Privacy-Aware Multidimensional Mobile Service Quality Prediction and Recommendation in Distributed Fog Environment

Algorithm 1

Three steps of approach.

Step 1 (build multi-dimensional user indices offline). For each , according to his/her observed QoS data

over dimensions , build his/her index (denoted by ) offline based on a LSH function family =

. Repeat the above index building process times so as to obtain hash tables.

Step 2 (search for similar friends of online). For each , compare and online; if

holds in any of the hash tables, then u can be regarded as similar with and put into set

.

Step 3 (service recommendation). For each ms (∈MS) never invoked by , predict its quality over

dimensions by , denoted by , based on set

obtained in Step 2. Finally, return the quality-optimal mobile service to .