Suicide prediction in workers using Naive Bayes

• DOI: 10.22533/at.ed.8182308129

• Palavras-chave: -

• Keywords: Suicide, Typical testors, Prediction.

• Abstract: The suicidal tendency is a noticeably big problem in Aguascalientes, Mexico, especially in the young and working population, which constitutes more than 50% of completed suicides. This brings with it a great affectation to the economy of the state, not only when the workers manage to commit suicide, but also when it is only attempted, since their work responsibilities are stopped. In this study, a database compiled by the psychology department of the Autonomous University of Aguascalientes with factors associated with mood is analyzed, in which we can find features associated with sleep disturbance, self- esteem problems, and even affectations in the weight. This database contains information on people with suicidal tendencies, and a control group. To identify the key features that suicidal people present, the total set of typical testors was obtained and the informational weight of each feature was obtained. In the same way, a predictor was made using a classifier based on the naive bayes theory, analyzing its effectiveness with the total set of features, and using only the best features, with an informational weight greater than 40%.