classification

Question 1

features

Answer 1

measurable quantities

Question 2

discriminatory power

Answer 2

quality of a feature that can help us classify

Question 3

feature space

Answer 3

the set of all features from the feature vector

Question 4

classification

Answer 4

process of organising things into groups or classes

Question 5

machine learning

Answer 5

the scienec of making

Question 6

features for image classification

Answer 6

form factor, Euler number

Question 7

What does a negative Euler number indicate?

Answer 7

number of holes is greater than the number of objects

Question 8

typical features of sound classification

Answer 8

waveform, frequency spectrum, formants(peaks in the spectrum), spectrogram, MFCC

Question 9

binarisation

Answer 9

converting an image or a signal into a binary format, typically consisting of only two intensity values, often black and white or 0 and 1

Question 10

rule based classifier should have…

Answer 10

mutually exclusive rules and those ruler should be exhaustive

Question 11

supervised classification
+
examples

Answer 11

relies on having a set of examples
template matching, decision tree, neural networks, naive Bayes

Question 12

unsupervised classification

Answer 12

doesn’t have a training set
clustering

Question 13

fingerprint

Answer 13

a feature that can be compared

Question 14

perceptron

Answer 14

a structure that contains a neuron with adjustable weights

Question 15

neural network

Answer 15

a multi-layer perceptron

Question 16

Hidden Markov Models

Answer 16

probabilistic reasoning algorithm that works on a set of temporal data

Question 17

reinforcement learning

Answer 17

between supervised and unsupervised

Question 18

classification error

Question 19

error rate

Answer 19

num of errors / num of attempts

Question 20

test set

Answer 20

has positive and negative instances

Question 21

true positive

Answer 21

an instance of the positive class that is correctly classifies

Question 22

true negative

Answer 22

an instance of the negative class that is correctly rejected

Question 23

false positive

Answer 23

an instance of the negative class that is wrongly classified

Question 24

false negative

Answer 24

an instance of the positive class that is wrongly rejected
dangerous!!

Question 25

accuracy

Answer 25

a=(tp+tn)/n

Question 26

precision

Answer 26

p=tp/(tp+fp)

Question 27

recall

Answer 27

r=tp/(tp+fn)

Question 28

confusion matrix

Answer 28

gives info on how frequently instances of class X are correctly classified as X or misclassified

Question 29

components of a classification system

Answer 29

sensing module, preprocessing mechanism, feature extraction mechanism, classifier, training set if needed