Factors Influencing the Phenomenon of Suicide (from Ideation to Action) Using the Decision Tree Algorithm

Mahna, Aminreza; Moheb, Naeimeh

Volume 29, Issue 9 (12-2022) RJMS 2022, 29(9): 345-353 | Back to browse issues page

Research code: 01

Ethics code: 01

Clinical trials code: 01

Mendeley

Zotero

RefWorks

Mahna A, Moheb N. Factors Influencing the Phenomenon of Suicide (from Ideation to Action) Using the Decision Tree Algorithm. RJMS 2022; 29 (9) :345-353
URL: http://rjms.iums.ac.ir/article-1-8523-en.html

Factors Influencing the Phenomenon of Suicide (from Ideation to Action) Using the Decision Tree Algorithm

Aminreza Mahna

, Naeimeh Moheb

Assistant Professor of Clinical Psychology Department , Islamic Azad University,Tabriz Branch ,Iran , moheb.n@yahoo.com

Abstract: (397 Views)

Background & Aims: Suicide is perhaps one of the saddest events in a person's life and one of the most painful problems in human society. The purpose of this research was to investigate factors affecting suicide from ideation to action based on Klonsky's theory in non-clinical populations. A person who can be said to be a masterpiece, intelligent, how can he decide to end his life and prefer not to live? Although many scientific and medical efforts are trying to protect people from diseases and injuries and perhaps increase their life span, on the other hand, we are witnessing an increase in the number of suicides and the unwillingness of a group of people to continue their lives. In recent decades, although studies have tried to reduce the number of suicides and somehow control this problem, we can practically see that the efforts have not had significant results, and relatively more modern humans have taken the path of giving up life than in the past and their primitive ancestors. At present, suicide is a major cause of concern and a serious problem for global health. The organization's statistics show that in 2023, worldwide, nearly 700,000 people will end their lives by suicide each year. Evidence suggests that for every adult who dies by suicide, there may be more than 20
Methods: : This research is basic in terms of its purpose and the method of carrying out descriptive research. The statistical population of this research was formed by all the people present in the social networks of Shahretbriez who could answer the questionnaire through the link to answer the questions. A sampling method was available and 638 were selected (based on about 20 people for each component of the model). The final sample was reduced to 607 people due to the failure of some participants in the final registration of answers. The entry criteria for this study were to have at least basic literacy to answer the questions and age (16-75), to have full consent, and also to be a volunteer for cooperation. To collect information, the questionnaires were placed in the dedicated link in social networks, and it was suggested that people participate in the research. In addition to ensuring the confidentiality of the participants, ethical considerations were observed in all stages of implementation. In this research, the data mining method and the decision tree technique were used to identify factors affecting suicide and provide a model to predict it. Finally, finding the answers to the study questions has been used. It is worth mentioning that in this study SPSS Modeler version 18 was used for data analysis.
Results: The research findings showed that there is a direct and significant relationship between the variables of mental pain, perception of being a burden, hopelessness, and neutral affiliation in the order of importance in the possibility of suicidal ideation, and also between the variables of suicidal thoughts, suicide readiness, despair, fearlessness about death, affiliation. There is a significant relationship between pain neutrality and pain tolerance in the order of importance in predicting suicide attempts. The parental bond variable did not have a significant effect on the prediction of ideation and suicide attempts. The probability of correct prediction in the training and evaluation groups in the C5.0 algorithm is 92.2 and 88.65, respectively, while on the other hand, the probability of correct prediction in the training and evaluation groups in the CART algorithm is 85.11 and 80, respectively. Two other indices show the optimality of the tree model (AUC and GINI indices) in the two training and experimental groups of the variable of suicidal ideation by separating the C5.0 and CART algorithms. According to the overall importance of the studied variables, the variables of mental pain and perception of burden were more effective than other variables in predicting suicidal ideation. In this way, with the increase in the amount of psychological pain, the perception of being a burden, disappointment, and neutral belongingness, the possibility of suicidal ideation increased. On the other hand, the possibility of suicidal ideation decreased in lower amounts of psychological pain and mild hopelessness. Also, among the two levels of interpersonal relationship variables (perception of burdensomeness and neutral affiliation), perception of burdensomeness was more effective than the variable of neutral belongingness in predicting suicidal ideation, so that as the score of perception of burdensomeness increases, the probability of suicidal ideation is higher and lower. As it went, its score would decrease. Therefore, it can be concluded that the variable of psychological pain and the perception of being a burden from the variables of interpersonal relationships had a more transient effect on the probability of suicidal ideation in the participants in the study. According to the results of the modeling of the prediction of suicidal ideation in the present study, it seems that Most of the variables considered in the study have been able to show a significant effect on the occurrence of this phenomenon, but at the same time, the results of the statistical technique used in the study indicate that the variables of parental bonding (care and control) included in the studied model have not been able to have a significant effect.
Conclusion: The variable of psychological pain showed the most importance in predicting the formation of suicidal ideation in the decision tree model, and also high values of psychological pain and despair, perception of being a burden, and neutral affiliation showed a higher probability of ideation in the model. In this study, among the three factors of suicide capacity, readiness to commit suicide, fearlessness of death, and pain tolerance, the variable of readiness to commit suicide showed the most importance in predicting suicide attempts. In terms of the decision tree model, it became important as a variable that was present in all the paths leading to suicide, along with suicidal ideation. In this study, among the three factors.

Keywords: suicide ideation, suicide attempt, pain, hopelessness, suicide

Full-Text [PDF 1190 kb] (151 Downloads)

Type of Study: Research | Subject: Clinical Psychiatry