Context awareness is currently being investigated for applicationsin different areas, including Mobile Computing. Many mobile devicesare already shipped with support for Bluetooth and Wireless LAN,making these technologies commonly available. It is thus possibleto exploit the wireless interfaces as sensors for deriving informationabout the device/user context. However, extracting features fromtypical Bluetooth or Wireless LAN properties is difficult becausenot only numerical, but also non-numerical features like the listof MAC addresses in range are important for context awareness. Inthis paper, we introduce a method to automatically classify thesehighly heterogeneous features with supervised or un-supervised classificationmethods. By defining two operators, a distance metric and an adaptionoperator, any feature can be used as input for the classifier andcan thus contribute to context detection.