LOAD FORECASTING IN HETEROGENEOUS TELECOMMUNICATION NETWORKS BASED ON THE DEVELOPED NEURAL MODEL

The paper addresses the relevant scientific challenge of traffic load forecasting in heterogeneous telecommunication networks using artificial neural networks. With the rapid deployment of 5G technologies, IoT devices, smart city infrastructures, and mobile cloud computing, there is a growing need for intelligent models capable of analyzing vast volumes of telecommunications data and accurately forecasting network load. One of the key challenges in heterogeneous networks is the uneven distribution of traffic across different communication technologies (e.g., LTE, Wi-Fi, NB-IoT), which requires adaptive resource management strategies. The proposed model is based on a multilayer perceptron (MLP) neural network trained on time series data representing network component loads. Input features include previous traffic load values, time-based characteristics, network type, and QoS levels. The model is trained using backpropagation and the mean squared error (MSE) loss function. The output of the model is a predicted load value for a given future time interval. The study provides a detailed algorithmic description of the model's operation, including its mathematical formulation, objective