Deep Learning with Tensorflow - Linear Regression with TensorFlow Video Lecture | Tensorflow: Learning made Easy - AI & ML

hello and welcome in this video we will
show you how to perform linear
regression using the tensorflow library
linear regression is a technique for
describing the relationship between two
or more variables simple linear
regression involves only two the
dependent variable is the value that we
want to analyze or predict the
independent variable is explanatory and
is used to predict the changes in the
dependent variable when there are
multiple independent variables in the
model the process is called multiple
linear regression when there are
multiple dependent variables the process
is called multivariate linear regression
let's look at a well-known simple linear
relation in this equation y is the
dependent variable and X is the
independent variable
so essentially Y is a function of X a
represents the slope or the gradient of
the linear function and B represents the
intercept both a and B are model
parameters that can be adjusted if you
plot the model you'll see that the
result is simply a line adjusting the
slope parameter will change the angle
that the line forms with the independent
variables axis the intercept parameter
changes the location where the line
crosses the dependent variables axis
before we move on we need to import some
libraries that we use throughout the
video now let's see an example of a
linear equation we can form a more
meaningful view of linear relations by
looking at the real world phenomena they
describe one example is the relationship
between speed and distance traveled
linear relations can also describe the
properties of different materials when
we want to perform simple linear
regression on a collection of data we
need to adjust the slope and intercept
parameters so that the resulting line
fits the data points accurately as
possible now let's perform linear
regression using tensor
before we start will generate a set of
random data points using a nonlinear
equation and then we'll attempt to
recreate the parameters with regression
analysis this is the equation we'll use
to generate random data points there's
nothing particularly special about this
equation we just need a data set to work
with for this example feel free to
change the parameters if you'd like here
is a sample of the data we just created
to start the regression we need to
initialize our variables a and B with
random values these initial values will
almost certainly result in a poor fit
but we'll adjust the parameters during
training
once we have these values we can then
define our linear function for linear
regression we need to specify a loss
value that will be minimized during
training will define the loss as the
squared error where error is the
difference between the models prediction
and the actual value from the data set
to compute the loss we can use the
function t f dot reduce underscore mean
parentheses this function calculates the
mean of a multi-dimensional tensor and
the result can have a different number
of dimensions than the input we then
need to define the optimizer method for
our purposes we'll use a simple gradient
descent with a learning rate of 0.5 the
minimize method will calculate the
gradients and apply them to our
variables and don't forget to initialize
the variables before executing a graph
and with that we are now ready to start
the optimization and run the graph at
this point you should understand how to
perform simple linear regression using
tensorflow thank you for watching this
video to practice and learn more go to
the lab and run the code for yourself