Suicide is recognized as a global factor contributing to both mortality and disability (Klonsky et al., 2016). According to the World Health Organization (2025), the number of individuals who attempt suicide significantly exceeds those who die by suicide, with more than 720,000 suicide deaths reported globally each year. Since 2003, South Korea has consistently exhibited the highest suicide rate among the member countries of the Organization for Economic Cooperation and Development (OECD), with the 2024 figure exceeding twice the OECD average (OECD, 2024). In South Korea, suicide ranks first in individuals aged 10−30 years and fifth among causes of death overall (Statistics Korea, 2022). Consequently, South Korea is recognized for its exceptionally high suicide rate.

Suicide represents a long-term progression, signifying an internal process initiated by reactions to one's environment (Hawton and Van Heeringen, 2000; Van Heeringen et al., 2000; Paykel et al., 1974). This process begins with hopelessness, often expressed as feelings that life lacks meaning or vague thoughts about dying. These may evolve into more persistent suicidal ideation, followed by serious contemplation of suicide, the development of a suicide plan, and ultimately, suicide attempts—potentially culminating in suicide completion. Intervention for suicide risk aims to impede the stages of this process. While risk factors related to the suicidal process, such as anhedonia (Nock and Kazdin, 2002), impulsiveness (Zouk et al., 2006), and high emotional reactivity (Nock et al., 2008a, Nock et al., 2008b), are well-documented, there is limited knowledge on how to classify individuals with various suicide risk factors into high-risk groups. Therefore, research is necessary to classify individuals at high risk of suicide and provide timely interventions.

The prevalence of psychiatric disorders among individuals who die by suicide is reportedly as high as 95 % (Litman, 1989), emphasizing a significant correlation between suicide and mental health conditions. Bipolar disorder and major depressive disorder include suicidal behavior as a symptom of their diagnostic criteria (American Psychiatric Association, 2013), with approximately 20 % of patients with bipolar disorder and 15 % of those with major depressive disorder dying by suicide (Frances et al., 1986; Jamison, 1986; Johns et al., 1986; Roy and Linnoila, 1986). Studies have shown that comorbidity with anxiety disorder is a significant predictor of suicide risk among patients with mood disorders (Antypa et al., 2016). Furthermore, individuals characterized by high neuroticism coupled with low extraversion, or high harm avoidance and novelty seeking, are at increased risk of suicide (Liu et al., 2017). However, standard logistic regression analysis used in previous studies only investigates the surface relationships between suicide risk factors, making it challenging to apply these findings directly in clinical practice (Ilgen et al., 2009).

Data mining approaches identify the most significant risk factor within the data and discern subgroups by grouping specific risk factors (Obenshain, 2004). In data mining, decision tree analysis is a technique for classifying data into subgroups, providing a visually understandable model for classifying variables related to a dependent variable. This method allows the examination of variables associated with suicide risk and provides insights into how these variables interact to contribute to suicide risk sequentially. Therefore, this study aimed to use data mining techniques to identify factors that optimally classify patients with mood disorders into high-risk suicide groups, providing foundational evidence for interventions during suicide crises.