Tools
Tools
Farrelly, Megan (2022) A Classification Approach to Identifying Female Victims of Intimate Partner Violence in Europe and the US. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration manual)
Download (4MB) | Preview |
Abstract
Violence against women is a global public health issue. However, it can be difficult to recognise if a woman is suffering abuse. Intimate partner violence is a subset of this abuse and few studies focus on identifying abuse caused by a partner or expartner using classification techniques. The studies that have been completed tend to apply the same models but report different performance metrics. Little work has been done to determine how to improve these models within this domain. To address these knowledge gaps, this study applied nine classification models, including ensemble, boosted, stacked and deep learning techniques, to determine which model was most appropriate to identify women suffering intimate partner violence. It was found that Random Forest returned the highest accuracy and AUC. XGBoost, Support Vector Machine and a stacked classifier also returned favourable metrics, while the deep learning techniques tended to perform poorly. It was found that reducing the number of features input into a Random Forest model reduced the average accuracy returned but maintained the error of the model. By reducing the number of data points input, accuracy was maintained but error increased. The findings suggest that geography plays a role in the rate of violence suffered, along with the number of people living in the same household. These are novel findings which may aid future classification studies and guide them when resources such as data availability are limited. Overall, it is hoped that this research will aid stakeholders such as healthcare professionals or women’s charities, to improve risk assessments or to identify women at risk of abuse using an appropriate model identified by this study. The recommendations made by this study could also alleviate the issues faced by women suffering violence by facilitating better government and business decisions at a community level.
| Item Type: | Thesis (Masters) |
|---|---|
| Subjects: | Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science H Social Sciences > HQ The family. Marriage. Woman > Domestic Violence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning R Medicine > RA Public aspects of medicine > Public Health System |
| Divisions: | School of Computing > Master of Science in Data Analytics |
| Depositing User: | Tamara Malone |
| Date Deposited: | 24 Jan 2023 14:54 |
| Last Modified: | 03 Mar 2023 12:29 |
| URI: | https://norma.ncirl.ie/id/eprint/6118 |
Actions (login required)
 |
View Item |