Last updated 2 years ago
If δk(x)\delta_k(x)δk(x)is linear, then DDD is also linear.
E(Yk∣X=x)=P(G∣X)E(Y_k|X=x)=P(G|X)E(Yk∣X=x)=P(G∣X) When we estimate this with XTβX^T\betaXTβ, the range can be out of range 0 to 1.
minBΣ∣∣yi−[(1,xiT)B]T∣∣2\min_{\mathbf{B}} \Sigma||y_i-[(1,x_i^T)\mathbf{B}]^T||^2minBΣ∣∣yi−[(1,xiT)B]T∣∣2
G^(x)=argmaxk∈gf^k(x)→G^(x)=argmink∣∣f^(x)−tk∣∣2\hat{G}(x)=argmax_{k \in g}\hat{f}_k(x) \rightarrow \hat{G}(x)=argmin_{k}||\hat{f}(x)-t_k||^2G^(x)=argmaxk∈gf^k(x)→G^(x)=argmink∣∣f^(x)−tk∣∣2
∑kG^(x)=1\sum_k\hat{G}(x)=1∑kG^(x)=1
