My Review of artificial neural networks
The concepts of Artificial Neural Networks
An ANN is a network of perceptions, a perceptron imitates the human brain cell. Like how a human brain is created with a network of cells, an ANN is created with a network of perceptions. The perceptron is called a node in the neural network. The node as in term is to represent perceptions.
Nodes are organized into multiple layers in a neural network.
A deep neural network usually has three or more layers. Each node has its own weights, bias, and activation functions. And each node is connected to all the nodes in the next layer, forming a dense network.
The nodes within a layer are not connected with each other. exceptions too are in advanced use cases.
The diagram above is an example neural network. Were each neural network having one input layer, one or more hidden layers, and one or more output layer.
In the input layer, there is one node for each independent variable. An example is to imagine, there are three. The hidden layer has three layers for this example. Layer one has four nodes, layer two has five nodes, and layer three has three nodes. The number of layers, and the nodes in each layer, are determined by experience and trials, and it will vary from case to case. The number of nodes in the output layer will vary based on the type of predictions. The output layer has two nodes. This arrangement of node represents the architecture of a given neural network.
How does the ANN work for predictions?
The inputs or independent variables are sent from the input layer to the network. Data may be pre-processed before using them. Inputs are passed on to the next layer. Each node is a perceptron containing weights, bias, and an activation function. The formula is applied to the weights and the outputs derived. This repeats for each node in the layer. The results from all the nodes in a layer are passed onto the next layer, and this process is repeated. As this process reaches the output layer, the final predictions will be derived.
