A supervised neural network can be presented as a black box learn and predict as following:
Being supervised problem, it requires labelled data.
This is done in 2 steps:
-
Forward propagation
-
Backward propagation
Initially, the error in the networks were high
After, 1000 iterations:
Hence, the network was improving during each iteration by adjusting the weights calculated using loss function.
Overall process can be summarized as:
Team Members: Roll No:
- Pramish Poudel 076bct047
- Prakash Chaulagain 076bct045
- Bikal Devkota 076bct0
- Denish Shrestha 076bct0