• Journal of Korean Society for Quality Management
• /
• v.42 no.2
• /
• pp.221-234
• /
• 2014

Purpose: The Purpose of this study is to develop a model for predicting agent churn group in the cosmetics industry. We develope two models, pattern model and matrix model, which are compared regarding the prediction accuracy of churn agents. Finally, we try to conclude if there is statistically significant difference between two models by empirical study. Methods: We develop two models using the part of RFM(Recency, Frequency, Monetary) method which is one of customer segmentation method in traditional CRM study. In order to ensure which model can predict churn agents more precisely between two models, we used CRM data of cosmetics company A in China. Results: Pattern model and matrix model have been developed. we find out that there is statistically significant differences between two models regarding the prediction accuracy. Conclusion: Pattern model and matrix model predict churn agents. Although pattern model employed the trend of monetary mount for six months, matrix model that used the amount of sales per month and the duration of the employment is better than pattern model in prediction accuracy.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.105-108
• /
• 1997

This paper proposes a new approach of balanced model reduction using matrix pencil. The algorithm presented in this paper is to convert full-rank high-order system into rank-deficient system using perturbation made by matrix pencil method. Then the system can be truncated to a low-order system that we want via balanced realization. We discuss the comparison with other methods and the various observations by simulations.

• International Journal of Automotive Technology
• /
• v.8 no.4
• /
• pp.483-491
• /
• 2007

A closed form solution of a composite mechanics system is performed for the investigation of elastic-plastic behavior in order to predict fiber stresses, fiber/matrix interfacial shear stresses, and matrix yielding behavior in short fiber reinforced metal matrix composites. The model is based on a theoretical development that considers the stress concentration between fiber ends and the propagation of matrix plasticity and is compared with the results of a conventional shear lag model as well as a modified shear lag model. For the region of matrix plasticity, slip mechanisms between the fiber and matrix which normally occur at the interface are taken into account for the derivation. Results of predicted stresses for the small-scale yielding as well as the large-scale yielding in the matrix are compared with other theories. The effects of fiber aspect ratio are also evaluated for the internal elastic-plastic stress field. It is found that the incorporation of strong fibers results in substantial improvements in composite strength relative to the fiber/matrix interfacial shear stresses, but can produce earlier matrix yielding because of intensified stress concentration effects. It is also found that the present model can be applied to investigate the stress transfer mechanism between the elastic fiber and the elastic-plastic matrix, such as in short fiber reinforced metal matrix composites.

• Journal of the Korean Data and Information Science Society
• /
• v.7 no.1
• /
• pp.79-85
• /
• 1996

This paper will show that the rank of the model matrix of a closed, n compartmental model with k sinks is n-k. This statement will be extended to include open compartmental models as a part of theorem.

• Korean Management Science Review
• /
• v.24 no.2
• /
• pp.177-196
• /
• 2007

For better predictions and classifications in customer recommendation, this study proposes an integrative model that efficiently combines the currently-in-use statistical and artificial intelligence models. In particular, by integrating the models such as Association Rule, Frequency Matrix, and Rule Induction, this study suggests an integrative prediction model. Integrated models consist of four models: ASFM model which combines Association Rule(A) and Frequency Matrix(B), ASRI model which combines Association Rule(A) and Rule Induction(C), FMRI model which combines Frequency Matrix(B) and Rule Induction(C), and ASFMRI model which combines Association Rule(A), Frequency Matrix(B), and Rule Induction(C). The data set for the tests is collected from a convenience store G, which is the number one in its brand in S. Korea. This data set contains sales information on customer transactions from September 1, 2005 to December 7, 2005. About 1,000 transactions are selected for a specific item. Using this data set. it suggests an integrated model predicting whether a customer buys or not buys a specific product for target marketing strategy. The performance of integrated model is compared with that of other models. The results from the experiments show that the performance of integrated model is superior to that of all other models such as Association Rule, Frequency Matrix, and Rule Induction.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.4
• /
• pp.727-733
• /
• 2011

To understand the stock market structure it is very important to extract meaningful information by analyzing the correlation matrix between stock returns. Recently there has been many studies on the correlation matrix using the Random Matrix Theory. In this paper we adopt this random matrix methodology to a single-factor model and we obtain meaningful information on the correlation matrix. In particular we observe the analysis of the correlation matrix using the single-factor model explains the real market data and as a result we confirm the usefulness of the single-factor model.

