Models & Definitions

Question 1

One vs Rest Classifier (OvR)

Answer 1

used for classification algorithms which are inherently binary (ex. logistic regression)

transforms multi-class classification targets into multiple binary classifications for each target class

ex. 5 prediction classes, OvR creates a binary True/False algorithm for each class

Is this city NY? Y/N, Boston? Y/N, DC? Y/N, Richmond? Y/N, Miami? Y/N

Question 2

Artificial Neural Network (ANN)

def. and comp.

Answer 2

computing systems which loosely model the activities of the human brain

components:

( 1 ) neurons: conceptual derivations of biological neuron, that accept inputs (data or outputs of other neurons) and produce a single output

( 2 ) connections + weights: connects neurons, and assigns weights (relative importance) to the outputs between neurons

( 3 ) propagation function: computes the input to a neuron from the outputs of its predecessor neurons

Question 3

Deep Neural Network (DNN)

def.

Answer 3

an ANN with multiple layers between the input and output layers

Question 4

Recurrent Neural Network (RNN)

Answer 4

Processes temporal information (data that comes in sequences). Uses the internal state memory to process varying lengths of sequences.

Used in unsegmented, handwriting, and speech recognition modeling task.

Question 5

Feedforward Neural Network

Answer 5

NN where the information only moves in one direction, forward, from the input nodes, to hidden nodes, then to the output nodes

there are no cycles or loops