Professional Documents
Culture Documents
Data Mning by Jaiwei Han Chapter 2
Data Mning by Jaiwei Han Chapter 2
Jiawei Han and Micheline Kamber Department of Computer Science University of Illinois at Urbana-Champaign www.cs.uiuc.edu/~hanj
December 19, 2012 Data Mining: Concepts and Techniques 1
What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation Further development of data cube technology From data warehousing to data mining
Data Mining: Concepts and Techniques 2
A decision support database that is maintained separately from the organizations 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 managements decision-making process.W. H. Inmon Data warehousing:
Data WarehouseSubject-Oriented
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.
Data WarehouseIntegrated
Constructed by integrating multiple, heterogeneous data sources relational databases, flat files, on-line transaction records Data cleaning and data integration techniques are applied. Ensure consistency in naming conventions, encoding structures, attribute measures, etc. among different data sources
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) Contains an element of time, explicitly or implicitly
But the key of operational data may or may not contain time element.
Data WarehouseNon-Volatile
operational environment.
Operational update of data does not occur in the data warehouse environment.
When a query is posed to a client site, a meta-dictionary is used to translate the query into queries appropriate for individual heterogeneous sites involved, and the results are integrated into a global answer set Complex information filtering, compete for resources
Information from heterogeneous sources is integrated in advance and stored in warehouses for direct query and analysis
Major task of traditional relational DBMS Day-to-day operations: purchasing, inventory, banking, manufacturing, payroll, registration, accounting, etc. Major task of data warehouse system Data analysis and decision making User and system orientation: customer vs. market
10
DBMS tuned for OLTP: access methods, indexing, concurrency control, recovery
Warehousetuned for OLAP: complex OLAP queries, multidimensional view, consolidation. 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
Data Mining: Concepts and Techniques 11
What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation Further development of data cube technology From data warehousing to data mining
12
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
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.
Data Mining: Concepts and Techniques 13
0-D(apex) cuboid
1-D cuboids
time,item
time,location
item,location item,supplier
location,supplier
time,supplier time,item,location
2-D cuboids
time,location,supplier
3-D cuboids
item,location,supplier
time,item,supplier
4-D(base) cuboid
time, item, location, supplier
December 19, 2012 Data Mining: Concepts and Techniques 14
15
item
Sales Fact Table time_key item_key branch_key
item_key item_name brand type supplier_type
branch
branch_key branch_name branch_type
location
location_key street city state_or_province country
location_key units_sold
dollars_sold
avg_sales
Measures
December 19, 2012 Data Mining: Concepts and Techniques 16
item
Sales Fact Table
item_key item_name brand type supplier_key
supplier
supplier_key supplier_type
time_key
item_key branch_key
branch
branch_key branch_name branch_type
location
location_key street city_key
location_key
units_sold dollars_sold avg_sales Measures
city
city_key city state_or_province country
17
item
Sales Fact Table time_key item_key branch_key
item_key item_name brand type supplier_type
item_key
shipper_key
from_location
location
location_key street city province_or_state country
branch
branch_key branch_name branch_type
Cube Definition (Fact Table) define cube <cube_name> [<dimension_list>]: <measure_list> Dimension Definition ( Dimension Table ) define dimension <dimension_name> as (<attribute_or_subdimension_list>) Special Case (Shared Dimension Tables) First time as cube definition define dimension <dimension_name> as <dimension_name_first_time> in cube <cube_name_first_time>
Data Mining: Concepts and Techniques 19
21
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.
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.
holistic: if there is no constant bound on the storage size needed to describe a subaggregate.
country
Germany
...
Spain
Canada
...
Mexico
city office
December 19, 2012
Frankfurt
...
