• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Annual review of communications >Statistical Billing of Telecom Traffic
【24h】

Statistical Billing of Telecom Traffic

机译：电信流量统计计费

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

A numerical approximation tool has been designed to forecast the revenue and gross margin provided by a utility service provider, in case tariff parameters are being changed. The utility service can be a mobile telephony service, Internet service, energy distribution, cargo, etc. A liberal (deregulated) market has been assumed.rnThe revenue and gross margin are calculated as empirical functions of tariff parameters. This process consists of two phases: First, the usage data is approximated in a high level of aggregation by means of a neuronal network. This regression model is calibrated by the past usage data (e.g., CDR) and by the existing tariffs. After calibrating, the usage data can be calculated-forecasted-as the function in tariff parameters. The aggregation level of usage data is the tariff plan and the customer segment. Second, the revenue, costs, and surplus are calculated from the approximated usage aggregates by a statistical billing method (aggregated billing).rnThe calibrated numerical model simulates the current price elasticity of demand within the small changes of tariff parameters. The main objective is to synthesize an optimized price schedule, providing the maximal gross margin to the carrier. It is expected that the optimal tariff parameters are near the current ones.rnThe idea follows the mathematical theory of optimal taxation originated by Frank P. Ramsey in 1927, further developed by James A. Mirrlees (1996 Nobel Prize for Economics), and published as the nonlinear pricing theory by R. Wilson in 1988.rnActually, the neuronal network model works with the multidimensional tariff bundles, currently used by the mobile telecom carriers. The aggregated billing process approximates the telecom indices average revenue per user (ARPU) and average margin per unit (AMPU) as the empirical functions of tariff parameters. The model incorporates the current price elasticity of demand in the mobile telecom market. This information is in fact obtained from past CDRs simultaneously recorded with different tariff plans.rnThe author has successfully experienced the neuronal network model with the tariff parameters of a real mobile carrier and with the artificial usage data, well simulating that usage data.rnThe submitted paper analyzes the telecom tariff model and the statistical billing of aggregated usage data.
机译：设计了一个数值近似工具，以预测公用事业服务提供商提供的收入和毛利率，以防改变关税参数。公用事业服务可以是移动电话服务，Internet服务，能源分配，货运等。已经假设了一个自由的（已取消关税的）市场。收入和毛利率是作为关税参数的经验函数计算的。该过程包括两个阶段：第一，使用数据通过神经元网络以较高的聚合程度进行近似。该回归模型由过去的使用数据（例如CDR）和现有的费率进行校准。校准后，可以将使用数据作为费率参数中的函数进行预测。使用数据的汇总级别是资费计划和客户群。其次，通过统计计费方法（合计计费）从近似的使用总量中计算出收入，成本和盈余。rnrn校准的数值模型在关税参数的微小变化内模拟了需求的当前价格弹性。主要目标是合成优化的价格计划，为承运人提供最大的毛利率。预计最优关税参数将接近当前的关税参数。该想法遵循的是最优税收的数学理论，该理论由弗兰克·P·拉姆齐（Frank P. Ramsey）于1927年提出，由詹姆斯·A·米尔雷斯（James A. Mirrlees）（1996年诺贝尔经济学奖）进一步发展，并以实际上，神经元网络模型适用于移动电信运营商当前使用的多维资费捆绑。汇总计费过程将电信指数的平均每用户收入（ARPU）和平均每单位利润率（AMPU）近似为费率参数的经验函数。该模型结合了移动电信市场当前的需求价格弹性。实际上，该信息是从同时记录有不同费率计划的CDR中获得的。rn作者已成功地体验了具有真实移动运营商的费率参数和人工使用数据的神经网络模型，并很好地模拟了该使用数据。分析电信资费模型和汇总使用数据的统计帐单。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号