Machine learning in the Internet of Things: A semantic-enhanced approach

Abstract

New Internet of Things (IoT) applications and services more and more rely on an intelligent understanding of the environment from data gathered via heterogeneous sensors and micro-devices. Though increasingly effective, Machine Learning (ML) techniques generally do not go beyond classification of events with opaque labels, lacking meaningful representations and explanations of taxonomies. This paper proposes a framework for a semantic-enhanced data mining on sensor streams, amenable to resource-constrained pervasive contexts. It merges an ontology-based characterization of data distributions with non-standard reasoning for a fine-grained event detection by treating the typical classification problem of ML as a resource discovery. Outputs of classification are endowed with machine-understandable descriptions in standard Semantic Web languages, while explanation of matchmaking outcomes motivates confidence on results. A case study on road and traffic analysis allowed to validate the proposal and achieve an assessment with respect to state-of-the-art ML algorithms.

Knowledge-Based Decision Support in Healthcare via Near Field Communication

“…In particular, first aid, ambulance transport and tele-homecare scenarios provide additional challenges: due to this reason they are necessary for a full validation—and possibly a refinement—of the framework from architectural and technological viewpoints. The interoperability between the patient’s NFC tag and wearable health monitoring sensors is a particularly meaningful development, as the latter may be exploited to update the health status by means of semantic-based data mining and annotation [ 47 ] in a fully decentralized and automated fashion. Prospective clinical studies are needed to assess the effective impact of the proposal on healthcare quality.…”

“…The interoperability between the patient’s NFC tag and wearable health monitoring sensors is a particularly meaningful development, as the latter may be exploited to update the health status by means of semantic-based data mining and annotation [ 47 ] in a fully decentralized and automated fashion.…”

“…SVN provided the best results in both tests; unfortunately, it is a black box technique, i.e., the trained model is not transparent and the outcome is a trivial classification label, without interpretable information annotation. This affects other machine learning approaches as well [ 47 ], preventing both outcome explainability and integration with knowledge-based systems like the one proposed in this paper. A semantic-based posture and gesture recognition system can be found in [ 48 ], based on popular three-dimensional (3D) motion sensors like Asus Xtion Pro Live: sequences of skeleton representations returned by the sensing device are annotated with respect to an OWL ontology, so that it undergoes semantic matchmaking with a knowledge base of attitude and gesture templates to retrieve the best matching ones.…”

A Review of Semantic Sensor Technologies in Internet of Things Architectures

“…Current high-level stream processors, with the capability of on arrival or nonarrival of data, deriving information for the system, are C-SPARQL, CQELS, and ACEIS [101]. With special architectures and techniques, machine learning based solutions can also be implemented [102]. (iii) Service configuration is the key in the vision of autonomic computing [103,104].…”

Intelligent Web Applications as Future Generation of Web Applications

“…In addition, Ruta M. et al (2018) presented a new method for a machine that is semantic enriched learning on assorted data streams within the Internet of Things (IOT). Mapping of raw data to concept labels that are ontology-based offers a lowlevel semantic understanding of the statistical information distribution, whereas the concept components' conjunctive aggregation enables automatic building of an important and rich exemplification of events in the course of the phase of model training [13].…”