Toronto
M. Wind
24
Specification of hierarchies
Schema hierarchy day < {month < quarter; week} < year
25
Multidimensional Data
Product
Product
City Office
Month
December 19, 2012 Data Mining: Concepts and Techniques 26
Date
3Qtr 4Qtr
Canada Mexico
sum
Country
TV PC VCR sum
sum
U.S.A
27
date
product,country
country
date, country
1-D cuboids
product,date
28
December 19, 2012
by climbing up hierarchy or by dimension reduction from higher level summary to lower level summary or detailed data, or introducing new dimensions project and select reorient the cube, visualization, 3D to series of 2D planes. drill across: involving (across) more than one fact table drill through: through the bottom level of the cube to its back-end relational tables (using SQL)
Data Mining: Concepts and Techniques 30
Pivot (rotate):
Other operations
CONTRACTS
AIR-EXPRESS TRUCK ORDER PRODUCT LINE
Time
ANNUALY QTRLY CITY SALES PERSON COUNTRY DAILY
Product
PRODUCT ITEM PRODUCT GROUP
DISTRICT
REGION Location
December 19, 2012
What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation Further development of data cube technology From data warehousing to data mining
32
Top-down view
allows selection of the relevant information necessary for the data warehouse exposes the information being captured, stored, and managed by operational systems consists of fact tables and dimension tables sees the perspectives of data in the warehouse from the view of end-user
Data Mining: Concepts and Techniques 33
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
Data Mining: Concepts and Techniques 34
Multi-Tiered Architecture
Metadata
other
sources
Operational Extract Transform Load Refresh
OLAP Server
DBs
Data Warehouse
Serve
Data Marts
Data Sources
December 19, 2012
Data Storage
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
Virtual warehouse A set of views over operational databases Only some of the possible summary views may be materialized
Data Mining: Concepts and Techniques 36
Data Mart
Data Mart
Model refinement
Model refinement
Relational OLAP (ROLAP) Use relational or extended-relational DBMS to store and manage warehouse data and OLAP middle ware to support missing pieces Include optimization of DBMS backend, implementation of aggregation navigation logic, and additional tools and services greater scalability Multidimensional OLAP (MOLAP) Array-based multidimensional storage engine (sparse matrix techniques) fast indexing to pre-computed summarized data Hybrid OLAP (HOLAP) User flexibility, e.g., low level: relational, high-level: array Specialized SQL servers specialized support for SQL queries over star/snowflake schemas
Data Mining: Concepts and Techniques 38
What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation Further development of data cube technology From data warehousing to data mining
39
Materialize every (cuboid) (full materialization), none (no materialization), or some (partial materialization) Selection of which cuboids to materialize
Cube Operation
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) ()
(city)
(item)
(year)
(city, item) (city, year) (item, year) (date, product, customer), (date,product),(date, customer), (product, customer), (date), (product), (customer) (city, item, year) ()
December 19, 2012 Data Mining: Concepts and Techniques 41
ROLAP-based cubing algorithms (Agarwal et al96) Array-based cubing algorithm (Zhao et al97) Bottom-up computation method (Beyer & Ramarkrishnan99) H-cubing technique (Han, Pei, Dong & Wang:SIGMOD01) Sorting, hashing, and grouping operations are applied to the dimension attributes in order to reorder and cluster related tuples Grouping is performed on some sub-aggregates as a partial grouping step
Aggregates may be computed from previously computed aggregates, rather than from the base fact table
Data Mining: Concepts and Techniques 42
This is not in the textbook but in a research paper Hash/sort based methods (Agarwal et. al. VLDB96) Smallest-parent: computing a cuboid from the smallest, previously computed cuboid Cache-results: caching results of a cuboid from which other cuboids are computed to reduce disk I/Os Amortize-scans: computing as many as possible cuboids at the same time to amortize disk reads Share-sorts: sharing sorting costs cross multiple cuboids when sort-based method is used Share-partitions: sharing the partitioning cost across multiple cuboids when hash-based algorithms are used
Data Mining: Concepts and Techniques 43
Partition arrays into chunks (a small subcube which fits in memory). Compressed sparse array addressing: (chunk_id, offset) Compute aggregates in multiway by visiting cube cells in the order which minimizes the # of times to visit each cell, and reduces memory access and storage cost.
