Predict Opioid-using Pharmacists from the Prescribing Practices Based on SVM Classification Model

DOI

10.2991/assehr.k.220110.050How to use a DOI?

Keywords

Opioids; Pharmacists; Prescription; Support vector machines

Abstract

The number of people dying from drug overdoses in the United States is increasing every year, and most of them are caused by opioids, so it becomes crucial to analyze the use of opioids. Since an important influence on opioid use is the physicians who use opioids, we studied physicians’ medication habits to obtain an analysis of physicians’ opioid use habits and predictions of physicians’ propensity to use opioids. We first analyzed physicians’ opioid use through the physician medication use dataset provided by CMS, and analyzed several opioids that are used more frequently to analyze the correlation of physicians’ behavior toward different types of opioid use. After that, through doctors’ medication habits, for whether doctors will use opioid drugs to make predictions, by constructing support vector machine models, comparing the classification effects of different kernel functions, and finally constructing a classification model with 85% prediction accuracy.

Copyright

© 2022 The Authors. Published by Atlantis Press SARL.

Open Access

This is an open access article under the CC BY-NC license.

Download article (PDF)

TY  - CONF
AU  - Ximing Ran
PY  - 2022
DA  - 2022/01/28
TI  - Predict Opioid-using Pharmacists from the Prescribing Practices Based on SVM Classification Model
BT  - Proceedings of the 2021 International Conference on Public Art and Human Development ( ICPAHD 2021)
PB  - Atlantis Press
SP  - 257
EP  - 260
SN  - 2352-5398
UR  - https://doi.org/10.2991/assehr.k.220110.050
DO  - 10.2991/assehr.k.220110.050
ID  - Ran2022
ER  -