Use efficiency study of a distributed algorithm for smart city computing systems

The main features of processing large amounts of data in smart city computing systems are analyzed. The advantages of distributed computing for improving the efficiency of information analysis are determined. Recommendation algorithms that allow processing large data faster and more accurately are studied. A distributed matrix factorization method for providing recommendations to users of smart city computing systems, which involves the exchange of public data between devices with local processing of private data, is proposed. The advantages of the proposed distributed recommendation model of a smart city computing system are determined: in particular, flexibility in changing parameters, higher accuracy of calculations due to the exchange of results between nodes, high speed of data processing, as well as scalability. Keywords algorithm, computing system, smart city, machine learning, Funk SVD, big data, distributed computing