c3 61 62 63 64 c2 45 46 47 48 c1 29 30 31 32 c0
b3
B 13
9 5 1 a0
14
15
16
b2 b1 b0
2 a1
3 a2
4 a3
60 44 28 56 40 24 52 36 20
44
c3 61 62 63 64 c2 45 46 47 48 c1 29 30 31 32 c0 B 13 14 15 16 28 24 2 3 4 20 40 36 52 60 44 56
b3
b2
9
5 1
b1
b0
a0
a1
a2
a3
45
c3 61 62 63 64 c2 45 46 47 48 c1 29 30 31 32 c0 B 13 14 15 16 28 24 2 3 4 20 40 36 52 60 44 56
b3
b2
9
5 1
b1
b0
a0
a1
a2
a3
46
Method: the planes should be sorted and computed according to their size in ascending order. See the details of Example 2.12 (pp. 75-78) Idea: keep the smallest plane in the main memory, fetch and compute only one chunk at a time for the largest plane Limitation of the method: computing well only for a small number of dimensions If there are a large number of dimensions, bottomup computation and iceberg cube computation methods can be explored
Data Mining: Concepts and Techniques 47
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
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 Retail 4 0 0 1 4 1 0 0 1 0 5 0 1 Dealer 5
Data Mining: Concepts and Techniques 48
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 a rather costly operation In data warehouses, join index relates the values of the dimensions of a start schema to rows in the fact table. E.g. fact table: Sales and two dimensions city and product A join index on city maintains for each distinct city a list of R-IDs of the tuples recording the Sales in the city Join indices can span multiple dimensions
Data Mining: Concepts and Techniques 49
transform drill, roll, etc. into corresponding SQL and/or OLAP operations, e.g, dice = selection + projection
50
Metadata Repository
Meta data is the data defining warehouse objects. It has the following kinds Description of the structure of the warehouse
schema, view, dimensions, hierarchies, derived data defn, data mart locations and contents
Operational meta-data
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 The mapping from operational environment to the data warehouse Data related to system performance
Business data
Data extraction:
Data cleaning:
get data from multiple, heterogeneous, and external sources detect errors in the data and rectify them when possible
Data transformation:
Load:
Refresh
What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation
Iceberg Cube
Computing only the cuboid cells whose count or other aggregates satisfying the condition: HAVING COUNT(*) >= minsup Motivation
Only a small portion of cube cells may be above the water in a sparse cube
Only calculate interesting datadata above certain threshold Suppose 100 dimensions, only 1 base cell. How many aggregate (non-base) cells if count >= 1? What about count >= 2?
Data Mining: Concepts and Techniques 54
BUC (Beyer & Ramakrishnan, SIGMOD99) Bottom-up vs. top-down? depending on how you view it! Apriori property: Aggregate the data, then move to the next level If minsup is not met, stop! If minsup = 1 compute full CUBE!
55
Partitioning
Usually, entire data set cant fit in main memory Sort distinct values, partition into blocks that fit Continue processing Optimizations Partitioning External Sorting, Hashing, Counting Sort Ordering dimensions to encourage pruning Cardinality, Skew, Correlation Collapsing duplicates Cant do holistic aggregates anymore!
