Evaluation of Market Competitiveness of Sport Competition Product Based on SVM

*Corresponding author.Yu Feng

Abstract. This article constructs evaluation index system based on Analytic Hierarchy Process, comparative analysis and so on. Further, it conducts a study on the market competitiveness evaluation of sports competition products in Guangdong Province by this evaluation index system. A novel method for market competitiveness evaluation of the sport competition product, i.e. SVM based market competitiveness evaluation method for sport competition product, is proposed. SVM is less dependent on sample data and the model is still good at generalization after learning limited samples. Compared with the fuzzy evaluation model, the model established by SMV is more advantageous in competitiveness evaluation for sport competition product. Empirical results show that this is an effective method to evaluate the sport competition product competitiveness. So, the article provides a theoretical basis for scientifically, systematically and effectively evaluating the different sports competitions and their market competitiveness.

Key words: Sport competition product, Market competitiveness, Evaluation index, SVM (Support Vector Machine).

1. Introduction

SVM method is a new learning method based upon statistical learning methods proposed by V. N. Vapnik et al. in 1960s, which began to gain wide attention until the middle of 1990s. And it also shows particular advantages and good application prospect in the solution of small samples, nonlinearity and high- dimension problems in pattern recognition. About the high efficiency and accuracy of SVM method in different categories of researches, foreign scholars have compared it with 16 existing classification methods and concluded that SVM method is the best. Now, this method has been widely applied in many fields, such as inspection of welding defects, computer network, multiple signal classification, decision-making and comprehensive evaluation.

2. Evaluation index system of market competitiveness of sport competition product

The specific index system is shown in Table 1.

3. Theoretical basis of Support Vector Machine

Consider a training set = with input vectors = and target labels . SVM binary classifier satisfies the following conditions in the non-linear case.

where represents the weight vector and represents the bias term. maps the input vectors into a high-dimensional feature space. It is general to introduce slack variables to permit misclassification. is the penalty parameter. Thus the optimization problem becomes

Finally, we get a nonlinear decision function in primal space for linearly non-separable case through converting the optimization problem into a dual QP-problem.

is kernel function satisfying Mercerp’s condition. Four common kernel function types of SVM are given as follows.

Linear kernel: = .

Polynomial kernel: = .

Radial basis kernel: = .

Sigmoid kernel: = .

where and are constants.

4. Empirical analysis

We select randomly 30 products data from sports competition products in Guangdong Province as learning samples to train the SVM as shown in Table 2. Based on the trained SVM model, the market competitiveness evaluation problem can be solved successfully and accurate results can be obtained as shown in Table 3.

References

[1] J. Drayer, D.A. Rascher and C.D. McEvoy, An examination of underlying consumer demand and sport pricing using secondary market data, Sport Management Review, vol.15, pp.448-460, 2012.

[2] K.K. Byon, J.J. Zhang and D.P. Connaughton, Dimensions of general market demand associated with professional team sports: Development of a scale, Sport Management Review, vol.13, pp.142-157, 2010.

[3] P. Danenas and G. Garsva, Credit risk evaluation modeling using evolutionary linear SVM classifiers and sliding window approach, Procedia Computer Science, vol.9, pp.1324-1333, 2012.

[4] L. Yu, X. Yao, S.Y. Wang and K.K. Lai, Credit risk evaluation using a weighted least squares SVM classifier with design of experiment for parameter selection, Expert Systems with Applications, vol.38, pp.15392-15399, 2011.

[5] E. Gumus, N. Kilic, A. Sertbas and O.N. Ucan, Evaluation of face recognition techniques using PCA, wavelets and SVM, Expert Systems with Applications, vol.37, pp.6404-6408, 2010.

[6] T.K. Wu, S.C. Huang and Y.R. Meng, Evaluation of ANN and SVM classifiers as predictors to the diagnosis of students with learning disabilities, Expert Systems with Applications, vol.34, pp.1846-1856, 2008.

[7] M.H. Horng, Performance evaluation of multiple classification of the ultrasonic supraspinatus images by using ML, RBFNN and SVM classifiers, Expert Systems with Applications, vol.37, pp.4146-4155, 2010.