scholarly journals Kernel-based data transformation model for nonlinear classification of symbolic data

2022 ◽  
Author(s):  
Xuanhui Yan ◽  
Lifei Chen ◽  
Gongde Guo
2017 ◽  
Vol 9 (7) ◽  
pp. 662 ◽  
Author(s):  
Yan Xu ◽  
Qian Du ◽  
Wei Li ◽  
Chen Chen ◽  
Nicolas Younan

2008 ◽  
Vol 41 (2) ◽  
pp. 14600-14605
Author(s):  
Geert Gins ◽  
Jef Vanlaer ◽  
Ilse Y. Smets ◽  
Jan F. Van Impe

2020 ◽  
Vol 10 (2) ◽  
pp. 1-9
Author(s):  
Michael Bobias Cahapay

A curriculum does not exist in a void; internal members play a key role in responding to the different forces that continually shape it. One of the approaches to evaluation is through internal evaluation from the perspective of the inside members who work with the curriculum. However, the internal evaluation may pose restricted evaluation due to the innate subjective human judgment. Considering these contexts, this paper performed a pilot internal evaluation of a selected aspect of a higher education curriculum using a triangulation mixed method design called the data transformation model. Based on the results, the evaluation using the data transformation model probed important points of agreement and discrepancy in the data sets. The implications for evaluation theory and curriculum practice are discussed. It is suggested that an extension of the current formative internal evaluation continuing the tradition of data transformative model but progressively focusing on larger aspects of the curriculum should be further conducted.


Energies ◽  
2020 ◽  
Vol 13 (7) ◽  
pp. 1551 ◽  
Author(s):  
Lixing Chen ◽  
Xueliang Huang ◽  
Hong Zhang

The accurate modeling of the charging behaviors for electric vehicles (EVs) is the basis for the charging load modeling, the charging impact on the power grid, orderly charging strategy, and planning of charging facilities. Therefore, an accurate joint modeling approach of the arrival time, the staying time, and the charging capacity for the EVs charging behaviors in the work area based on ternary symmetric kernel density estimation (KDE) is proposed in accordance with the actual data. First and foremost, a data transformation model is established by considering the boundary bias of the symmetric KDE in order to carry out normal transformation on distribution to be estimated from all kinds of dimensions to the utmost extent. Then, a ternary symmetric KDE model and an optimum bandwidth model are established to estimate the transformed data. Moreover, an estimation evaluation model is also built to transform simulated data that are generated on a certain scale with the Monte Carlo method by means of inverse transformation, so that the fitting level of the ternary symmetric KDE model can be estimated. According to simulation results, a higher fitting level can be achieved by the ternary symmetric KDE method proposed in this paper, in comparison to the joint estimation method based on the edge KDE and the ternary t-Copula function. Moreover, data transformation can effectively eliminate the boundary effect of symmetric KDE.


Sign in / Sign up

Export Citation Format

Share Document