Data Mining: Concepts and Techniques 56
Drawbacks of BUC
Does not obtain good performance with dense CUBEs Overly skewed data or a bad choice of dimension ordering reduces performance Cannot compute iceberg cubes with complex measures
CREATE CUBE Sales_Iceberg AS SELECT month, city, cust_grp,
AVG(price), COUNT(*)
FROM Sales_Infor CUBEBY month, city, cust_grp HAVING AVG(price) >= 800 AND COUNT(*) >= 50
December 19, 2012 Data Mining: Concepts and Techniques 57
Non-Anti-Monotonic Measures
(Mar, *, *, 600, 1800) fails the HAVING clause (Mar, *, Bus, 1300, 360) passes the clause
City Tor Tor Tor Mon Van Cust_grp Edu Hld Edu Bus Edu Prod Printer TV Camera Laptop HD Cost 500 800 1160 1500 540 Price 485 1200 1280 2500 520
CREATE CUBE Sales_Iceberg AS SELECT month, city, cust_grp, AVG(price), COUNT(*) FROM Sales_Infor CUBEBY month, city, cust_grp HAVING AVG(price) >= 800 AND COUNT(*) >= 50
58
Top-k Average
59
Computing top-k avg is costly with large k Binning idea 50 Avg (c) >= 800 Large value collapsing: use a sum and a count to summarize records with measure >= 800
If count>=800, no need to check small records One bin covers a range, e.g., 600~800, 400~600, etc. Register a sum and a count for each bin
Data Mining: Concepts and Techniques 60
61
Accumulate quant-info for cells to compute average iceberg cubes efficiently Three pieces: sum, count, top-k bins Use top-k bins to estimate/prune descendants Use sum and count to consolidate current cell weakest
Approximate avg50()
Anti-monotonic, can be computed efficiently
strongest
real avg50()
Anti-monotonic, but computationally costly
Data Mining: Concepts and Techniques
avg()
Not antimonotonic
62
Can we compute iceberg cube more efficiently? Top-k H-cubing: an efficient method to compute iceberg cubes with average measure
Header table
Quant-Info Sum:2285
Prod Printer TV Camera Laptop HD Cost 500 800 1160 1500 540
Side-link
Jan
Mar
Jan
Feb
Tor
Price 485 1200 1280 2500 520
Van
Tor
Mon
Quant-Info
Sum: 1765 Cnt: 2 bins
Q.I.
Q.I.
Q.I.
64
Properties of H-tree
Construction cost: a single database scan Completeness: It contains the complete information needed for computing the iceberg
cube
Side-link
Jan.
Feb.
Tor.
Quant-Info Sum: 1765 Cnt: 2 bins
Data Mining: Concepts and Techniques
Van. Q.I.
Tor. Q.I.
Mon. Q.I.
66
Quant-Info Sum:2285
Side-link
Jan. Q.I.
Feb. Q.I.
Tor.
Top-k OK mark: if Q.I. in a child passes top-k avg threshold, so does its parents. No binning is needed!
Data Mining: Concepts and Techniques 67
Van.
Tor.
Mont.
edu
hhd
bus
Jan Q.I.
Mar Q.I.
Jan Q.I.
Feb Q.I.
Tor
Van
Tor
Mon
68
Properties of H-Cubing
Space cost
an H-tree
a stack of up to (m-1) header tables
One database scan Main memory-based tree traversal & side-links updates Top-k_OK marking
69
Runtime (second)
top-k BUC
0.10%
70
Count threshold
December 19, 2012 Data Mining: Concepts and Techniques
Key point: find a function which is weaker but ensures certain anti-monotonicity
Examples
Memory-hog? what if the cubing is too big to fit in memory?projection and then cubing
Data Mining: Concepts and Techniques 72
Condensed Cube
W. Wang, H. Lu, J. Feng, J. X. Yu, Condensed Cube: An Effective Approach to Reducing Data Cube Size. ICDE02. Icerberg cube cannot solve all the problems
Suppose 100 dimensions, only 1 base cell with count = 10. How many aggregate (non-base) cells if count >= 10? Only need to store one cell (a1, a2, , a100, 10), which represents all the corresponding aggregate cells Adv.
Condensed cube
What is a data warehouse? A multi-dimensional data model Data warehouse architecture Data warehouse implementation
Information processing
supports querying, basic statistical analysis, and reporting using crosstabs, tables, charts and graphs
Analytical processing
Data mining
supports associations, constructing analytical models, performing classification and prediction, and presenting the mining results using visualization tools.
High quality of data in data warehouses DW contains integrated, consistent, cleaned data Available information processing structure surrounding data warehouses ODBC, OLEDB, Web accessing, service facilities, reporting and OLAP tools OLAP-based exploratory data analysis mining with drilling, dicing, pivoting, etc. On-line selection of data mining functions integration and swapping of multiple mining functions, algorithms, and tasks.
