CN113486233A - Content recommendation method, device and medium - Google Patents

Content recommendation method, device and medium Download PDF

Info

Publication number
CN113486233A
CN113486233A CN202010850633.7A CN202010850633A CN113486233A CN 113486233 A CN113486233 A CN 113486233A CN 202010850633 A CN202010850633 A CN 202010850633A CN 113486233 A CN113486233 A CN 113486233A
Authority
CN
China
Prior art keywords
target
content
identity
emotional state
voice information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010850633.7A
Other languages
Chinese (zh)
Inventor
陈维强
蒋鹏民
高雪松
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Qingdao Hisense Electronic Industry Holdings Co Ltd
Original Assignee
Qingdao Hisense Electronic Industry Holdings Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Qingdao Hisense Electronic Industry Holdings Co Ltd filed Critical Qingdao Hisense Electronic Industry Holdings Co Ltd
Priority to CN202010850633.7A priority Critical patent/CN113486233A/en
Publication of CN113486233A publication Critical patent/CN113486233A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3329—Natural language query formulation or dialogue systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/3332—Query translation
    • G06F16/3334—Selection or weighting of terms from queries, including natural language queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3343—Query execution using phonetics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G06F16/353—Clustering; Classification into predefined classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/903—Querying
    • G06F16/9032—Query formulation
    • G06F16/90332—Natural language query formulation or dialogue systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9532—Query formulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Acoustics & Sound (AREA)
  • Human Computer Interaction (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The application discloses a content recommendation method, content recommendation equipment and a content recommendation medium, which are used for solving the problems of weak man-machine interaction dynamic sense and low intelligence in the prior art. The method comprises the following steps: determining the intention of the received voice information and the contained keywords, if the keywords do not contain the necessary types of keywords corresponding to the intention, acquiring the target identity of the target user inputting the voice information according to the voice information and the identity of the user registered in advance, identifying the target emotional state of the target user according to the voice information, and determining and recommending the target content connected with the target identity in the target emotional state according to the emotional state connected between the identity and the content stored in the family knowledge map. According to the method and the device, when the keywords obtained through the intelligent device do not contain the necessary type keywords contained in the intention, the corresponding recommended content is found through searching the family knowledge graph, and the interaction feeling of man-machine interaction is enhanced.

