Chapter 1.

Introduction
hy Data Mining?

 What Is Data Mining?

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology is Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 Summary

1
 Why Data Mining?

 The Explosive Growth of Data: from terabytes to petabytes

 Data collection and data availability
 Automated data collection tools, database systems, Web,
computerized society
 Major sources of abundant data
 Business: Web, e-commerce, transactions, stocks, …
 Science: Remote sensing, bioinformatics, scientific simulation, …
 Society and everyone: digital cameras, YouTube, social media
 We are drowning in data, but starving for knowledge!
 Data mining—Automated analysis of massive data sets to discover
knowledge

2
 Chapter 1. Introduction
 Why Data Mining?

 What Is Data Mining?

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology is Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 Summary

3
 What Is Data Mining?

 Data mining (knowledge discovery from data)

 Extraction of interesting (non-trivial, implicit, previously
unknown and potentially useful) patterns or knowledge from
huge amount of data
 Alternative names
 Knowledge discovery (mining) in databases (KDD), knowledge
extraction, data/pattern analysis, business intelligence, etc.
 Is everything “data mining”? Differenciate
 Simple search and query processing
 (Deductive) expert systems

4
 Knowledge Discovery (KDD) Process
 This is a view from typical
database systems and data
Pattern Evaluation
warehousing communities
 Data mining plays an essential
role in the knowledge discovery
process Data Mining

Task-relevant Data

Data Warehouse Selection

Data Cleaning

Data Integration

Databases
5
 KDD Process: A Typical View from ML and
Statistics

Input Data Data Pre- Data Post-

Processing Mining Processing

Data integration Pattern discovery Pattern evaluation

Normalization Association & correlation Pattern selection
Feature selection Classification Pattern interpretation
Clustering
Dimension reduction Pattern visualization
Outlier analysis
…………

 This is a view from typical machine learning and statistics communities

6
 Chapter 1. Introduction
 Why Data Mining?

 What Is Data Mining?

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology is Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 Summary

7
 Data Mining: On What Kinds of Data?
 Database-oriented data sets and applications
 Relational database, data warehouse, transactional database
 Advanced data sets and advanced applications
 Data streams and sensor data
 Time-series data, temporal data, sequence data (incl. bio-sequences)
 Structure data, graphs, social networks and multi-linked data
 Object-relational databases
 Heterogeneous databases and legacy databases
 Spatial data and spatiotemporal data
 Multimedia database
 Text databases
 The World-Wide Web

8
 Chapter 1. Introduction
 Why Data Mining?

 What Is Data Mining?

 A Multi-Dimensional View of Data Mining

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology Are Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 A Brief History of Data Mining and Data Mining Society

 Summary
9
Data Mining Function: (1) Generalization

 Information integration and data warehouse construction

 Data cleaning, transformation, integration, and
multidimensional data model
 Data cube technology
 Scalable methods for computing (i.e., materializing)
multidimensional aggregates
 OLAP (online analytical processing)
 Multidimensional concept description: Characterization
and discrimination
 Generalize, summarize, and contrast data
characteristics

10
 Data Mining Function: (2) Association and
Correlation Analysis
 Frequent patterns (or frequent itemsets)
 What items are frequently purchased together by a
customer
 Association, correlation vs. causality
 A typical association rule
 Bread  Peanut Butter [0.5%, 75%] (support, confidence)
 Support reflects utility while confidence reflects certainty of the conclusion
 High confidence value need not necessarily indicate strong
correlation between the items.
 If 80% transactions has Peanut Butter, the above rule reflects negative
association between the two.

 Additional correlation metrics like ‘Lift’ are used to mine such

patterns and rules efficiently in large datasets. 11
Data Mining Function: (3) Classification

 Classification and label prediction

 Construct models (functions) based on labelled training examples
 Describe and distinguish classes or concepts for future prediction
 E.g., classify tourist locations based on (climate, affordability,
activities, # days, etc), or estimate the cost of used cars based
on (mileage, age, model, fuel type, etc.)
 Apply the models to predict class labels for unknown entities
 Typical methods
 Decision trees, naïve Bayesian classification, support vector
machines, neural networks, rule-based classification, pattern-
based classification, logistic regression, …
 Typical applications:
 Credit card fraud detection, direct marketing, diagnosing diseases,
etc.
12
Data Mining Function: (4) Cluster Analysis

 Unsupervised learning (i.e., Class label is unknown)

