An Algorithm for Non-Parametric Pattern Recognition

Sebestyen, George; Edie, Jay

doi:10.1109/pgec.1966.264473

Cited by 49 publications

(8 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, more homogeneous clusters can be formed, by assuming for the CH election process the node degree as primary parameter and the lowest ID as secondary parameter. The Sebastyen algorithm is another alternative [3,4]: this strategy allows to create homogeneous clusters by distributing CHs in a fairly uniform fashion across the network. This algorithm assigns a given set of nodes to appropriate clusters by implementing a decision rule based on the minimum Euclidean distance.…”

Section: Overview Of Related Workmentioning

confidence: 99%

“…1, where several sensor nodes collect and can send data to a unique collector that usually acts as a gateway. 4 The resulting topology consists of disjointed clusters, whose CHs gather, aggregate, process and deliver local data to a gateway (GW) via single-hop communication. As a result, a two tier hierarchical topology is assumed to achieve the benefits described above.…”

Section: Proposed Approachmentioning

confidence: 99%

“…If power adaptation is enabled, an intra-cluster communication cost is adopted 3. Only in the case of group mobility model, flat routing outperforms hierarchical routing since the mobile nodes are well aggregated 4. This represents the typical situation of wide area dense WSNs.…”

mentioning

confidence: 99%

See 2 more Smart Citations

A distributed clustering scheme with self nomination: proposal and application to critical monitoring

et al. 2014

View full text Add to dashboard Cite

Clustering is a well known methodology to optimize the use of the resources, to lower the congestion and to improve the reliability in self-organized networks as the wireless sensor networks. This paper deals with the proposal of a novel clustering approach based on a low complexity distributed cluster head election based on a two-stage process. In particular, a suitable objective function is introduced in order to take into account the number of 1-hop neighbours (i.e., node degree) and the residual node energy. It is shown in the paper that the proposed protocol achieves remarkable performance improvements with respect to different alternatives, especially in the case of unpredictable scenarios. Moreover, the proposed protocol exhibits self-organize capabilities that are of special interest for critical monitoring applications, in particular when the effect of nodes mobility is significant.

show abstract

Section: Overview Of Related Workmentioning

confidence: 99%

Section: Proposed Approachmentioning

confidence: 99%

See 1 more Smart Citation

A distributed clustering scheme with self nomination: proposal and application to critical monitoring

et al. 2014

View full text Add to dashboard Cite

show abstract

“…and ^ = * ± > a ± w . The procedure had been applied'to Sebestyen and Edie's data [54] consisting of 168 two-dimensional vectors representing six classes, Table 3 and compare favorably with the MSE solution. Table 2 gives the results of the MSE solution up to 33 features.…”

Section: Theoremmentioning

confidence: 99%

Some applications of generalized inverse to pattern recognition (Ph.D. Thesis abstr.)

Al-Alaoui¹

1976

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

“…Histogram methods can be regarded as the earliest nonparametric estimators of a density function. On the basis of a partition of the feature space into cells of fixed or variable size, histogram methods involve the problem of their application to high-dimensional spaces, where the scarcity of available data makes it likely to find empty cells [1]. Finding the -nearest neighbors ( -NNs) represents another way to compute a density function estimate.…”

Section: Introductionmentioning

confidence: 99%

Multisource data classification with dependence trees

Datcu

Melgani²,

Piardi³

et al. 2002

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

In order to apply a statistical approach to the classification of multisource remote-sensing data, one of the main problems to face lies in the estimation of probability distribution functions. This problem arises out of the difficulty of defining a common statistical model for such heterogeneous data. A possible solution is to adopt nonparametric approaches, which rely on the availability of training samples without any assumption about the related statistical distributions. The purpose of this paper is to investigate the suitability of the concept of dependence trees for the integration of multisource information through estimation of probability distributions. First, this concept, introduced by Chow and Liu, is used to provide an approximation of a probability distribution defined in an-dimensional space by a product of 1 probability distributions defined in two-dimensional (2-D) spaces; this approximation corresponds, in terms of graph theoretical interpretation, to a tree of dependence. For each land cover class, a dependence tree is generated by minimizing an appropriate closeness measure. Then, a nonparametric estimation of the second-order probability distributions is carried out through the Parzen window approach, based on the implementation of 2-D Gaussian kernels. In this way, it is possible to reduce the complexity of the estimation, while capturing a significant part of the interdependence among variables. A comparison with other multisource data fusion methods, namely, the multilayer perceptron (MLP) method, the-nearest neighbor (-NN) method, and a Bayesian hierarchical classifier (BHC), is made. Experimental results obtained on multisensor [airborne thematic mapper (ATM) and synthetic aperture radar (SAR)] and multisource (experimental synthetic aperture radar (E-SAR) and a textural feature) data sets show that the proposed fusion method based on dependence trees is able to provide a classification accuracy similar to those of the other methods considered, but with the advantage of a reduced computational load.

show abstract

An Algorithm for Non-Parametric Pattern Recognition

Cited by 49 publications

References 4 publications

A distributed clustering scheme with self nomination: proposal and application to critical monitoring

A distributed clustering scheme with self nomination: proposal and application to critical monitoring

Some applications of generalized inverse to pattern recognition (Ph.D. Thesis abstr.)

Multisource data classification with dependence trees

Contact Info

Product

Resources

About