Scribd
Upload
10 views

Week 3 - 1-Linear Regression

Linear regression fits a line to a data set of observations to predict unobserved values. It works by minimizing the squared error between each data point and the line using least squares. R-squared measures how well the linear regression line fits the data, ranging from 0 to 1, with higher values indicating more of the variance is captured by the model.

Uploaded by

Noah Byrne
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
10 views

Week 3 - 1-Linear Regression

Linear regression fits a line to a data set of observations to predict unobserved values. It works by minimizing the squared error between each data point and the line using least squares. R-squared measures how well the linear regression line fits the data, ranging from 0 to 1, with higher values indicating more of the variance is captured by the model.

Uploaded by

Noah Byrne
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 8

Linear Regression

Fit a line to a data set


of observations
Use this line to
predict unobserved
values
I don’t know why they call it
“regression”. It’s really
misleading. You can use it to
predict points in the future, the
past, whatever. Usually, time has
nothing to do with it.
Linear Regression: How does it
work?
Usually using “least squares”
Minimises the squared-error between each
point and the line
Remember the slope-intercept equation
of a line? Y=mx + b
The slope is the correlation between the
two variables times the standard deviation
in Y, all divided by the standard deviation
in X.
 Nice how the standard deviation has some
real mathematical meaning?
The intercept is the mean of Y minus
the slope times the mean of X
But python will do all that for you.
Linear Regression: How does it
work?
Least squares minimises the sum of the squared
errors.
This is the same as maximizing the likelihood of the
observed data if you start thinking of the
probabilities and probability distribution functions
This is something called “maximum
likelihood estimation”
More than one way to do it
Gradient Descent is an alternate
method to least squares.
Basically iterates to find the line
that best follows the contours
defined by the data.
Can make sense when dealing
with 3D data.
Easy to try in Python and just
compare the results to least
squares
But usually least squares is
a perfectly good choice.
Measuring error with r-squared
How do we measure how well our line fits our
data?
R-squared (aka coefficient of
determination) measures:

The fraction of the total variation in


Y that is captured by the model
Computing r-squared
Interpreting r-squared
Ranges from 0 to 1
0 is bad (none of the variance is captured), 1 is
good (all of the variance is captured).

LinearRegression.ipynb

You might also like