Description

Content recommendation method, device and medium
Technical Field
The present application relates to the field of human-computer interaction technologies, and in particular, to a content recommendation method, device, and medium.
Background
With the rapid development of artificial intelligence and the increasing popularization of various intelligent terminals, human-computer interaction plays an increasingly important role. The most common way of interaction today is voice interaction, and man-machine interactive dialog systems are applied in various intelligent terminals. The most important thing in the man-machine interaction based on the voice interaction is that the intelligent terminal can fully understand the intention of the user, and determines corresponding content according to the necessary types of keywords corresponding to the intention and recommends the content to the user.
Due to the limitation of the current technology, the current interpersonal interaction system cannot ensure that the intelligent terminal completely understands the intention of the user because a voice signal is input into the intelligent terminal, and under the condition that the intention of the user is not completely understood, the intelligent terminal generally directly outputs prompt information of the contents which are not found. Therefore, the man-machine interaction method in the prior art can seriously affect the use experience of the user, and can cause the problems of weak interaction between the user and the intelligent terminal, low intelligence of the intelligent terminal and the like.
Disclosure of Invention
The application provides a content recommendation method, device, equipment and medium, which are used for solving the problems of weak interaction and low intelligence of a user and an intelligent terminal in interpersonal interaction in the prior art.
In a first aspect, the present application provides a content recommendation method, including:
determining the intention of the received voice information and the contained keywords;
if the keyword does not contain the necessary type keyword corresponding to the intention, acquiring a target identity of a target user inputting the voice information according to the voice information and a pre-registered identity of the user, and identifying a target emotional state of the target user according to the voice information;
and determining and recommending target content connected with the target identity by the target emotional state according to the emotional state connected between the identity and the content stored in the knowledge graph.
In a second aspect, the present application further provides a content recommendation apparatus, including:
first determining means for determining an intention of the received voice message and a keyword included therein;
the extraction device is used for acquiring a target identity of a target user inputting the voice information according to the voice information and a pre-registered identity of the user if the keyword does not contain the necessary type of keyword corresponding to the intention, and identifying the target emotion state of the target user according to the voice information;
and the second determining device is used for determining and recommending the target content connected with the target identity mark in the target emotional state according to the emotional state connected between the identity mark and the content stored in the knowledge graph.
In a third aspect, the present application further provides a smart device, which at least comprises a processor and a memory, wherein the processor is configured to implement the steps of the content recommendation method as described above when executing the computer program stored in the memory.
In a fourth aspect, the present application further provides a smart device, including:
the sound pickup is used for collecting voice information;
a player for playing;
a controller to perform:
determining the intention and the keywords of the voice information collected by a sound pickup, and if the keywords do not contain the necessary types of keywords corresponding to the intention, acquiring the target identity and the target emotional state of the target user inputting the voice information according to the voice information;
and determining target content connected with the target identity by the target emotion according to the connection emotional state between the identity and the content stored in the knowledge graph, and controlling the player to play the target content.
In a fifth aspect, the present application further provides a computer-readable storage medium storing a computer program, which when executed by a processor, implements the steps of the content recommendation method as described above.
The embodiment of the application provides a method, a device, equipment and a medium for recommending content, wherein the method comprises the following steps: after the received voice information intention and the contained keywords are determined, if the keywords do not contain the necessary types of keywords corresponding to the intention, the target identity of the target user inputting the voice information is obtained according to the voice information and the identity of the user registered in advance, the target emotional state of the target user is identified according to the voice information, the target content connected with the target identity in the target emotional state is determined and recommended according to the emotional state connected between the identity and the content stored in the knowledge map, and therefore the method can avoid the situation that the user intention is not understood in the human-computer interaction process, prompt information of the unseen content is directly output, interaction and mutual inductance of human-computer interaction are enhanced, and the intellectualization of a human-computer interaction system is improved.
Drawings
In order to more clearly illustrate the technical solutions of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a schematic diagram of a content recommendation process provided in some embodiments of the present application;
FIG. 2 is a schematic diagram of a family knowledge graph structure provided in some embodiments of the present application;
FIG. 3 is a flowchart of a knowledge-graph based content recommendation process provided by some embodiments of the present application;
FIG. 4 is a flow diagram of knowledge-graph based content creation and update provided by some embodiments of the present application;
FIGS. 5a-5c are diagrams of examples of knowledge-graph based content recommendations provided by some embodiments of the present application;
FIG. 6 is a diagram of a specific content recommendation process provided by some embodiments of the present application;
fig. 7 is a schematic structural diagram of a content recommendation device according to some embodiments of the present application;
fig. 8 is a schematic structural diagram of a smart device according to some embodiments of the present application.
Detailed Description
In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the accompanying drawings, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
After determining the intention of the received voice information and the contained keywords, if the keywords do not contain the keywords corresponding to the intention, the intelligent equipment further determines the target identification mark and the target emotional state of the target user inputting the voice information, and then determines and recommends the target content connected with the target identification mark in the target emotional state according to the emotional state relation between the identification mark and the content stored in the knowledge graph. Therefore, when the keywords acquired by the intelligent equipment do not contain the necessary types of keywords corresponding to intentions, the keywords are recommended according to the prestored knowledge graph and the user identity and emotional state based on the use record of the user aiming at the content, so that the interactive feeling of man-machine interaction is improved, and the use feeling of the user is improved.
In order to enhance the interaction and mutual inductance of human-computer interaction and improve the intelligence of human-computer interaction, the application provides a content recommendation method, a device, equipment and a medium.
Fig. 1 is a schematic diagram of a content recommendation process provided in the present application, where the content recommendation process includes the following steps:
s101: the intention of the received voice message and the included keywords are determined.
The content recommendation provided by the embodiment of the application is applied to intelligent equipment, and the intelligent equipment can be equipment such as an intelligent sound box and a smart phone and can also be a server.
In the application, the voice information received by the intelligent device may be the voice information acquired by the intelligent device itself, or the voice information received by other intelligent devices. After the intelligent device acquires the voice information, the intention of the voice information and the contained keywords are determined.
Specifically, when the intelligent device obtains the intention of the voice information and the included keywords, the voice information is firstly converted into text information through voice recognition, and then the text information is subjected to semantic analysis, wherein the semantic analysis process can be that the intelligent device performs semantic analysis by itself or that the text information is sent to other devices for semantic analysis. After the semantics of the voice information are obtained, adopting a fasttext text classification algorithm to identify the intention, determining the intention and determining the keywords contained in the intention.
The method for converting the voice information into the text information, performing semantic analysis, determining the intention and identifying the keywords contained in the intention based on the text information belongs to the prior art, and is not repeated herein.
The semantic meaning refers to a word that can embody the purpose of the text message, for example, "play song" in the text message of "play a junjie song" is the semantic meaning of the text message. The keywords refer to words that can represent part of semantics of the text information, for example, "zhou jeren" and "blue and white porcelain" in the text information of "playing zhou jeren" are keywords of the text information. In the present application, the keyword may be a verb, a place noun, a name of a person, a name of a brand, or the like.
