Two-phase linear reconstruction measure-based classification for face recognition

Effective management of customer’s knowledge leads to efficient Customer Relationship Management (CRM). To accurately predict customer’s behaviour, clustering, especially K-means, is one of the most important data mining techniques used in customer relationship management marketing, with which it is possible to identify customers’ behavioural patterns and, subsequently, to align marketing strategies with customer preferences so as to maintain the customers. However, it has been observed in various studies on K-means clustering that customers with different behavioural indicators in clustering may seem to be the same, implying that customer behavioural indicators do not play any significant role in customer clustering. Therefore, if the level of customer participation depends on behavioural parameters such as their satisfaction, it can have a negative effect on the K-means clusters and has no acceptable result. In this paper, customer behavioural features—malicious feature—is considered in customer clustering, as well as a method for finding the optimal number of clusters and the initial values of cluster centres to obtain more accurate results. Finally, according to the organizations’ need to extract knowledge from customers’ views through ranking customers based on factors affecting customer value, a method is proposed for modelling their behaviour and extracting knowledge for customer relationship management. The results of the evaluation of the customers of Hamkaran System’s Company show that the improved K-means method proposed in this paper outperforms K-means in terms of speed and accuracy.

Determination of Customer Satisfaction using Improved K-means algorithm

A self-adaptive k-means classifier for business incentive in a fashion design environment

Using kernel sparse representation to perform coarse-to-fine recognition of face images

As an important application in optical imaging, palmprint recognition is interfered by many unfavorable factors. An effective fusion of blockwise bi-directional two-dimensional principal component analysis and grouping sparse classification is presented. The dimension reduction and normalizing are implemented by the blockwise bi-directional two-dimensional principal component analysis for palmprint images to extract feature matrixes, which are assembled into an overcomplete dictionary in sparse classification. A subspace orthogonal matching pursuit algorithm is designed to solve the grouping sparse representation. Finally, the classification result is gained by comparing the residual between testing and reconstructed images. Experiments are carried out on a palmprint database, and the results show that this method has better robustness against position and illumination changes of palmprint images, and can get higher rate of palmprint recognition.

/pdf/an-efficient-classification-method-based-on-principal-1osv81ctlh.pdf

An efficient classification method based on principal component and sparse representation.

Novel classification rule of two-phase test sample sparse representation

https://isiarticles.com/bundles/Article/pre/pdf/46604.pdf

Qingxiang Feng

Papers

Novel classification rule of two-phase test sample sparse representation

Global linear regression coefficient classifier for recognition