You are on page 1of 4

How do neural networks work?

Ex. Property valuation


Look at the neural network that takes in some parameters about a
property and value set and the thing here

Important part in neural networks is training them up.

Pretend: Work with a neural network that we're is already trained up.

Parameters of the property:


1. Area (feet sq)
2. Number of bedrooms INPUT LAYER
3. Distance to city (miles)
4. Age of property

We always begin with a layer of input variables. These are different factors assembled in a
single row of data, represented below on the left-hand side.

Their values go through the weighted synapses straight


over to the output layer. All four will be analyzed, an
activation function will be applied, and the results will be
produced.

This is comprehensive enough on a basic level.


But there is a way to amplify the power of the Neural
Network and increase its accuracy by a very simple
addition to the system.
Basic form a neural network
1. input layer
2. output layer, so no hidden layers and our output layer is the price that we're
predicting

Input variables
1. weighted up by the synapse
2. output layer will be calculated

 Calculated as the weighted sum


of all of the inputs
Activation functions
logistic regression
squared
Most of the machine learning algorithms that exist can be represented in this format.

Diagrammatic representation of how you deal with the variables by changing the weights
the formulas.

Extra advantage that gives us lots of flexibility and power which is where that increase in
accuracy comes from and that power is the hidden layers.

Ex. Property valuation: imagine that we're plugging in a property.


We begin with the four variables on the left and the top neuron of the hidden layer in the
middle. All four variable all be connected to the neuron by synapses.
However, not all of the synapses are weighted. They will either have a 0 value or non
0 value.
Four variables on the left
1. top neuron on the hidden layer, neurons have synapses connecting each one of
them to the top neuron in the hidden layer and those simulates have weight
2. Some weights will have a non-zero value some weights will have a 0 value
because basically not all inputs will be valid or not all inputs will be important for
every single neuron sometimes

x1 and x3 the area and the distance to the city in


miles are important for that neuron whereas bedrooms
and age are not.

Why? The further away you get from the city the
cheaper real estate becomes and therefore the space
in square feet of properties becomes larger. For the
same price you can get a larger property.

Their distance from the city have an unfair square foot area, higher than average
(abnormal).

neuron might be picking lesser,


activation function - will activate only when the certain criteria is met

Therefore, this neuron is focused on area and distance to the city. as long as we
3 Parameters
- area
- bedrooms
- age of the property

Ex. That specific city in those suburbs that this neural network has been trained in. Perhaps
there's a lot of families with kids with two or more children who are looking for large
properties with lots of bedrooms.

1. modern families
2. social demographic shift
3. growth in employment and jobs for the younger side of population
4. population demographics
5. prefer newer properties
6. so they want the age of the property to be lower and hence from the training that
this neural network has undergone it
knows
when there is a property with a large area and with lots of bedrooms
at least three bedrooms for the parents
for the first child for the second child
for at least three bedrooms
guest room when a new property with higher area and lots of bedrooms that is
valued in that market that is valuable

Neurons know what I'm looking for


 doesn't care about the distance of the city in miles whatever it is
 as long as it has high area lots of bedrooms

Helps with the evaluation of the property combines them into a new attribute and therefore
it's more precise.
1 Parameter:

 older property issues (less valuable) because it's worn out


 falling apart more maintenance is required so the price drops
in terms of the price of the real estate
 if a property is over a certain age that could indicate that it's
a historic property
- if a property is under 100 years old then the older it is the less
valuable it is but as soon as it’s over 100 years it becomes a
historic property (tells a story, history)
 some people value that

Otherwise if it's under 100 years old then it won't


good example of the rectifier function being applied
Zero to up, 0 years to 100 yrs = higher price
neurons
 contributes or maybe it this detracts from the price neuron picked up all four
parameters

Power of hidden layer situation


 allows you to increase the flexibility of your neural network
 allows the neural network to look for very specific things

 Each one of this neurons by itself cannot predict the price but together they have
super powers and they predict the price.
 they can do an accurate job if trained

You might also like