Transcript www.cs.laurentian.ca

Data Mining:
Concepts and Techniques
(3rd ed.)
— Chapter 4 —
Jiawei Han, Micheline Kamber, and Jian Pei
University of Illinois at Urbana-Champaign &
Simon Fraser University
©2010 Han, Kamber & Pei. All rights reserved.
1
July 17, 2015
Data Mining: Concepts and Techniques
2
Chapter 4: Data Warehousing and On-line
Analytical Processing

Data Warehouse: Basic Concepts

Data Warehouse Modeling: Data Cube and OLAP

Data Warehouse Design and Usage

Data Warehouse Implementation

Data Generalization by Attribute-Oriented
Induction

Summary
3
What is a Data Warehouse?

Defined in many different ways, but not rigorously.

A decision support database that is maintained separately from
the organization’s operational database

Support information processing by providing a solid platform of
consolidated, historical data for analysis.

“A data warehouse is a subject-oriented, integrated, time-variant,
and nonvolatile collection of data in support of management’s
decision-making process.”—W. H. Inmon

Data warehousing:

The process of constructing and using data warehouses
4
Data Warehouse—Subject-Oriented

Organized around major subjects, such as customer,
supplier, product, sales

Focusing on the modeling and analysis of data for
decision makers, not on daily operations or transaction
processing

Provide a simple and concise view around particular
subject issues by excluding data that are not useful in
the decision support process
5
Data Warehouse—Integrated


Constructed by integrating multiple, heterogeneous data
sources
 relational databases, flat files, on-line transaction
records
Data cleaning and data integration techniques are
applied to
 Ensure consistency in naming conventions, encoding
structures, attribute measures, etc. among different
data sources


E.g., Hotel price: currency, tax, breakfast covered, etc.
When data is moved to the warehouse, it is
converted.
6
Data Warehouse—Time Variant

The time horizon for the data warehouse is significantly
longer than that of operational systems



Operational database: current value data
Data warehouse data: provide information from a
historical perspective (e.g., past 5-10 years)
Every key structure in the data warehouse


Contains an element of time, explicitly or implicitly
But the key of operational data may or may not
contain “time element”
7
Data Warehouse—Nonvolatile

A physically separate store of data transformed from the
operational environment

Operational update of data does not occur in the data
warehouse environment

Does not require transaction processing, recovery,
and concurrency control mechanisms

Requires only two operations in data accessing:

initial loading of data and access of data
8
Operational Databases Vs Data Warehouses


Online Transaction Processing (OLTP)
 Operational databases perform online
transaction and query processing
 Day-to-day operations, e.g. purchasing,
inventory, manufacturing, banking, payroll
Online Analytical Processing (OLAP)
 Data Warehouse systems
 Data analysis and decision making
 Organize and present data in various formats
9
OLTP vs. OLAP
10
Why a Separate Data Warehouse?

High performance for both systems



Warehouse—tuned for OLAP: complex OLAP queries,
multidimensional view, consolidation
Different functions and different data:




DBMS— tuned for OLTP: access methods, indexing, concurrency
control, recovery
missing data: Decision support requires historical data which
operational DBs do not typically maintain
data consolidation: DS requires consolidation (aggregation,
summarization) of data from heterogeneous sources
data quality: different sources typically use inconsistent data
representations, codes and formats which have to be reconciled
Note: There are more and more systems which perform OLAP
analysis directly on relational databases
11
Data Warehouse: A Multi-Tiered Architecture
12
Data Warehouse: A Multi-Tiered Architecture



Bottom-tier
 Warehouse database server – relational database system
 Back-end tools and utilities feed data from operational
databases – extract, clean, transform, load and refresh
 Data extracted using APIs called gateways – ODBC,
OLEDB, JDBC
 Metadata repository
Middle tier
 OLAP server – relational OLAP (ROLAP) or
multidimensional OLAP (MOLAP)
Top tier
 Front-end client layer – query, reporting tools, analysis
tools, data mining tools
13
Three Data Warehouse Models


Enterprise warehouse
 collects all of the information about subjects spanning
the entire organization
Data Mart
 a subset of corporate-wide data that is of value to a
specific groups of users. Its scope is confined to
specific, selected groups, such as marketing data mart


Independent vs. dependent (directly from warehouse) data mart
Virtual warehouse
 A set of views over operational databases
 Only some of the possible summary views may be
materialized
14
Top-down and bottom-up approaches

Top-down development of enterprise warehouse



Systematic solution, minimizes integration problems
Expensive, time consuming, lacks flexibility
Bottom-up design, development, deployment of independent
data marts


Flexibility, low cost, rapid return of investment
Problems in integrating disparate data marts into a consistent
enterprise data warehouse
15
Data warehouse - Third approach

Incremental and evolutionary method of development of
warehouse
 Define high-level corporate data model
 Provides corporate-wide, consistent, integrated view of
data among different subjects and users
 Independent data marts can be implemented in
parallel with enterprise warehouse
 Distributed data marts can be constructed to integrate
different data marts
 Finally, a multitier data warehouse is constructed –
enterprise warehouse custodian of all warehouse data

Then distributed to the various dependent data marts
16
Data warehouse construction
17
Extraction, Transformation, and Loading (ETL)





Data extraction
 get data from multiple, heterogeneous, and external
sources
Data cleaning
 detect errors in the data and rectify them when possible
Data transformation
 convert data from legacy or host format to warehouse
format
Load
 sort, summarize, consolidate, compute views, check
integrity, and build indicies and partitions
Refresh
 propagate the updates from the data sources to the
warehouse
18
Metadata Repository

