CREATED BY GHUASIN_RASHED

Ch. 6

Frequent pattern: a that occurs frequently in a data .1

set

?Why Is Freq. Pattern Mining Important .2

.An intrinsic and important property of datasets

:Foundation for many essential data mining tasks
Association, correlation, and causality analysis
Sequential, structural patterns
Classification

Apriori pruning principle: If there is any itemset .3

which is infrequent, its superset should not be
!generated/tested

?Further Improvement of the Apriori Method .4

 Major computational challenges
 Multiple scans of transaction database
 Huge number of candidates
 Tedious workload of support counting for
candidates
 Improving Apriori: general ideas
 Reduce passes of transaction database scans
 Shrink number of candidates
 Facilitate support counting of candidates

5. Benefits of the FP-tree Structure ?

 Completeness
 Preserve complete information for frequent
pattern mining
 Never break a long pattern of any transaction
 Compactness
 Reduce irrelevant info—infrequent items are gone
 Items in frequency descending order: the more
frequently occurring, the more likely to be shared Never
be larger than the original database
 CREATED BY GHUASIN_RASHED

6. What about if FP-tree cannot fit in

memory?
 DB projection

7. Parallel projection vs. partition projection

techniques ?
Parallel projection
Project the DB in parallel for each frequent
item
Parallel projection is space costly
All the partitions can be processed in parallel
Partition projection
Partition the DB based on the ordered frequent
items
Passing the unprocessed parts to the
subsequent partitions

8. Advantages of the Pattern Growth Approach ?

Divide-and-conquer
A good open-source implementation and refinement
of FPGrowth

Ch. 7

Redundancy Filtering: Some rules may be redundant .9

due to “ancestor” relationships between items

Categorical Attributes: finite number of possible .10

values, no ordering among values—data cube
approach
 CREATED BY GHUASIN_RASHED

Quantitative Attributes: Numeric, implicit ordering .11

among values—discretization, clustering, and
gradient approaches

?Techniques can be categorized by how numerical attributes .12

 Static discretization based on predefined concept
hierarchies
 Dynamic discretization based on data distribution
 Clustering: Distance-based association
 One dimensional clustering then association

Negative and Rare Patterns .13

 Rare patterns: Very low support but interesting
E.g., buying Rolex watches
 Negative patterns Since it is unlikely that one buys
Ford Expedition
Constraint-Based Frequent Pattern Mining . 14

 Pattern space pruning constraints

 Anti-monotonic: If constraint c is violated, its further
mining can be terminated
 Monotonic: If c is satisfied, no need to check c again
 Succinct: c must be satisfied, so one can start with
the data sets satisfying c
 Convertible: c is not monotonic nor anti-monotonic,
but it can be converted into it if items in the
transaction can be properly ordered
 Data space pruning constraint
 Data succinct: Data space can be pruned at the
initial pattern mining process
 Data anti-monotonic: If a transaction t does not
satisfy c, t can be pruned from its further mining

Why Is Pattern-Fusion Efficient?

 A bounded-breadth pattern tree traversal
 Ability to identify “short-cuts” and take “leaps”
 Gearing toward colossal patterns
 Catching outliers
 CREATED BY GHUASIN_RASHED

Ch. 8
15. Supervised vs. Unsupervised Learning ?

 Supervised learning (classification)

Supervision: The training data (observations,
measurements, etc.) are accompanied by labels
indicating the class of the observations
 New data is classified based on the training set
 Unsupervised learning (clustering)
The class labels of training data is unknown
Given a set of measurements, observations, etc.
with the aim of establishing the existence of
classes or clusters in the data

16. Classification—A Two-Step Process ?

 Model construction: describing a set of
predetermined classes
 Each tuple/sample is assumed to belong to a
predefined class, as determined by the class label
attribute
 The set of tuples used for model construction
is training set
 The model is represented as classification
rules, decision trees, or mathematical formulae
 Model usage: for classifying future or unknown
objects
 Estimate accuracy of the model
 The known label of test sample is
compared with the classified result from the model
 Accuracy rate is the percentage of test
set samples that are correctly classified by the model
 Test set is independent of training set
(otherwise overfitting)
 If the accuracy is acceptable, use the model
to classify new data
 Note: If the test set is used to select models, it is
called validation (test) set
 CREATED BY GHUASIN_RASHED

17. Algorithm for Decision Tree Induction

 Basic algorithm (a greedy algorithm)

 Tree is constructed in a top-down recursive divide-
and-conquer manner
 At start, all the training examples are at the root
 Attributes are categorical (if continuous-valued,
they are discretized in advance)
 Examples are partitioned recursively based on
selected attributes
 Test attributes are selected on the basis of a
heuristic or statistical measure
 Conditions for stopping partitioning
 All samples for a given node belong to the same
class
 There are no remaining attributes for further
partitioning – majority voting is employed for
classifying the leaf
 There are no samples left

