Author(s)

Rui Yan, Ming Wang, Jiajing Li

Corresponding Author

Rui Yan

Abstract

The United States is in the midst of an opioid drug epidemic, which the Center for Disease Control has classified as a top public health concern, calling it the worst drug epidemic in US history. To address this problem, we analysed all US Census socia-economic data as well as the NFLIS data by using neutral network algorithm. And then, we use two ways of clustering to find the most possible locations to produce and output the opioid drug. In addition to it, we proposed several efficient strategy to solve this problem. After testing, our programs have heterogeneous effects on heroin crime incidents depending on counties pre-policy levels of prescriptions opioids per capital, with an 49-89% decrease in heroin crime within more opioid-dense conties.

Keywords

Neutral network algorithm, principle component analysis, Cluster analysis, mix-effect linear model.