S102: and if the keywords do not contain the necessary types of keywords corresponding to the intentions, acquiring a target identity of a target user inputting the voice information according to the voice information and the identity of a pre-registered user, and identifying the target emotional state of the target user according to the voice information.
In order to ensure that when the intelligent device acquires the voice information, corresponding content can be recommended for a user inputting the voice information, after the intention and the contained keywords of the voice information are acquired in the application, whether the keywords contain the necessary keywords corresponding to the intention or not is judged based on the intention and the keywords, and if the keywords contain the necessary keywords corresponding to the intention and the information content contained in the voice information is complete, the content can be determined and recommended directly according to the necessary types of keywords corresponding to the intention.
In order to determine whether a necessary keyword corresponding to an intention is included for each intention, a necessary type of the keyword corresponding to the intention is stored for each intention in advance, and therefore, when the intention is recognized and the keyword included in the voice information is determined, it is possible to determine whether the keyword of the necessary type is included in the voice information according to the necessary type of the keyword corresponding to the intention.
For example, when the speech information is "coming yinjiang blue and white porcelain", the recognized intention is "coming music", and the necessary type of the keyword corresponding to the intention of "coming music" is the music name, and the keywords are "yinjiang" and "blue and white porcelain", and at this time, the keyword includes the necessary keyword corresponding to the intention, so that the information content included in the speech information is complete. When the voice information is 'put a song of a lin junjie', the extracted intention is 'put a song', the necessary type of the keyword corresponding to the intention is a music name, 'put a song', and the keyword identified based on the voice information is 'lingjie', at this moment, the keyword is judged to be free from the necessary keyword corresponding to the intention, and the song to be played can not be determined to be exactly which song of the lin junjie, so the information content contained in the voice information is incomplete.
If it is determined that the keywords included in the voice message do not include the necessary types of keywords corresponding to the intention, in order to recommend corresponding content to the user, the intelligent device in the present application stores the identity of the user that is registered in advance. Specifically, in the application, the smart device performs voiceprint registration on a user capable of providing a service, and acquires an identity of a pre-registered user. If the intelligent equipment is the intelligent household equipment, voiceprint registration can be carried out on family members of the family where the intelligent household equipment is located. The voiceprint registration is carried out based on a voiceprint recognition algorithm, and after the voiceprint registration is completed, the voiceprint features obtained through the voiceprint registration are stored in the intelligent device, and the identity corresponding to the voiceprint features is correspondingly stored.
When the fact that the keywords contained in the voice information do not contain the necessary types of keywords corresponding to the intention is determined, the intelligent device matches the voiceprint features of the received voice information with the stored registered voiceprint features, and when the fact that the matched registered voiceprint features exist is determined, the identity corresponding to the matched registered voiceprint features can be determined to be the target identity.
In order to accurately determine the recommended content, in the present application, the target emotional state of the target user of the speech information is also identified according to the speech information. When the target emotional state is determined, the emotional state of the voice information can be identified and determined according to an emotion recognition classification algorithm.
The process of determining the emotional state based on the emotion recognition classification algorithm is the prior art, and is not described herein.
S103: and determining and recommending target content connected with the target identity by the target emotional state according to the emotional state connected between the identity and the content stored in the knowledge graph.
In the present application, in order to ensure the accuracy of recommended content, a knowledge graph is constructed in advance, and a node corresponding to a user having a pre-registered identity and a node corresponding to content connected with the node corresponding to the user having the identity in a certain emotional state are stored in the knowledge graph.
If the intelligent equipment that carries out content recommendation that this application provided is intelligent home equipment, for example intelligent audio amplifier etc. because its family in place is fixed after intelligent home equipment is installed, consequently this intelligent home equipment also can only carry out content recommendation for the family member who is located in the family, consequently can only save family's knowledge map this moment, this moment this family's knowledge map includes: the identity, emotional state and content of the pre-registered members in the family. And the identity marks are connected with various contents in a family knowledge graph in a certain emotional state. The identity may be the name of the family member or the name of the family member, and the various contents may be recommended contents required by the user, such as the names of stories and songs.
Therefore, after the target identity and the target emotional state of the target user inputting the voice information are determined, the target content corresponding to the node connected with the target emotional state between the nodes of the target identity is determined according to the nodes corresponding to the identity stored in the knowledge graph and the nodes corresponding to the content connected with the nodes corresponding to the user of the identity in a certain emotional state, and the target content is recommended.
After the received intention of the voice information and the contained keywords are determined in the application, if the keywords do not contain the necessary types of keywords corresponding to the intention, the target identity of the target user inputting the voice information is obtained according to the voice information and the identity of the user registered in advance, the target emotional state of the target user is identified according to the voice information, the target content connected with the target identity in the target emotional state is determined and recommended according to the emotional state connected between the identity and the content stored in the knowledge map, and therefore the method can avoid that the user intention is not understood by intelligent terminal equipment in the man-machine interaction process, prompt information of the unseen content is directly output, interaction of man-machine interaction is enhanced, and the intellectualization of a human-machine interaction system is improved.
For accurate content recommendation, on the basis of the above embodiments, in this application, when determining that the keywords include the necessary types of keywords corresponding to the intention, the method further includes:
and determining and recommending the content according to the necessary type of keywords corresponding to the intention.
When the keywords contained in the voice information are determined to contain the necessary types of keywords corresponding to the intention, the information content contained in the voice information is complete, and at this time, the intelligent device can directly find the corresponding content according to the voice information and recommend the content to the user.
Specifically, when recommending content according to the fact that the information content of the voice information is determined to be complete, since the intention of the voice information is known, the content can be determined and recommended according to the necessary type of keywords corresponding to the intention.
Fig. 2 is a schematic diagram of a family knowledge graph structure provided in some embodiments of the present application, where "dad", "mom", "son" and "daughter" are all identifiers, "song 1", "song 2", "story 1" and "story 2" are all contents, a connection relationship exists between a part of the identifiers and the contents, such as "mom" and "song 1", and a connection relationship does not exist between a part of the identifiers and the contents, such as "mom" and "story 1". The identity with the connection relation is connected with the content in an emotional state, for example, the connection relation exists between the 'mom' and the 'song 1', and the corresponding target emotional state is optimistic, that is, when the target identity is identified as 'mom' and the target emotional state is 'optimistic', the content searched in the family knowledge graph based on the target identity and the target emotional state is 'song 1', and if the content is the only searched content, the content is used as the target content.
Fig. 3 is a flowchart of a knowledge-graph-based content recommendation process according to some embodiments of the present application:
s301: the intelligent device receives the voice information.
S302: and obtaining the intention of the voice information and the contained keywords.
S303: whether the keyword includes the necessary type of keyword corresponding to the intention is determined, if yes, S304 is executed, and if no, S305 is executed.
S304: and determining and recommending the content according to the necessary type of keywords corresponding to the intention.
S305: and determining the target identity and the target emotional state of the target user inputting the voice information.
S306: and determining whether the content connected with the target identity in the target emotional state exists according to the emotional relation between the identity and the content stored in the knowledge graph, if so, performing S307, and otherwise, performing S308.
