Some Researches And Applications On Risk-adjusted CUSUM Chart

Statistical process control refers to the analysis and evaluation of a certain process of the research object by means of mathematical statistical methods,the abnormal process is found in time through feedback information,and corresponding effective measures are taken to make the process at a normal level.Control charts are the most important tool for statistical process control,where the cumulative sum(CUSUM)chart is more sensitive when monitoring small drifts.In the old days,many monitoring processes,people usually only monitored the result data,but in the actual situation,there are many inevitable objective conditions that will affect the ultimate result.If these objective factors are not considered in the actual monitoring process,it will increase the probability of false alarms in the control charts,and cause waste of resources in order to find the cause of the false alarm.So we can’t monitor all situations with commonly used standards,but take these influencing factors into account in the chart.Using the riskadjusting CUSUM chart to solve practical problems,this paper mainly explores the following two aspects:(1)An uncontrollable flu event poses a threat to public health.The control chart can monitor the flu and give early warning of an outbreak.The number of flu cases per day in a given area is generally counted at the same time interval,usually following a Poisson distribution.In past surveillance,we often assumed that the daily number of influenza cases followed the same Poisson distribution in the same area,that is,the Poisson mean parameter is a constant.But in real life,the number of people infected with influenza every day is not the same distribution,will be affected by such factors as temperature and air pollution,ignoring these factors may make the control chart false alarm and thus affect the decision-making of the disease control departments.Taking these factors into account,a CUSUM control chart for risk adjustment of Poisson distribution is proposed in this paper,reduce the waste of resources in searching for the specific causes of runaway influenza.It can be seen that the risk-adjusted Poisson CUSUM control chart can predict and monitor the spread of influenza timely and accurately,it is a good tool for public health.(2)An important indicator of neonatal health is neonatal birth weight,which studies have shown is closely related to certain chronic diseases,it is of great significance to detect the abnormal weight in time so as to avoid the occurrence of some diseases.The weight of the newborn usually follows the normal distribution.Because the weight of the newborn is related to the sex,the age of the pregnant mother and the degree of pregnancy,so the weight of every newborn does not follow the same normal distribution,it is not reasonable to monitor the weight of different newborns by the same standard.Therefore,a risk-adjusted CUSUM control chart for a normal distribution is proposed,which can avoid the misjudgment caused by not considering the influence factors of newborn weight,it is an effective tool for medical researchers to find out the cause of abnormal weight of newborn and diagnose the pregnant mother and her newborn.