18. Comparing Attribute Selection Measures ?

The three measures, in general, return good results but

 Information gain:
 biased towards multivalued attributes
 Gain ratio:
 tends to prefer unbalanced splits in which one
partition is much smaller than the others
 Gini index:
 biased to multivalued attributes
 has difficulty when # of classes is large
 tends to favor tests that result in equal-sized
partitions and purity in both partitions
19. Other Attribute Selection Measures ?
 CHAID: a popular decision tree algorithm,
measure based on χ2 test for independence
 CREATED BY GHUASIN_RASHED

 C-SEP: performs better than info. gain and gini

index in certain cases
 G-statistic: has a close approximation to χ2
distribution
 MDL (Minimal Description Length) principle (i.e.,
the simplest solution is preferred):

20. Overfitting and Tree Pruning ?

Overfitting: An induced tree may over fit the
training data
Two approaches to avoid overfitting
Prepruning: Halt tree construction early ̵ do not
split a node if this would result in the goodness
measure falling below a threshold
Postpruning: Remove branches from a “fully
grown” tree—get a sequence of progressively
pruned trees

21. Enhancements to Basic Decision Tree Induction?

Allow for continuous-valued attributes
Handle missing attribute values
Attribute construction

?Why is decision tree induction popular .22

 relatively faster learning speed (than other
classification methods)
 convertible to simple and easy to understand
classification rules
 can use SQL queries for accessing databases
 comparable classification accuracy with other
methods

?Bayesian Classification: Why .23

A statistical classifier: performs probabilistic prediction, i.e.,

predicts class membership probabilities
.Foundation: Based on Bayes’ Theorem
 CREATED BY GHUASIN_RASHED

Performance: A simple Bayesian classifier, naïve Bayesian

classifier, has comparable performance with decision tree and
selected neural network classifiers
Incremental: Each training example can incrementally
increase/decrease the probability that a hypothesis is correct —
prior knowledge can be combined with observed data
Standard: Even when Bayesian methods are computationally
intractable, they can provide a standard of optimal decision making
against which other methods can be measured

? Naïve Bayes Classifier .24

Advantages
Easy to implement
Good results obtained in most of the cases
Disadvantages
Assumption: class conditional independence, therefore loss
of accuracy
Practically, dependencies exist among variables
?How to Learn-One-Rule .25

..Start with the most general rule possible

Adding new attributes by adopting a greedy depth-first
..strategy
Rule-Quality measures: consider both coverage and accuracy
..Rule pruning based on an independent set of test tuples
How can we measure accuracy? Other metrics to .26
?consider
 Use validation test set of class-labeled
tuples instead of training set when
assessing accuracy
 Methods for estimating a classifier’s accuracy:
Holdout method, random subsampling
Cross-validation
Bootstrap
 CREATED BY GHUASIN_RASHED

 Comparing classifiers:
Confidence intervals
Cost-benefit analysis and ROC Curves

? Evaluating Classifier Accuracy: Bootstrap .27

Bootstrap
Works well with small data sets
Samples the given training tuples uniformly with
replacement

? Issues Affecting Model Selection .28

 Accuracy
 classifier accuracy: predicting class label
 Speed
 time to construct the model (training
time)
 time to use the model
(classification/prediction time)
 Robustness: handling noise and missing
values
 Scalability: efficiency in disk-resident
databases
 Interpretability
 understanding and insight provided by
the model
? Popular ensemble methods .29
 Bagging: averaging the prediction over a collection of
classifiers
 Boosting: weighted vote with a collection of classifiers
 Ensemble: combining a set of heterogeneous
classifiers

?How boosting works .30

 CREATED BY GHUASIN_RASHED

 Weights are assigned to each training tuple

 A series of k classifiers is iteratively learned
 After a classifier Mi is learned, the weights are
updated to allow the subsequent classifier
 The final M* combines the votes of each individual
classifier, where the weight of each classifier's vote is
a function of its accuracy

31. Typical methods for imbalance data in 2-class classification?

 Oversampling: re-sampling of data from

positive class
 Under-sampling: randomly eliminate tuples
from negative class
 Threshold-moving: moves the decision
threshold, t, so that thrare class tuples are easier
to classify, and hence, less chance of costly false
negative errors
 Ensemble techniques: Ensemble multiple
classifiers introduced above

Ch. 9

32. How Are Bayesian Networks Constructed?

 Subjective construction: Identification of (direct)
causal structure
 Synthesis from other specifications
 Learning from data