S307: and determining and recommending target content according to the content connected with the target identity in the target emotional state.
S308: and outputting prompt information of the content which is not found.
For accurate content recommendation, on the basis of the above embodiments, in the present application, the method further includes:
judging whether a target node corresponding to the content exists in the knowledge graph or not, and if not, creating a node corresponding to the content;
acquiring a target identity of a target user inputting the voice information, identifying a target emotional state of the target user according to the voice information, judging whether the target identity is connected with the content in the target emotional state in the knowledge graph, if not, connecting a node corresponding to the target identity with a node corresponding to the content, recording the content connected with the target emotional state by the target identity, and if so, updating the frequency of the content connected with the target emotional state by the target identity.
In the application, if the content of the input voice information is complete, after determining and recommending the content according to the necessary type of keywords corresponding to the intention of the voice information, in order to ensure the accuracy of subsequent output, in the application, whether a node corresponding to the content exists in the knowledge graph is judged, if not, the node corresponding to the content is created, the target identity and the target emotional state of a target user who inputs the voice information are obtained, the node corresponding to the target identity information is determined, the node corresponding to the target identity is connected with the node corresponding to the content, the content connected by the target emotional state of the target identity is recorded, and the recommendation time and the recommendation times corresponding to the connection relation are recorded.
If the node corresponding to the content exists in the knowledge graph, judging whether the target identity in the knowledge graph is connected with the content in a target emotion state or not according to the acquired target identity and the target emotion state, and if so, directly updating the recommended times of the content connected with the target identity in the target emotion state within a set time length; otherwise, connecting the node corresponding to the target identity with the node corresponding to the content, recording the content of the target emotion state connection of the target identity, and recording the recommendation time and recommendation times corresponding to the connection relation.
FIG. 4 is a knowledge-graph-based content creation and update flow diagram provided by some embodiments of the present application:
s401: the intelligent device receives the voice information.
S402: and obtaining the intention of the voice information and the contained keywords.
S403: and if the keywords contain the keywords of the necessary type corresponding to the intention, determining the content according to the keywords of the necessary type corresponding to the intention and recommending the content.
S404: and determining the target identity and the target emotional state of the target user inputting the voice information.
S405: and judging whether a target node corresponding to the content exists in the knowledge graph, if so, executing S406, and if not, executing S409.
S406: and judging whether the target identity mark in the knowledge graph is connected with the content by the target emotional state, if so, executing S407, otherwise, executing S408.
S407: and updating the recommended times of the target identity mark in the set time length of the content connected with the target emotional state.
S408: and connecting the node corresponding to the target identity with the node corresponding to the content, and recording the content of the target identity connected with the target emotional state.
S409: and creating a target node corresponding to the content, connecting the node of the target identity identifier and the target node corresponding to the content in the target emotion state, and executing step S407.
In order to accurately obtain target content connected by a target identity in a target emotional state, on the basis of the foregoing embodiments, in the present application, the determining, according to an emotional state connected between an identity and content stored in a knowledge graph, the target content connected by the target emotional state with the target identity includes:
determining candidate content connected with the target identity by the target emotional state in the knowledge graph according to the target identity and the target emotional state;
if more than two candidate contents exist, determining the recommended times of each candidate content in a set time length according to the recommended times of the content connected in an emotional state by the identity marks stored in the knowledge graph in the set time length;
and determining the target content according to the recommended times corresponding to each candidate content.
In the following embodiments, the smart device is taken as an example of a smart home device, and the knowledge graph stored at this time may be considered as a family knowledge graph.
When the target content is searched based on the knowledge graph according to the determined target identity and the target emotional state, if more than two contents connected with the target identity in the target emotional state are searched, in order to ensure the accuracy of content recommendation, each content is used as a candidate content, and the target content is determined in the more than two candidate contents.
In order to accurately recommend content, the knowledge graph of the present application further stores the recommended times of each candidate content connected by the target identity in the target emotional state within a set time length.
Therefore, in the process of determining the target content, the target content may also be determined according to the recommended times of each candidate content connected by the target identity in the target emotional state within the set time length, specifically, the recommended times of each candidate content in the target emotional state within the set time length is searched for, and the target content is determined according to the recommended times of each candidate content. For example, the candidate content that is recommended the most number of times may be used as the target content.
After the target identity identification mark and the target emotion state are obtained, when content is searched based on the knowledge graph, and only one content in the knowledge graph is searched according to the target identity mark and the target emotion state, the searched content is used as the target content and recommended to the target user.
In order to accurately obtain recommended content, on the basis of the foregoing embodiments, in this application embodiment, the determining target content according to the recommended times corresponding to each candidate content includes:
if the recommended times corresponding to each candidate content are the same, determining the total recommended times of each candidate content in a set time length according to the recommended times of the emotional-state-connected content in the set time length by the identity marks stored in the family knowledge graph, and determining the candidate content corresponding to the maximum value of the total times as the target content;
and if the recommended times corresponding to each candidate content are different, taking the candidate content corresponding to the maximum value of the recommended times as the target content.
When searching for content based on a knowledge graph according to a determined target identity and a target emotion state, if more than two contents connected with the target identity in the target emotion state are searched, each connected content is used as a candidate content, in order to ensure the accuracy of content recommendation, the target content can be determined according to the recommended times of each candidate content connected with the target identity in the target emotion state within a set time length, and when the recommended times within the set time length corresponding to each candidate content are different, the candidate content corresponding to the maximum recommended times within the set time length is used as the target content.
However, the situation that the recommended times of each candidate content connected in the target emotional state according to the target identity within the set time length are the same may occur, in order to determine the target content, the total recommended times of each candidate content within the set time length are determined, and the candidate content corresponding to the maximum value of the total times in each candidate content is taken as the target content.
Specifically, the total number of times that the candidate content is recommended in a set time length is the total number of times that the candidate content is recommended in any emotional state by all the identifiers in a specific time length.
Fig. 5a-5c are diagrams of examples of knowledge-graph based content recommendation provided by some embodiments of the present application, and a content recommendation process is described in detail in connection with the family knowledge graph:
based on fig. 5a, when the target identity of the target user of the acquired input voice information is "mom" and the acquired target emotion state is "happy", the contents of connection of the target identity in the target emotion state, which are "song 1" and "song 2", are found based on the family knowledge graph. In order to determine the target content, the recommended times of the two contents of song 1 and song 2 which are stored in the family knowledge graph and are connected with the target identity of mom and the target emotional state of happy are known within the set time length, the recommended times of the song 1 and the song 2 within the set time length are respectively 3 times and 4 times, the recommended times of the two contents within the set time length are different, and therefore the song 2 with the most recommended times is taken as the target content and recommended.
