Potpourri of current database trends - PowerPoint PPT Presentation

1 / 6
About This Presentation
Title:

Potpourri of current database trends

Description:

Lecture 22, CS470. 1. Potpourri of current database trends. Data mining. Uses. Goals. Methods ... Marketing: Analyzing data on consumer behavior and predicting ... – PowerPoint PPT presentation

Number of Views:36
Avg rating:3.0/5.0
Slides: 7
Provided by: chitrad
Category:

less

Transcript and Presenter's Notes

Title: Potpourri of current database trends


1
Potpourri of current database trends
  • Data mining
  • Uses
  • Goals
  • Methods
  • The WWW and databases
  • Specialized databases
  • Multimedia
  • Mobile databases
  • Geographic Information Systems
  • Biological databases
  • Digital Libraries

2
Data Mining
  • Uses
  • Marketing Analyzing data on consumer behavior
    and predicting strategies to maximize profit
  • Science Predicting correlations between
    biomolecules
  • Finance Which of the applicants are eligible for
    loans?
  • Manufacturing Maximize number of cars produced
    per dollar
  • Health Care Maximize number of patients seen per
    day
  • Goals
  • Prediction
  • Identification
  • Classification
  • Optimization

3
Data Mining
  • Methods
  • Association Rules If customer buys milk, he also
    buys eggs (Conclusion Places eggs next to milk
    in store)
  • Sequential pattern algorithms A week later,
    customer will return to buy a battery for his
    camcorder
  • Time series pattern algorithms When should I
    buy, when should I sell stock?
  • Classification rules Class A of applicants is
    eligible for a mortgage loan while class B is not
  • Machine learning Use known data to predict
    unknown data

4
  • WWW Databases (Fig 27.1)
  • Alternative to CGI scripts
  • ASP, JSP, ODBC, JDBC ..
  • Alternative to RDBMS
  • XML databases
  • Flat file databases
  • Mobile Databases
  • DDBMS with a twist (Fig 27.4)
  • Geographical Information Systems (Fig 27.5)

5
Multimedia Databases
  • Properties
  • Heterogeneous data (Video, Audio, Images,
    Animations..) - Data varies in both type and
    length
  • Synchronous delivery required
  • Approaches
  • BLOBs (Binary large object)
  • Alternatives to text-based indexing (Image
    processing)
  • Do not support compute intensive queries

6
Digital libraries
  • Virtual library
  • Millions of species
  • Species share some properties and not others
  • DNA sequence of 100 known
  • Traditionally, non-standardized flat file
    retrieval systems
  • Newer approaches try to include other kinds of
    data
  • and provide more functionality

Biological Databases
Write a Comment
User Comments (0)
About PowerShow.com