Method, computer device and storage medium for mining point of interest competitive relationship

Patent metadata

Field	Value
Publication number	US-11232116-B2
Application number	US-201916443633-A
Country	US
Kind code	B2
Filing date	Jun 17, 2019
Priority date	Aug 27, 2018
Publication date	Jan 25, 2022
Grant date	Jan 25, 2022

Abstract

A method, computer device and storage medium for mining a point of interest competitive relationship are disclosed. The method includes: for a first POI to be processed, obtaining a set of second POIs serving as mining objects of the first POI; for each second POI in the set, forming a POI pair with the second POI and the first POI, determining a relationship evaluation index of the POI pair according to user's search operations on a map for POIs, and judging whether the two POIs in the POI pair are in a competitive relationship according to the relationship evaluation index. The technical solution of the present disclosure may be applied to improve the accuracy of the processing results.

First claim

What is claimed is:

1. A method for mining a point of interest (POI) competitive relationship, comprising:
for a first POI to be processed, obtaining a set of second POIs serving as mining objects of the first POI; and
for each second POI in the set, forming a POI pair with the second POI and the first POI, determining a relationship evaluation index of the POI pair according to user's search operations on a map for POIs, and judging whether the two POIs in the POI pair are in a competitive relationship according to the relationship evaluation index,
wherein the determining the relationship evaluation index of the POI pair comprises:
counting times of occurrence of a first case in a most recent preset time period, the first case including: a same user sequentially searches for the first POI and the second POI on the map, and a time interval of the two search actions is less than a preset second threshold;
counting times of occurrence of a second case in a most recent preset time period, the second case including: a same user sequentially searches for the second POI and the first POI on the map, and a time interval of the two search actions is less than the second threshold; and
taking two counting results as the relationship evaluation index of the POI pair.

2. The method according to claim 1 , wherein the obtaining a set of second POIs serving as mining objects of the first POI comprises:
selecting, according to a predetermined rule, second POIs similar to the first POI from POIs on the map recorded in a database, and obtaining the set of second POIs.

3. The method according to claim 2 , wherein the predetermined rule comprises at least one of:
a second POI belongs to a same category as the first POI; and
a distance between a second POI and the first POI is less than a preset first threshold.

4. The method according to claim 1 , wherein the judging whether the two POIs in the POI pair are in a competitive relationship according to the relationship evaluation index comprises:
under the condition that the two counting results are both greater than a preset third threshold, determining that the two POIs in the POI pair are in the competitive relationship.

5. The method according to claim 1 , wherein the judging whether the two POIs in the POI pair are in a competitive relationship according to the relationship evaluation index comprises:
taking the relationship evaluation index at least as a feature of the POI pair, inputting the feature of the POI pair into a machine learning model obtained by pre-training to obtain a competitive relationship score of the POI pair;
and judging whether the two POIs in the POI pair are in the competitive relationship according to the competitive relationship score.

6. The method according to claim 5 , wherein the judging whether the two POIs in the POI pair are in the competitive relationship according to the competitive relationship score comprises:
comparing the competitive relationship score with a preset fourth threshold, and under the condition that the competitive relationship score is greater than the fourth threshold, determining that the two POIs in the POI pair are in the competitive relationship.

7. The method according to claim 5 , wherein the judging whether the two POIs in the POI pair are in the competitive relationship according to the competitive relationship score comprises:
after respectively obtaining the competitive relationship scores of the POI pairs in which each second POI in the set of second POIs lies, sorting the competitive relationship scores in a descending order, and under the condition that a competitive relationship score of any POI pair ranks at top K, determining that the two POIs in said any POI pair are in the competitive relationship, K being a positive integer and smaller than the number of the second POIs in the set of second POIs.

8. The method according to claim 5 , wherein the training to obtain the machine learning model comprises:
obtaining a POI pair as a positive sample and a POI pair as a negative sample, wherein two POIs in the POI pair as the positive sample are in the competitive relationship, and two POIs in the POI pair as the negative sample are not in the competitive relationship;
obtaining a feature of the POI pair as the positive sample and a feature of the POI pair as the negative sample, the features obtained each at least including a relationship evaluation index; and
based on the features obtained, obtaining the machine learning model by training based on the positive sample and the negative sample.

9. A computer device, comprising a memory, a processor and a computer program which is stored on the memory and runs on the processor, wherein the processor, upon executing the program, implements a method for mining a point of interest (POI) competitive relationship, wherein the method comprises:
for a first POI to be processed, obtaining a set of second POIs serving as mining objects of the first POI; and
for each second POI in the set, forming a POI pair with the second POI and the first POI, determining a relationship evaluation index of the POI pair according to user's search operations on a map for POIs, and judging whether the two POIs in the POI pair are in a competitive relationship according to the relationship evaluation index,
wherein the determining the relationship evaluation index of the POI pair comprises:
counting times of occurrence of a first case in a most recent preset time period, the first case including: a same user sequentially searches for the first POI and the second POI on the map, and a time interval of the two search actions is less than a preset second threshold;
counting times of occurrence of a second case in a most recent preset time period, the second case including: a same user sequentially searches for the second POI and the first POI on the map, and a time interval of the two search actions is less than the second threshold; and
taking two counting results as the relationship evaluation index of the POI pair.

10. A non-transitory computer-readable storage medium on which a computer program is stored, wherein the program, when executed by a processor, implements a method for mining a point of interest (POI) competitive relationship, wherein the method comprises:
for a first POI to be processed, obtaining a set of second POIs serving as mining objects of the first POI; and
for each second POI in the set, forming a POI pair with the second POI and the first POI, determining a relationship evaluation index of the POI pair according to user's search operations on a map for POIs, and judging whether the two POIs in the POI pair are in a competitive relationship according to the relationship evaluation index,
wherein the determining the relationship evaluation index of the POI pair comprises:
counting times of occurrence of a first case in a most recent preset time period, the first case including: a same user sequentially searches for the first POI and the second POI on the map, and a time interval of the two search actions is less than a preset second threshold;
counting times of occurrence of a second case in a most recent preset time period, the second case including: a same user sequentially searches for the second POI and the first POI on the map, and a time interval of the two search actions is less than the second threshold; and
taking two counting results as the relationship evaluation index of the POI pair.

Assignees

• Baidu online network technology beijing co ltd

Inventors

• Zhou Jingbo
• Duan Jianguo
• Jiang Airong
• Xiong Hui

Classifications

• G06N20/00

  Machine learning · CPC title

• G06F16/29

  Geographical information databases · CPC title

• G06Q30/0201 Primary

  Market modelling; Market analysis; Collecting market data · CPC title

• G06F16/2465 Primary

  Query processing support for facilitating data mining operations in structured databases · CPC title