Based on the graph shown in fig. 5b, when the acquired target identification is "mom" and the target emotion state is "happy", two contents, namely "song 2" and "song 1", are found based on the family knowledge map. "song 2" and "song 1" are both recommended 4 times within a set length of time, but the total number of times "song 2" is recommended is greater than the total number of times "song 1" is recommended, and therefore "song 2" is targeted and recommended.
Based on fig. 5c, when the obtained target identification mark is "mom" and the target emotional state is "happy", based on the family knowledge graph, two candidate contents are found. The recommended times of the set time lengths of "song 2" and "song 1" are not recommended, that is, both are 0 times, but the total recommended times of "song 2" is less than the total recommended times of listening to "song 1", and therefore "song 1" is targeted and recommended. If the total recommended times of "song 2" and "song 1" are both also 0, an optional one of "song 2" and "song 1" is recommended to the user as the target content.
In order to facilitate the determination of the target content according to the recommended times within the set time length, on the basis of the foregoing embodiments, in this application embodiment, after the target content connected with the target identity in the target emotional state is determined and recommended, the method further includes:
and updating the target identity marks stored in the family atlas according to the recommendation date of the target content connected by the target emotion and the recommendation times within a set time length.
After the target content connected by the target identity with the target emotion is determined, the target content is recommended to a target user, and in order to ensure the accuracy of subsequent recommendation, in the application, the recommendation date of the target content connected by the target emotion and the recommendation times within a set time length of the target identity stored in the family atlas are updated.
In order to facilitate the user to know the searched state, on the basis of the foregoing embodiments, in this application, the method further includes:
and if the target content connected by the target identity in the target emotional state is not found, outputting prompt information of the not found content.
And searching target content connected with the target identity in the target emotional state according to the emotional state connected between the identity and the content stored in the family knowledge map, and if the target content is not searched, outputting prompt information of the content which is not searched. And after receiving the prompt message, the user can select whether to interact with the intelligent equipment for the next time.
The following describes the content recommendation process provided in the embodiment of the present application in detail with reference to a specific embodiment.
FIG. 6 is a schematic diagram of a specific content-based recommendation process provided by some embodiments of the present application;
s601: and acquiring a target identity and a target emotional state.
Specifically, the intelligent device receives voice information, determines the intention of the voice information and contained keywords, matches the voice information with the pre-registered voice information when the contained keywords do not contain the necessary types of keywords corresponding to the intention, and obtains the target identity according to the matching condition.
S602: and determining the content connected with the target identity by the target emotional state in the knowledge graph according to the target identity and the target emotional state.
S603: and if more than two contents exist, taking each content as a candidate content, and inquiring the recommended times of each candidate content which is stored in the family knowledge graph and connected by the target identity in the target emotional state within a set time length.
S604: and judging whether the recommended times of each candidate content in the set time length are equal, if not, executing S605, and if so, executing S606.
S605: and taking the candidate content corresponding to the maximum recommended times as the target content.
S606: and searching the recommended total times of each candidate content in a set time length based on the knowledge graph, and determining the candidate content corresponding to the maximum value of the total times as the target content.
The information completeness determination process provided by the embodiment of the present application is described in detail below with reference to a specific embodiment.
Figure BDA0002644608910000141
As shown in the above table, when the speech information is "next celadon of zhongjilun", the corresponding question method is extracted as "next character", the speech information is converted into text information through speech recognition, then the text information is subjected to semantic analysis, the obtained semantic information is "played song", the intention corresponding to the speech information is "next music", and the keyword of the necessary type corresponding to the intention is the name of the music. And entering a corresponding domain module for processing according to the intention, wherein the domain module corresponding to the intention is 'music'. The parameter corresponding to the played music is < star index ═ 1'/> < star index ═ 2'/>, keywords "zhou jieren" and "blue and white porcelain" are extracted according to the parameter, and it is determined that the keywords include the necessary type of keywords corresponding to the intention, so the information is complete.
When the voice information is 'telling' the story, the corresponding method is abstracted to 'telling', the voice information is converted into text information through voice recognition, then semantic analysis is carried out on the text information, the obtained semantic information is 'telling', the intention corresponding to the voice information is 'telling', and the key words of the necessary types corresponding to the intention are the names of the stories. And entering a corresponding domain module for processing according to the intention, wherein the domain module corresponding to the intention is a "store". However, since there is no corresponding parameter for extracting the keyword, the keyword is not obtained, and thus the information is incomplete.
Fig. 7 is a schematic structural diagram of a content recommendation device according to some embodiments of the present application, where the content recommendation device includes:
a first determining module 701, configured to determine an intention of the received voice message and a keyword included in the received voice message;
an extracting module 702, configured to, if it is determined that the keyword does not include the keyword of the necessary type corresponding to the intention, obtain a target identity of a target user who inputs the voice information according to the voice information and a pre-registered identity of the user, and identify a target emotion state of the target user according to the voice information;
the second determining module 703 is configured to determine and recommend target content connected with the target identity in the target emotional state according to an emotional state of connection between the identity and the content stored in the family knowledge graph.
In a possible embodiment, the apparatus further comprises:
the first processing module is used for judging whether a target node corresponding to the content exists in the knowledge graph or not, and if not, a node corresponding to the content is created;
and the second processing module is used for acquiring a target identity of a target user inputting the voice information, identifying a target emotional state of the target user according to the voice information, judging whether the target identity is connected with the content in the target emotional state in the knowledge graph, if not, connecting a node corresponding to the target identity with a node corresponding to the content, recording the content connected with the target emotional state by the target identity, and if so, updating the frequency of the content connected with the target emotional state by the target identity.
In a possible implementation manner, the second determining module 703 is specifically configured to determine, according to the target identity and the target emotional state, candidate content in the family knowledge graph, which is connected to the target identity in the target emotional state; if more than two candidate contents exist, determining the recommended times of each candidate content in a set time length according to the recommended times of the content connected in an emotional state by the identity marks stored in the family knowledge graph in the set time length; and determining the target content according to the recommended times corresponding to each candidate content.
In a possible implementation manner, the second determining module 703 is specifically configured to determine, if the recommended times corresponding to each candidate content are the same, the total times that each candidate content is recommended within a set time length according to the times that the content connected in an emotional state is recommended within the set time length by the identity stored in the family knowledge graph, and determine the candidate content corresponding to the maximum value of the total times as the target content; and if the recommended times corresponding to each candidate content are different, taking the candidate content corresponding to the maximum value of the recommended times as the target content.
In a possible embodiment, the apparatus further comprises:
an updating module 704, configured to determine and recommend the target content connected with the target identity in the target emotion state, and update the recommendation times of the target identity stored in the family graph and the target content connected with the target emotion in the set time length.
The second determining module 703 is further configured to output a prompt message of the not-found content if the target content connected by the target emotion state to the target identity is not found.
On the basis of the foregoing embodiments, the present application further provides an intelligent device, where the intelligent device includes:
the sound pickup is used for collecting voice information;
a player for playing;
a controller to perform:
determining the intention and the keywords of the voice information collected by a sound pickup, and if the keywords do not contain the necessary types of keywords corresponding to the intention, acquiring the target identity and the target emotional state of the target user inputting the voice information according to the voice information;
