发明名称 METHOD AND DEVICE FOR PREDICTING TRAFFIC WITH A NEURAL NETWORK
摘要 The invention concerns a method and a device for predicting, on successive prediction windows (Tp), traffic (5) on a telecommunication network, the device comprising: means for characterising (10, 11, 12) past traffic permanently producing data (U tau ,i) characterising past traffic on respective observation windows (T1); predicting means (100) comprising a neural network (110), and producing, at times (to, to+Tp, to+2Tp, ), an information (Di) characterising traffic on a next prediction window (Tp), derived from a series of said values (U tau ,i) characterising past traffic; learning means (200) training the neural network (110) on a permanent basis.
申请公布号 WO0225452(A3) 申请公布日期 2002.06.27
申请号 WO2001FR02964 申请日期 2001.09.25
申请人 FRANCE TELECOM;CLEROT, FABRICE;GOUZIEN, PASCAL 发明人 CLEROT, FABRICE;GOUZIEN, PASCAL
分类号 G06N3/02;H04L12/24 主分类号 G06N3/02
代理机构 代理人
主权项
地址