Answer Happy

Please Fill in the gap with correct answers.
1. The optimization method used by linear discrimination is
called ________.
2. 4-bit vector quantization corresponds to how many
reference vectors? __________________
3. A machine learning algorithm consists of a model, a loss
function, and an _________ procedure.
4. The two categories of dimension reduction methods are
________ and ________.
5. In practice, the best way to find the optimal complexity
of a model is ________.
6. According to Bayes' rule, posterior=prior x likelihood /
________.
7. A sample has 20 positive instances. If an algorithm
correctly recognizes 15 of them, then the false negative rate of
this algorithm is ________%.
8. In stochastic gradient descent, update=(________) x
(expected output - actual output) x (________)