?Classification by Backpropagation .33

Backpropagation: A neural network learning
algorithm
Started by psychologists and neurobiologists to
develop and test computational analogues of neurons
A neural network: A set of connected input/output units
where each connection has a weight associated with it During the
 CREATED BY GHUASIN_RASHED

learning phase, the network learns by adjusting the weights so as to

be able to predict the correct class label of the input

? Neural Network as a Classifier .34

 Weakness
 Long training time
 Require a number of parameters typically best
determined empirically, e.g., the network topology or
“structure.”
 Poor interpretability: Difficult to interpret the
symbolic meaning behind the learned weights and of
“hidden units” in the network
 Strength
 High tolerance to noisy data
 Ability to classify untrained patterns
 Well-suited for continuous-valued inputs and outputs
 Successful on an array of real-world data, e.g., hand-
written letters
 Algorithms are inherently parallel
 Techniques have recently been developed for the
extraction of rules from trained neural networks
35.Discriminative Classifiers ?
 Advantages
 Prediction accuracy is generally high
 Robust, works when training examples contain errors
 Fast evaluation of the learned target function
 Criticism
 Long training time
 Difficult to understand the learned function (weights)
 Not easy to incorporate domain knowledge

36.SVM vs. Neural Network ?

 CREATED BY GHUASIN_RASHED

Support Vector Machine Neural Network

(SVM )
Deterministic algorithm Nondeterministic
algorithm
Nice generalization Generalizes well but
properties doesn’t have strong
mathematical foundation
Hard to learn – learned Can easily be learned in
in batch mode using incremental fashion
quadratic programming
techniques
Using kernels can learn To learn complex
very complex functions functions—use
multilayer perceptron
(nontrivial)

37.Why Associative Classification effective?

 It explores highly confident associations among

multiple attributes and may overcome some
constraints introduced by decision-tree induction
 Associative classification has been found to be often
more accurate than some traditional classification
methods

38.Lazy vs. Eager Learning ?

a.Lazy learning (e.g., instance-based learning): Simply stores training
data (or only minor processing) and waits until it is given a test tuple
b.Eager learning (the above discussed methods): Given a set of training
tuples, constructs a classification model before receiving new (e.g.,
test) data to classify

39. Genetic Algorithms (GA) ?

Genetic Algorithm: based on an analogy to biological evolution

An initial population is created consisting of randomly generated rules
40.Semi-Supervised Classification ?
Semi-supervised: Uses labeled and unlabeled data to build a classifier .
 CREATED BY GHUASIN_RASHED

Ch 10

41.What is Cluster Analysis?

Cluster: A collection of data objects similar (or related) to one another
within the same group dissimilar (or unrelated) to the objects in other
groups
Cluster analysis (or clustering, data segmentation, …)
Finding similarities between data according to the
characteristics found in the data and grouping similar data
objects into clusters .

42.Basic Steps to Develop a Clustering Task ?

 Feature selection
 Proximity measure
 Clustering criterion
 Clustering algorithms
 Validation of the results
 Interpretation of the results

43.What Is Good Clustering ?

 A good clustering method will produce high quality

clusters
 high intra-class similarity: cohesive within
clusters
 low inter-class similarity: distinctive between
clusters
 The quality of a clustering method depends on
 the similarity measure used by the method
 its implementation, and
 Its ability to discover some or all of the hidden
patterns
 CREATED BY GHUASIN_RASHED

44.What are the Considerations for Cluster Analysis?

 Partitioning criteria
 Separation of clusters
 Similarity measure
 Clustering space
45.Major Clustering Approaches ?
 Partitioning approach:
 Hierarchical approach:
 Density-based approach:
 Grid-based approach:
 Model-based:
 Frequent pattern-based
 User-guided or constraint-based:
 Link-based clustering
46.What Is the Problem of the K-Means Method?
 The k-means algorithm is sensitive to outliers
 K-Medoids: Instead of taking the mean value of the object in a
cluster as a reference point, medoids can be used, which is the
most centrally located object in a cluster

47.Distance between Clusters ?

 Single link: smallest distance between an element in one cluster
and an element in the other
 Complete link: largest distance between an element in one
cluster and an element in the other
 Average: avg distance between an element in one cluster and an
element in the other
 Centroid: distance between the centroids of two clusters
 Medoid: distance between the medoids of two clusters

48.Algorithmic and Probabilistic hierarchical clustering?

 CREATED BY GHUASIN_RASHED

 Algorithmic hierarchical clustering

 Nontrivial to choose a good distance measure
 Hard to handle missing attribute values
 Optimization goal not clear: heuristic, local search
 Probabilistic hierarchical clustering
 Use probabilistic models to measure distances between
clusters
 Generative model: Regard the set of data objects to be
clustered as a sample of the underlying data generation
mechanism to be analyzed
 Easy to understand, same efficiency as algorithmic
agglomerative clustering method, can handle partially
observed data

? Two parameters for the Density-Based Clustering .49

 Eps: Maximum radius of the neighbourhood

 MinPts: Minimum number of points in an
Eps-neighbourhood of that point

50.Major features of Density-Based Clustering ?

