Medical card data imputation and patient psychological and behavioral profile construction

Missing data is a typical problem for many hands-on tasks and researches, which has required human intervention and contributed to an increase in errors during algorithms application that demand for a large number of metrics. Solving this particular problem is essential for medicine and healthcare, because it allows more easily diagnosing certain types of diseases, improving medical service quality, etc. The main approach for medical data imputation is to automate this process at all stages, beginning from finding the NA (Not Available) or missing Data, to the completion of the analysis and insertion of lost information entity. The proposed methods of mathematical computing and modeling, statistical functions, data-flow diagrams during the imputation, and the use of computer programming tools should be implemented ni the medical-field to improve and address the missing data issue. The evaluation of key characteristics (algorithin's error, number of imputed data, datasets dimensionality) helped to determine the factors for obtaining the most accurate result with the help of various algorithms and functions. The study is useful for the medical industry in general, since it will eliminate the missing data values in patient medical records by applying statistical methods and artificial intelligence, which will significantly shorten the automation of large datasets processing and thcilitate their descriptive and exploratory analysis during further data discovery to identify certain patterns and features. (C) 2019 The Authors. Published by Elsevier B.V.