公开(公告)号：US20160217383A1

公开(公告)日：2016-07-28

申请号：US14902303

申请日：2014-12-16

Applicant: BEIJING BAIDU NETCOM SCIENCE AND TECHNOLOGY CO., LTD.

Inventor： Jiewei ZHANG ,  Xiao ZHANG ,  Jian HE ,  Jing HUANG

IPC: G06N7/00 ,  G06F17/50

CPC classification number: G06N7/005 ,  G06F17/50 ,  G06Q10/04

Abstract: Method and apparatus for forecasting a characteristic information change. Historical characteristic data in at least one computation period and current incremental data of multiple pieces of first characteristic information corresponding to a pre-estimation model is acquired. The current incremental data is used for indicating a ratio of characteristic data on a day immediately before a forecasting day of each of the multiple pieces to the historical characteristic data in the at least one computation period of each of the multiple pieces of first characteristic information. A first change information on the forecasting day of second characteristic information is determined by a forecasting process using the pre-estimation model based on the historical characteristic data and the current incremental data. Based on the first change information, change pre-estimation information on the forecasting day of the second characteristic information is determined to prompt a user to execute a corresponding operation.

Abstract translation: 用于预测特征信息变化的方法和装置。 获取至少一个计算周期中的历史特征数据和与预估模型对应的多条第一特征信息的当前增量数据。 当前的增量数据用于指示在多个第一特征信息中的每一个的至少一个计算周期中紧接在每个多个片段的预测日之前的一天的特征数据与历史特征数据的比率。 通过使用基于历史特征数据和当前增量数据的预估模型的预测处理来确定第二特征信息的预测日的第一变化信息。 基于第一改变信息，确定改变第二特征信息的预测日的预估信息，以提示用户执行相应的操作。