Data Mining Applications in Risk Research: a Systematic Literature Review

Abstract

Despite the rising literature on data mining (DM) approaches, there is a lack of a complete literature review and categorization system within risk research. This paper presents the first recognized academic literature review on the application of data mining tools in risk research provides an up-to-date SCOPUS literature database. Based on bibliometric analysis, 5422 papers related torisk were identified from a total of 77,410 studies on data mining and thoroughly analyzed. Each of the selected 5422 papers was classified into four risk categories: global risk, public health risk, molecular and biomedical risk, and pharmaceutical risk. Each primary risk category was further subdivided to highlight the specific research focuses within each domain. Global risks encompass business, environmental, and social risks. Scholars have predominantly focused on the banking, market, and construction sectors within business risk, while environmental risk includes catastrophe-related risks. Social risks encompass areas such as education, traffic safety, and transportation concerns. Clinical data is usually employed in public health risk research, while various radiomic databases are utilized in genetic and molecular biology research. In pharmaceutical research, DM is primarily used to detect adverse drug effects. According to the findings of this review, the fields of computer science and medicine received the most significant research attention. The review also discusses limitations and provides a roadmap to guide future research, aiming to enhance knowledge development related to the application of data mining techniques in risk-related studies.