Analysis and Prediction of Vertical Office Network Bandwidth Using a Backpropagation-Based Neural Network

Bandwidth is a time-series data, as its measurement is conducted continuously over a specific period.Bandwidth data records the extent of network utilization within a given timeframe [8], enabling the analysis of usage patterns based on historical trends.This analysis allows vertical offices to predict potential future network bandwidth requirements, which can serve as a reference for decision-making and as a basis for optimizing network bandwidth capacity planning more efficiently [9], [10].The Backpropagation-Based Neural Network algorithm is a supervised learning method designed to solve complex tasks such as identification, prediction, and pattern recognition [11], [12], [13].The predictive capability of the Backpropagation algorithm is developed through a process of learning based on historical data.This process allows the algorithm to identify and internalize complex patterns within the data, thereby enabling it to produce predictive results for future data [14], [15].The study of electricity consumption forecasting in North Sumatra [16] highlights the increasing demand for electrical energy, requiring accurate predictions to support effective policy-making.Forecasting is crucial for ensuring sufficient supply, as electricity demand is influenced by population growth, customer categories, and regional economic activity.The Backpropagation Neural Network method was applied using six input variables, including