Reliable Information Processing Sample Clauses
Reliable Information Processing. The dynamicity and heterogeneity of IoT environments involves changes and noise in the data, especially when dealing with data created by people. The methods for information extraction and data processing, however, require accuracy and trust issues to be taken into account. This module measures and processes accuracy and reputation in data acquisition, federation and aggregation. It integrates techniques for monitoring and testing, ensuring reliable information processing. For example, it provides fault tolerance mechanisms when malfunctioning or disappearing sensor are detected, or providing conflict resolution strategies when data analysis result in conflicting information. Provenance also plays an important role for Smart cities applications. These applications acquire data from heterogeneous sources, some of them more reliable (e.g. government data), and some of them less reliable (e.g. crowd sourced data). Based on user or application preferences, the application provider could choose to use less reliable data in various cases, e.g. it needs more up to date information. The Reliable Information Processing module performs quality analysis to assert the reliability of the data. To enable the process of annotating data with QoI a list of QoI parameters is defined in Section 5.2 and an ontology is developed that is described in Section 5.3.
Reliable Information Processing. The process of QoI and provenance annotation of raw stream data, which is used within the Reliable Information Processing component in the CityPulse framework, is described in detail in the following sections. In Section
5.1 the general idea of QoI and provenance annotation is explained and first parameters for the annotation are defined or extracted from literature. In Section 5.2 we highlight the parameters that are used to measure quality in detail. Finally we present the information model that is developed to represent an application independent view of quality and provenance information for data streams in Section 5.3. While Quality of Service (QoS) has been widely studied in sensor networks and has well defined measurable properties (e.g. throughput, jitter or packet loss), which are inherited from the field of network communications, the Quality of Information (QoI) is not well defined and sometimes difficult to measure. In this section, we present quality and provenance annotation process of data streams, and highlight the parameters that are used to obtain corresponding measurements. In order to explicitly represent the reliability of the information and the adequacy of the data in the Semantic Web environment, we also develop an information model, which uses an existing ontology to describe the provenance of the processed information.