CN107886949B - Content recommendation method and device - Google Patents
Content recommendation method and device Download PDFInfo
- Publication number
- CN107886949B CN107886949B CN201711195319.4A CN201711195319A CN107886949B CN 107886949 B CN107886949 B CN 107886949B CN 201711195319 A CN201711195319 A CN 201711195319A CN 107886949 B CN107886949 B CN 107886949B
- Authority
- CN
- China
- Prior art keywords
- user
- recommendation
- interactive
- knowledge base
- target user
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 74
- 230000002452 interceptive effect Effects 0.000 claims abstract description 111
- 239000013598 vector Substances 0.000 claims description 32
- 230000003993 interaction Effects 0.000 claims description 22
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 238000000556 factor analysis Methods 0.000 claims description 5
- 238000001228 spectrum Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 description 9
- 238000013135 deep learning Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001568 sexual effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/225—Feedback of the input speech
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The embodiment of the invention provides a content recommendation method and a content recommendation device, wherein the method comprises the following steps: acquiring voice request data of a target user, and extracting corresponding interactive recommendation features in the voice request data; according to the interactive recommendation characteristics, acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users. The method greatly facilitates the use of the user, and combines acoustics, semantics, user attributes and speaker information requested by the target user, thereby further improving the accuracy of content recommendation.
Description
Technical Field
The present invention relates to the technical field of speech signal processing and natural language processing, and more particularly, to a content recommendation method and apparatus.
Background
With the coming of the wave of "artificial intelligence", more and more intelligent devices move into the visual field of ordinary users, and users increasingly rely on using intelligent devices to realize their own requirements, such as amazon's Echo, Google Home's intelligent sound box, hundreds of intelligent robots, household intelligent televisions, etc., when these intelligent devices are used, generally, a plurality of users share the same intelligent device, such as family members use the same intelligent television to watch television, the same robot serves a plurality of different clients, etc.; the intelligent recommendation method on the existing intelligent equipment can definitely greatly improve the stickiness of the user to the intelligent equipment and reserve the user for the intelligent equipment, when the content is recommended for the user, the user generally needs to register an account in advance, the intelligent equipment puts the data of the same user together, and the content which is interested by the user is recommended according to the historical information of the user each time; however, for the unregistered user, the recommended content cannot be provided, and the user experience is poor.
In summary, it is desirable to provide a more convenient and accurate content recommendation scheme in the prior art.
Disclosure of Invention
In order to solve the above problems, embodiments of the present invention provide a content recommendation method and apparatus that overcome the above problems or at least partially solve the above problems.
According to a first aspect of embodiments of the present invention, there is provided a content recommendation method, including:
acquiring voice request data of a target user, and extracting corresponding interactive recommendation features in the voice request data;
according to the interactive recommendation characteristics, acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users.
According to a second aspect of embodiments of the present invention, there is provided a content recommendation apparatus including:
the acquisition module is used for acquiring voice request data of a target user and extracting corresponding interactive recommendation characteristics in the voice request data;
the recommendation module is used for acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user according to the interactive recommendation characteristics and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users.
According to a third aspect of embodiments of the present invention, there is provided an electronic apparatus, including:
at least one processor; and
at least one memory communicatively coupled to the processor, wherein:
the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the content recommendation method provided by any of the various possible implementations of the first aspect.
According to a fourth aspect of the present invention, there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the content recommendation method provided in any one of the various possible implementations of the first aspect.
The embodiments of the present invention provide a content recommendation method and apparatus, which greatly facilitate the use of users and combine acoustics, semantics, user attributes and speaker information requested by target users, thereby further improving the accuracy of content recommendation.
Drawings
Fig. 1 is a flowchart illustrating a content recommendation method according to an embodiment of the present invention;
FIG. 2 is a block diagram of a content recommendation device according to an embodiment of the present invention;
fig. 3 is a block diagram of an electronic device according to an embodiment of the invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
The existing recommendation method generally requires a user to register an account in advance; when a user needs to interact with the intelligent equipment, a corresponding account needs to be logged in first, and after the identity of the user is determined, content which the user is interested in is recommended according to historical information of the target user using the intelligent equipment; if a user needs to pre-input a plurality of sentences of voice data for voiceprint registration, when the user uses the intelligent equipment, voiceprint recognition is carried out according to the voice data requested by the user, and after the user identity is determined, contents which are interested by the user are recommended based on the historical information of the target user; if the user does not register in advance, the content cannot be accurately recommended to the user; therefore, the existing method has high requirements on users, and particularly has poor user experience for non-login users.