and determining target content connected with the target identity by the target emotion according to the connection emotional state between the identity and the content stored in the knowledge graph, and controlling the player to play the target content.
The method comprises the steps that after voice information of a user is collected by a sound pick-up of the intelligent equipment, the intention and key words of the voice information are recognized, when the key words are determined to contain the key words of the necessary types corresponding to the intention, a player is directly controlled to play the content determined by the key words of the necessary types corresponding to the intention, when the key words are determined not to contain the key words of the necessary types corresponding to the intention, the target identity identification and the target emotional state of the target user are determined according to the obtained voice information of the target user, the target content connected with the target identity identification in the emotional state is determined based on a knowledge graph, and then the player is controlled to play the target content.
The principle of content recommendation performed by the intelligent device is the same as that of content recommendation performed by the above embodiment, and repeated parts are not described again.
On the basis of the foregoing embodiments, the present application further provides an intelligent device, as shown in fig. 8, including: the system comprises a processor 801, a communication interface 802, a memory 803 and a communication bus 804, wherein the processor 801, the communication interface 802 and the memory 803 are communicated with each other through the communication bus 804.
The memory 803 has stored therein a computer program which, when executed by the processor 701, causes the processor 801 to perform the steps of:
determining the intention of the received voice information and the contained keywords;
if the keyword does not contain the necessary type keyword corresponding to the intention, acquiring a target identity of a target user inputting the voice information according to the voice information and a pre-registered identity of the user, and identifying a target emotional state of the target user according to the voice information;
and determining and recommending target content connected with the target identity by the target emotional state according to the emotional state connected between the identity and the content stored in the family knowledge graph.
Further, the processor 801 is further configured to determine and recommend content according to the necessary types of keywords corresponding to the intention.
Further, the processor 801 is further configured to determine whether a target node corresponding to the content exists in the knowledge graph, and if not, create a node corresponding to the content; acquiring a target identity of a target user inputting the voice information, identifying a target emotional state of the target user according to the voice information, judging whether the target identity is connected with the content in the target emotional state in the knowledge graph, if not, connecting a node corresponding to the target identity with a node corresponding to the content, recording the content connected with the target emotional state by the target identity, and if so, updating the frequency of the content connected with the target emotional state by the target identity.
Further, the processor 801 is further configured to determine, according to the target identity and the target emotional state, candidate content in the family knowledge graph, which is connected with the target identity in the target emotional state; if more than two candidate contents exist, determining the recommended times of each candidate content in a set time length according to the recommended times of the content connected in an emotional state by the identity marks stored in the family knowledge graph in the set time length; and determining the target content according to the recommended times corresponding to each candidate content.
Further, the processor 801 is further configured to determine, if the recommended times corresponding to each candidate content are different, a total recommended time of each candidate content within a set time length according to the recommended times of the content connected in an emotional state by the identity identifier stored in the family knowledge graph within the set time length, and determine the candidate content corresponding to the maximum value of the total recommended times as the target content; and if the recommended times corresponding to each candidate content are the same, taking the candidate content corresponding to the maximum value of the recommended times as the target content.
Further, the processor 801 is further configured to update the target identity stored in the family graph with a recommendation date and a recommendation number of times within a set time length of the target content connected by the target emotion.
Further, the processor 801 is further configured to, if the target content connected by the target identifier in the target emotional state is not found, output a prompt message indicating that the content is not found.
The communication bus mentioned in the above smart device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface 802 is used for communication between the above-described smart device and other devices.
The Memory may include a Random Access Memory (RAM) or a Non-Volatile Memory (NVM), such as at least one disk Memory. Alternatively, the memory may be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, including a central processing unit, a Network Processor (NP), and the like; but may also be a Digital instruction processor (DSP), an application specific integrated circuit, a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like.
On the basis of the foregoing embodiments, the present application further provides a computer-readable storage medium, in which a computer program executable by a smart device is stored, and when the program is run on the smart device, the smart device is caused to perform the following steps:
determining the intention of the received voice information and the contained keywords;
if the keyword does not contain the necessary type keyword corresponding to the intention, acquiring a target identity of a target user inputting the voice information according to the voice information and a pre-registered identity of the user, and identifying a target emotional state of the target user according to the voice information;
and determining and recommending target content connected with the target identity by the target emotional state according to the emotional state connected between the identity and the content stored in the family knowledge graph.
Further, when it is determined that the keywords include the necessary types of keywords corresponding to the intention, the method further includes: and determining and recommending the content according to the necessary types of keywords corresponding to the intentions.
Further, the determining, according to the emotional state of the connection between the identity and the content stored in the family knowledge graph, the target content connected with the target identity in the target emotional state includes: determining candidate content connected with the target identity in the target emotion state in the family knowledge graph according to the target identity and the target emotion state;
if more than two candidate contents exist, determining the recommended times of each candidate content in a set time length according to the recommended times of the content connected in an emotional state by the identity marks stored in the family knowledge graph in the set time length; and determining the target content according to the recommended times corresponding to each candidate content.
Further, the determining the target content according to the recommended times corresponding to each candidate content includes:
if the recommended times corresponding to each candidate content are the same, determining the total recommended times of each candidate content in a set time length according to the recommended times of the emotional-state-connected content in the set time length by the identity marks stored in the family knowledge graph, and determining the candidate content corresponding to the maximum value of the total times as the target content; and if the recommended times corresponding to each candidate content are different, taking the candidate content corresponding to the maximum value of the recommended times as the target content.
Further, after determining and recommending the target content connected with the target identity in the target emotional state, the method further comprises: and updating the target identity marks stored in the family atlas according to the recommendation date of the target content connected by the target emotion and the recommendation times within a set time length.
Further, the method further comprises: and if the target content connected by the target identity in the target emotional state is not found, outputting prompt information of the not found content.
After determining the intention of the received voice information and the contained keywords, judging if the keywords do not contain the necessary types of keywords corresponding to the intention, acquiring the target identity of the target user inputting the voice information according to the voice information and the identity of the user registered in advance, identifying the target emotional state of the target user according to the voice information, and determining and recommending the target content connected with the target identity in the target emotional state according to the emotional state connected between the identity and the content stored in the family knowledge graph, so that the method can avoid the problem of poor user experience caused by directly outputting the prompt information of the unseen content when the acquired voice information is incomplete due to the fact that the necessary types of keywords corresponding to the intention are not contained in the keywords acquired by the intelligent equipment in the man-machine interaction process, the mutual inductance of human-computer interaction is enhanced, and the intellectualization of the interpersonal interaction system is improved.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims (10)

