Supervised Machine Learning Approach Discovers Protective Sequence for Avoiding Sexual Victimization in Criminal Suit Documents

摘要: Previous research has studied effective self-protective behaviors, such as a victim’s physical resistance leading to the avoidance of sexual victimization. However, there are few studies on behavioral sequences, an offender’s violence followed by resistance. Our study aims clarify these sequences through supervised machine learning approach. The samples consisted 88 official documents assaults involving women, committed male offenders incarcerated in Japanese local prison. These crimes were classified completed or attempted cases based judges’ evaluations. All phrases each crime description also partitioned and coded according Penal Code. support vector identified most likely behaviors predict cases. Approximately 90% correctly predicted identification behavior sequences. sequence assault. general Victims’ need be interpreted from perspectives rather than single action perspective. methodologies may extract more effectively other methodologies. is fundamental part during Training focused protective contributes improvement training rape rates.