You are on page 1of 34

O R K S, CO NN E C T I O N ,

NETW
RELA T I O N S H I P
1
NETWORKS
AN INTERCONNECTED OR
INTERRELATED CHAIN, GROUP, OR
SYSTEM
2
CONNECTIONS
THE ACT OF CONNECTING; THE
STATE OF BEING CONNECTED
3
RELATIONSHIP
THE RELATION CONNECTIONS
OR BINDING PARTICIPANTS IN A
RELATIONSHIP. 4
WHAT IS NEURAL
NETWORK? 5
NEURAL NETWORKS
-a branch of "artificial intelligence".
*ARTIFICIAL NEURAL NETWORK
-is a system loosely modeled based on the human
brain 6
HOW NEURAL
NETWORKS WORK? 7
A SIMPLE NEURAL NETWORK INCLUDES AN INPUT
LAYER, AN OUTPUT (OR TARGET) LAYER AND, IN
BETWEEN, A HIDDEN LAYER. THE LAYERS ARE
CONNECTED VIA NODES, AND THESE CONNECTIONS
FORM A “NETWORK” – THE NEURAL NETWORK – OF

INTERCONNECTED NODES. 8
9
THE INPUT LAYER CONSISTS OF ONE OR MORE FEATURE
VARIABLES (OR INPUT VARIABLES OR INDEPENDENT
VARIABLES) DENOTED AS X1, X2, …, XN. THE HIDDEN LAYER
CONSISTS OF ONE OR MORE HIDDEN NODES OR HIDDEN UNITS.
A NODE IS SIMPLY ONE OF THE CIRCLES IN THE DIAGRAM
ABOVE. SIMILARLY, THE OUTPUT VARIABLE CONSISTS OF ONE
OR MORE OUTPUT UNITS. 10
11
Traning is a very important part of
neural network which I will
discuss in the next article. But as
of now just we will see the
application of neural network how
it works practically.
1. Area ( in square feet).
2. Number of Bedrooms.
3. Distance to City (in miles).
4. Age of the property.
BASIC NEURAL NETWORK
basic neural network, these input variables
are just weighted up with synapses and the
output is calculated. In that case, we get an
output as a price. And here you can use any
activation function in the output layer for
predicting the output.
Without having a hidden layer neural networks perform most of the
operations. So this shows how much a powerful neural network is.
STANDARD NEURAL NETWORK
WE HAVE THE FLEXIBILITY
AND POWER TO INCREASE
ACCURACY. AND THAT
POWER IS A HIDDEN LAYER.
SYNAPSES HAVE SOME WEIGHTS. SOME WEIGHTS MAY HAVE ZERO VALUE AND SOME WEIGHTS MAY
HAVE NON ZERO VALUE. THAT MEANS NOT ALL INPUTS ARE VALID AND IMPORTANT FOR EVERY SINGLE
NEURON. SOMETIMES SOME INPUTS ARE NOT VALID FOR SOME SPECIFIC NEURON.
Neurons are connected with the help of synapses
So it means that the farther away you get
from the city the cheaper real state becomes.
And therefore the space or area in square feet
of the house becomes larger. So for the same
price you get a larger property as farther
away you go from the city. That’s normal!.
That means it is looking for properties that are not so far
from the city but have a large area. So the activation
function will activate and it will fire up only when the certain
criteria are met. And then this neuron performs calculations
inside itself and it combines these two input variables. And
that contributes to the price in the output layer.
Therefore this top neuron doesn’t care about bedrooms
and age of the property because it is focused on these two
parameters.
Here we have got three parameters feeding into this neuron. That is area,
number of Bedrooms, and Age of the property.
The reason is that in specific cities neural networks may
be trained for those families who have two or more
children. And these families are looking for large and new
properties with lots of bedrooms. They are not looking for
older properties. So they want the age of the property to be
lower. When this neuron finds a property with a new, large
area and with lots of bedrooms, it will fire up. So neuron
picks up this property.
Perhaps if a property is over a certain age that can
indicate that it’s a historic property. For instance if a
property is under 100 years old then its less
valuable. But as soon as it jumps over 100 years old
all of a sudden it becomes a historic property.
Because it tells a story and some people like that. In
fact, a lot of people feel proud to live in such a
property.
Therefore this neuron as soon as it sees a property over 100
years old it will fire up and contribute to the overall price.
And if the property is under 100 years old then it will not
work. This is the good example of the Rectifier function. So
here you have got 0 until a certain point let’s say 100 years
and after 100 years old the older the property gets the
higher the value. And higher the contribution of this neuron
to the overall price.
The neural network could have picked up
things that we wouldn’t have thought of
ourselves. For instance, they can pick up a
bedroom and city maybe that is in
combination and somehow contribute to
the price. It may be not as strong as other
neurons but still, it can contribute.
As a single ant can’t
build a colony but a
thousand of ants can
build.

You might also like