Please use this identifier to cite or link to this item:
|Title: ||Missing Values Estimation and Characteristic Modes Modeling of DNA Microarray Data|
|Authors: ||Cheng-Fa Cheng;Meng-Lin Wu|
|Contributors: ||NTOU:Department of Communications Navigation and Control Engineering|
|Keywords: ||Microarray;Characteristic modes;Gene expression;Missing values estimation;RBF NN|
|Issue Date: ||2011-10-21T02:36:04Z
|Publisher: ||2006 Automatic Control Conference|
|Abstract: ||Abstract:DNA microarray data often contains multiple missing values that can significantly affect the performance of statistical and machine learning algorithms. In this paper, imputation methods based on the fixed rank approximation algorithm (FRAA) are proposed to estimate the missing values in the gene expression data. Then applications of Radial Basis Function Neural Networks (RBF NN) to the characteristic modes modeling problem of gene expression will be investigated. The corresponding parameters of RBF NN (number of neurons, and their respective centers and radii) can be determined automatically. This task is often done by hand, or based in hillclimbing methods which are highly dependent on initial values. Genetic Algorithm (GA) will be employed to assist the search for the optimal RBF NN structure, and then the output layer weights can be learnt using the usual least mean square (LMS) algorithm. Furthermore, extra time points chosen halfway between the original points will be added to improve the modeling approximation of RBF NN. Three interpolation, linear interpolation, piecewise cubic Hermite interpolation and Cubic spline interpolation, are used to determine the locations of such extra time points. Finally, the Yeast Cell Cycle data set is illustrated to show the validation of the proposed modeling methods|
|Appears in Collections:||[通訊與導航工程學系] 演講及研討會|
Files in This Item:
There are no files associated with this item.
All items in NTOUR are protected by copyright, with all rights reserved.