Generation method of teenager suicide risk prediction model and prediction system

Abstract

The invention provides a method for generating a suicide risk prediction model, which comprises the following steps of: 1, using a psychological health questionnaire on an online platform, and carrying out questionnaire survey on a part of middle and primary schools in our city; 2, after the questionnaire is recycled, depression, anxiety, gender, age and the like serve as independent variables, suicide high risks serve as dependent variables, binary logistic regression is used, and high-risk factors influencing suicide behaviors are analyzed. And 3, taking the suicide high risk as a prediction target, taking depression, anxiety, gender, age and the like as risk factors, establishing a mathematical model of sample data by using an XGBoost machine learning method, and performing effective prediction on the suicide high risk teenagers. In addition, performance comparison with seven different machine learning methods including Gaussian naive Bayes, a support vector machine, a random forest and the like is carried out on the same data set, and the accuracy and reliability of the XGBoost model are further proved. According to the invention, the data acquisition is simple and convenient, the identification of the suicide possibility of the teenagers is realized by using the simple and understandable questionnaire and artificial intelligence technology, and powerful support is provided for prevention and intervention of psychological crisis.

Description

technical field [0001] The discovery relates to the fields of psychiatry, psychology, epidemiology and artificial intelligence, and more specifically relates to a generation and prediction system of a suicide risk prediction model. Background technique [0002] The latest domestic statistical and epidemiological surveys show that the overall prevalence of psychological abnormalities among children and adolescents in my country is 15.6%, among which the prevalence of depression is about 2.8-4.6%. Depression is the main cause of adolescent suicide. According to data from the Global Burden of Disease Study released by the World Health Organization (WHO), the risk of suicide caused by depression peaks in people aged 10-24, ranking first among all causes of death in this age group. Suicide not only endangers the life safety of young people, but also causes huge psychological trauma to relatives and friends, and creates a huge economic burden on the whole society. The study found...

AI Technical Summary

Problems solved by technology

[0003] There are currently few tools for predicting individual suicide risk for children and adolescents, including the Columbia-Suicide Severity Rating Scale for Children (C-SSRS), the PANSI scale (Positive and Negative Suicide Ideation) and the SIQ-JR scale (Suicidal Ideation Questionnaire - Junior ), through these tools, it is possible to assess the possibility of suicidal behavior of the work object in the future, but these methods are still seldom used in my country, and all rely on the self-report of the subjects, and require the active cooperation and active response of the subjects. and costly

At present, some overseas researchers have developed suicide risk prediction tools based on the visit data of the healthcare system, but due to their low sensitivity and low positive predictive value, most people think that these tools have no clinical value, so they are currently less used , more at the scientific level

Moreover, the current predictive models are mainly used in cross-enterprise health care systems, including the US Department of Defense, the US Department of Veterans Affairs, and large medical institutions. There is no suicide prediction model for ordinary school children and adolescents.

Images