Meta data is the data defining warehouse objects. It stores:

Description of the structure of the data warehouse


Operational meta-data


schema, view, dimensions, hierarchies, derived data defn, data
mart locations and contents
data lineage (history of migrated data and transformation path),
currency of data (active, archived, or purged), monitoring
information (warehouse usage statistics, error reports, audit trails)
The algorithms used for summarization

Measure and dimension definition, data on granularity, partitions,
subject areas, aggregation, summarization, predefined queries and
reports
19
Metadata Repository

The mapping from operational environment to the data warehouse



Source databases and their content, gateway descriptions, data
partitions, data extraction, cleaning, transformation rules and
defaults, data refresh and purging rules, security (user
authorization and access control)
Data related to system performance
 Indices and profiles that improve data access and retrieval and
performance, rules for timing and scheduling of refresh , update
and replication cycles
Business data

July 17, 2015
business terms and definitions, ownership of data, charging
policies
Data Mining: Concepts and Techniques
20
Chapter 4: Data Warehousing and On-line
Analytical Processing

Data Warehouse: Basic Concepts

Data Warehouse Modeling: Data Cube and OLAP

Data Warehouse Design and Usage

Data Warehouse Implementation

Data Generalization by Attribute-Oriented
Induction

Summary
21
From Tables and Spreadsheets to
Data Cubes

A data warehouse is based on a multidimensional data model
which views data in the form of a data cube

A data cube, such as sales, allows data to be modeled and viewed in
multiple dimensions

Dimension tables, such as item (item_name, brand, type), or
time(day, week, month, quarter, year)

Fact table contains measures (such as dollars_sold) and keys
to each of the related dimension tables
22
2-D and 3-D Data Cubes


2-D Data: Dimensions – time, item, Measure – dollars_sold, Location –
Vancouver
3-D Data: Dimensions – time, item, location, Measure – dollars_sold
23
3-D and 4-D Data cubes

3-D cube


Dimensions – time, item, location
4-D cube as a series of 3-D cubes

Dimensions – time, item, location, supplier
24
Data cube – multidimensional data storage





Actual physical storage may differ from its logical
representation
Data cubes – n-dimensional
In data warehousing literature, an n-D base cube is called
a base cuboid.
The top most 0-D cuboid, which holds the highest-level of
summarization, is called the apex cuboid.
The lattice of cuboids forms a data cube.
July 17, 2015
Data Mining: Concepts and Techniques
25
A Sample Data Cube
2Qtr
3Qtr
4Qtr
sum
U.S.A
Canada
Mexico
Country
TV
PC
VCR
sum
1Qtr
Date
Total annual sales
of TVs in U.S.A.
sum
26
Cube: A Lattice of Cuboids
all
time
0-D (apex) cuboid
item
time,location
time,item
location
supplier
item,location
time,supplier
1-D cuboids
location,supplier
2-D cuboids
item,supplier
time,location,supplier
3-D cuboids
time,item,location
time,item,supplier
item,location,supplier
4-D (base) cuboid
time, item, location, supplier
27
Cuboids Corresponding to the Cube
all
0-D (apex) cuboid
product
product,date
date
country
product,country
1-D cuboids
date, country
2-D cuboids
3-D (base) cuboid
product, date, country
28
Conceptual Modeling of Data Warehouses

Modeling data warehouses: dimensions & measures

Star schema: A fact table in the middle connected to a
set of dimension tables

Snowflake schema: A refinement of star schema
where some dimensional hierarchy is normalized into a
set of smaller dimension tables, forming a shape
similar to snowflake

Fact constellations: Multiple fact tables share
dimension tables, viewed as a collection of stars,
therefore called galaxy schema or fact constellation
29
Example of Star Schema
time
item
time_key
day
day_of_the_week
month
quarter
year
Sales Fact Table
time_key
item_key
branch_key
branch
location_key
branch_key
branch_name
branch_type
units_sold
dollars_sold
avg_sales
item_key
item_name
brand
type
supplier_type
location
location_key
street
city
state_or_province
country
Measures
30
Example of Snowflake Schema
time
time_key
day
day_of_the_week
month
quarter
year
item
Sales Fact Table
time_key
item_key
branch_key
branch
location_key
branch_key
branch_name
branch_type
units_sold
dollars_sold
avg_sales
Measures
item_key
item_name
brand
type
supplier_key
supplier
supplier_key
supplier_type
location
location_key
street
city_key
city
city_key
city
state_or_province
country
31
Example of Fact Constellation
time
time_key
day
day_of_the_week
month
quarter
year
item
Sales Fact Table
time_key
item_key
item_name
brand
type
supplier_type
item_key
location_key
branch_key
branch_name
branch_type
units_sold
dollars_sold
avg_sales
Measures
time_key
item_key
shipper_key
from_location
branch_key
branch
Shipping Fact Table
location
to_location
location_key
street
city
province_or_state
country
dollars_cost
units_shipped
shipper
shipper_key
shipper_name
location_key
shipper_type 32
Schemas for Data Warehouse


Data warehouses
 Fact constellation schema commonly used
 It can model multiple, interrelated subjects
Data Mart
 Department subset of the data warehouse
focusing on selected subjects
 Star or snowflake schema commonly used
 Geared towards modeling single subjects
July 17, 2015
Data Mining: Concepts and Techniques
33
Dimensions – Concept Hierarchies





