Search

Abstract: Taiwan''s banking industry has experienced greatly structural changes since the implementing of Financial Holding Company Act in July 2001. The paper investigates whether Taiwan''s commercial banks establishing or joining in financial holding companies (FHCs) could promote their efficiency and productivity, as well as the determinants of efficiency and productivity changes of commercial banks. The paper applies a data envelopment analysis approach for calculating bank efficiency and a Malmquist total factor productivity index for measuring productivity change of banks. The results show that except for pure technical efficiency, other efficiencies and productivity of commercial banks do not improved because of their establishment of or joining in FHCs; on the contrary, it is just because of their better efficiency that they can firstly establish or join in FHCs. The results also reveal that, in the aspect of determinants of bank efficiency, bank size and overdue ratio have significant negative relations to technical efficiency of establishing or joining in FHCs’ banks, while equity-to-total asset ratio and loan-to-deposit ratio have significant positive relations to their technical efficiency; moreover, bank size and overdue ratio have significant negative relations to technical efficiency of banks with no established FHCs, while business diversification and loan-to-deposit ratio have significant negative relations to scale efficiency for these banks. Finally, in the aspect of determinants of bank productivity changes, overdue ratio has a significant negative relation to productivity growth of establishing or joining in FHCs’ banks, while business diversification has significant negative relations to both technological growth and productivity growth for banks with established and no established FHCs; furthermore, equity-to-total asset ratio and loan-to-deposit ratio have slightly significant negative relations to technical efficiency growth of banks with no established FHCs, while business diversification has a slightly significant positive contribution to their technical efficiency growth. [Copyright &y& Elsevier]

[Display omitted] • The small dataset problem is an important issue in manufacturing and academia. • A new Newton-VSG approach generates virtual samples with non-linear distributions. • The suggested method's efficacy is verified using two real datasets. • Paired Wilcoxon test elucidates the significance of differences among four VSG methods. In today's highly competitive environment, modeling the relationship between inputs and outputs using limited data for a management system at the early stages is important, but difficult. The virtual sample generation (VSG) method has been proposed in many studies to explore potential information to improve prediction performance of learning models for small datasets. However, those studies in general must assume an underlying distribution such as a linear triangular membership function or a Gaussian distribution to generate virtual samples. Thus, previous VSG methods may not effectively upgrade learning performance when the assumed distribution is not elastic enough for small datasets. To address this issue, in this paper, we proposed a novel VSG method called Newton-VSG method to generate virtual samples for small datasets with non-linear and asymmetric distributions. In the suggested method, we used Newton's method to estimate the minimum value of data range and the shape of data distribution. Further, we developed two deep learning models including Siamese network (SN) model for screening virtual sample input values and bagging auto-encoder (AE) model for predicting virtual sample output to ensure the quality of virtual samples. One real dataset for solidification cracking susceptibility test data and an other real dataset obtained from TFT-LCD process of a leading company in Taiwan were used to demonstrate the efficacy of the proposed method. On the partial least square regression (PLSR) and the back propagation neural network (BPNN) predictive models, we compared the proposed method with three state-of-the-art VSG methods in items of the mean absolute error (MAE) and the root mean squared error (RMSE). The experimental results demonstrated that the proposed method outperforms the other three VSG methods in prediction accuracy for small datasets. [ABSTRACT FROM AUTHOR]

Abstract: The use of dimension reduction techniques has attracted considerable attention owing to information explosion. Without considering the underlying phenomena of interest, traditional dimension reduction approaches aim to search a feature set for optimizing performance. In recommending entertainment videos, beyond the successful recommendations, marketing strategy can be benefited from interpreting precise social context information accurately. Therefore, how to find an easy-to-explain feature set to achieve optimal prediction performance becomes an important issue. In this paper, we propose a three-phase feature synthesis approach to search heuristically optimal feature set within exponential easy-to-explain features. The first phase performs feature selection by screening low-informative features, the second phase shrinks the high-dependent feature subset, and the third phase enhances the dominated features. An implemented social recommendation system and the 11 months purchasing data from the largest commercial entertainment video Web shop in Taiwan are adopted to evaluate the effectiveness and efficiency of the proposed feature synthesis method in the experiments. The experimental results show that our approach can obtain the interpretable clustering results as well as improve the recommendation. [Copyright &y& Elsevier]