DSS

1
DSS DATA MANAGEMENT

• IS3301 Session 3
• Prof. Mark Nissen

2
Agenda

• GLSC DSS
• DSS Elements Examples
• Group Health Data Warehouse
• Data Warehouse Exercise
• Data Management
• Data Mining

3
GLSC DSS

• Gotaas-Larsen Shipping Corp
• DSS for planning? (F 3.7)
• Internal external data?
• User data input?
• Sim model?
• Sensitivity analysis?
• What-if vs. trial error?
• Strategic planning - errors catastrophic!

4
DSS Characteristics

• Support semi(un)structured DM
• Support individual group DM
• Support variety of DM processes/styles
• Support, not replace, DMer
• Support non-IT wizards
• Users build/mod systems themselves

5
DSS Elements

• Overview (F 3.2)
• Data management system (F 3.3)
• Model management system (F 3.4)
• User interface system (DGMS)
• Knowledge management system (AI)
• User - integral part of system

6
DSS Examples

• Marketing (p. 91) - POS
• APL Workflow (p. 93) - Web
• Glaxo forecasting (p. 93) - data warehouse
• TN forecasting (p. 95) - scenarios decision
  sensitivities
• UGA courses (p. 96) - Web DBMS
• Mil housing (p. 104) - hetero inputs

7
Group Health Data Warehouse

• Hospitals, centers, insurance
• Data warehouse
• TPS input
• SAS Warehouse
• Feeds DSS EIS
• Performance measurement
• Trend analysis
• Contract proposal

8
Data Warehouse Exercise

• 3 data sources - sales
• Marketing - commit (who, what, why, )
• Contracts - definitize (who, what, )
• Accounting - collection (what, , when)
• 2 data uses
• Case flow modeling forecasting
• Market segmentation
• Queries/analyses slow TPS
• Integrate data to support OLAP

9
Data Warehouse

• Overview (F 4.1)
• Many data types sources?
• Data integrity utility key (T 4.1)
• Cleaning, clustering summarizing
• Intranet integration multimedia
• Typical 3-tier architecture (F 4.3)
• Only 15 of data is digital?
• Mil analog?

10
OnLine Analytical Processing

• OLAP vs. OLTP?
• Separate DSS from TPS - why?
• OLAP ad-hoc queries, stat analyses, data viz
  summaries, by user
• Generally feed OLAP from warehouse

11
Data Mining

• Knowledge discovery in DBs
• KDD conference in Summer (AAAI)
• Excellent tutorials
• Data Mining
• Vast data sources - beyond human ability
• Clean filter, then analyze
• Search for info nuggets
• Examples (p.130)
• Pattern recognition task - ANN ES