Chapter 8

Question 1

supervised learning(classification)

Answer 1

the training data (observations, measurments, etc.) are accompanied by labels indicating the class of  the observations
new data is classified based on the training set

Question 2

Unsupervised learning (clusterng)

Answer 2

THe class labels of training data is unknown
Given a set of measurments, observations, etc. with the aim of establishing the existence of classes or clusters in the data.

Question 3

Classification

Answer 3

predicts categorical calss labesl (discrete or nominal)
Classifies data based on the training set and the vaslues (class labels) in a classifying attribute and uses it in classifying new data

Question 4

Numeric prediction

Answer 4

models continous-valued functions, i.e., predicts unknown or missing values
credit loan aproval
medical diagnosis
fruad detection
web page categorization

Question 5

Learning step (model construction)

Answer 5

describing a ser of predetermined classes
each tuple/sample is assumed to belong to a predefined class, as determined by the class ;abe; attribute
the setr of tuples ued for model construction is the traning set

Question 6

Classification step (model usage)

Answer 6

for classifying future or unknown objects

Question 7

Estimate accuracy

Answer 7

the known label of test sample is compared with the classified reult from the model

Question 8

acurracy rate

Answer 8

is the percentage of test set samples thgat are correctly classified by the model

Question 9

test set

Answer 9

independent of traning set (otherwise overfitting)

Question 10

ckassify new data

Answer 10

if the accuracy is acceptable

Question 11

validation test set

Answer 11

if the test set is used to select modeles

Question 12

decision tree induction

Answer 12

the learning of decison trees form class labled training tuples

Question 13

internal node

Answer 13

denotes a test on an attriubute

Question 14

branch

Answer 14

represents an outcome of the test

Question 15

leaf node or terminal node

Answer 15

holds a class label

Question 16

root node

Answer 16

top most node in a tree

Question 17

d-data partition

Answer 17

initially it is the complete ser of training tuples and their associated class labels

Question 18

attribute-list

Answer 18

list ofattrivutes describes the tuples

Question 19

attribute_selection_method

Answer 19

a heuristic procedure for selecting the attribute that best disctiminates the given tuples according to class

Question 20

gini index

Answer 20

enforce the resulting tree to be binary

Question 21

information cange

Answer 21

allows multiway splits

Question 22

determine splitting criterion

Answer 22

idelly the resulting partitions at each branch are as pure as possible

Question 23

pure partition

Answer 23

when all the tuples in it belong to the same class

Question 24

discrete valued partitioning

Answer 24

each variable has a partition or line to it

Question 25

continous valued partition

Answer 25

there is a slit point resaulting in two partitions either below or over the split

Question 26

discrete-valued and a binary tree must be produced

Answer 26

A question is asked and the outcome is the partition or line

Question 27

attribute selection measures

Answer 27

splitting rules

Question 28

The splitting attribute

Answer 28

the attribute habving the best score fot the measure, either maxiamise or minimize, is chosen for the given tuple.

Question 29

entropy

Answer 29

is the state of disorder, confusion, surprise. uncertainty, and disorganization ( second law of thermal dynamics, ethropy increases over time)

Question 30

The more heterogenous the event the more uncertainty

Answer 30

less heterogeneous, more homogenous thge event, less uncertainty

Question 31

x-axis

Answer 31

the probability of the event

Question 32

y-axis

Answer 32

is the heterogeneity or the imurity denoted by H(X) - Entropy (impuraty measure)

Question 33

uncertantiy chjanges depending on the likelihood of an event

Answer 33

pr(x=0) no uncertainty, pr(x=0.0) maximum uncerainty, pr(x=1) no uncertainty

Question 34

higher entropy

Answer 34

higher uncertainty

Question 35

lower entropy

Answer 35

lower uncertainty

Question 36

info(D)

Answer 36

is just the average amount of infromation needed to identify the class label of a tuple in D

Question 37

High ingformation gain

Answer 37

minimizes the information needed to classify truple in the resulting partitions, and reflects the least randomness or impurity in the partitions

Question 38

information gained

Answer 38

defined as the differnce between the priginal information required ( based on just the proportion of classes) and the new requirement (obtained after partitioning on A)

Question 39

Gain(a)

Answer 39

tells us how much would be ained by brachinbg on A. it is the expected reduction in the information requirement caused by knowing the value of A. THe attribute A with the highest information gain is chosen as the splitting attribute node N

Question 40

gain ratio for atrribute selection (c4.5)

Answer 40

information gain meaure is biased towars atrributes with a large number of values. IT uses gain ratio to overcome the problem (normalization to information gain)

Question 41

maximum gain ratio

Answer 41

the attribute is slected as the splitting attribute

Question 42

gini index measure the impuritty of D

Answer 42

if a data set D contins examples form n classes =, giuni index , gini(d) is define The gini index considers a binary split for each attribute

Question 43

Smallest gini(D)

Answer 43

the larges reduction in impuity is chosen to split the node

Question 44

information gain

Answer 44

biased towards multivaalued attributes ( attributes with a large nu,ber of values)

Question 45

Gain ratio

Answer 45

tends to prefer unbalanced splits in which one p[partition is much smaller than the others

Question 46

gigni index

Answer 46

biased to multivalued attributes has difficulty when number of classes is large ttends to favor test that result in equal-sized portions and purity in both partitions

Question 47

evaluation metrics

Answer 47

how can we measure accuracy? other metrics to consider?

Question 48

postive tuples samples

Answer 48

tuples of the main class of interest

Question 49

negative tuples

Answer 49

all other tuples

Question 50

true positve

Answer 50

the postitve tuples that were correctly labled by the classsifier

Question 51

ture negatives

Answer 51

the negative tuples that were correctly labled by the classifier

Question 52

false positve

Answer 52

the negaive tuples that were incorrectyl labled as psoitive

Question 53

false negative

Answer 53

the poistive tiuples that were mislabled as negaive

Question 54

ressubstitution error

Answer 54

error rate on training set insted of a test set

Question 55

classifier accuracy or recognition rate

Answer 55

percentage of test set tuples that are correctly classified (TP + TN)/All

Question 56

error rate

Answer 56

misclassification rate; 1-accuracy | (fp+fn)/All

Question 57

Class imbalance problem

Answer 57

``` one calss may be rare, e.g. fraud or cancer significant majority of the negative class and minority of the positive class ```

Question 58

sensitivity

Answer 58

true pisitve recognition rate; tp/p

Question 59

specifitity

Answer 59

true negatice recognition rate; tn/n

Question 60

precision

Answer 60

exactness, what % of tuples that the classifier labeled as positive are actually psitive

Question 61

recall

Answer 61

completeness, what % of positive tuples did the classifier label as positive, the perfect score is 1.0

Question 62

perfect percision score

Answer 62

1.0 for a class C means that every tuple that the classifier leaves as belonging to class C does indeed being to class C. it does not tell us anything about the number of class C tuples that the classifier mislabled

Question 63

perfect recall score

Answer 63

1.0 for C means that very item from class C was labeled as such, but ut does not tell us how many thoeth tuples were incorrectly labeled as belonging to class C.

Question 64

f measeure

Answer 64

combine precision and recall in one formula, harmonic mean of precision and recall gives equal weight to ptrecision and recall

Question 65

accuracy

Answer 65

classifier accuracy, predicting class label

Question 66

speed

Answer 66

time to sonstruct the model, taining time | time to use the model. classification or prediction time

Question 67

robustness

Answer 67

handling noise and missing values

Question 68

scalability

Answer 68

efficieny in disk resident database

Question 69

interpretability

Answer 69

undersatading and insight provided by the model