Concept Hierarchies – sequence of mappings from a set
of low-level concepts to higher-level concepts
Many concept hierarchies are implicit within the database
systems
Attributes in a dimension may be related by a total order
(hierarchy) or partial order (lattice)
Total order: attributes of the dimension location
 Street < city < province_or_state < country
Partial order: attributes of dimension time
 Day < {month < quarter; week} < year
July 17, 2015
Data Mining: Concepts and Techniques
34
A Concept Hierarchy:
Dimension (location)
all
all
Europe
region
country
city
office
Germany
Frankfurt
...
...
...
Spain
North_America
Canada
Vancouver ...
L. Chan
...
...
Mexico
Toronto
M. Wind
35
Concept Hierarchy



(a) location as a hierarchy (total order)
(b) time as a lattice (partial order)
Schema hierarchy – a concept hierarchy that is a total or
a partial order among attributes in a database schema
36
Multidimensional Data
Sales volume as a function of product, month,
and region
Dimensions: Product, Location, Time
Hierarchical summarization paths
Industry Region
Year
Category Country Quarter
Product

Product
City
Office
Month Week
Day
Month
37
Concept Hierarchy


Set-grouping hierarchy – concept hierarchy defined by
discretizing or grouping values for a given dimension or
attribute
A total or partial order can be defined among groups of
values
38
View of Warehouses and Hierarchies
Specification of hierarchies

Schema hierarchy
day < {month <
quarter; week} < year

Set_grouping hierarchy
{1..10} < inexpensive
39
Data Cube Measures: Three Categories

Distributive: if the result derived by applying the function to
n aggregate values is the same as that derived by applying
the function on all the data (without partitioning), the
function can be computed in a distributed manner


Algebraic: if it can be computed by an algebraic function
with M arguments (where M is a bounded integer), each of
which is obtained by applying a distributive aggregate
function


E.g., count(), sum(), min(), max()
E.g., avg(), min_N(), max_N(), standard_deviation()
Holistic: if there is no constant bound on the storage size
needed to describe a subaggregate.

E.g., median(), mode(), rank()
40
Typical OLAP Operations




Roll up (drill-up): summarize data
 by climbing up hierarchy or by dimension reduction
Drill down (roll down): reverse of roll-up
 from higher level summary to lower level summary or
detailed data, or introducing new dimensions
Slice and dice: select and project
 Slice – select on one dimension resulting in a subcube
 Dice – select two or more dimensions to get subcube
Pivot (rotate):
 Rotate the axes in the cube for visualization, transform
3D cube to a series of 2D planes
41
Typical OLAP Operations


Other operations
 drill across: executes queries involving (across) more
than one fact table
 drill through: the bottom level of a data cube to its
back-end relational tables (using SQL)
 Ranking the top N or bottom N items in lists,
computing averages, growth rates, interests, internal
rates of return, depreciation, currency conversions and
statistical functions
OLAP offers analytical modeling capabilities
 Calculation engine for deriving ratios, variance,
computing measures across multiple dimensions
42
Fig. 3.10 Typical OLAP
Operations
43
OLAP Vs Statistical Databases

Statistical database is a database system that is designed
to support statistical applications
 SDBs focus on socieconomic applications
 Privacy issues regarding concept hierarchies


Summarized socioeconomic data – view corresponding lowlevel data is controversial
OLAP is targeted for business applications
 Designed to handle huge amounts of data efficiently
44
A Star-Net Query Model
Customer Orders
Shipping Method
Customer
CONTRACTS
AIR-EXPRESS
ORDER
TRUCK
PRODUCT LINE
Time
Product
ANNUALY QTRLY
DAILY
PRODUCT ITEM PRODUCT GROUP
CITY
SALES PERSON
COUNTRY
DISTRICT
REGION
Location
Each circle is
called a footprint
DIVISION
Promotion
Organization
45
Browsing a Data Cube



Visualization
OLAP capabilities
Interactive manipulation
46
Chapter 4: Data Warehousing and On-line
Analytical Processing

Data Warehouse: Basic Concepts

Data Warehouse Modeling: Data Cube and OLAP

Data Warehouse Design and Usage

Data Warehouse Implementation

Data Generalization by Attribute-Oriented
Induction

Summary
47
Design of Data Warehouse: A Business
Analysis Framework

Four views regarding the design of a data warehouse

Top-down view



allows selection of the relevant information necessary for the
data warehouse
This information matches the current and future business
needs
Data source view


exposes the information being captured, stored, and
managed by operational systems
Information documented at various level of details and
accuracy, from individual data source tables to integrated
data source tables
48
Design of Data Warehouse: A Business
Analysis Framework

Data warehouse view




consists of fact tables and dimension tables
Information stored inside data warehouse – precalculated
totals and counts,
information regarding source, data and time of origin for
historical context
Business query view

sees the perspectives of data in the warehouse from the view
of end-user
49
Data Warehouse Design Process


Top-down, bottom-up approaches or a combination of both

Top-down: Starts with overall design and planning (mature)

Bottom-up: Starts with experiments and prototypes (rapid)
From software engineering point of view



Waterfall: structured and systematic analysis at each step before
proceeding to the next
Spiral: rapid generation of increasingly functional systems, short
turn around time, quick turn around
Typical data warehouse design process

Choose a business process to model, e.g., orders, invoices, etc.

