CSCI 343 Quiz 3 Flashcards
random forests is a(n) ? method
ensemble
for each tree in random forests, you should
randomly choose a subset of features and build a decision tree using only those
decision rules
if, then rules interpreted from a tree (ex: if income > 70, then risk)
build trees with
training data
test tree strength (quality) with
validation
you can combine trees with high percentages to make
better trees
NN stands for
neural network
NN is named this way because it is
a model of a brain
single neuron structure
inputs with weights -> neuron -> output
the neuron contains some function
f(Σxiwi)
weights are originally
randomly assigned
deep neural nets have
a bunch of hidden layers
recurrent neural nets have
cycles
model of a neural net is the
structure of the net and learned weights
activation function at node’s idea is to make the output
non-linear
sigmoid function
f(y) = 1 / (1 + e^-y)
which node activation function is most common
sigmoid
ReLU stands for
rectified linear unit
ReLU function
f(y) = max(0, y)
in NN, the work is in (training/testing)
training
structure of a neural net
input layer, 0 or more hidden layers, output layer
NN are ? graphs
directed, acyclic
typically the number of inputs in a neural net is equivalent to
the number of features
it is typical for a neural net to be fully
connected
