A quick survey around the room was made to gather Data Mining experiences and expectations:
Knowledge management applications:
- using Lucene to process documents for context-driven textual vectors
Web log and traffic pattern analysis applications:
- using SSAS to analyze and model data-mining heuristics for web traffic flow analysis and business modeling
Robotics and cognitive learning application
- data-mining techniques can be used for robotic navigation and complex terrain mapping
Discussion ensued about keeping the T/A data separate from the historical warehouse data by using multi-sourced scheduled ETL's in order to capture historical snapshots of production data at regular intervals.
More topics included multi-Dimensional data mapping, various OLAP hierarchies, and Data Warehousing in general:
- OLAP (dimensioned and cubed historical data for modeling and analysis)
- DW (cross-domain aggrated historical data)
- ODS (operational data store - consolidated de-normalized domain-specific data snapshots )
- OLTP (real-time normalized data)
Discussion continued with the Kimball group, the Kimball method, star-schema, facts tables, and training models to analyze new data to come up with new business rules and relationships pulled from that data. Also mentioned that MS Team System comes with SSAS built in ( but possibly not turned on) in order to do better Bug tracking and failure mode analysis to the project source code trees.
Tools:
- Hyperion
- Cognos
- Seibel Analytics
- SSAS - Sql Server Analysis Services
- SSIS - Sql Server Integration Services (ETL tools)
- MS Business Intelligence Developer Studio (VStudio)
- BITS - Microsoft Business Intelligence Toolset
- Microsoft Data Warehouse Toolkit.
- WECA (OS project)
- Mindstorm Robotics Toolkit
- MS SharePoint Report Builder and Report Services
- MS Performance Point (dashboard) tool
Additional Resources:
Blogs - Mathew Podwysocki: Functional Programming and Collective Intelligence - Part III
Comments (0)
You don't have permission to comment on this page.