Nowadays, one of the main challenges in the smart cities is mining high-level semantics from low-level activities. In this context, real-time data streams are continuously produced and analysed by efficient and effective algorithms, which are able to handle complexities related to big data, in order to enable the core functions of Decision Support Systems in the smart city. These algorithms should receive input data coming from different city domains (or pillars) and process, aggregate and reason over them in a way that it is possible to find hidden correlations among different and heterogeneous elements (e.g., traffic, weather, cultural events) along space and time dimensions. This paper proposes the online implementation and deployment of Temporal Fuzzy Concept Analysis on a distributed real-time computation system, based on Apache Storm, to face with big data stream analysis in the smart city context. Such online distributed algorithm is able to incrementally generate the timed fuzzy lattice that organizes the knowledge on several and cross-domain aspects of the city. Temporal patterns, of how situations evolve in the city, can be elicited by both exploring the lattice and observing its growth in order to obtain actionable knowledge to support smart city decision-making processes.
|Titolo:||Distributed online Temporal Fuzzy Concept Analysis for stream processing in smart cities|
|Data di pubblicazione:||2017|
|Appare nelle tipologie:||1.1 Articoli su Rivista|