Research And Development Of Feature Modeling Algorithms For Staging And Survival Of Head And Neck Squamous Cell Carcinoma

With the development of society and the Internet,people will generate more and more data in their daily activities.Many researchers have used this data to do a lot of research.For example,some recommendation algorithms on shopping websites can use a consumer's past consumption records to predict the products that this consumer may need.Merchants can also use the consumer record of many consumers to help develop the right product package.Data has become an important resource in today's society,and data can be used to mine a lot of important information.Then the human body itself has a lot of data.So whether you can use this data to mine some important information.Common genetic data includes transcriptome data,methylation data,etc.Transcriptome data is primarily used to study the type and amount of various RNA(usually m RNA)molecules produced in a single cell,or a cell population of a particular type of cell,tissue,organ,or developmental stage.Transcriptome data measure the abundance and type of genes expressed in a particular sample.The abundance of m RNA refers to the average number of each m RNA molecule in each cell.Methylation generally means that DNA cytosine is methylated in a specific manner under the action of an enzyme.A number of studies have found that DNA methylation levels affect the development and progression of human tumors in some way.The level of DNA methylation can be used to judge the staging of tumors and the prognosis of tumors,which is of great significance for the treatment of tumors.More than 90% of cancers in head and neck cancer are head and neck squamous cell carcinoma.Head and neck squamous cell carcinoma is one of the top ten common cancers in the world and has a low survival rate.It would be important to find a genetic marker associated with head and neck squamous cell carcinoma and to effectively predict the stage of the disease and the survival time of the patient.This article mainly uses the transcriptome data and methylation data for staging and survival time prediction.Methylation data and transcriptome data belong to primary head and neck squamous cell carcinoma.First,both methylation data and transcriptome data have tens of thousands of features.We used the designed algorithm to select features in the combined dataset,that is,to select features that are useful for distinguishing staging of head and neck squamous cell carcinoma from a large number of features.At the same time,because the data sample distribution is not balanced,that is,the number of samples in the IV period is the largest,accounting for about 60% of the total sample size.Therefore,we designed the "OROO" method to solve this problem and improve the prediction effect.In this paper,after feature extraction and parameter optimization,154 features were selected from the original tens of thousands of features.Among the 154 features,144 features belong to transcriptome data,and 10 features belong to methylation data.Finally,we predicted the staging of head and neck squamous cell carcinoma with high accuracy,with an average absolute error of 0.027 and an accuracy of 97.98%.It also proves that the features we have chosen are closely related to the staging of head and neck squamous cell carcinoma.Next,we used the selected features to select the 31 features again,and predicted the survival time of the samples with an average absolute error of 14.175 months.It also shows that these 31 features are closely related to survival time.This experiment is of great significance for the study of head and neck squamous cell carcinoma and the development of appropriate treatment measures.Finally,the binary classification of staging of head and neck squamous cell carcinoma was performed using the SFMC method proposed in this paper.And compared with the current binary classification effect of staging of head and neck squamous cell carcinoma,our experimental method has achieved better results.