 Discover clusters of arbitrary shape
 Handle noise
 One scan
 Need density parameters as termination
condition

51. Strength and Weakness of CLIQUE ?

 Strength
 automatically finds subspaces of the highest
dimensionality such that high density clusters exist
in those subspaces
 insensitive to the order of records in input and does
not presume some canonical data distribution
 scales linearly with the size of input and has good
scalability as the number of dimensions in the data
increases
 Weakness
 CREATED BY GHUASIN_RASHED

 The accuracy of the clustering result may be

degraded at the expense of simplicity of the
method

?Measuring Clustering Quality .52

kinds of measures: External, internal and relative 3

External: supervised, employ criteria not inherent to
the dataset
Internal: unsupervised, criteria derived from data itself
Relative: directly compare different clusterings, usually
those obtained via different parameter settings for the
same algorithm

Ch 11

?The k-means algorithm has two steps at each iteration .53

 Expectation Step (E-step): Given the

current cluster centers, each object is
assigned to the cluster whose center is
closest to the object: An object is
expected to belong to the closest
cluster
 Maximization Step (M-step): Given
the cluster assignment, for each cluster,
the algorithm adjusts the center so that
the sum of distance from the objects
assigned to this cluster and the new
center is minimized
?The EM (Expectation Maximization) Algorithm .54
The (EM) algorithm: A framework to
approach maximum likelihood or maximum
 CREATED BY GHUASIN_RASHED

a posteriori estimates of parameters in

.statistical models
 E-step assigns objects to clusters
according to the current fuzzy clustering
or parameters of probabilistic clusters
 M-step finds the new clustering or
parameters that maximize the sum of
squared error (SSE) or the expected
likelihood
? Advantages and Disadvantages of Mixture Models .55

 Strength
 Mixture models are more general than
partitioning and fuzzy clustering
 Clusters can be characterized by a small
number of parameters
 The results may satisfy the statistical
assumptions of the generative models
 Weakness
 Converge to local optimal (overcome: run
multi-times w. random initialization)
 Computationally expensive if the number of
distributions is large, or the data set contains
very few observed data points
 Need large data sets
 Hard to estimate the number of clusters

? Methods of Clustering High-Dimensional Data .56

Subspace-clustering: Search for clusters existing in

subspaces of the given high dimensional data space

Dimensionality reduction approaches: Construct a

much lower dimensional space and search for clusters
.there

Graph Clustering: Challenges of Finding Good .57

? Cuts
 High computational cost
 CREATED BY GHUASIN_RASHED

 Sophisticated graphs
 High dimensionality
 Sparsity
58. Two Approaches for Graph Clustering ?
 Use generic clustering methods for high-dimensional data
 Designed specifically for clustering graphs

Ch 12

?What Are Outliers and its types .59

Outlier: A data object that deviates significantly from the
normal objects as if it were generated by a different mechanism
Global outlier (or point anomaly)
Contextual outlier (or conditional outlier)
Collective Outliers

? Challenges of Outlier Detection .60

..Modeling normal objects and outliers properly
..Application-specific outlier detection
..Handling noise in outlier detection
..Understandability

?Two ways to categorize outlier detection methods .61

a) Based on whether user-labeled examples of outliers

..can be obtained
b) Based on assumptions about normal data and outlier

Clustering-Based Method: Strength and .62

?Weakness

 Strength
 Detect outliers without requiring any
labeled data
 Work for many types of data
 Clusters can be regarded as
summaries of the data
 Once the cluster are obtained, need
only compare any object against the
 CREATED BY GHUASIN_RASHED

clusters to determine whether it is an

outlier (fast)
 Weakness
 Effectiveness depends highly on the
clustering method used—they may
not be optimized for outlier detection
 High computational cost: Need to first
find clusters
 A method to reduce the cost: Fixed-
width clustering

Ch 13

? Statistical data mining methods .63

Regression: predict the value of a response (dependent) variable
from one or more predictor (independent) variables where the
variables are numeric
Generalized Linear Model: allow a categorical response
variable (or some transformation of it) to be related to a set of
predictor variables

Analysis of Variance: Analyze experimental data for two or more

populations described by a numeric response variable and one or
more categorical variables
Mixed-Effect Models: For analyzing grouped data, i.e. data that
can be classified according to one or more grouping variables

Factor Analysis: determine which variables are

combined to generate a given factor

Discriminant Analysis: predict a categorical response

variable, commonly used in social science

Survival Analysis: Predicts the probability that a patient

undergoing a medical treatment would survive at least to time

? Integration of visualization and data mining .64

 data visualization
 data mining result visualization
 data mining process visualization
 interactive visual data mining
 ‫‪CREATED BY GHUASIN_RASHED‬‬

‫تمنياتي للجميع بالسداد والتوفيق‬