# what is Actually Forward Propagation?

First step of Training neural network let us start understanding using dataset

let's understand through examples:

Data Set overview:

let take an example and as marked in diagram :

let us consider the first-row of dataset and build a neural network so it will easy to visualize

Each and independent feature First row of a dataset will passed to input-layer to a neural network act as an input

start as forward propagation by randomly intialising weight of all neurons this weight affected by edges connecting to neurons

weight are thought of more apropriately between two layers since edges connecting two layers

as shown in the diagram will show red-colored

let's discuss about **first neuron of first hidden** layer this neuron will receive **input** from a previous layer and **combines** with own set of weights and **represented by red edges** After **applying some Activation function** calculates **output for hidden activation f**or this neuron .

hidden Activation will serve as input for next layer neurons similarly for all hidden activation layer for all neurons are calculated

when hidden activation are calculated this is said to have calculated hidden activation for first layer

hidden Activation layer will serve as input to these neurons as described below in diagram

once last hidden layer is calculated and combined by a finite set of weight between last hidden layer and output layer to produce output single row of observation

actual value will be not validated with predicted value as you notice we have calculated activation sequentially neuron that means or one after the other

one optimization we can make is to calculate in parallel since calculation is independent of each other that can be computed in parallel easily.

we can calculate the first hidden layer easily as shown in the diagram below:

how we can compute parallel easily first layer then calculate the second layer simultaneously then finally calculate the output as shown in diagram below

final summary and definition derived for Forwarding propagation:

**Definition:**

*series of calculation from the input layer to the output layer is called** forward propagation.*