A FORECASTING METHOD FOR PUBLIC SWITCHED TELEPHONE NETWORK CALLS USING NEURAL NETWORK

摘要

A method for forecasting calls with the neural network in the PSTN(Public Switched Telephone Network) is provided to offer a convenient and precise forecasting system, offer a continuous effect for improving the forecasting system by simply adding and deleting variables, and improve call quality by preventing over-call and keeping the stable exchange network. If each exchange transmits data to a data collecting system, the data collecting system collects a call processing result by each exchange and stores the collected call processing results to a database. A data analysis system performs learning for forecasting by using the neural network and a call forecasting system outputs a forecasting result based on a learning model.