The method comprises the steps of collecting a large amount of user request voice data in advance, clustering the collected voice data according to the interactive recommendation characteristics of each piece of voice data, automatically constructing a recommendation knowledge base, and recommending contents for a target user according to the recommendation knowledge base; the recommended content may be a television series, a movie, a song, news, purchased goods, transacted business, and the like; the method is mainly applied to a scene that a plurality of users use the same intelligent device, the intelligent device needs to recommend proper content for each user, for example, family members use various household intelligent devices such as an intelligent television, an intelligent sound box, an intelligent air conditioner and the like together, the intelligent device needs to recommend content interested by each family member according to the specific situation of each family member, or the intelligent device can be an intelligent robot serving the public.
In view of the above situation, fig. 1 is a schematic flowchart illustrating an overall content recommendation method according to an embodiment of the present invention. For convenience of description, the embodiment of the present invention takes an execution subject as an example of an intelligent device. In general, the method includes the following steps.
S1, acquiring voice request data of a target user, and extracting corresponding interactive recommendation features in the voice request data.
The method comprises the steps of firstly obtaining voice request data newly sent by a current target user, and obtaining at least one item of interactive recommendation feature from the voice request data, so that corresponding recommendation content can be obtained from a pre-constructed recommendation knowledge base corresponding to the current target user according to the extracted interactive recommendation feature in the subsequent steps, and the recommendation content is pushed to the current target user.
S2, acquiring corresponding recommended content from a pre-constructed recommended knowledge base corresponding to the target user according to the interactive recommendation characteristics, and sending the recommended content to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users.
The recommendation knowledge base is constructed by collecting interaction information between a large number of other historical users and the intelligent equipment executed by the scheme in advance, wherein the interaction information comprises user IDs, user voice data, interaction recommendation characteristics corresponding to the user voice data and contents recommended by the intelligent equipment for the historical users correspondingly. Specifically, the recommendation knowledge base is constructed in advance according to a large number of voice data of interaction between the historical users and the intelligent device, and the voice data requested by each historical user, the corresponding interaction recommendation characteristics of the voice data, and the cluster type of the voice data of each historical user are stored. And different clustering categories generated in the recommendation knowledge base respectively form a corresponding relation model, and the corresponding relation model is established based on interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user, wherein the interactive recommendation features extracted from sample voice request data of each historical user and the recommendation contents corresponding to each historical user are included in the corresponding user category, each historical user belonging to the user category, the interactive recommendation features extracted from sample voice request data of each historical user and the recommendation contents corresponding to each historical user.
It is further noted that the interaction recommendation feature can include at least one of: acoustic information, semantic information, user attribute information, speaker information of the user and the use condition of the recommended content by the user.
Specifically, acquiring corresponding recommendation content from a pre-constructed recommendation knowledge base corresponding to the current target user according to the extracted interactive recommendation features, and pushing the recommendation content to the current target user, specifically including: the method comprises the steps of firstly determining a cluster category of a current target user in a recommendation knowledge base according to interactive recommendation characteristics of voice request data of the current user, and then recommending at least one piece of recommended content corresponding to other user requests under the cluster category to the current target user.
According to the method, a recommendation knowledge base is constructed by collecting a large amount of voice interaction information of historical users and intelligent equipment in advance, the user request category corresponding to the current target user request in the recommendation knowledge base is found according to the interaction recommendation characteristics contained in the current target user request voice data, content recommendation is carried out, the current target user does not need to register an account in advance, the method is convenient for the user to use, meanwhile, various characteristic parameters such as acoustics, semantics, user attributes and speaker information in the user request voice data are combined during recommendation, and the recommendation accuracy is further improved.
On the basis of the above specific embodiment of the present invention, a content recommendation method is provided, where the correspondence model includes the following four information and association relations therebetween: the method comprises the following steps of obtaining a user category, historical users belonging to the user category, interactive recommendation features extracted from sample voice request data of the historical users and recommendation contents corresponding to the historical users.
On the basis of the above specific embodiment of the present invention, a content recommendation method is provided, in which the recommendation knowledge base is constructed by the following steps: acquiring interactive recommendation characteristics from sample voice request data of each historical user, clustering the interactive recommendation characteristics of each historical user, and correspondingly acquiring each user category and each historical user belonging to each user category according to clustering results; at least one user category and each corresponding historical user belonging to the user category, interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user form a corresponding relation model.
And the interactive recommendation features extracted from the sample voice request data of the historical users and the recommendation contents corresponding to the historical users form a corresponding relation model.
Specifically, different clustering categories generated in the recommendation knowledge base respectively form a corresponding relationship model, and the corresponding relationship model is established based on interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user, wherein the interactive recommendation features extracted from sample voice request data of each historical user, the interactive recommendation features belonging to the user category, and the recommendation contents corresponding to each historical user are included in the corresponding user category, each historical user belonging to the user category, the interactive recommendation features extracted from sample voice request data of each historical user, and the recommendation contents corresponding to each historical user.
