https://scholars.lib.ntu.edu.tw/handle/123456789/521811
Title: | Health outcome prediction using multiple perturbations | Authors: | WEN-CHUNG LEE | Keywords: | big data; biostatistics; data mining; outcome prediction; predictive analytics | Issue Date: | 2020 | Publisher: | Lippincott Williams and Wilkins | Journal Volume: | 99 | Journal Issue: | 2 | Source: | Medicine (United States) | Abstract: | Public health workers and medical practitioners are frequently required to make predictions regarding various health outcomes. However, a prediction with nearly 100% certainty is seldom possible.If a person has a health outcome of concern or is in the process of developing the outcome, many attributes of that person may undergo subtle changes - the perturbations. We propose a method, namely "prediction using multiple perturbations" and investigate its asymptotic properties when the number of attributes tends to infinity. This is a proof-of-concept study.The proposed method can predict the health outcome of a person to near certainty if personal data with billions or trillions of attributes can be collected and 4 conditions (described subsequently in this paper) are met.Collecting personal data with billions or trillions of attributes may someday become possible in the current era of big data. If such information can be obtained, theoretically we can predict the health outcome of a person to near certainty. ? 2020 the Author(s). Published by Wolters Kluwer Health, Inc. |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85077759051&doi=10.1097%2fMD.0000000000018664&partnerID=40&md5=869f0f4621dc466400e43e5ab1794eae https://scholars.lib.ntu.edu.tw/handle/123456789/521811 |
ISSN: | 0025-7974 | DOI: | 10.1097/MD.0000000000018664 | SDG/Keyword: | adult; article; big data; biostatistics; data mining; human; prediction; proof of concept; algorithm; prognosis; Algorithms; Humans; Prognosis |
Appears in Collections: | 流行病學與預防醫學研究所 |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.