Transcript Models Created by Data Mining - Home

Data Mining
Models Created by Data
Mining
•
•
•
•
•
•
Linear Equations
Rules
Clusters
Graphs
Tree Structures
Recurrent Patterns
2
Knowledge Discovery in
Databases (KDD)
•
•
•
•
•
Select target data
Preprocess data
Transform (if necessary)
Data mine information
Interpret discovered structures
3
Dependant and Independent
Variables
• Dependant Variable - Attribute to be
predicted.
• Independent Variable - Attributes used
for making the prediction.
4
Fields Contributing to Data
Mining
•
•
•
•
•
•
•
•
Database Technology
Statistics
Machine Learning
High Performance Computing
Pattern Recognition
Neural Networks
Data Visualization
Information Retrieval
5
Applications of Data Mining
•
•
•
•
Decision Making
Process Control
Information Management
Query Processing
6
Methods of Data Reduction
•
•
•
•
Drill-down analysis
Clustering
Aggregation
Simple Tabulation
7
Exploratory Data Analysis
(EDA)
•
•
•
•
•
•
Distributions of Variables
Correlation Matrices
Multi-way Frequency Tables
Cluster Analysis
Classification Trees
Other multivariate techniques
8
Statistical Methods Used in
Data Mining
• Regression Analysis
• Standard Distribution
• Cluster Analysis
9
Industries Using Data Mining
•
•
•
•
•
•
Banking
Insurance
Medicine
Retail
Security
Sciences
10
Financial Uses of Data Mining
• Fraud Detection
• Money Laundering Detection
• Risk Management
11
Medical Uses of Data Mining
• Chemical Compounds
• Genetic Material
• Predictive Treatment Models
12
Retail Uses of Data Mining
• Direct Marketing
• Store Design
• Store Operations
13
Security Uses of Data Mining
•
•
•
•
Assess crime patterns
Homeland Security
Identification of suspicious activities
Pre-screening
14
Scientific Uses of Data Mining
• Image analysis
• Classification of large data sets
15
Other Novel Uses for Data
Mining
• NBA’s Advanced Scout Program
• Firefly
16
Predictive Analytics
• An advanced form of data mining that
makes prediction models for the
behavior of variables in large data sets.
• Highly specialized for each application
17
Uses of Predictive Analytics
• Cost-Benefit Analysis
• Predicting Customer Behavior
• Reducing Costs
18
Financial Uses of Predictive
Analytics
• Credit Ratings
• Economic Prediction Models
• Federal Reserve
19
Text Mining
• Extracts data from unstructured data
sets
• Allows for data mining of large data sets
that are not databases
20
Sentiment Analysis
• Uses semantic techniques and
keywords to detect favorable and
unfavorable opinions toward specific
subjects.
21
Privacy Concerns with Data
Mining
• Big Brother
• Puts too much power into the hands of
Governmental Security Forces
22
False Positives in Data Mining
for Security Reasons
• Costs the people and the Government
• Subject of controversy and civilian
mistrust
23
Data Mining as Another Tool
for Security
• Government doesn’t wish to interfere in
civilian life
• Actual intrusions of privacy incur legal
costs
• Useful for correlating with other sources
of data
24
Visual and Speech Processing
• Examining large amounts of real-time
input for specific data and relationships
between data
• Requires a certain amount of predictive
modeling
25
Data Mining is an Essential
Use of Computers
• It makes the previously impossible
possible
• Powerful tool for progress and
understanding
• Lasting Impact
26