Automatic Domestic Violence Detection through Textual Data using Deep Learning Techniques

سال انتشار: 1403
محل انتشار: سومین کنفرانس بین المللی و هشتمین کنفرانس ملی کامپیوتر، فناوری اطلاعات و کاربردهای هوش مصنوعی
کد COI اختصاصی: CEITCONF08_024
زبان مقاله: انگلیسی
تعداد مشاهده: 160

نویسندگان

Master’s student of IT engineering, Computer engineering and Information Technology Department, Shiraz University of Technology, Shiraz, Iran, ۷۱۵۵۷-۱۳۸۷۶

Pirooz Shamsinejadbabki

Assistant Professor, Computer engineering and Information Technology Department, Shiraz University of Technology, Shiraz, Iran, ۷۱۵۵۷-۱۳۸۷۶

چکیده

Domestic violence is a serious social issue that remains hidden in many societies, and identifying it is difficult due to various factors such as fear, low self-confidence, or lack of awareness of legal rights among victims. Given the importance of this issue, research has been conducted on identifying domestic violence using artificial intelligence and machine learning techniques. This issue, as a social and psychological problem in various communities, requires precise and efficient tools for detection and analysis. In this study, the aim of this study is to explore the application of machine learning in analyzing content related to domestic violence in conversational data collected from family members. Due to the lack of access to real and collected data from conversations between family members, textual data from subtitles of movies containing domestic violence scenes was extracted, and two deep learning models, CNN and LSTM, were utilized for predicting and identifying domestic violence. The dataset consists of ۳۶۹ samples, with ۱۶۴ representing non-violent scenes (happiness, excitement, calmness, e.g.) and ۲۰۵ representing violent scenes. The sampling method was random. Data analysis was performed using Python and various machine learning libraries, including TensorFlow and Keras. The CNN model was employed to identify complex features in the textual content, while the LSTM model was used to preserve long-term dependencies in the data. The results revealed that the CNN model achieved higher accuracy in identifying content patterns related to domestic violence, while the LSTM model performed better in preserving the semantic dependencies of the sentences. Overall, the final accuracy of the models for CNN and LSTM reached ۸۸.۸% and ۸۸%, respectively. These findings indicate the high efficiency of deep learning models in detecting domestic violence in textual data. It can be concluded that these models could serve as useful tools in identifying and preventing domestic violence in today's world.

کلیدواژه ها

Domestic violence, machine learning, text analysis, CNN, LSTM

مقالات مرتبط جدید

اطلاعات بیشتر در مورد COI

COI مخفف عبارت CIVILICA Object Identifier به معنی شناسه سیویلیکا برای اسناد است. COI کدی است که مطابق محل انتشار، به مقالات کنفرانسها و ژورنالهای داخل کشور به هنگام نمایه سازی بر روی پایگاه استنادی سیویلیکا اختصاص می یابد.

کد COI به مفهوم کد ملی اسناد نمایه شده در سیویلیکا است و کدی یکتا و ثابت است و به همین دلیل همواره قابلیت استناد و پیگیری دارد.