Kaggle Dataset: https://www.kaggle.com/datasets/ardamavi/sign-language-digits-dataset
It is also called deep neural network or deep learning.
What is neural network: It is basically taking logistic regression and repeating it at least 2 times.
In logistic regression, there are input and output layers. However, in neural network, there is at least one hidden layer between input and output layer. What is deep, in order to say "deep" how many layer do I need to have: When I ask this question to my teacher, he said that ""Deep" is a relative term; it of course refers to the "depth" of a network, meaning how many hidden layers it has. "How deep is your swimming pool?" could be 12 feet or it might be two feet; nevertheless, it still has a depth--it has the quality of "deepness". 32 years ago, I used two or three hidden layers. That was the limit for the specialized hardware of the day. Just a few years ago, 20 layers was considered pretty deep. In October, Andrew Ng mentioned 152 layers was (one of?) the biggest commercial networks he knew of. Last week, I talked to someone at a big, famous company who said he was using "thousands". So
I prefer to just stick with "How deep?"" Why it is called hidden: Because hidden layer does not see inputs(training set) For example you have input, one hidden and output layers. When someone ask you "hey my friend how many layers do your neural network have?" The answer is "I have 2 layer neural network". Because while computing layer number input layer is ignored.
It is solved with the example of logistic regression application. It is a sequel. Just in case, I added the Logistic regression application to github. Good coding already
Size of layers and initializing parameters weights and bias
Forward propagation
Loss function and Cost function
Backward propagation
Update Parameters
Prediction with learnt parameters weight and bias
Create Model