1. A method for recommending content, the method comprising:
determining the intention of the received voice information and the contained keywords;
if the keyword does not contain the necessary type keyword corresponding to the intention, acquiring a target identity of a target user inputting the voice information according to the voice information and a pre-registered identity of the user, and identifying a target emotional state of the target user according to the voice information;
and determining and recommending target content connected with the target identity by the target emotional state according to the emotional state connected between the identity and the content stored in the knowledge graph.
2. The method of claim 1, wherein when determining that the keywords include the necessary types of keywords corresponding to the intention, the method further comprises:
and determining and recommending the content according to the necessary types of keywords corresponding to the intentions.
3. The method of claim 2, further comprising:
judging whether a target node corresponding to the content exists in the knowledge graph or not, and if not, creating a node corresponding to the content;
acquiring a target identity of a target user inputting the voice information, identifying a target emotional state of the target user according to the voice information, judging whether the target identity is connected with the content in the target emotional state in the knowledge graph, if not, connecting a node corresponding to the target identity with a node corresponding to the content, recording the content connected with the target emotional state by the target identity, and if so, updating the frequency of the content connected with the target emotional state by the target identity.
4. The method of claim 1, wherein determining the target content connected with the target emotion state according to the emotion state of the connection between the identity and the content stored in the knowledge graph comprises:
determining candidate content connected with the target identity by the target emotional state in the knowledge graph according to the target identity and the target emotional state;
if more than two candidate contents exist, determining the recommended times of each candidate content in a set time length according to the recommended times of the content connected in an emotional state by the identity marks stored in the knowledge graph in the set time length;
and determining the target content according to the recommended times corresponding to each candidate content.
5. The method according to claim 4, wherein the determining the target content according to the recommended times corresponding to each candidate content comprises:
if the recommended times corresponding to each candidate content are the same, determining the total recommended times of each candidate content in a set time length according to the recommended times of the emotional-state-connected content in the set time length by the identity marks stored in the knowledge graph, and determining the candidate content corresponding to the maximum value of the total times as the target content;
and if the recommended times corresponding to each candidate content are different, taking the candidate content corresponding to the maximum value of the recommended times as the target content.
6. The method of claim 1, wherein after determining and recommending the target content associated with the target emotional state to the target identity, the method further comprises:
and updating the target identity marks stored in the family atlas according to the recommendation date of the target content connected by the target emotion and the recommendation times within a set time length.
7. The method of claim 1, further comprising:
and if the target content connected by the target identity in the target emotional state is not found, outputting prompt information of the not found content.
8. An intelligent device, characterized in that the intelligent device comprises a processor for implementing the steps of the method according to any one of claims 1-7 when executing a computer program stored in a memory.
9. A smart device, the smart device comprising:
the sound pickup is used for collecting voice information;
a player for playing;
a controller to perform:
determining the intention and the keywords of the voice information collected by a sound pickup, and if the keywords do not contain the necessary types of keywords corresponding to the intention, acquiring the target identity and the target emotional state of the target user inputting the voice information according to the voice information;
and determining target content connected with the target identity by the target emotion according to the connection emotional state between the identity and the content stored in the knowledge graph, and controlling the player to play the target content.
10. A computer-readable storage medium, characterized in that it stores a computer program which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
CN202010850633.7A 2020-08-21 2020-08-21 Content recommendation method, device and medium Pending CN113486233A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010850633.7A CN113486233A (en) 2020-08-21 2020-08-21 Content recommendation method, device and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010850633.7A CN113486233A (en) 2020-08-21 2020-08-21 Content recommendation method, device and medium

