ART2 neural network - based gene expression data analysis


The third level of Class 8 to 10 banner and the Class A ll, class of 12 stage ... results all levels of the E stage E stage of the third level and the P - stage , two levels of P - phase of all levels of P - stage the Lan level and stage A and data set imported into the traditional ART network , the network parameters remain unchanged under the classification of the results from table 3 . Table 3 traditional ART2 neural network operating results in the different parameters of similarity Meng quit the results obtained by the traditional ART in different levels of alert function analysis of genes which point in time can not be obtained stable class class necessarily linked table 2 shows that , due to the use of distance vigilance parameter ,ugg greece, the same stable class of genes similar point in time . 3 Conclusion gene expression profiling data analysis of an important emerging field as bioinformatics . Attracted much attention over the years many scientists have used various methods of gene expression data analysis system in this experiment , we mainly use the adaptive resonance theory network to complete the analysis of gene expression data , we improve the traditional network of ART2 neural network structure , to solve the problem of genes with different expression levels on the same phase can not be separated ,franklin marshall, and the improved ART2 neural network can basically different gene expression levels to distinguish between the gene which point in time , get a better result . We believe that the future direction of efforts : (1) to further discuss the use of neural network in the gene expression data analysis area,mbt sko priser, the availability and expansion of neural networks ,nike free run, gene expression data analysis. , We hope to get the gene expression data analysis system of the correct rate higher ; (2) further study of the neural network based on the biological characteristics of the data set we used in this article are obtained through the experiment with the biological characteristics of the data , we think we can further the study of neural network structures and algorithms , so that it can reflect the biological characteristics of the data . [
],doudoune moncler pas cher,[