Data processing apparatus and method for automatically generating a classification component

Abstract

Data processing apparatus operative to generate a classification component is disclosed. The data processing apparatus is configured to provide a template classifier bank comprising a plurality of classifier modules, each classifier module operative to receive training data comprising data elements having one of two or more known class affiliations and to output a class affiliation estimate value for each input data element. The data processing apparatus is further configured to derive a combination of the class affiliation estimate values providing a highest correlation to the two or more known class affiliations, and to generate a classification component formed of a resultant classifier bank comprising a combination of the plurality of classifier modules corresponding to the combination of estimate values providing the highest correlation.

Description

FIELD[0001]The present invention relates to data processing apparatus and a method, in particular but not exclusively for automatically generating a classification component.Introduction[0002]There are many different applications (domains) where the automated classification of data is utilised or desirable. One example is homeland security, where different types of security threats may be present either simultaneously or separately. The disparate nature of the threats in the homeland security domain makes new demands on the requirements for automated decision making. Threats in homeland security are broadly characterized as Nuclear, Radiological, Biological or Chemical. Each of these domains is characterized by a wide variety of different sensors and different operational regimes. Consequently, the resultant threat profiles are associated with widely different data characteristics. Furthermore, the nature of threats is expected to be constantly evolving as a consequence of the inter...

AI Technical Summary

Benefits of technology

[0007]By using such an arrangement, a classification system may be implemented which comprises many different classifiers not all of which are necessary or desirable for a range of classification applications to which the system will be put, but the system can be automatically configured with an optimum combination of classifier modules for a particular type of classification application to which the system is to be put. Thus, a generic classifier template may be used in order to cope with unanticipated classification problems, and a suitable classification bank derived from that generic template for particular classification applications. This saves time, cost and effort in designing classification systems.

[0022]Typically, said data elements are representative of source data of said known class affiliation. That is, said data elements are a reduced set of data elements which do not directly represent the source data but are derived from it. This means that there are less data points which can reduce the source of errors.

[0029]remove from said sub-set data elements having a contribution to the representation of said multivariate data set by said sub-set less than a second threshold value to form a final data set representative of said multivariate data set with a reduced number of sources of error.

[0030]Such an arrangement provides a feature selection and extraction process which reduces the amount of data to a data set having fewer members than the original data set which may reduce the sources of error.

[0046]One or more embodiments of the present invention may provide a technology which underlies a decision support development environment for threat detection such as the Advanced Threat Engine™ developed by Radiation Watch Limited of Cowes, UK (RWL)—The Threat Engine™ may provide a framework for providing for the automatic construction of decision support systems. This automatic generation capability allows the system to reconfigure its internal complexity to match the requirements of a current threat domain. The Threat Engine™ may provide a generic decision support mechanism, which can be integrated into large scale command and control applications or provide stand alone classification systems.

Problems solved by technology

This saves time, cost and effort in designing classification systems.

Images