Previous discussions in other publications have often revolved around the notion of the “Trouble with Data”. But the output of data is often numbers in a report or table. The notion of the “trouble with data” can also be applied to the “trouble with numbers”. For instance, how are numbers interpreted and what kinds of insights are being inferred from the numbers. The data or source information itself behind these numbers is entirely correct but the numbers themselves can be misleading. What do I mean by this? One good practical example is correlation analysis where the trained mathematician would clearly
Already receive the Machine Learning Times emails? The Machine Learning Times now requires legacy email subscribers to upgrade their subscription - one time only - in order to attain a password-protected login and gain complete access.