摘要
由一名新闻记者-机器人与机器学习的工作人员新闻编辑每日新闻-调查人员讨论机器学习的新发现。根据NewsRx e Ditors在俄罗斯莫斯科的新闻报道，Research称：“本文提出了一种联合使用统计和机器学习(ML)模型的方法，以解决历史事件的精确重建、正在进行的事件的实时检测和未来服务质量相关事件的预测问题，以促进现代网络的未来发展。”"介绍了深高斯混合模型(DGMM)的回归模型."这项研究的财政支持来自RUDN大学科学计划的资助系统。我们的新闻记者从俄罗斯科学院的研究中得到一句话：“首先，本文利用葡萄牙移动运营商的真实数据和公共蜂窝业务数据,对基于有限正态混合的数据进行了初步聚类,并将这些信息作为有监督ML算法的输入,这是电信网络领域概率信息ML方法的基本概念。本文将该方法与随机森林、支持向量机回归、梯度boosting和LSTM等方法进行了比较,并以广义Gamma(GG)分布参数为基础的向量自回归作为一个基准,证明基于DGMM的回归比LS TM的回归快6.82～22.8倍。基于DGMM的回归方法对最重要的流量特征(平均流量和总流量、用户数量)能得到较好的结果,对于MAPE和RMSE指标,分别比统计AL方法的结果高46.7%(RMSE)和91.5%(MAPE)(中位数分别提高28.0%和80.1%);对于最大似然估计法分别达到13.0%(RMSE)和35.7%(MAPE)(中位数分别为0.39%和2.5%),因此,在计算速度和精度方面,使用概率信息方法似乎是最佳的.
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators discuss new findings in Machine Learning. According to news reporting out of Moscow, Russia, by NewsRx e ditors, research stated, "The paper proposes an approach to the joint use of sta tistical and machine learning (ML) models to solve the problems of the precise r econstruction of historical events, real-time detection of ongoing incidents, an d the prediction of future quality of service -related occurrences for prospecti ve development of the modern networks. For forecasting, a regression version of the deep Gaussian mixture model (DGMM) is introduced." Financial support for this research came from RUDN University Scientific Project s Grant System. Our news journalists obtained a quote from the research from the Russian Academy of Sciences, "First, the preliminary clustering based on the finite normal mixt ures is performed. This information is then used as an input for some supervised ML algorithm. It is the basic concept of the probability -informed ML approach in the field of telecommunications networks. Using the real -world datasets from a Portuguese mobile operator as well as public cellular traffic data, the artic le compares this approach with methods such as random forests, support vector ma chine regression, gradient boosting and LSTM. Vector autoregression,informed by the parameters of the generalized gamma (GG) distribution, which has also been successfully used to reconstruct past traffic patterns, is also used as a benchm ark. We demonstrate that DGMM-based regression is 6.82-22.8 times faster than LS TM for the dataset. Moreover, DGMM-based regression can achieve better results f or the most important traffic characteristics (average and total traffic, the nu mber of users). For metrics MAPE and RMSE, it surpasses the results of statistic al methods up to 46.7% (RMSE) and 91.5% (MAPE) (medi an increases are 28.0% and 80.1%, respectively), as w ell as for ML methods up to 13.0% (RMSE) and 35.7% ( MAPE) (median increases are 0.39% and 2.5%, respectiv ely). Thus, the use of a probability -informed approach for telecommunication da ta seems optimal for the computational speed and accuracy trade-off."
NETL
NSTL