Supervised Learning - Classification Week 9 Challenge

Ans 1. A Neural Network is a computational model inspired by the way biological neural networks in the human brain work. It is a type of ML algorithm used primarily for pattern recognition, data classification, and regression tasks.

Ans 2. Deep learning is a subset of machine learning that focuses on algorithms inspired by the structure and function of the brain, specifically ANN. It is called "deep" learning because these networks consist of multiple layers that progressively extract higher-level features from raw input data.

Ans 3. The activation function in a neural network introduces non-linearity into the model. Without it, the network would behave like a linear regression model. It helps the network to understand more intricate relationships between inputs and outputs. Popular activation functions include ReLU, Sigmoid.

Ans 4. Types of activation function:

a) Sigmoid Activation Function:-The Sigmoid function is a smooth, S-shaped curve that squashes the input values between 0 and 1. It's commonly used in the output layer for binary classification tasks.

Sigmoid(x)=1+e−x1​

Range: (0, 1)

b) ReLU (Rectified Linear Unit) Activation Function:-The ReLU is one of the most popular activation functions, especially for deep learning. It returns the input directly if it's positive, otherwise, it returns 0.

ReLU(x)=max(0,x)

Range: [0, ∞)

Ans 5. Backpropagation is a fundamental algorithm used to train artificial neural networks by optimizing their weights and biases. It is part of the process called supervised learning and is essential for improving a model's accuracy during training.