7.6 () Consider the logistic regression model with a target variable t ∈ {−1, 1}. If we define p(t = 1|y) = σ(y) where y(x) is given by (7.1), show that the negative log likelihood, with the addition of a quadratic regularization term, takes the form

(7.47).