A comparative analysis of structural risk minimization by support vector machines and nearest neighbor rule

Karaçali, B.; Ramanath, Rajeev; Snyder, Wesley E.

doi:10.1016/j.patrec.2003.09.002

Cited by 30 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This test has been done for real data and the results obtained from the test show that ANN acts better than the other two methods. B. Karacali et al [17] have comp ared SVM and KNN methods by using error rate and finally by combining these two methods and by using the power of SVM and simplicity of k-NN have expanded a synthesis classifier which has the advantages of two methods. M. O'Farrell et al [18] have compared k-NN and ANN in classification of spectral data.…”

Section: Related Work; Background and Motivationmentioning

confidence: 99%

An experimental investigation of the effect of discrete attributes on the precision of classification methods

Entezari‐Maleki

Iranmanesh

Minaei-Bidgoli

2009

2009 International Conference on Information and Communication Technologies

View full text Add to dashboard Cite

Abstract:In this paper, the precision of logistic regression, naïve-Bayes and linear data classification methods, with regard to the Area Under Curve (AUC) metric have been compared. The effect of the parameters including size of the dataset, kind of the independent attributes, number of the discrete attributes and their values have been investigated. From the results, it can be concluded that in datasets consisting of both discrete and continuous attributes, the AUC of the three mentioned classifiers are the same. With increasing the number of the discrete attributes, the AUC of logistic regression is increased and the precision related to this classifier become more than the other two classifiers. Also considering the impact of the discrete attributes it can be seen that with increasing the number of values in discrete attributes the AUC related to the logistic regression classifier increases and linear classifier's AUC decreases, but the AUC of the naïve-Bayes classifier remains constant. Therefore, the results of this research can help data miners in selecting the most efficient classifier by considering the characteristics of the datasets.

show abstract

Section: Related Work; Background and Motivationmentioning

confidence: 99%

An experimental investigation of the effect of discrete attributes on the precision of classification methods

Entezari‐Maleki

Iranmanesh

Minaei-Bidgoli

2009

2009 International Conference on Information and Communication Technologies

View full text Add to dashboard Cite

show abstract

“…This test has been done for real data and results of the test show that ANN could act better than the other two methods. Karacali et al [17] have compared SVM and k-NN methods using error rate, and finally by combining these two methods and using the power of SVM and simplicity of k-NN have gained a synthesis classifier which has the advantages of the two methods. O'Farrell et al [18] have compared k-NN and ANN in classification of the spectral data.…”

Section: Related Workmentioning

confidence: 99%

Comparison of Classification Methods Based on the Type of Attributes and Sample Size

Entezari‐Maleki¹,

Rezaei²,

Minaei‐Bidgoli³

2009

JCIT

View full text Add to dashboard Cite

In this paper, the efficacy of seven data classification methods; Decision Tree (DT), k-Nearest Neighbor (k-NN), Logistic Regression (LogR), Naïve Bayes (NB), C4.5, Support Vector Machine (SVM) and Linear Classifier (LC) with regard to the Area Under Curve (AUC) metric have been compared. The effects of parameters including size of the dataset, kind of the independent attributes, and the number of the discrete and continuous attributes have been investigated. Based on the results, it can be concluded that in the datasets with few numbers of records, the AUC become deviated and the comparison between classifiers may not do correctly. When the number of the records and the number of the attributes in each record are increased, the results become more stable. Four classifiers DT, k-NN, SVM and C4.5 obtain higher AUC than three classifiers LogR, NB and LC. Among these four classifiers, C4.5 provides higher AUC in the most cases. As a comparison among three classifiers LogR, NB and LC, it can be said that NB provides the best AUC among them and classifiers LogR and NB have the same results, approximately.

show abstract

“…By mapping the signal data into higher dimension space, data points in the space tended to be more disperse, which made it easier for the model to classify the samples into different clusters. Besides, SVM is based on the principle of structural risk minimization, which will enhance the robust of model significantly [5]. However the complex of SVM not only brought the steady and robust of the model, but also the difficulty of parameter selecting.…”

Section: Introductionmentioning

confidence: 99%

Research on Pattern Recognition Method for Honey Nectar Detection by Electronic Nose

Ning-jing

Shi

Zhao

et al. 2015

Computer and Computing Technologies in Agriculture VIII

View full text Add to dashboard Cite

Electronic nose (e-nose) utilizes the gas sensors array to absorb the volatile organic compounds (VOCs) of samples to classify them into different clusters, and it is noted by the sensitive of the sensors. However, limited by the methodologies of the pattern recognition, this kind of advantage had not been exploited fully. The research studied on different types of pattern recognition method, and selected the optimum method for the detection of samples with little nuance, exemplified by honey nectar detection for including rape honey, linden honey and acacia honey. It was found that support vector machine (SVM, non-linear prediction model) showed better performance than linear discriminate analysis (LDA, linear prediction model) for the classification of tiny different samples, especially when multi-group detection was involved. After the optimized method was selected, the key points of the SVM model were analyzed, and two key parameters were displayed, which were kernel parameter and penalty parameter. Three algorithms, including grid searching (GS), particle swarm optimization (PSO) and genetic algorithm (GA) were applied to find the appropriate parameter values. The results showed parameters optimized by genetic algorithm (kernel parameter and penalty parameter is 0.11 and 14.38 respectively) led to the optimal model, whose training accuracy was 98.78% and prediction accuracy was 97.5%. The results suggested that in the method of SVM with parameters selected by GA, e-nose could handle well the discrimination of similar samples like honey nectar detection.

show abstract

A comparative analysis of structural risk minimization by support vector machines and nearest neighbor rule

Cited by 30 publications

References 16 publications

An experimental investigation of the effect of discrete attributes on the precision of classification methods

An experimental investigation of the effect of discrete attributes on the precision of classification methods

Comparison of Classification Methods Based on the Type of Attributes and Sample Size

Research on Pattern Recognition Method for Honey Nectar Detection by Electronic Nose

Contact Info

Product

Resources

About