Question: You are preparing for a competition on using the perception algorithm without computational aid Consider the following two-point 20 dataset: 1. Positive class ( +1):


You are preparing for a competition on using the perception algorithm without computational aid Consider the following two-point 20 dataset: 1. Positive class ( +1): (1,3) 2. Negative class (-1): (-1, 4) Starting with wp " (0,0), how many updates will you have to perform to w until convergence? Write down the sequence of each updated w; (w1,"02...), ) by iterating the data points in the order: [(1, 3), (-1, 4)]- Problem 2 Suppose your boss Celeste gives you a fully trained (i.e. converged) Perceptron with weight vector w, the training set used to train it (Dry,), and a separate test act (D. ). Here's e Rd and ye {-10, +10). She wants to know whether the test error is greater than the training error
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
