Spatial Voting with Data Modeling
Citations Over TimeTop 10% of 2010 papers
Abstract
We use the Spatial Voting (SV) process for fusing spatial positions in a 2-D grid. This process yields a centroid and covariance estimate as the basis of robust cluster identification. We calculate a series of geospatial features unique to the identified cluster and attempt to identify unique and consistent features to enable automated target recognition. We define the geospatial features and outline our process of deriving a decision architecture populated with Data Models. We attempt to identify the support vectors of the feature space and enable the smallest subsample of available exemplars to be used for extracting the analytical rule equations. We present details of the decision architecture derivation process. We construct ambiguity resolvers to further sieve and classify mislabeled sensor hits by deriving a new resolver Data Model that further processes the output from the previous layer. In this fashion through a cascade filter, we are able to demonstrate unique classification and full assignment of all available examples even in high dimensional spaces.
Related Papers
- → Some asymptotic properties of kriging when the covariance function is misspecified(1989)51 cited
- → Selection of covariance functions in Gaussian process-based soft sensors(2014)2 cited
- → On the Family of Covariance Functions Based on ARMA Models(2021)1 cited
- → Certain Periodically Correlated Multicomponent Locally Stationary Processes(2015)1 cited
- Directional Random Effect Model and Its Covariance Structure(2004)