Choose the grain (atomic level of data) of the business process

Choose the dimensions that will apply to each fact table record

Choose the measure that will populate each fact table record
50
Data Warehouse Development:
A Recommended Approach
Multi-Tier Data
Warehouse
Distributed
Data Marts
Data
Mart
Data
Mart
Model refinement
Enterprise
Data
Warehouse
Model refinement
Define a high-level corporate data model
51
Data Warehouse Design Process



Goals of Data warehouse implementation – specific,
achievable, measurable
 Determine the time and budget allocations, subset of
organization to be modeled, number of data sources
selected, number and types of departments to be
served
Initial deployment
 Initial installation, roll-out planning, training, and
orientation, platform upgrades and maintenance
Data warehouse administration
 Data refreshment, data source synchronization,
planning for disaster recovery, managing access
control and security, managing data growth,
performance
52
Data Warehouse Usage

Three kinds of data warehouse applications

Information processing



supports querying, basic statistical analysis, and reporting
using crosstabs, tables, charts and graphs
Analytical processing

multidimensional analysis of data warehouse data

supports basic OLAP operations, slice-dice, drilling, pivoting
Data mining

knowledge discovery by finding hidden patterns , associations,
constructing analytical models, performing classification and
prediction, and presenting the mining results using
visualization tools
53
From Online Analytical Processing (OLAP)
to Online Analytical Mining (OLAM)

Why online analytical mining?
 High quality of data in data warehouses
 DW contains integrated, consistent, cleaned data
 DW constructed by such preprocessing serves as
valuable source of high-quality data for OLAP and
data mining
 Available information processing structure surrounding
data warehouses
 Information processing and data analysis
infrastructure surrounding DW


Accessing, integration, consolidation, transformation of
multiple heterogeneous databases
ODBC, OLEDB, Web accessing, service facilities, reporting
and OLAP analysis tools
54
From Online Analytical Processing (OLAP)
to Online Analytical Mining (OLAM)


OLAP-based exploration of multidimensional data
 Multidimensional data mining provides facilities for
mining on different subsets of data and at varying
levels of abstraction, by drilling, filtering, dicing,
pivoting, etc.
 Data/knowledge visualization tools enhance the
power of data mining
On-line selection of data mining functions
 Integration of OLAP with data mining functions
allows to select desired data mining functions and
swap data mining tasks dynamically
55
Chapter 4: Data Warehousing and On-line
Analytical Processing

Data Warehouse: Basic Concepts

Data Warehouse Modeling: Data Cube and OLAP

Data Warehouse Design and Usage

Data Warehouse Implementation

Data Generalization by Attribute-Oriented
Induction

Summary
56
The “Compute Cube” Operator

Cube definition and computation in DMQL
define cube sales [item, city, year]: sum (sales_in_dollars)
compute cube sales

Transform it into a SQL-like language (with a new operator cube by,
introduced by Gray et al.’96)
()
SELECT item, city, year, SUM (amount)
FROM SALES

CUBE BY item, city, year
Need compute the following Group-Bys
(city)
(city, item)
(city, item, year),
(city,item),(city, year), (item,year),
(city), (item), (year)
()
(item)
(city, year)
(year)
(item, year)
(city, item, year)
57
The “Compute Cube” Operator



The compute cube operator computes aggregates over all
subsets of the dimensions specified in the operation
A cube operator on n dimensions is equivalent to a
collection of group by statements, one for each subset of
the n dimensions
Precomputation – compute all or some cuboids in
advance
 Leads to fast response time, avoids some redundant
computation
 Storage space may explode if all of the cuboids in a
data cube are precomputed
 Curse of dimensionality – storage requirements
excessive when many dimensions have associated
concept hierarchies, each with multiple levels
58
Efficient Data Cube Computation

Data cube can be viewed as a lattice of cuboids

The bottom-most cuboid is the base cuboid

The top-most cuboid (apex) contains only one cell

How many cuboids in an n-dimensional cube with Li
levels for dimension i?
n
T   (Li 1)
i 1



If cube has 10 dimensions, each dimension has 5
levels (including all), total cuboids is 510  9.8x106
Size of each cuboid also depends on the cardinality
(i.e. #of distinct values) of each dimension
As #of dimensions, #of concept hierarchies, cardinality
increases, storage space will exceed size of input
59
Efficient Data Cube Computation

Materialization of data cube

No materialization: do not precompute any of the
“nonbase” cuboids


Full materialization: precompute all of the cuboids


Leads to computing expensive multidimensional aggregates on
the fly, which is extremely slow
Requires huge amounts of memory space
Partial materialization: selectively compute a
proper subset of the whole set of possible cuboids

Subcube – only some of the cells may be precomputed for
various cuboids
60
Partial materialization of cuboids


Partial materialization of cuboids or subcubes should
consider three factors:
 Identify the subset of cuboids or subcubes to
materialize
 Exploit the materialized cuboids or subcubes during
query processing
 Efficiently update the materialized cuboids or subcubes
during load and refresh
The selection of the subset of cuboids or subcubes should
take into account
 Queries in the workload, their frequencies, their
accessing costs, workload characteristics, cost of
incremental updates, total storage requirements
61
Partial materialization of cuboids



A popular approach is to materialize the set of cuboids on
which other frequently referenced cuboids are based
Iceberg cube – data cube that stores only those cube
cells whose aggregate value (e.g. count) above some
minimum support threshold
Shell cube – precomputing the cuboids for only a small
number of dimensions (3 to 5) of a data cube
 Queries on additional combinations of the dimensions
can be computed on-the-fly
62
Indexing OLAP Data: Bitmap Index






