This page lists down **40 regression (linear / univariate, multiple / multilinear / multivariate) interview questions** (in form of **objective questions**) which may prove helpful for **Data Scientists / Machine Learning** enthusiasts. Those appearing for interviews for machine learning / data scientist freshers / intern / beginners position would also find these questions very helpful and handy enough to quickly brush up / check your knowledge and prepare accordingly.

### Practice Tests on Regression Analysis

These **interview questions** are split into four different **practice tests with questions and answers** which can be found on following page:

- Linear, Multiple regression interview questions and answers – Set 1
- Linear, Multiple regression interview questions and answers – Set 2
- Linear, Multiple regression interview questions and answers – Set 3
- Linear, Multiple regression interview questions and answers – Set 4

### Regression Topics covered in these Practice Tests

Some of the following topics have been covered in these questions:

- Introduction to linear (univariate) and multi-linear / multiple (multivariate) regression
- Concepts related with coefficient of determination vis-a-vis pearson correlation coefficient
- Evaluation of regression models using different techniques such as t-tests, analysis of variance f-tests
- Sum of squares calculations and related concepts
- Concepts related with R-squared, adjusted R-squared

### Linear / Multi-linear Regression Questions and Answers

- In ________ regression, there is _______ dependent variable and ________ independent variable(s)
- Simple linear, one, multiple
- Multiple, multiple, one
- Simple linear, one, one
- Multiple, one, multiple

- In ________ regression, there is _______ dependent variable and ________ independent variable(s)
- Simple linear, multiple, one
- Simple linear, one, multiple
- Multiple, one, multiple
- Multiple, multiple, multiple

- It is OK to add independent variables to a multi-linear regression model as it increases the explained variance of the model and makes model more effcient
- True
- False

- Linear or multilinear regression helps in predicting _______
- Continuous valued output
- Discrete valued output

- Regression analysis helps in studying __________ relationship between variables.
- Deterministic
- Statistical

- Regression analysis helps in doing which of the following?
- Causal analysis
- Effects in forecasting
- Forecasting trends
- All of the above

- The best fit line is achieved by finding values of the parameters which minimizes the sum of __________
- Prediction errors
- Squared prediction errors

- Best fit line is also termed as _______
- Maximum squares regression line
- Least squares regression line

- Which of the following can be used to understand the statistical relationship between dependent and independent variables in linear regression?
- Coefficient of determination
- Correlation coefficient
- Both of the above
- None of the above

- It is absolutely OK to state that correlation does imply causation
- True
- False

- The value of coefficient of determination, R-squared, is _________
- Less than 0
- Greater than 1
- Between 0 and 1

- Which of the following can be used to understand the positive or negative relationship between dependent and independent variables
- Coefficient of determination
- Pearson correlation coefficient

- The goal of the regression model is to achieve the R-squared value ________
- Closer to 0
- Closer to 1
- More than 1
- Less than 1

- Pearson correlation coefficient is __________ to coefficient of determination
- Directly proportional
- Inversely proportional

- Pearson correlation coefficient does always have positive value
- True
- False

- Value of Pearson correlation coefficient near to zero represents the fact there is a stronger relationship between dependent and independent variables
- True
- False

- Population correlation coefficient and sample correlation coefficient are one and the same
- True
- False

- The value of Pearson correlation coefficient falls in the range of _________
- 0 and 1
- 0 and -1
- -1 and 1
- 1 and 2

- The value of correlation coefficient and R-squared remains same for all samples of data
- True
- False

- The large value of R-squared can be safely interpreted as the fact that estimated regression line fits the data well.
- True
- False

- The value of R-squared does not depend upon the data points; Rather it only depends upon the value of parameters
- True
- False

- The value of correlation coefficient and coefficient of determination is used to study the strength of relationship in ________
- Samples only
- Both Samples and Population
- Population only

- Which of the following tests can be used to determine whether a linear association exists between the dependent and independent variables in a simple linear regression model?
- T-test
- ANOVA F-test
- Both of the abov
- None of the abovee

- In order to estimate population parameter, the null hypothesis is that the population parameter is ________ to zero?
- Equal
- Not equal

- Which of the following can be used for learning the value of parameters for regression model for population and not just the samples?
- Hypothesis testing
- Confidence intervals
- Both of the above
- None of the above

- The value of R-Squared _________ with addition of every new independent variable?
- May increase or decrease
- Always increases
- Always decreases

- In order to reject the null hypothesis while estimating population parameter, p-value has to be _______
- More than 0.05
- Less than 0.05

- The value of ____________ may increase or decrease based on whether a predictor variable enhances the model or not
- R-squared
- Adjusted R-squared

- The value of Adjusted R-squared _________ if the predictor variable enhances the model less than what is predicted by chance?
- Increases
- Decreases

- In regression model t-tests, the value of t-test statistics is equal to ___________?
- Coefficient divided by Standard error of coefficient
- Standard error of coefficient divided by coefficient
- Coefficient plus standard error of coefficient

- In ANOVA test for regression, degrees of freedom (regression) is _________
- Equal to number of parameters being estimated
- One more than the number of parameters being estimated
- One less than the number of parameters being estimated

- In ANOVA test for regression, degrees of freedom (regression) is _________
- Equal to number of predictor variables
- One more than the number of predictor variables
- One less than the number of predictor variables

- For SST as sum of squares total, SSE as sum of squared errors and SSR as sum of squares regression, which of the following is correct?
- SST = SSR – SSE
- SST = SSR + SSE
- SST = SSR/SSE

- The value of coefficient of determination is which of the following?
- SSR / SST
- SSE / SST

- Mean squared error can be calculated as _______
- Sum of squares error / degrees of freedom
- Sum of squares regression/ degrees of freedom
- Sum of squares total/ degrees of freedom

- Sum of Squares Regression (SSR) is ________
- Sum of Squares of predicted value minus average value of dependent variable
- Sum of Squares of Actual value minus predicted value
- Sum of Squares of Actual value minus average value of dependent variable

- Sum of Squares Error (SSE) is ________
- Sum of Squares of predicted value minus average value of dependent variable
- Sum of Squares of Actual value minus predicted value
- Sum of Squares of Actual value minus average value of dependent variable

- Sum of Squares Total (SST) is ________
- Sum of Squares of predicted value minus average value of dependent variable
- Sum of Squares of Actual value minus predicted value
- Sum of Squares of Actual value minus average value of dependent variable

- ______ the value of sum of squares regression (SSR), better the regression model
- Greater
- Lesser

- The objective for regression model is to minimize ______ and maximize ______
- SSR, SSE
- SSE, SSR
- SSR, SST
- SSE, SST

Hope you would find above set of **questions** along with **practice tests **related with linear / multiple rergression useful for next / upcoming **interviews** in relation with **data scientist / machine learning engineer** position.

- Correlation Concepts, Matrix & Heatmap using Seaborn - September 29, 2020
- Beta Distribution Explained with Python Examples - September 24, 2020
- Bernoulli Distribution Explained with PythonExamples - September 23, 2020