点击次数：

论文类型：会议论文

发表时间：2021-01-09

摘要：Nowadays, due to the development of information technology and the expanding network of knowledge covered all the aspects of fields, the progressively increasing number of data could be accessed in various approaches. However, while bid data has become a significant wealth for different disciplines included clinical medicine, it also poses an overwhelming issue which is enlargement of the scale of the missing data, owing to various circumstance. Moreover, at times, for some missingness, it is impossible to retest them, considering the cost invested in and effectiveness to impute them. To be more specific, taking a glimpse of current situation in the dataset of hospital system, clinical datasets, which mostly consist of longitudinal datasets derived from quantities of patients' multivariable laboratory test results which generally are impersonal, are usually dropped out prematurely while others may miss one or more assessments. Instead of deleting missing values, it has been recommended to impute them and ensure the imputation value could be generated in a fully automatic way, which researchers have been putting a great deal of efforts into since it was proposed. As a result, thanks to the challenge organized by Yuan Luo, there is an opportunity, considering about the particularity of clinical data, to discuss about the availability of developed methods in data imputation and the improvements for these diverse methods. ? 2019 IEEE.