Bayesian Analysis Of Sensitivity Feature Ratio Based On Multi-option Non-randmized Model

Today's society is increasingly inclusive and open,but there are still various conflicts of ideas and ideas.These problems or phenomena are showing a diversified trend.In order to resolve conflicts and build a harmonious society,we must first have a comprehensive and accurate understanding of them.However,in the course of the investigation,we often encounter respondents who are unwilling to answer or give false answers to certain problems.Answer,this will affect the accuracy of the final survey results.It is noted that most of these issues involve the privacy or grey interests of respondents,often referred to as sensitivity issues.For example,whether the test is cheating,whether there has been derailment,whether it has participated in gambling,etc.,or even some political issues.Many statisticians at home and abroad have conducted a lot of research on the related technologies of the investigation of sensitivity problems.The focus of these studies is to protect the privacy of respondents,reduce their insecurities,improve loyalty and cooperation,and obtain scientific and comprehensive investigation results.This paper introduces a randomized answering model and a non-randomized answering model for sensitivity problems.These are two typical types of sensitivity survey techniques.In the randomized answering model,the Mangat model,the two-stage randomized answering model,and the randomized answering model of the sensitivity problem of multiple classifications are introduced in the Warner model,the non-correlated randomized answering model,and the other two stages are not related.The problem randomization answering model is based on the non-correlation problem randomization answering model and the two-stage randomized answering model.The multi-classified non-correlation problem randomization answering model is a combination of the non-correlation problem randomization answering model and the characteristics and advantages of the randomized answering model of multiple classification sensitivity problems.Triangular models,crossover models and parallel models,as well as multi-option non-randomized answering models are typical non-randomized answering models,each with its own characteristics and applicable context.The main work of this thesis is based on the multi-option non-randomized answering model,which divides the sexual orientation into four categories,namely,homosexuality,heterosexuality,bisexuality and asexuality,and introduces the unrelated problem birth month to the privacy of the respondent.Protection also divides the non-related problems into four categories,so as to conduct a sample survey and analysis on the sexual orientation of students in a university.