Mobile traffic forecasting based on the relative position of multi-scale average lines

摘要

Both of the mobile call traffic and the stock price forecasting have some common character form the angle of statistical view. Thus, an experiential method of stock price forecasting, namely the relative position of multi-scale average lines, is reformed and imported into the traffic forecasting. Firstly, the reformed method chooses scales of average lines, which optimal reflect the change direction and base volume character. Secondly, these average lines are fitted with some interpolating means. After comparing the lines each other, the change tendency and volume are predicted by analyzing the relative position of the lines. Some primary experiments illustrate that the measure is feasible and valuable.