Serco Inc.: Misclassification Essay

In many fields, such as medical, credit scoring and quality control research, one can obtain binary data that can occasionally be misclassified. In an example of medical field, a healthy patient could be falsely diagnosed as having a disease or an unhealthy patient could be falsely diagnosed as not having a disease. Bross (1954) was first discovered the problem of ignoring the misclassification that could conclude an extremely biased in the results from the binary data. Tenenbein (1970) introduced two methods to correct the bias that occurs during the misclassification of the data. The first method that he suggested to collect the data from the training session by using double sample scheme; the second method is to gather the prior sufficient

In my company, Serco Inc., is an operation center that controls retirement fund. One of the processes is to define the right beneficiaries when analysts are making a validity determination base on the existing applications and resources. Falsely defining the wrong beneficiaries can cause lawsuits and lead to millions dollars loss for the government agency. In Serco Inc., there are two types of employees to process beneficiary forms. The first type is analyst level, and many of them only hold a high school diploma and the analysts’ works can lead to many data misclassifications but they are not expensive. The second type is verification level, and major of them hold a bachelor degree with some sort of extensive working experience and the verifiers can have errorless in data classification but they are much more expensive. Due to the company’s economic issue, we need to hire certain amount of the employees to complete the job on time and satisfy the contracting agreement. We can’t hire all verifiers; hence, development a relative risk for the project is an essential to my