Research On Discovery Algorithm Of Glioblastoma Subtypes Based On Patter Recognition

Nowadays,data analysis in other fields has become widespread.Especially in the fields of biology and medicine through computer science and technology,there are many data that require urgent analysis,and they cannot be uniformly and effectively implemented through traditional medical methods,or through traditional computer technology it cannot be efficient to explore the professional information behind the data,and many biological information that is instructive to clinical medicine is hidden under them.Similarly,the rapid development of bioinformatics data also requires a noval and more efficient processing method to solve the current problems.Glioblastoma also known as GBM is a difficult cancer.Researching on glioblastoma-related biological information,this paper try to use data mining and pattern recognition technologies to explore the knowledge hiding behind the genetic information.Considering the processing problems existing on the GBM data,how to effectively solve it becomes more and more necessary.This paper conducts multiple GBM data and analysis them,including firstly using conventional experience-based methods for data preprocessing,using the density characteristics of data features to show GBM subtypes more user-friendly and clearly visible,and at the same time presenting a method based on rLDA and the MRMR data reduction process,apply the RELRED algorithm to reduce the data dimension and perform redundant processing.Finally,in order to clearly discover the feature subsets of GBM subtypes,in other words,the relevant features which can guide the significance for GBM classification,this paper proposes the SJNMF algorithm based on feature recognition.The SJNMF algorithm decomposes the joint matrix uniformly in the iterative process of the algorithm.The result of the decomposition is closely related to the tagging information of the classified data.In the end,not only the requirement for feature reduction of the high dimension bioinformatics data is achieved,but also the extraction of important genetics features is completed.The combination of feture reduction and feature selection algorithms makes it possible to dig deeper into GBM data.The project uses the relevant data sets in the TCGA cancer database.The TCGA data is analyzed according to the data processing flow in this paper.By comparing with other commonly algorithms.The results show that the advantages of RELRED processing flow and the SJNMF algorithm in computeing speed and calculation accuracy...