (average of outcomes in training
set)
.Two measures that can
2) Prediction Accuracy Measures (MAE (mean
be used to assess
absolute error), ME (mean error), RMSE)
predictive performance
are ___________ .
MAE and ME for training
RMSE for validation
