In this post we will learn some basics about Artificial Neural Network(ANN) like Introduction to ANN, Architecture of ANN, Types of ANN and Advantages of ANN.
INTRODUCTION
- An artificial neural network is a computional model based on the structure and functionality that resembles biological neural network.
- In ANN the neurons are interconnected to each other which are called as nodes.
- An ANN is configured for specific application such as pattern recognition, data classification through a learning process.
- ANN is a machine learning model that are inspired by the structure and function of human brain.
ARCHITECTURE
The architecture of ANN consists of various layers like, Input Layer, Hidden Layer, Output Layer.
- INPUT LAYER : Input layer is the first layer of a artificial neural network. It accepts the inputs from several formats provided by the user.
- HIDDEN LAYER : The inputs received in the input layer are sent to hidden layer to perform all calculations. The Hidden layer performs all the required calculations on the received data and discover patterns observe and analyze them and finally produces the desired output.
- OUTPUT LAYER : The calculated output of the given dataset is conveyed by the output layer. The main task of output layer is to show the output that is calculated in the hidden layer.
ADVANTAGES OF ANN
- ANN’s can model and learn complex relationships between input and output data.
- ANN’s are excel at recognition and extraction of pattern from large datasets.
- ANN can generalize the training data that to make predictions on unknown data.
- ANN can automatically extract features on raw data reducing the need of manual feature engineering.
- ANN can handle large datasets making them suitable for big data applications.
TYPES OF ANN
The types of ANN are as follows :-
- FEED FORWARD NEURAL NETWORK
- CONVOLUTIONAL NEURAL NETWORK(CNN)
- RECURRENT NEURAL NETWORK(RNN)
FEED FORWARD NEURAL NETWORK : This neural network is the most basic type of ANN where the information flows only in one direction from the input layer through one or more hidden layers to the output layers.
CONVOLUTIONAL NEURAL NETWORK(CNN) : CNN is a type of neural network which is specially designed for image and video processing tasks. They employ convolutional layers that extract relevant features from input data and pooling layers to down sample the extracted features.
RECURRENT NEURAL NETWORK(RNN) : RNN’s are designed to process sequential data where information flows in loops allowing the networks to have memory. Each neuron in RNN receives memory not only from previous layer but also previous time stop. RNN’s are mostly used for language processing speech recognition time series analysis.
