Violence Detection in Schools Based on Multi Fusion Sensor and Optimized Relief-F Algorithm

In elementary schools, bullying is a widespread social issue. Early school violence is considered more destructive than other school violence around the world. It is advisable to use fusion and advanced relief-F instructions as a multi-censor-based approach for early school violence identity. Two important Motion sensors are collecting data in the detection of violence in elementary schools and our daily practices. There are nine different categories of violence. The qualities and features of the time and frequency of a province are restored and filtered by an improved relief-F system. The authors create a two-level classification next. The first-level assortment separates the random forest jump function from others. In the previous paper, the author used an end wood assortment; However, in the recommended work, a random forest assortment is used, and the next level uses a network with radial basic functions to identify the remaining 9 types of functions. An end layer fusion system is used later to combine the results of the recognition of the two sensors. According to our research, everyday life can be recognized with 98.1% accredited accuracy, while elementary school violence can be recognized with an average of 85.9% accuracy.