Tag based Women Violence Prediction System using Deep Learning Techniques

In this Research article women violence Prediction system using Deep Learning Techniques to do novel approach for prediction of women violence. Now a day women violence increased day by day and it need to monitor the ratio of women violence in specific type of violence. This research article helps to predict the various types of women violence happens in throughout the globe. A system which is designed to predict specific category of women violence through the deep learning techniques. Convolution Neural network, Transfer Learning and MobileNetV2 algorithms used for women violence prediction. The model predict the Human Trafficking, Emotional Abuse, Gender Based violence, Sexual Harassment, Intimate Partner violence, Child molestation and Harassment, Brutality and Domestic violence etc. In this research article, tag based image retrieving technique used for 944 image dataset. Dataset consists of 9 category images that show the prediction of violence. Each women violence processed 105 tag based images approximately for prediction. With the help of this research one can easily predict the trend of women violence and take decision to minimize women violence rate. Preprocessing steps need for shaping the images uniquely, rescale the size of image, image annotation and clustering. Model building is the important process which needs creation of feature set and applying the method for tag based image classification. The experimental results of CNN model shows accuracy of 76% and transfer model shows the 75%. MobileNetV2 algorithms helps for device-based data processing and easy access the content that helps to user they can easily predict the women violence. This research article help to use the image datasets for women violence prediction to researcher sand studies prominent methods applied in deep learning algorithms to predict women violence.