Improved k-nearest neighbor classification

Wu, Yingquan; Ianakiev, Krassimir; Govindaraju, Venu

doi:10.1016/s0031-3203(01)00132-7

Cited by 173 publications

(65 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…5 Example of signal points generated by a change of radius offset r off , and angle offset θ off the sensor deployment strategy that is compared with the two previous classifiers, the probabilistic Bayesian classifier in [27] and the k-NN method in [33]. This comparison experiment is performed in two different conditions: (a) the signals are generated irrespective of the location of sensors and (b) the signals are generated as being influenced by installed location such as the three cases described in "Efficient location tracking strategy".…”

Section: Resultsmentioning

confidence: 99%

SensDeploy: efficient sensor deployment strategy for real-time localization

Lee

Shin

2017

Hum. Cent. Comput. Inf. Sci.

View full text Add to dashboard Cite

In order to estimate the location of the user, the previous studies introduced many sensor deployment methods applying the sensor network. The important issues to consider when placing the sensors are a configuration cost and detection area of a sensor network. In other words, the sensors consisting the network should be optimally deployed by taking into account the coverage and connectivity of them. In this paper, a sensor signal is modeled as the Gaussian distribution based signal points group, and signal points in overlapping region between two different sensors are classified by e-SVM (support vector machine) method as each sensor group. In addition, a trilateration technique is used for estimating the position of a moving object. At this time, we efficiently deploy the sensors with f-Apriori method to maintain the connectivity between the sensors as well as to apply the trilateration. The proposed method can be utilized for optimal placement of sensors if we know a detection range of one sensor. In this paper, we introduce more effective and adaptive deployment method to consist the sensor network as taking into account the cost and the situation.

show abstract

Section: Resultsmentioning

confidence: 99%

SensDeploy: efficient sensor deployment strategy for real-time localization

Lee

Shin

2017

Hum. Cent. Comput. Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…for example: the family of four instance reduction algorithms denoted respectively IRA1-IRA4 [9], the All k-NN method [43]) try to eliminate unwanted training examples using some removal criteria that need to be fulfilled. The same principle has been mentioned in [52]. The authors of [52] conclude that if many instances of the same class are found in an area, and when the area does not include instances from the other classes, then an unknown instance can be correctly classified when only selected prototypes from such area is used.…”

Section: Related Workmentioning

confidence: 88%

Cluster-based instance selection for machine classification

Czarnowski

2011

Knowl Inf Syst

View full text Add to dashboard Cite

Instance selection in the supervised machine learning, often referred to as the data reduction, aims at deciding which instances from the training set should be retained for further use during the learning process. Instance selection can result in increased capabilities and generalization properties of the learning model, shorter time of the learning process, or it can help in scaling up to large data sources. The paper proposes a cluster-based instance selection approach with the learning process executed by the team of agents and discusses its four variants. The basic assumption is that instance selection is carried out after the training data have been grouped into clusters. To validate the proposed approach and to investigate the influence of the clustering method used on the quality of the classification, the computational experiment has been carried out.Keywords Machine learning · Data mining · Instance selection · Multi-agent system IntroductionLearning from examples remains the most important paradigm of the machine learning. The problem of learning from data, according to [7], can be formulated as follows: Given a dataset D, a set of hypotheses H , a performance criterion P, the learning algorithm L outputs a hypothesis h ∈ H that optimizes P. The data D consists of N training examples, also called instances. Each example is described by a set A of n attributes. The goal of learning is to produce a hypothesis that optimizes the performance criterion. In the pattern classification application, h is a classifier (i.e. decision tree, artificial neural network, naive Bayes, k-nearest neighbor, etc.) that has been induced based on the training set D.Research works in the field of machine learning have resulted in the development of numerous approaches and algorithms for classification problems [46,51]. One of the recent focuses of such research includes methods of selecting relevant information to be used within the I. Czarnowski (B)

show abstract

“…A variety of improved selection schemes have been proposed which aim at retaining relevant information contained in the data set, see e.g. [11] and references therein.…”

Section: Nearest-neighbor Classifiersmentioning

confidence: 99%

Distance Measures for Prototype Based Classification

Biehl¹,

Hammer

Villmann

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The basic concepts of distance based classification are introduced in terms of clear-cut example systems. The classical k-NearestNeigbhor (kNN) classifier serves as the starting point of the discussion. Learning Vector Quantization (LVQ) is introduced, which represents the reference data by a few prototypes. This requires a data driven training process; examples of heuristic and cost function based prescriptions are presented. While the most popular measure of dissimilarity in this context is the Euclidean distance, this choice is frequently made without justification. Alternative distances can yield better performance in practical problems. Several examples are discussed, including more general Minkowski metrics and statistical divergences for the comparison of, e.g., histogram data. Furthermore, the framework of relevance learning in LVQ is presented. There, parameters of adaptive distance measures are optimized in the training phase. A practical application of Matrix Relevance LVQ in the context of tumor classification illustrates the approach.

show abstract

Improved k-nearest neighbor classification

Cited by 173 publications

References 12 publications

SensDeploy: efficient sensor deployment strategy for real-time localization

SensDeploy: efficient sensor deployment strategy for real-time localization

Cluster-based instance selection for machine classification

Distance Measures for Prototype Based Classification

Contact Info

Product

Resources

About