BADM 211 Midterm 2 – Questions With Authentic
Answers
Save
Terms in this set (84)
In a dataset, rows Records
represent
In a dataset, columns Variables (features)
represent
Numerical variable can Continuous or integer
be either
Two types of categorical Ordinal and non-ordinal
variables are
We need to dummy code Non-ordinal categorical variables
for
Two ways of normalizing Z-score, or subtract min and divide by range
data are
The main reason for To avoid overfitting and to predict new data
doing data partitioning is
Scatter plots represent Relationship between 2 numerical variables
Bar chart represents Relationship between categorical and numerical
Histogram is used to Visualize the entire distribution of a variable
, A way to predict numerical
outcomes
Uses predictors (X1, ..., Xp) to
explain outcome (Y)
What is multiple linear
regression?
Ex.) house characteristics to
predict selling price,
advertising expense to predict
sales, product demand to
predict valuation
1) Noise follows normal distribution
2) There is a linear relationship between predictors
What are the four main
and outcome variables
assumptions in linear
3) Cases (records) are independent of each other
regression?
4) Variance of the error does not depend on the
values of the predictors (homoskedasticity)
***Y/N: If assumptions of YES. On average, the sample error is going to be 0
MLR hold true, then does
mean error (ME) = 0?
**What is the naive Mean of outcome variable
predictor in regression?
****Which of the Noise follows a poisson distribution
following is not an
assumption of linear
regression models?
What do beta The weights given to each predictor
coefficients represent?
***T/F: In a MLR, the False
coefficients (B) are the
predictor (independent)
variables.
How do we estimate the By using test data (validation set)
beta coefficients of a
MLR model? Minmizing error in the training set
Answers
Save
Terms in this set (84)
In a dataset, rows Records
represent
In a dataset, columns Variables (features)
represent
Numerical variable can Continuous or integer
be either
Two types of categorical Ordinal and non-ordinal
variables are
We need to dummy code Non-ordinal categorical variables
for
Two ways of normalizing Z-score, or subtract min and divide by range
data are
The main reason for To avoid overfitting and to predict new data
doing data partitioning is
Scatter plots represent Relationship between 2 numerical variables
Bar chart represents Relationship between categorical and numerical
Histogram is used to Visualize the entire distribution of a variable
, A way to predict numerical
outcomes
Uses predictors (X1, ..., Xp) to
explain outcome (Y)
What is multiple linear
regression?
Ex.) house characteristics to
predict selling price,
advertising expense to predict
sales, product demand to
predict valuation
1) Noise follows normal distribution
2) There is a linear relationship between predictors
What are the four main
and outcome variables
assumptions in linear
3) Cases (records) are independent of each other
regression?
4) Variance of the error does not depend on the
values of the predictors (homoskedasticity)
***Y/N: If assumptions of YES. On average, the sample error is going to be 0
MLR hold true, then does
mean error (ME) = 0?
**What is the naive Mean of outcome variable
predictor in regression?
****Which of the Noise follows a poisson distribution
following is not an
assumption of linear
regression models?
What do beta The weights given to each predictor
coefficients represent?
***T/F: In a MLR, the False
coefficients (B) are the
predictor (independent)
variables.
How do we estimate the By using test data (validation set)
beta coefficients of a
MLR model? Minmizing error in the training set
