Scribd
Upload
32 views

Regression Analysis

This document summarizes key metrics and concepts from regression analysis hands-on cards, including: - The adjusted R-squared value was 0.483 and R-squared was 0.484 for an OLS regression in Python Statsmodels. - There were 506 observations in the dataset and the estimated coefficient for variable RM was 9.1. - For a multiple linear regression in MLR Hands On, the constant term was 36.49 and R-squared was 0.741. - It examines concepts like standard error, p-values, multi-collinearity, and how adding more variables can increase model accuracy but also diminish returns.

Uploaded by

Sujit Kumar
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
32 views

Regression Analysis

This document summarizes key metrics and concepts from regression analysis hands-on cards, including: - The adjusted R-squared value was 0.483 and R-squared was 0.484 for an OLS regression in Python Statsmodels. - There were 506 observations in the dataset and the estimated coefficient for variable RM was 9.1. - For a multiple linear regression in MLR Hands On, the constant term was 36.49 and R-squared was 0.741. - It examines concepts like standard error, p-values, multi-collinearity, and how adding more variables can increase model accuracy but also diminish returns.

Uploaded by

Sujit Kumar
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 2

Regression Analysis

Based on the hands on card “ OLS in Python Statsmodels” What is the adjusted R sq
value ? -- .483

Based on the hands on card “ OLS in Python Statsmodels” What is the value of R sq ?
-- 0.484

Based on the hands on card “ OLS in Python Statsmodels”What is the value of the
constant term ? -- -34.67

Based on the hands on card “ OLS in Python Statsmodels” How many observations are
there in the dataset ? - 506

Based on the hands on card “ OLS in Python Statsmodels” What is the value of the
estimated coef for variable RM ? - 9.1

Based on the hands on card “MLR Hands On” What is the value of the estimated coef
for the constant term ? - 36.49

Based on the hands on card “MLR Hands On” what is the value of R sq ? -- 0.741

Based on the hands on card “MLR Hands On” Perform a correlation among all the
independent variables . What is the correlation between variables NOX and DIS ? --
-0.76923

Based on the hands on card “MLR Hands On” What is the P>|t| value for the 'INDUS'
variable -- 0.731

Based on the hands on card “MLR Hands On” What is the standard error for the
constant term ? -- 5.104

-----------

__________ means predicting one variable from another. -- Regress

It is OK to discard theoretical considerations for Statistical Measures -- false

By adding multiple variables in Multi Variable Regression , the model accuracy


_____________ -- Increases

It is advised to go for a simpler model while fitting multiple regression for a


dataset - true

It is advised to omit a term that is highly correlated with another while fitting a
Multiple Regression Model -- true

The SSE depends on the number of observations in the data set -- True

pr(>|t|) term signifies how likely the estimated value is zero -- true

When two or more variables are correlated in a Multiple Regression Model , it is


called as -- multi collinearity

Regression can show causal relationship between two variables. -- false

When more variables are added in Multi Variable Regression the marginal improvement
decreases as each variable is added. This term is called ? -- law of diminishing
returns
What is the quantity that measures the strength of relationship between two
variables ? -- corelation

What is the term that represents the difference between actual and predicted value
called ? -- error

What is the process of removing the mean and dividing the value by the standard
deviation -- standardisation

In Multi Variable regression you predict one variable using more than one variable
-- true

What is the sum of standard error for the baseline model ? -- error

Sum of Squared error is a measure of standard for a Regression Line - true

__________ is a unit less quantity - Rsquare

What is the good range of correlation values to include in the regression model --
-0.7 to 0.

R Square Value can be greater than zero -- false

What is the process of rescaling the values in the range [0,1] -- normalisation

Arithmetic Mean can be used as a prediction measure. -- true

SSE is _________ for the Line of Best Fit and _______ for the baseline model --
small, big

What is the formula for root means square error ? --- sqrt(SSE/n)

is the basic property of the model of best fit -- minimise error

You might also like