I liked that the mathematical explanation is very clear. Also, for the python implementation you wrote the code from scratch, rather than copy past it, and walked the viewer through each line. Thank you.
He is updating the weights and biases for each data sample. So at each iteration he makes the prediction with the updated weight. This is stochastic gradient. It can be done in the other way as you said also. The weights will be updated once after an epoch in that case.
during fittig linear_output = np.dot(x_i,self.weight) +self.bias during prediction linear_output = np.dot(X,self.weight) +self.bias X and x_i are two type of different object, so during the fitting linear_output is the prediction is done on one item, and during the prediction is a prediction of a list of item right?
I liked that the mathematical explanation is very clear. Also, for the python implementation you wrote the code from scratch, rather than copy past it, and walked the viewer through each line. Thank you.
The ending escalated very quickly, lol
Thank you! Clear and precise.
What's different in between this and Logit? Both scripts feel the same?
And how would you implement the multiclass one ?
why you did not write the loop in vectorized form like what you did in regression models?
what is the difference?
He is updating the weights and biases for each data sample. So at each iteration he makes the prediction with the updated weight. This is stochastic gradient. It can be done in the other way as you said also. The weights will be updated once after an epoch in that case.
Thnx buddy! Clean explanation
during fittig
linear_output = np.dot(x_i,self.weight) +self.bias
during prediction
linear_output = np.dot(X,self.weight) +self.bias
X and x_i are two type of different object, so during the fitting linear_output is the prediction is done on one item, and during the prediction is a prediction of a list of item right?
Thank you for sharing!!
Thanks for watching!
Nice explanation
thanks for help
Love you man...😘
didnt know messi was into teaching python
accuracy 100%?? that should be suspicious
not in a binary classifier