Q:

Could somebody indicate to me where I can find test reports, case study or research information regarding data mining quality in relation to business intelligence reporting accuracy.

I heard there are some reports available that indicate that in practice about 80 percent accuracy can be reached. While BI is used to make business decisions, these experiences make me worry.

A:

Larissa Moss’ Answer: And worry you should. If your organization can achieve 80 percent accuracy, you are fortunate because 80 percent accuracy is actually quite high. In my experience, some of the large organizations that have many old operational systems on mainframes have such enormous data quality problems that some of their most critical data elements (from a decision support perspective) have no more than 50 percent accuracy. Besides the poor quality of the data, the tremendous amount of redundancy also creates a problem for data mining. While some organizations attempt to cleanse their data and reduce the redundancy, most organizations find that effort overwhelming in time and cost; and they do little or nothing about it. Some organizations hope to find the "silver bullet" solution in a data cleansing tool. But tools cannot perform the analysis that is needed to find and diagnose the data quality problems (data domain violations are relatively unique in every organization), Data analysis is, and always will remain, a human effort; hence, the overwhelming amount of time and money. The true business impact of poor quality data has not been studied and published enough for business executives to comprehend the associated risk and exposure to their organizations. I recommend Larry English’s book: Improving Data Warehouse and Business Information Quality, John Wiley & Sons, 1999. ISBN 0-471-25383-9.

Joe Oates’ Answer: Data mining is a statistical tool. Statistics do not guarantee 100 percent accuracy. If a tool, such as data mining can provide 80 percent accuracy in prediction, that is very good. Without sophisticated statistical tools, when one guesses, the best that can be hoped for is 50 percent accuracy. In business, an organization that is 5 percent better in predicting product and market trends than their competition, enjoys a huge advantage.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access