Index on a particular column
Each value in the column has a bit vector: bit-op is fast
The length of the bit vector: # of records in the base table
The i-th bit is set if the i-th row of the base table has the value for
the indexed column
not suitable for high cardinality domains
A recent bit compression technique, Word-Aligned Hybrid (WAH),
makes it work for high cardinality domain as well [Wu, et al. TODS’06]
Base table
Cust
C1
C2
C3
C4
C5
Region
Asia
Europe
Asia
America
Europe
Index on Region
Index on Type
Type RecIDAsia Europe America RecID Retail Dealer
Retail
1
1
0
1
1
0
0
Dealer 2
2
0
1
0
1
0
Dealer 3
1
0
0
3
0
1
4
0
0
1
4
1
0
Retail
0
1
0
5
0
1
Dealer 5
63
Compression of Bitmap Indices

Bitmap indexes must be compressed to reduce I/O costs
and minimize CPU usage—majority of the bits are 0’s


Two compression schemes:

Byte-aligned Bitmap Code (BBC)

Word-Aligned Hybrid (WAH) code
Time and space required to operate on compressed
bitmap is proportional to the total size of the bitmap

Optimal on attributes of low cardinality as well as those of
high cardinality.

WAH out performs BBC by about a factor of two
64
Indexing OLAP Data: Join Indices



Join index: JI(R-id, S-id) where R (R-id, …)  S (S-id, …)
Traditional indices map the values to a list of record ids
 It materializes relational join in JI file and speeds up
relational join
In data warehouses, join index relates the values of the
dimensions of a star schema to rows in the fact table.
 E.g. fact table: Sales and two dimensions location and
item
A join index on location maintains for each distinct
location a list of R-IDs of the tuples recording the
Sales in the location
Join indices can span multiple dimensions to form
composite join indices


65
Indexing OLAP Data: Join Indices
Linkages between a sales fact
table and dimension tables for
location and item
Join index tables based on
linkages between the sales fact
table and dimension tables for
location and item
66
Efficient Processing OLAP Queries

Determine which operations should be performed on the available cuboids

Transform any selection, projection, drill-down, roll-up (group-by), etc. into
corresponding SQL and/or OLAP operations, e.g., slice, dice = selection
and/or projection operations on a materialized cuboid

Determine which materialized cuboid(s) should be selected for OLAP op.

This involves identifying all of the materialized cuboids that may potentially
be used to answer the query

Pruning the above set using knowledge of “dominance” relationships among
the cuboids

Estimating the costs of using the remaining materialized cuboids

Selecting the cuboid with the least cost
67
Efficient Processing OLAP Queries

Example

Let the query to be processed be on {brand, province_or_state} with the
condition “year = 2010”, and there are 4 materialized cuboids available:
1) {year, item_name, city}
2) {year, brand, country}
3) {year, brand, province_or_state}
4) {item_name, province_or_state} where year = 2010
Which should be selected to process the query?

Finer-granularity data cannot be generated from coarser-granularity data

So, cuboid 2 cannot be used as country is more general concept than
province_or_state
68
Efficient Processing OLAP Queries

Cuboids 1, 3 and 4 can be used to process the query as







they have the same set or superset of the dimensions in the query
the selection clause in the query can imply the selection in the cuboid
The abstraction levels for the item and location dimensions in these
cuboids are at a finer level than brand and province_or_state,
respectively
Cuboid 1 would cost most – item_name and city at lower level than
brand and province_or_state
If few year values associated with items in the cube, but there are
several item_names for each brand, then cuboid 3 is smaller than
cuboid 4
If efficient indices available for cuboid 4 then cuboid 4 is a better
choice
Explore indexing structures and compressed vs. dense array structs in
MOLAP
69
OLAP Server Architectures

Relational OLAP (ROLAP)




Intermediate servers between a relational back-end server and
client front-end tools
Use relational or extended-relational DBMS to store and manage
warehouse data and OLAP middleware
Include optimization for each DBMS backend, implementation of
aggregation navigation logic, and additional tools and services

Greater scalability than MOLAP technology

E.g. DSS server of Microstrategy
Multidimensional OLAP (MOLAP)

Support multidimensional views of data through array-based
multidimensional storage engine

Map multidimensional views directly to data cube array structures

Fast indexing to pre-computed summarized data
70
OLAP Server Architectures


with multidimensional data stores, the storage utilization may be
low if the data set is sparse
Two-level storage representation:



Sparse subcubes employ compression technology for efficient storage
utilization
Hybrid OLAP (HOLAP) (e.g., Microsoft SQLServer)

Combine ROLAP and MOLAP

Greater scalability of ROLAP and faster computation of MOLAP


Denser subcubes are identified and stored as array structures
Large volumes of detail data stored in relational database, while
aggregations kept in MOLAP store
Specialized SQL servers (e.g., Redbricks)

Specialized support for SQL queries over star/snowflake schemas
in a read-only environment
71
Storage of ROLAP





ROLAP uses relational tables to store data for OLTP
Base fact table – fact table associated with base cuboid
Base fact table stores data at the abstraction level indicated
by the join keys in the schema for given data cube
Summary fact tables - Aggregated data stored in fact tables
Some summary fact tables store both base fact table data
and aggregated data
Summary fact table :Single table for base and summary facts
72
Chapter 4: Data Warehousing and On-line
Analytical Processing

Data Warehouse: Basic Concepts

Data Warehouse Modeling: Data Cube and OLAP

Data Warehouse Design and Usage

Data Warehouse Implementation

Data Generalization by Attribute-Oriented
Induction

Summary
73
Data Generalization

Data generalization summarizes data
by replacing relatively low-level values (numeric values for age)
with higher-level concepts (young, middle, senior)
 or by reducing the number of dimensions (removing birth-date
and telephone numbers for group of students)
Ex: individual customer transactions  prefer to view data
summarized by customer groups according to geographic regions,
frequency of purchases per group, customer income
Concept description – form of data generalization, refers to a
collection of data e.g. frequent_buyers, graduate_students
 generates descriptions for the characterization and comparison of
the data
 Characterization provides a concise and succinct summarization of
the given collection of data
 Concept or class comparison (discrimination) provides descriptions
comparing two or more collections of data



