204.5.6 Neural Network Intuition

Link to the previous post : https://course.dvanalyticsmds.com/204-5-5-practice-implementing-intermediate-outputs-in-python/

Before going further into neural network algorithm, we need to understand and break down how the algorithm is working.

Final Output

$y = o u t (h) = g (\sum W_{j} h_{j})$ $h_{j} = o u t (x) = g (\sum w_{(} j k) x_{k})$

$y = o u t (h) = g (\sum W_{j} g (\sum w_{(} j k) x_{k}))$

So h is a non linear function of linear combination of inputs – A multiple logistic regression line.
Y is a non linear function of linear combination of outputs of logistic regressions.
Y is a non linear function of linear combination of non linear functions of linear combination of inputs.
We find W to minimize $\sum_{i = 1}^{n} [y_{i} - g (\sum W_{j} h_{j})]^{2}$
We find $W_{j}$ and $w_{j} k$ to minimize $\sum_{i = 1}^{n} [y_{i} - g (\sum W_{j} g (\sum w_{(} j k) x_{k}))]^{2}$
Neural networks is all about finding the sets of weights $W_{j}$ and $w_{j} k$ using Gradient Descent Method.

The neural networks methodology is similar to the intermediate output method explained above.
But ,we will not manually subset the data to crate the different models.
The neural network technique automatically takes care of all the intermediate outputs using hidden layers.
It works very well for the data with non-linear decision boundaries.
The intermediate output layer in the network is known as hidden layer.
In Simple terms, neural networks are multi layer nonlinear regression model.
If we have sufficient number of hidden layers, then we can estimate any complex non-linear function.

Why are they called hidden layers?

A hidden layer “hides” the desired output.
Instead of predicting the actual output using a single model, build multiple models to predict intermediate output
There is no standard way of deciding the number of hidden layers.

Algorithm is all about finding the weights/coefficients
We randomly initialize some weights; Calculate the output by supplying training input; If there is an error the weights are adjusted to reduce this error.

The next post is about the neural network algorithm .

Link to the next post : https://course.dvanalyticsmds.com/204-5-7-the-neural-network-algorithm/

21st June 2017