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

hello and welcome in this video we will
provide an overview of logistic
regression as well as a basic
implementation in tensorflow logistic
regression is an important tool for
machine learning and a great stepping
stone for learning about more advanced
models and algorithms you might be
familiar with linear regression which is
a useful tool for estimating continuous
values like housing prices for example
but if we need to perform classification
and determine which class a data point
is likely to belong to then we need to
use a different type of model this is
where logistic regression comes in
logistic regression provides a formula
that predicts the likelihood that a
given input belongs to a certain class
the model makes these predictions by
analyzing the data spheres which are a
set of independent variables that
describe the data intuitively logistic
regression takes in a series of inputs
and passes them through a set of three
functions first the inputs are
multiplied by a set of values known as
the weights the weight values represent
the relative amount of emphasis to put
on each feature of the input the model
improves its classification ability by
gradually updating these weights
throughout the training process the
weight multiplied value is then fed into
another function which adds on an
additional value called a bias a bias is
an important value that increases the
flexibility of our model beyond what a
set of weights could do alone and like
the weights the bias is also updated
throughout the training process our last
step is to feed the resulting sum into
something called a sigmoid function the
sigmoid function will map the sum onto a
curve that represents the probability of
the input belonging to a certain class
after loading your data set it's easy to
create this three function model with
tensorflow
we first need to implement the weight
multiplication which we'll do by using
the mat mol function
Matt mole simply stands for matrix
multiplication and this function
receives our input and the weights as
parameters we then take the result of
the matrix multiplication step and use
the add function to add on the bias and
finally we pass the sum through the n n
dot sigmoid function at this point you
should understand the purpose of
logistic regression and how to implement
a three function model in tensorflow
thank you for watching this video