 Group data to form new categories (i.e., clusters) of similar
entities,
 e.g., cluster houses to find distribution patterns or neighborhoods
 Principle: Maximizing intra-cluster similarity & minimizing inter-
cluster similarity
 Typical methods:
 Partitional Clustering eg: K-Means, K-medoids
 Hierarchical clustering eg: AGNES, DIANA
 Density based clustering eg: DBSCAN, OPTICS
 Applications: customer segmentation, taxonomy formation, topic
identification by document clustering, image quantization, pattern
recognition, etc.
13
Data Mining Function: (5) Outlier Analysis
 Outlier analysis
 Outlier: A data object that does not comply with the general
behavior of the data
 Noise or exception? ― One person’s garbage could be another
person’s treasure
 Methods: by product of clustering or regression analysis, …
 Useful in fraud detection, rare events analysis

14
 Evaluation of Patterns
 Interesting patterns represent knowledge
 Are all mined patterns interesting?
 One can mine tremendous amount of “patterns” and knowledge
 Patterns are interesting if they are:
 Easily understood
 Valid on new or unknown data with a high degree of certainty
 Potentially useful and Novel
 Evaluation of mined patterns  Directly mine only interesting
patterns / knowledge using:
 some objective measures like typicality, support, confidence are used for
descriptive tasks while precision, recall, accuracy, etc are used for
predictive tasks.
 Novelty, timeliness and actionability are subjective assessments 15
 Chapter 1. Introduction
 Why Data Mining?

 What Is Data Mining?

 What is Data Warehousing?

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology is Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 Summary

16
Data Mining: Confluence of Multiple Disciplines

Machine Pattern Statistics

Learning Recognition

Info Retrieval Visualization

Data Mining

Algorithm Database High-Performance

Technology Computing

17
Why Confluence of Multiple Disciplines?

 Tremendous amount of data

 Algorithms must be highly scalable to handle tera-bytes of data
 High-dimensionality of data
 Micro-array may have tens of thousands of dimensions
 High complexity of data
 Data streams and sensor data
 Time-series data, temporal data, sequence data
 Structure data, graphs, social networks and multi-linked data
 Heterogeneous databases and legacy databases
 Spatial, spatiotemporal, multimedia, text and Web data
 Software programs, scientific simulations
 New and sophisticated applications

18
 Chapter 1. Introduction
 Why Data Mining?

 What Is Data Mining?

 What is Data Warehousing?

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology is Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 Summary

19
 Applications of Data Mining
 Business Intelligence systems: Customer Relationship Management,
Predictive analytics for specific contexts, OLAP support for better
understanding business scenario
 Web page analysis: Search engines for web page classification,
clustering using PageRank & HITS algorithms, context–aware Query
recommendations
 Collaborative Filtering & Recommender systems
 Market Basket analysis to targeted marketing
 Medical data analysis: disease diagnosis, anomaly detection in medical
images, microarray data analysis
 Weather modelling and prediction of future climatic conditions

20
 Chapter 1. Introduction
 Why Data Mining?

 What Is Data Mining?

 What is Data Warehousing?

 What Kind of Data Can Be Mined?

 What Kinds of Patterns Can Be Mined?

 What Technology is Used?

 What Kind of Applications Are Targeted?

 Major Issues in Data Mining

 Summary

21
 Major Data Mining Issues related to …
 Mining Methodology
 Mining various, possibly new, kinds of knowledge
 Mining knowledge in (the subspaces of) a multi-dimensional space
 Data mining: An interdisciplinary effort (eg: Q&A sys need NLP, Info
Retrieval and Mining)
 Boosting the power of discovery in a networked environment (info
sharing among semantically linked heterogeneous data sources)
 Handling noise, uncertainty, and incompleteness of data; sometimes
incorrect data due to attackers
 Pattern evaluation and pattern- or constraint-guided mining (to focus
mining on specific topics or aspects of interest, context-aware RSs, etc.)
 User Interaction
 Interactive mining(dynamically change focus based on previous results)
 Incorporation of background knowledge (domain specific relationships)
 Presentation and visualization of data mining results 22
 Major Data Mining Issues related to …

 Efficiency and Scalability

 Efficiency and scalability of data mining algorithms
 Parallel, distributed, stream, and incremental mining methods
 Diversity of data types
 Handling complex types of data
 Mining dynamic, networked, and global data repositories
 Data mining and society
 Social impacts of data mining
 Privacy-preserving data mining
 Invisible data mining

23