Predictive model based on machine learning for the detection of physically mistreated women in the Peruvian scope

Nemias Saboya, A. Angel Sullon, Omar L. Loaiza

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Violence against women in recent years in Peru showed alarming results, where 65.0 % of them were victims of some type of violence at some time in their lives. The study consisted of the elaboration of a predictive model that allows the recognition of the future physically mistreated woman. The data used was from a public survey of national scope. The methodology specified the data obtained; three supervised learning models Random Forest Classifier, Decision Tree Classifier, and Extra Trees Classifier were developed with the intention of buying the results and selecting that of the best performance. Regarding the results, Random Forest Classifier was the best model to obtain a precision of 51.0 % and a recall of 40.0 %, above those obtained by the other algorithms used. Subsequently, the best model passed to a calibration process where the average score was 0.7182, the precision was 0.76, and the recall was 0.31; concluding that the model classifies 76.0% of cases as physically mistreated women of which 31.0 % are effectively mistreated.

Original languageEnglish
Title of host publicationICAAI 2019 - 2019 the 3rd International Conference on Advances in Artificial Intelligence
PublisherAssociation for Computing Machinery
Pages18-23
Number of pages6
ISBN (Electronic)9781450372534
DOIs
StatePublished - 26 Oct 2019
Event3rd International Conference on Advances in Artificial Intelligence, ICAAI 2019 - Istanbul, Turkey
Duration: 26 Oct 201928 Oct 2019

Publication series

NameACM International Conference Proceeding Series

Conference

Conference3rd International Conference on Advances in Artificial Intelligence, ICAAI 2019
Country/TerritoryTurkey
CityIstanbul
Period26/10/1928/10/19

Fingerprint

Dive into the research topics of 'Predictive model based on machine learning for the detection of physically mistreated women in the Peruvian scope'. Together they form a unique fingerprint.

Cite this