TY - JOUR

T1 - Mining Sensor Data in Larger Physical Systems

AU - O'Leary, Paul

AU - Harker, Matthew

AU - Ritt, Roland

AU - Habacher, Michael

AU - Landl, Katharina

AU - Brandner, Michael

PY - 2016

Y1 - 2016

N2 - This paper presents a framework for the collection, management and mining of sensordata in large cyber-physical systems. Particular emphasis has been placed on mathematicalmethods, data structures and implementations which enable the real-time solution of inverseproblems associated with the system in question. That is, given a system model, to obtainan estimate for the phenomenological cause of the sensor observation. This enables the useof causality, rather than mere correlation, when computing measures of signicance duringmachine learning and knowledge discovery in very large data sets. The model is an abstractrepresentation of a real physical system establishing the relationships between cause and eects.The pertinent behaviour of the model is captured in the form of equations, e.g., dierentialequations. The inverse solution of these model-equations, within certain constraints, permit usto establish the semantic reference between the sensor observation and its cause. Without thissemantic reference there can be no physically based knowledge discovery.Embrechts pyramid of knowledge is addressed and shown that it will not suce for futuredevelopments. The issue of information content is addressed more formally than in most datamining literature. Additionally the Epistemology for the emergent-perceptive portion of speechis presented and a prototype implementation with experimental results in data mining arepresented. A lexical symbolic analysis of sensor data is implemented

AB - This paper presents a framework for the collection, management and mining of sensordata in large cyber-physical systems. Particular emphasis has been placed on mathematicalmethods, data structures and implementations which enable the real-time solution of inverseproblems associated with the system in question. That is, given a system model, to obtainan estimate for the phenomenological cause of the sensor observation. This enables the useof causality, rather than mere correlation, when computing measures of signicance duringmachine learning and knowledge discovery in very large data sets. The model is an abstractrepresentation of a real physical system establishing the relationships between cause and eects.The pertinent behaviour of the model is captured in the form of equations, e.g., dierentialequations. The inverse solution of these model-equations, within certain constraints, permit usto establish the semantic reference between the sensor observation and its cause. Without thissemantic reference there can be no physically based knowledge discovery.Embrechts pyramid of knowledge is addressed and shown that it will not suce for futuredevelopments. The issue of information content is addressed more formally than in most datamining literature. Additionally the Epistemology for the emergent-perceptive portion of speechis presented and a prototype implementation with experimental results in data mining arepresented. A lexical symbolic analysis of sensor data is implemented

KW - data mining

KW - entropy

KW - linear dierential operators

KW - lexical analysis

KW - data mining

KW - entropy

KW - linear dierential operators

KW - lexical analysis

UR - http://www.sciencedirect.com/science/article/pii/S2405896316316561

U2 - http://dx.doi.org/10.1016/j.ifacol.2016.10.093

DO - http://dx.doi.org/10.1016/j.ifacol.2016.10.093

M3 - Article

VL - 49

JO - IFAC-PapersOnLine

JF - IFAC-PapersOnLine

SN - 2405-8963

IS - 20

T2 - 17th IFAC Symposium on Control, Optimization and Automation in Mining, Mineral and Metal Processing

Y2 - 31 August 2016 through 2 September 2016

ER -