Hands-On Artificial Intelligence for IoT

Cross-entropy loss function

Logistic regression aims to find weights W and bias b, so that each input vector, X_i, in the input feature space is classified correctly to its class, y_i. In other words, y_i and should have a similar distribution for the given x_i. We first consider a binary classification problem; in this case, the data point y_i can have value 1 or 0. Since logistic regression is a supervised learning algorithm, we give as input the training data pair (X_i, Y_i) and let be the probability that P(y=1|X=X_i); then, for p training data ...