CorSub: Corpus-Based Semantic Topic Matching for Subscriptions

The number of smart devices is growing exponentially in both industrial and private environments. The Message Queuing Telemetry Transport (MQTT) is an established messaging protocol in the Internet of Things. However, the increased volume of communication also leads to larger amounts of data that need to be processed. In this work we propose CorSub, a novel approach as a future solution for designing intelligent infrastructure for information and communication, that combines word and sentence embeddings with a specific preprocessing to enable semantic subscription of MQTT Topics. The performance evaluation shows that CorSub’s time complexity is \(O(n \log n)\) and its average accuracy of 94%, which demonstrates the suitability of CorSub for a timely and accurate flexible subscriptions in MQTT.