Modified Convolution Neural Network (MCNN) based Multi-Class Identification for Domestic Violence Online Posts

  • Neeraj Varshney et al.


Domestic violence (DV) isn't just a significant wellbeing and welfare issue yet in addition an infringement of mortal rights. As of late, domestic violence crisis support (DVCS) bunches dynamic via web-based networking media have demonstrated fundamental in the help administrations gave to exploited people and their families. Profound learning models with the embedding’s approach have just exhibited unrivaled outcomes in online content characterization assignments. The programmed substance order will be addressing the issues related to versatility along with permitting DVCS gatherings to mediate in a split second with the precise help required. Previous work, DV used a Convolutional Neural Network (CNNs) algorithm to characterize the internet based life posts as either enlightening or non-instructive. The order precision of CNNs prepared on pre-prepared as well as DV-explicit embedding’s was then analyzed. In CNN group, each degree of preparing assesses the total blunder however assesses it after a given interim layer. In spite of the fact that the time lock diminishes, the aggregate mistake increments. To improve characterization precision in DV multi-class distinguishing proof proposed Modified Convolutional Neural Network (MCNN) calculation. The programmed substance arrangement will be addressing the issues related adaptability and permit DVCS gatherings to intercede right away with the definite help required. MCNN which permits the DVCS gatherings to proficiently deal with the high-volume and high speed information, assessing the issue idea finally reacting in a flash. The exact proof on a ground truth dataset utilized for results measurements are exactness, recall, F-measure and precision.

How to Cite
et al., N. V. (2019). Modified Convolution Neural Network (MCNN) based Multi-Class Identification for Domestic Violence Online Posts. International Journal of Control and Automation, 12(6), 179 - 189. Retrieved from