Publications (1)

Publication Number Publication Date
CN113486233A true CN113486233A (en) 2021-10-08

Family

ID=77932721

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010850633.7A Pending CN113486233A (en) 2020-08-21 2020-08-21 Content recommendation method, device and medium

Country Status (1)

Country Link
CN (1) CN113486233A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115356939A (en) * 2022-08-18 2022-11-18 青岛海尔科技有限公司 Control command transmission method, control device, storage medium, and electronic device

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115356939A (en) * 2022-08-18 2022-11-18 青岛海尔科技有限公司 Control command transmission method, control device, storage medium, and electronic device

Similar Documents

Publication Publication Date Title
JP7150770B2 (en) Interactive method, device, computer-readable storage medium, and program
CN107146612B (en) Voice guidance method and device, intelligent equipment and server
US9582757B1 (en) Scalable curation system
US20180052824A1 (en) Task identification and completion based on natural language query
CN109165302B (en) Multimedia file recommendation method and device
CN108364650B (en) Device and method for adjusting voice recognition result
CN109979450B (en) Information processing method and device and electronic equipment
CN103345467A (en) Speech translation system
CN111261151B (en) Voice processing method and device, electronic equipment and storage medium
CN106486120B (en) Interactive voice response method and answering system
CN111522909B (en) Voice interaction method and server
CN106407393B (en) information processing method and device for intelligent equipment
CN111312230B (en) Voice interaction monitoring method and device for voice conversation platform
CN111210842A (en) Voice quality inspection method, device, terminal and computer readable storage medium
CN111178081B (en) Semantic recognition method, server, electronic device and computer storage medium
CN111462741B (en) Voice data processing method, device and storage medium
CN111354357A (en) Audio resource playing method and device, electronic equipment and storage medium
CN106649253A (en) Auxiliary control method and system based on post verification
WO2021135561A1 (en) Skill voice wake-up method and apparatus
JP2020009440A (en) Method and device for generating information
CN111414512A (en) Resource recommendation method and device based on voice search and electronic equipment
CN111309857A (en) Processing method and processing device
JP2019091416A5 (en)
CN115509485A (en) Filling-in method and device of business form, electronic equipment and storage medium
CN112182046A (en) Information recommendation method, device, equipment and medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination