偏差检测是一项在大型数据集中发现异常数据记录的任务。
Deviation detection is defined as the task of finding unusual data records in large datasets.
例如，热岛效应似乎在无风的夜晚最显著，但人们在无风夜晚与在有风夜晚记录下两组数据，比较之后，发现双方的趋势并无很大差别。
The heat-island effect is likely to be strongest on still nights, for example, yet trends from data recorded on still nights are not that different from those from windy ones.
接下来的小节将提供一个例子，以逐步演示如何用 InfoSphere Warehouse 发现离群值，以及如何为各个数据记录赋予偏差度。
The following section provides a step-by-step example of how to find outliers with InfoSphere Warehouse and how to assign deviation degrees to individual data records.
应用推荐