Architecture of OLAM
Data Mining: Concepts and Techniques 76
An OLAM Architecture
Mining query
User GUI API
Mining result
OLAM Engine
Data Cube API
OLAP Engine
Layer3
OLAP/OLAM
Layer2
MDDB
Meta Data
Filtering&Integration
MDDB
Database API
Data cleaning
Filtering
Layer1 Databases
December 19, 2012
Data Repository
77
Hypothesis-driven
pre-compute measures indicating exceptions, guide user in the data analysis, at all levels of aggregation
Exception: significantly different from the value anticipated, based on a statistical model Visual cues such as background color are used to reflect the degree of exception of each cell
Data Mining: Concepts and Techniques 78
Parameters
Computation of exception indicator (modeling fitting and computing SelfExp, InExp, and PathExp values) can be overlapped with cube construction Exception themselves can be stored, indexed and retrieved like precomputed aggregates
Data Mining: Concepts and Techniques 79
80
Multi-feature cubes (Ross, et al. 1998): Compute complex queries involving multiple dependent aggregates at multiple granularities Ex. Grouping by all subsets of {item, region, month}, find the maximum price in 1997 for each group, and the total sales among all maximum price tuples
Continuing the last example, among the max price tuples, find the min and max shelf live, and find the fraction of the total sales due to tuple that have min shelf life within the set of all max price tuples
Data Mining: Concepts and Techniques 81
Cube-Gradient (Cubegrade)
Analysis of changes of sophisticated measures in multi-dimensional spaces Query: changes of average house price in Vancouver in 00 comparing against 99 Answer: Apts in West went down 20%, houses in Metrotown went up 10% Cubegrade problem by Imielinski et al. Changes in dimensions changes in measures Drill-down, roll-up, and mutation
Data Mining: Concepts and Techniques 82
Serious challenges
83
Significance constraint Csig: (cnt100) Probe constraint Cprb: (city=Van, cust_grp=busi, prod_grp=*) Gradient constraint Cgrad(cg, cp): (avg_price(cg)/avg_price(cp)1.3)
Probe cell: satisfied Cprb
Base cell
Aggregated cell Siblings
December 19, 2012
cid c1 c2 c3 c4
Yr 00 * * *
Ancestor
A LiveSet-Driven Algorithm
Compute probe cells using Csig and Cprb
Pushing selection deeply Set-oriented processing for probe cells Iceberg growing from low to high dimensionalities Dynamic pruning probe cells during growth Incorporating efficient iceberg cubing method
85
Summary
Star schema, snowflake schema, fact constellations A data cube consists of dimensions & measures
OLAP operations: drilling, rolling, slicing, dicing and pivoting OLAP servers: ROLAP, MOLAP, HOLAP Efficient computation of data cubes
Partial vs. full vs. no materialization Multiway array aggregation Bitmap index and join index implementations Discovery-drive and multi-feature cubes From OLAP to OLAM (on-line analytical mining)
Data Mining: Concepts and Techniques 86
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. VLDB96
D. Agrawal, A. E. Abbadi, A. Singh, and T. Yurek. Efficient view maintenance in data warehouses. SIGMOD97.
R. Agrawal, A. Gupta, and S. Sarawagi. Modeling multidimensional databases. ICDE97 K. Beyer and R. Ramakrishnan. Bottom-Up Computation of Sparse and Iceberg CUBEs.. SIGMOD99. S. Chaudhuri and U. Dayal. An overview of data warehousing and OLAP technology. ACM SIGMOD Record, 26:65-74, 1997. OLAP council. MDAPI specification version 2.0. In http://www.olapcouncil.org/research/apily.htm, 1998.
G. Dong, J. Han, J. Lam, J. Pei, K. Wang. Mining Multi-dimensional Constrained Gradients in Data Cubes. VLDB2001
J. Gray, S. Chaudhuri, A. Bosworth, A. Layman, D. Reichart, M. Venkatrao, F. Pellow, and H. Pirahesh. Data cube: A relational aggregation operator generalizing group-by, cross-tab and sub-totals. Data Mining and Knowledge Discovery, 1:29-54, 1997.
Data Mining: Concepts and Techniques 87
References (II)
J. Han, J. Pei, G. Dong, K. Wang. Efficient Computation of Iceberg Cubes With Complex Measures. SIGMOD01
www.cs.uiuc.edu/~hanj
Work to be done
Add MS OLAP snapshots! A tutorial on MS/OLAP Reorganize cube computation materials Into cube computation and cube exploration
90