74
Concept Description for Large Data Sets


Complex data types and aggregation
 Many OLAP systems confine dimensions (attributes) to nonnumeric data and measures (aggregate functions) to numeric
data
 Existing data types: numeric, non-numeric, spatial, text, image
 Aggregation of attributes in a database – sophisticated data
types, e.g. collection of non-numeric data, merging of spatial
regions, composition of images, integration of texts, grouping of
object pointers
User-control Vs. automation
 OLAP operations (drill-down, roll-up, slice, dice) are user
controlled
 Desirable to have automated process to help users determine
which dimensions should be included in the analysis, and the
degree of generalization to obtain interesting summarization
75
Attribute-Oriented Induction (AOI) for
Data Characterization



AOI approach to concept description Proposed in 1989 (KDD
‘89 workshop)
Data cube based materialized views of the data, which is
precomputed in a data warehouse
AOI – query-oriented, generalization-based, online data
analysis technique




Collect the task-relevant data (initial relation) using a
relational database query
Perform generalization by attribute removal or attribute
generalization
Apply aggregation by merging identical, generalized
tuples and accumulating their respective counts
Different forms of presentation – charts, rules, etc.
76
Attribute-Oriented Induction: An Example
Example: Describe general characteristics of graduate
students in the Big University database
Data Mining query in data mining query language DMQL:
use Big_University_DB
mine characteristics as “Science_students”
in relevance to name, gender, major, birth_place,
birth_date, residence, phone#, gpa
from student
where status in “graduate”

Step 1: Data focusing should be performed before
attribute-oriented induction, i.e. specification of the taskrelevant data
77
Attribute-Oriented Induction: An Example




Specifying the set of relevant attributes may be difficult for the user
User selects few attributes
 Ex: dimension birth_place defined by attributes city,
province_or_state, country
 User specifies only city
 To allow generalization on birth_place dimension, other attributes
should also be included
 i.e. system automatically includes province_or_state, country; helps
generalize city to higher conceptual level
User selects too many attributes
 E.g. in relevance to * (all attributes specified by from included)
 Relevance analysis: correlation-based, entropy-based methods,
attribute subset selection
where status in “graduate”: concept hierarchy exists for status
 Organize primitive level data into higher conceptual levels
 M.Sc., M.A., M.B.A., Ph.D., B.Sc., B.A. – graduate and undergraduate 78
Attribute-Oriented Induction: An Example

Step 1. Data mining query transformed into relational query
for collection of task-relevant set of data
use Big_University_DB
Select name, gender, major, birth_place, birth_date,
residence, phone#, gpa
from student
where status in {“M.Sc.”, “M.A.”, “M.B.A.”, “Ph.D.” }

Each tuple in figure (next slide) is a conjunction of
attribute-value pairs

Step 2. Perform attribute-oriented induction on initial working
relation by (1) attribute removal (2) attribute generalization

Step 3. Present results in generalized relation, cross-tab, or
rule forms
79
Class Characterization: An Example
Name
Initial
Working
Relation
(taskrelevant)
Gender
Jim
Woodman
Scott
Lachance
Laura Lee
…
M
F
…
Removed
Retained
Major
Residence
Phone #
GPA
M
M
F
…
Birth_date
Vancouver,BC, 8-12-76
Canada
CS
Montreal, Que, 28-7-75
Canada
Physics Seattle, WA, USA 25-8-70
…
…
…
3511 Main St.,
Richmond
345 1st Ave.,
Richmond
687-4598
3.67
253-9106
3.70
125 Austin Ave.,
Burnaby
…
420-5232
…
3.83
…
Sci,Eng,
Bus
City
Removed
Excl,
VG,..
CS
Gender Major
Prime
Generalized
Relation
Birth-Place
Science
Science
…
Country
Age range
Birth_region
Age_range
Residence
GPA
Canada
Foreign
…
20-25
25-30
…
Richmond
Burnaby
…
Very-good
Excellent
…
Count
16
22
…
Birth_Region
Canada
Foreign
Total
Gender
M
16
14
30
F
10
22
32
Total
26
36
62
80
Basic Principles of Attribute-Oriented Induction




Data focusing: task-relevant data, including dimensions,
and the result is the initial relation
Attribute-removal: remove attribute A if there is a large set
of distinct values for A but (1) there is no generalization
operator on A, or (2) A’s higher level concepts are
expressed in terms of other attributes
Attribute-generalization: If there is a large set of distinct
values for A, and there exists a set of generalization
operators on A, then select an operator and generalize A
How large is a “large set of distinct values for an attribute”

Attribute generalization threshold control

Generalized relation threshold control
81
Attribute-Oriented Induction –
generalization control

Attribute generalization threshold control:

(1) One generalization threshold for all attributes

(2) One threshold for each attribute



If #of distinct values in an attribute > attribute threshold
=> perform further attribute removal or attribute
generalization
Attribute threshold typically 2-8, specified/default
Generalized relation threshold control:



Sets a threshold for the generalized relation
If #of distinct tuples in the generalized relation >
threshold => perform further generalization
Either preset within a range 10-30, or set by an expert
82
Attribute-Oriented Induction –
quantitative/statistical information







