Risk early warning method and apparatus

Patent metadata

Field	Value
Publication number	US-9953517-B2
Application number	US-201615246404-A
Country	US
Kind code	B2
Filing date	Aug 24, 2016
Priority date	Mar 24, 2016
Publication date	Apr 24, 2018
Grant date	Apr 24, 2018

Abstract

The present application discloses a risk early warning method and apparatus. An implementation of the method includes: monitoring, in real time, search traffic for a predetermined location from users using an online map within a preset period; determining whether the search traffic exceeds a preset search traffic threshold; and sending early warning information of a crowd gathering risk if the search traffic exceeds the preset search traffic threshold. The implementation effectively utilizes the map search traffic capable of reflecting the intention of users and realizes the early warning for the crowd gathering risk.

First claim

What is claimed is:

1. A risk early warning method, comprising:
monitoring, in real time, search traffic for a predetermined location from users using an online map within a preset period;
determining whether the search traffic exceeds a search traffic threshold, the search traffic threshold being set by
recording peak values of the search traffic for the predetermined location within the preset period every day, wherein the peak values are random variables;
determining a probability distribution consistent with the peak values; and
setting the search traffic threshold according to a mean and a mean square error of the probability distribution; and

sending early warning information of a crowd gathering risk if the search traffic exceeds the search traffic threshold.

2. The method according to claim 1 , further comprising:
introducing a search traffic time sequence and a positioning traffic time sequence into a pre-trained prediction model to obtain positioning traffic of mobile devices at the predetermined location after the preset period, wherein the search traffic time sequence is a time sequence of the search traffic for the predetermined location from the users using the online map, and the positioning traffic time sequence is a time sequence of positioning traffic of mobile devices at the predetermined location.

3. The method according to claim 1 , wherein the setting of the search traffic threshold according to a mean and a mean square error of the probability distribution comprises:
obtaining a weight coefficient of the mean square error of the probability distribution based on historical search traffic and historical positioning traffic; and
setting a sum of a product of the weight coefficient and the mean square error and the mean of the probability distribution as the search traffic threshold.

4. The method according to claim 2 , wherein the prediction model is trained by the following steps:
extracting search traffic feature information and positioning traffic feature information from a historical search traffic time sequence and a historical positioning traffic time sequence, respectively; and
training the prediction model used for predicating the positioning traffic of mobile devices at the predetermined location within a future set period by using a machine learning method based on time information, the search traffic feature information and the positioning traffic feature information.

5. A risk early warning apparatus, comprising:
at least one processor; and
a memory storing instructions, which when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:
monitoring, in real time, search traffic for a predetermined location from users using an online map within a preset period;
recording peak values of the search traffic for the predetermined location within the preset period every day, wherein the peak values are random variables;
determining a probability distribution consistent with the peak values; and
setting a search traffic threshold according to a mean and a mean square error of the probability distribution determining whether the search traffic exceeds the search traffic threshold; and
sending early warning information of a crowd gathering risk if the search traffic exceeds the search traffic threshold.

6. The apparatus according to claim 5 , wherein the operations further comprises:
introducing a search traffic time sequence and a positioning traffic time sequence into a pre-trained prediction model to obtain positioning traffic of mobile devices at the predetermined location after the preset period, wherein the search traffic time sequence is a time sequence of the search traffic for the predetermined location from the users using the online map, and the positioning traffic time sequence is a time sequence of positioning traffic of mobile devices at the predetermined location.

7. The apparatus according to claim 5 , wherein the setting the search traffic threshold according to a mean and a mean square error of the probability distribution comprises:
obtaining a weight coefficient of the mean square error of the probability distribution based on historical search traffic and historical positioning traffic; and
setting a sum of a product of the weight coefficient and the mean square error and the mean of the probability distribution as the search traffic threshold.

8. The apparatus according to claim 6 , the prediction model is trained by the following steps:
extracting search traffic feature information and positioning traffic feature information from a historical search traffic time sequence and a historical positioning traffic time sequence, respectively; and
training the prediction model used for predicating the positioning traffic of mobile devices at the predetermined location within a future set period by using a machine learning method based on time information, the search traffic feature information and the positioning traffic feature information.

9. A non-transitory storage medium storing one or more programs, the one or more programs when executed by an apparatus, causing the apparatus to perform a risk early warning method, comprising:
monitoring, in real time, search traffic for a predetermined location from users using an online map within a preset period;
determining whether the search traffic exceeds a search traffic threshold, the search traffic threshold being set by
recording peak values of the search traffic for the predetermined location within the preset period every day, wherein the peak values are random variables;
determining a probability distribution consistent with the peak values; and

setting the search traffic threshold according to a mean and a mean square error of the probability distribution; and
sending early warning information of a crowd gathering risk if the search traffic exceeds the search traffic threshold.

Assignees

• Baidu online network technology beijing co ltd
• Baidu Online Networks Tech Beijing Co Ltd

Inventors

• Zhou Jingbo
• Pei Hongbin
• Wu Haishan

Classifications

• H04W4/023 Primary

  using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds · CPC title

• G08B31/00 Primary

  Predictive alarm systems characterised by extrapolation or other computation using updated historic data · CPC title

• G08B21/02

  Alarms for ensuring the safety of persons · CPC title

• G08B25/14

  Central alarm receiver or annunciator arrangements · CPC title

• G08B21/182 Primary

  Level alarms, e.g. alarms responsive to variables exceeding a threshold · CPC title