Privacy-preserving data release is an increasingly important problem in today's ubiquitous computing. Performing data processing on edge devices saves CPU time, network traffic and improves latency. The savings are especially significant for edge devices with low-power and limited resources. In addition, processing and privacy-preserving release of the data on the edge devices prevents sensitive data from leaving the devices at all. Thus, solving a bunch of potential problems in data protection. We propose a usage of query-independent, similitude models for privacy-preserving data release on the edge devices. Those models build a compact representation of the data for any possible query, as opposite to a common method of queryspecific data release. Compared to the query-specific models, similitude models can be combined and subtracted when a new edge devices appear or shutdown. In addition, a data oriented approach allows further processing in a fog, near the edge devices.