Important to accumulate count and other aggregate values
in the induction process
Aggregate function count is associated with each database
tuple
Initialize count = 1 for each tuple in initial working relation
Through attribute removal and attribute generalization,
tuples may be generalized => groups of identical tuples
Identical tuples are merged to form one tuple
Count of new tuple = #of tuples merged
Other popular aggregate functions to associate with each
tuple – sum and avg
83
Attribute-Oriented Induction - Example





Example: Big_University_DB
name: large #of distinct values and no generalization
operation defined on it => remove this attribute
gender: only 2 distinct values => retained, no generalization
major: concept hierarchy defined as {arts&science,
engineering, business}.
 Attribute generalization threshold = 5
 #of distinct values > 20
 major is generalized by climbing the concept hierarchy
birth_place: large #of distinct value => generalize
 Concept hierarchy defined as city < province_or_state <
country
84
Attribute-Oriented Induction - Example

If #of distinct values for country > threshold =>
remove birth_place

Because even though a generalization operator exists for it,
generalization threshold would not be satisfied
If #of distinct values for country < threshold =>
generalize to birth_country
birth_date: suppose a hierarchy exists that can generalize
birth_date to age and age to age_range
 Let #of age ranges is small with respect to the
attribute generalization threshold
 Generalization should take place
residence: defined by attributes number, street,
residence_city, residence_province_or_state,
residence_country



85
Attribute-Oriented Induction - Example
#of distinct values for number and street will be high
 Attributes number and street should be removed
 Generalize to residence_city which contains fewer
distinct values
phone#: too many distinct values, so remove
gpa: suppose concept hierarchy exists that groups values
into numeric intervals {3.75-4.00, 3.5-3.75,…}, which are
grouped into descriptive values {excellent, good,…}
 The attribute can therefore be generalized



86
87
Attribute-Oriented Induction: Basic
Algorithm




InitialRel: Query processing of task-relevant data, deriving
the initial relation.
PreGen: Based on the analysis of the number of distinct
values in each attribute, determine generalization plan for
each attribute: removal? or how high to generalize?
PrimeGen: Based on the PreGen plan, perform
generalization to the right level to derive a “prime
generalized relation”, accumulating the counts.
Presentation: User interaction: (1) adjust levels by drilling,
(2) pivoting, (3) mapping into rules, cross tabs,
visualization presentations.
88
Presentation of Generalized Results

Generalized relation:


Cross tabulation:


Relations where some or all attributes are generalized, with counts
or other aggregation values accumulated.
Mapping results into cross tabulation form (similar to contingency
tables).

Visualization techniques:

Pie charts, bar charts, curves, cubes, and other visual forms.
Quantitative characteristic rules:
Mapping generalized result into characteristic rules with quantitative
information associated with it, e.g.,
grad ( x)  male( x) 
birth_ region( x) "Canada"[t :53%] birth_ region( x) " foreign"[t : 47%].

89
Mining Class Comparisons

Comparison: Comparing two or more classes (or concepts)

Target and contrasting classes must be comparable in the sense that
they share similar dimensions and attributes


E.g. three classes person, address, item are not comparable
whereas sales in the last three years are comparable classes
Method:

Data collection: collect relevant data by query processing and
partition the set into the target class and the contrasting classes

Dimension relevance analysis: if many dimensions, select only the
highly relevant dimensions for further analysis. Correlation and
entropy-based measures can be used for this step

Synchronous generalization: Generalization is performed on the
target class to the level controlled by a user/expert specified
dimension threshold, resulting in prime target class relation

Generalize contrasting class(es) to the same high level concepts
90
Mining Class Comparisons


Compare tuples with the same high level descriptions
Presentation of the derived comparisons: The resulting class
comparison description can be visualized in the form of tables,
graphs, and rules

Presentation includes “contrasting” measure such as count%

The user can adjust the comparison description by applying drilldown, roll-up, and other OLAP operations to the target and
contrasting classes
91
Mining Class Comparisons

Example: compare general properties between graduate
students and undergraduate students at Big_University_DB
use Big_University_DB
mine comparison as “grad_vs_undergrad_students”
in relevance to name, gender, major, birth_place, birth_date, residence,
phone#, gpa
for “graduate_student”
where status in “graduate”
versus “undergraduate_students”
where status in “undergraduate”
analyze count%
from student
92
Mining Class Comparisons

First, the query is transformed into two relational queries
that collect two sets of task-relevant data – one for the
initial target class working relation and the other for the
initial contrasting class working relation.
Initial working relations: the target class (graduate students)
Initial working relations: the contrasting class (undergraduate students)
93
Mining Class Comparisons


Second, dimension relevance analysis can be performed,
when necessary, on the two classes of data.
 After this analysis, irrelevant or weakly relevant
dimensions, such as name, gender, birth_place,
residence, and phone# are removed from the resulting
classes.
Third, synchronous generalization is performed:
Generalization is performed on the target class to the
levels controlled by user/expert-specified dimension
thresholds, forming the prime target class relation.
 The contrasting class is generalized to the same levels
as those in the prime target class relation, forming the
prime contrasting class(es) relation.
94
Mining Class Comparisons


Finally, the resulting class comparison is presented in the
form of tables, graphs, and/or rules
Contrasting measures such as count% compare between
target class and contrasting class
Prime generalized
relation for the
target class
(graduate students)
Prime generalized
relation for the
contrasting class
(undergraduate
students)
95
Concept Description vs. Cube-Based OLAP
Similarity:
 Data generalization
 Presentation of data summarization at multiple levels of
abstraction
 Interactive drilling, pivoting, slicing and dicing
 Differences:
 OLAP has systematic preprocessing, query independent,
and can drill down to rather low level
 AOI has automated desired level allocation, and may
perform dimension relevance analysis/ranking when
there are many relevant dimensions
 AOI works on the data which are not in relational forms