• Journal of applied mathematics & informatics
• /
• v.8 no.2
• /
• pp.651-658
• /
• 2001

The projection matrix plays an important role in the linear model theory. In this paper we derive an algebraic relationship between the projection matrices of submatrices of the design matrix. Using this relationship we can easily obtain the projection matrices of any submatrices of the design matrix. Also we show that every projection matrix can be obtained as a linear combination of Kronecker products of identity matrices and matrices with all elements equal to 1.

• Journal of Mechanical Science and Technology
• /
• v.20 no.11
• /
• pp.1891-1897
• /
• 2006

In this study, a new model to predict the effective elastic constants of composites with spherical fillers is proposed. The original Eshelby model is extended to a finite filler volume fraction without using Mori-Tanaka's mean field approach. When single filler is embedded in the matrix, the effective elastic constants of the composite are computed. The composite is in turn considered as a new matrix, where new single filler is again embedded in the matrix. The predicted results by the present model with a series of embedding procedures are compared with those by Mori-Tanaka, self-consistent, and generalized self-consistent models. It is revealed through parametric studies such as stiffness ratio of the filler to the matrix and filler volume fraction that the present model gives more accurate predictions than Mori-Tanaka model without using the complicated numerical scheme used in self-consistent and generalized self-consistent models.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.9
• /
• pp.4684-4705
• /
• 2019

Collaborative filtering recommender systems are vulnerable to shilling attacks in which malicious users may inject biased profiles to promote or demote a particular item being recommended. To tackle this problem, many robust collaborative recommendation methods have been presented. Unfortunately, the robustness of most methods is improved at the expense of prediction accuracy. In this paper, we construct a robust Bayesian probabilistic matrix factorization model for collaborative filtering recommender systems by incorporating the detection of user anomaly rating behaviors. We first detect the anomaly rating behaviors of users by the modified K-means algorithm and target item identification method to generate an indicator matrix of attack users. Then we incorporate the indicator matrix of attack users to construct a robust Bayesian probabilistic matrix factorization model and based on which a robust collaborative recommendation algorithm is devised. The experimental results on the MovieLens and Netflix datasets show that our model can significantly improve the robustness and recommendation accuracy compared with three baseline methods.

• Earthquakes and Structures
• /
• v.22 no.4
• /
• pp.387-399
• /
• 2022

To study the seismic damage of masonry structures and understand the characteristics of the multi-intensity region, according to the Dujiang weir urbanization of China Wenchuan earthquake, the deterioration of 3991 masonry structures was summarized and statistically analysed. First, the seismic damage of multistory masonry structures in this area was investigated. The primary seismic damage of components was as follows: Damage of walls, openings, joints of longitudinal and transverse walls, windows (lower) walls, and tie columns. Many masonry structures with seismic designs were basically intact. Second, according to the main factors of construction, seismic intensity code levels survey, and influence on the seismic capacity, a vulnerability matrix calculation model was proposed to establish a vulnerability prediction matrix, and a comparative analysis was made based on the empirical seismic damage investigation matrix. The vulnerability prediction matrix was established using the proposed vulnerability matrix calculation model. The fitting relationship between the vulnerability prediction matrix and the actual seismic damage investigation matrix was compared and analysed. The relationship curves of the mean damage index for macrointensity and ground motion parameters were drawn through calculation and analysis, respectively. The numerical analysis was performed based on actual ground motion observation records, and fitting models of PGA, PGV, and MSDI were proposed.