Method, apparatus, device, and storage medium for intention recommendation

Patent metadata

Field	Value
Publication number	US-2021209109-A1
Application number	US-202117207503-A
Country	US
Kind code	A1
Filing date	Mar 19, 2021
Priority date	Jun 29, 2020
Publication date	Jul 8, 2021
Grant date	—

Abstract

The present application discloses a method, an apparatus, a device, and a storage medium for intention recommendation, which relates to the field of big data, artificial intelligence, intelligent search, information flow and deep learning technologies in the field of computer technologies. A specific implementation scheme includes: receiving an intention query request carrying an intention keyword and a user identification, determining a first recommendation list according to the intention keyword and a pre-configured intention repository, where the intention repository includes at least one tree-shaped intention set, and each tree-shaped intention set includes at least one graded intention, processing intentions in the first recommendation list by using intention strategy information corresponding to the user identification to obtain a target recommendation list and output it.

First claim

What is claimed is:

1 . A method for intention recommendation, comprising:
receiving an intention query request, wherein the intention query request carries an intention keyword and a user identification; determining a first recommendation list according to the intention keyword and a pre-configured intention repository, wherein the intention repository comprises at least one tree-shaped intention set, and each of the at least one tree-shaped intention set comprises at least one graded intention; processing intentions in the first recommendation list by using intention strategy information corresponding to the user identification to obtain a target recommendation list; and outputting the target recommendation list.

2 . The method according to claim 1 , wherein the determining a first recommendation list according to the intention keyword and a pre-configured intention repository comprises:
performing semantic analysis on the intention keyword to determine a target intention of a user corresponding to the user identification querying the pre-configured intention repository according to the target intention to determine level information and tree information of where the target intention is located; and obtaining the first recommendation list according to the level information and the tree information of where the target intention is located and a tree-shaped intention set corresponding to the tree information.

3 . The method according to claim 2 , wherein the obtaining the first recommendation list according to the level information and the tree information of where the target intention is located and a tree-shaped intention set corresponding to the tree information comprises:
generating an initial recommendation list corresponding to the target intention; determining at least one associated intention of the target intention according to the level information and the tree information of where the target intention is located; and obtaining the first recommendation list by adding the at least one associated intention of the target intention to the initial recommendation list sequentially in an order of a same level, a superordinate level, and a tree level.

4 . The method according to claim 1 , wherein the processing intentions in the first recommendation list by using intention strategy information corresponding to the user identification to obtain a target recommendation list comprises:
determining the intention strategy information corresponding to the user identification, wherein the intention strategy information is used to indicate effective range information and effective sequence information of intentions; determining an intention set that is effective, in the first recommendation list according to the effective range information of intentions indicated by the intention strategy information; and sorting all intentions in the intention set according to the effective sequence information of intentions indicated by the intention strategy information to obtain the target recommendation list.

5 . The method according to claim 1 , wherein the method further comprising:
processing the intention query request and a predetermined similarity matrix based on a collaborative filtering algorithm to obtain a second recommendation list; and the processing intentions in the first recommendation list by using intention strategy information corresponding to the user identification to obtain a target recommendation list comprises: processing the intentions in the first recommendation list and intentions in the second recommendation list by using the intention strategy information corresponding to the user identification to obtain the target recommendation list.

6 . The method according to claim 5 , wherein the second recommendation list is obtained by processing the intention query request and a predetermined user behavior similarity matrix by using a user behavior-based collaborative filtering algorithm, and/or, processing the intention query request and a predetermined intention similarity matrix by using an intention-based collaborative filtering algorithm.

7 . The method according to claim 5 , wherein the method further comprising:
obtaining a user data set; generating a user behavior similarity matrix and/or an intention similarity matrix according to the user data set, wherein the user behavior similarity matrix is used to characterize a behavior similarity between users involved in the user data set, and the intention similarity matrix is used to characterize a similarity between intentions involved in the user data set; and storing the user behavior similarity matrix and/or intention similarity matrix.

8 . An electronic device, comprising:
at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions thereon, and the instructions, when executed by the at least one processor, cause the at least one processor to: receive an intention query request, wherein the intention query request carries an intention keyword and a user identification; determine a first recommendation list according to the intention keyword and a pre-configured intention repository, wherein the intention repository comprises at least one tree-shaped intention set, and each of the at least one tree-shaped intention set comprises at least one graded intention; process intentions in the first recommendation list by using intention strategy information corresponding to the user identification to obtain a target recommendation list; and output the target recommendation list.

9 . The electronic device according to claim 8 , wherein the at least one processor is configured to:
perform semantic analysis on the intention keyword to determine a target intention of the user corresponding to the user identification; query the pre-configured intention repository according to the target intention to determine level information and tree information of where the target intention is located; and obtain the first recommendation list according to the level information and the tree information of where the target intention is located and a tree-shaped intention set corresponding to the tree information.

10 . The electronic device according to claim 9 , wherein
the at least one processor is configured to: generate an initial recommendation list corresponding to the target intention; determine at least one associated intention of the target intention according to the level information and the tree information of where the target intention is located; and obtain the first recommendation list by adding at least one associated intention of the target intention to the initial recommendation list sequentially in an order of a same level, a superordinate level, and a tree level.

11 . The electronic device according to claim 8 , wherein the at least one processor is configured to:
determine the intention strategy information corresponding to the user identification, wherein the intention strategy information is used to indicate effective range information and effective sequence information of intentions; determine an intention set that is effective, in the first recommendation list according to the effective range information of intentions indicated by the intention strategy information; and sort all intentions in the intention set according to the effective sequence information of intentions indicated by the intention strategy information to obtain the target recommendation list.

12 . The electronic device according to claim 8 , wherein the at least one processor is further configured to:
process the

Assignees

• Beijing Baidu Netcom Sci & Tech Co Ltd

Inventors

• Zhang Yang
• Yang Shuangquan
• Xie Yi
• Liu Chang

Classifications

• G06F16/24542 Primary

  Plan optimisation · CPC title

• G06F16/9536 Primary

  Search customisation based on social or collaborative filtering · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06F16/9035

  Filtering based on additional data, e.g. user or group profiles · CPC title