96
Chapter 4: Data Warehousing and On-line
Analytical Processing

Data Warehouse: Basic Concepts

Data Warehouse Modeling: Data Cube and OLAP

Data Warehouse Design and Usage

Data Warehouse Implementation

Data Generalization by Attribute-Oriented
Induction

Summary
97
Summary

Data warehousing: A multi-dimensional model of a data warehouse




A data cube consists of dimensions & measures
Star schema, snowflake schema, fact constellations
OLAP operations: drilling, rolling, slicing, dicing and pivoting
Data Warehouse Architecture, Design, and Usage

Multi-tiered architecture

Business analysis design framework
Information processing, analytical processing, data mining, OLAM (Online
Analytical Mining)
Implementation: Efficient computation of data cubes

Partial vs. full vs. no materialization

Indexing OALP data: Bitmap index and join index

OLAP query processing

OLAP servers: ROLAP, MOLAP, HOLAP



Data generalization: Attribute-oriented induction
98
References (I)









S. Agarwal, R. Agrawal, P. M. Deshpande, A. Gupta, J. F. Naughton, R. Ramakrishnan,
and S. Sarawagi. On the computation of multidimensional aggregates. VLDB’96
D. Agrawal, A. E. Abbadi, A. Singh, and T. Yurek. Efficient view maintenance in data
warehouses. SIGMOD’97
R. Agrawal, A. Gupta, and S. Sarawagi. Modeling multidimensional databases. ICDE’97
S. Chaudhuri and U. Dayal. An overview of data warehousing and OLAP technology.
ACM SIGMOD Record, 26:65-74, 1997
E. F. Codd, S. B. Codd, and C. T. Salley. Beyond decision support. Computer World, 27,
July 1993.
J. Gray, et al. Data cube: A relational aggregation operator generalizing group-by,
cross-tab and sub-totals. Data Mining and Knowledge Discovery, 1:29-54, 1997.
A. Gupta and I. S. Mumick. Materialized Views: Techniques, Implementations, and
Applications. MIT Press, 1999.
J. Han. Towards on-line analytical mining in large databases. ACM SIGMOD Record,
27:97-107, 1998.
V. Harinarayan, A. Rajaraman, and J. D. Ullman. Implementing data cubes efficiently.
SIGMOD’96
99
References (II)










C. Imhoff, N. Galemmo, and J. G. Geiger. Mastering Data Warehouse Design:
Relational and Dimensional Techniques. John Wiley, 2003
W. H. Inmon. Building the Data Warehouse. John Wiley, 1996
R. Kimball and M. Ross. The Data Warehouse Toolkit: The Complete Guide to
Dimensional Modeling. 2ed. John Wiley, 2002
P. O'Neil and D. Quass. Improved query performance with variant indexes.
SIGMOD'97
Microsoft. OLEDB for OLAP programmer's reference version 1.0. In
http://www.microsoft.com/data/oledb/olap, 1998
A. Shoshani. OLAP and statistical databases: Similarities and differences.
PODS’00.
S. Sarawagi and M. Stonebraker. Efficient organization of large
multidimensional arrays. ICDE'94
P. Valduriez. Join indices. ACM Trans. Database Systems, 12:218-246, 1987.
J. Widom. Research problems in data warehousing. CIKM’95.
K. Wu, E. Otoo, and A. Shoshani, Optimal Bitmap Indices with Efficient
Compression, ACM Trans. on Database Systems (TODS), 31(1), 2006, pp. 1-38.
100
July 17, 2015
Data Mining: Concepts and Techniques
101
Chapter 4: Data Warehousing and On-line
Analytical Processing






Data Warehouse: Basic Concepts

(a) What Is a Data Warehouse?

(b) Data Warehouse: A Multi-Tiered Architecture

(c) Three Data Warehouse Models: Enterprise Warehouse, Data Mart, ad Virtual Warehouse

(d) Extraction, Transformation and Loading

(e) Metadata Repository
Data Warehouse Modeling: Data Cube and OLAP

(a) Cube: A Lattice of Cuboids

(b) Conceptual Modeling of Data Warehouses

(c) Stars, Snowflakes, and Fact Constellations: Schemas for Multidimensional Databases

(d) Dimensions: The Role of Concept Hierarchy

(e) Measures: Their Categorization and Computation

(f) Cube Definitions in Database systems

(g) Typical OLAP Operations

(h) A Starnet Query Model for Querying Multidimensional Databases
Data Warehouse Design and Usage

(a) Design of Data Warehouses: A Business Analysis Framework

(b) Data Warehouses Design Processes

(c) Data Warehouse Usage

(d) From On-Line Analytical Processing to On-Line Analytical Mining
Data Warehouse Implementation

(a) Efficient Data Cube Computation: Cube Operation, Materialization of Data Cubes, and Iceberg Cubes

(b) Indexing OLAP Data: Bitmap Index and Join Index

(c) Efficient Processing of OLAP Queries

(d) OLAP Server Architectures: ROLAP vs. MOLAP vs. HOLAP
Data Generalization by Attribute-Oriented Induction

(a) Attribute-Oriented Induction for Data Characterization

(b) Efficient Implementation of Attribute-Oriented Induction

(c) Attribute-Oriented Induction for Class Comparisons

(d) Attribute-Oriented Induction vs. Cube-Based OLAP
Summary
102