In addition, the recommendation knowledge base is specifically constructed by the following steps.
Collecting a large amount of sample data, wherein the sample data is interactive information of each historical user and intelligent equipment; the interactive information comprises request voice data of each historical user when interacting with the intelligent equipment each time, and the intelligent equipment recommends contents to each historical user. The sample data can be directly collected from the intelligent equipment or the server of the intelligent equipment according to the ID number of the intelligent equipment used by the user.
And step two, extracting interaction recommendation characteristics of the request voice data when each piece of sample data interacts with the intelligent equipment according to a large amount of collected sample data. The interaction recommendation feature can include a combination of at least one or more of: acoustic features, semantic features, user attribute features, user speaker features of the user request voice data, and recommended content of whether the user uses the intelligent device. The method for extracting the acoustic feature, the semantic feature and the user attribute feature of the user request voice data may be any extraction method in the prior art, and this embodiment is not specifically limited herein; whether the user uses the recommended content of the intelligent device can be directly determined according to the specific operation of the user on the recommended content in the user intelligent device.
And step three, clustering the collected voice data requested by each user according to the interactive recommendation characteristics of the voice data requested by the user. The recommended features obtained in the second step are in a vector form, and a vector clustering method (such as a K-means method, which is not specifically limited in this embodiment) in the prior art may be adopted to cluster the recommended features of the voice data of each user, and divide the user request voice data of each user into a plurality of clustering categories, where the categories are user categories.
And step four, constructing a recommendation knowledge base according to the cluster type (user type) to which the voice data of each historical user request belongs. Specifically, each piece of historical user request voice data, corresponding interactive recommendation features and corresponding cluster types are stored in a recommendation knowledge base, and each cluster type (user type) and corresponding historical users belonging to the user type form a corresponding relation model together with the interactive recommendation features extracted from sample voice request data of the historical users and recommendation contents corresponding to the historical users.
On the basis of the foregoing specific embodiments of the present invention, a content recommendation method is provided, where according to the interactive recommendation feature, a recommendation knowledge base corresponding to the target user is pre-constructed to obtain corresponding recommended content and send the recommended content to the target user, and the method includes: calculating the distance between the interactive recommendation characteristics of the target user and the clustering center of the interactive recommendation characteristics of the sample voice request data of each historical user in the recommendation knowledge base; and recommending at least one recommended content corresponding to the user category to which the target user belongs in the recommendation knowledge base to the target user by taking the user category to which the history user with the minimum distance belongs in the recommendation knowledge base as the user category to which the target user belongs in the recommendation knowledge base.
On the basis of the foregoing specific embodiments of the present invention, a content recommendation method is provided, where according to the interactive recommendation feature, a recommendation knowledge base corresponding to the target user is pre-constructed to obtain corresponding recommended content and send the recommended content to the target user, and the method includes:
calculating the distance between the interactive recommendation characteristics of the target user and the clustering center of the interactive recommendation characteristics of the sample voice request data of each historical user in the recommendation knowledge base; and taking the user category to which the historical user with the minimum distance belongs in the recommendation knowledge base as the user category to which the target user belongs in the recommendation knowledge base, acquiring at least one sample voice request data with the highest similarity to the interactive recommendation feature of the target user from a corresponding relation model in which the user category is located in the recommendation knowledge base, and recommending the recommendation content corresponding to the sample voice request data to the target user.
On the basis of the above specific embodiment of the present invention, a content recommendation method is provided, where the interactive recommendation feature at least includes at least one of the following: acoustic characteristics of user request voice data, semantic characteristics of the user request voice data, user attribute characteristics and user speaker characteristics; correspondingly, the interactive recommendation feature is extracted through at least one of the following steps:
acquiring a frequency spectrum characteristic corresponding to user voice request data as an acoustic characteristic;
carrying out voice recognition on user voice request data to obtain a corresponding recognition text, and obtaining semantic features according to the recognition text;
inputting the acoustic features of the user into a pre-constructed user attribute feature determination model to obtain user attribute features; the user attribute features include at least one of: gender, age, and occupation;
and respectively extracting the user speaker feature vector TV-vector and the user speaker feature vector CNN-vector based on an all-variable factor analysis method and a pre-constructed convolutional neural network model according to the acoustic features.
The method for extracting the acoustic feature, the semantic feature and the user attribute feature of the user request voice data may be any extraction method in the prior art, and this embodiment is not specifically limited herein; in the following examples, some specific implementation manners for extracting the above-mentioned interactive features from the user voice data are illustrated, and are not limited specifically.
Extracting acoustic features of user request voice data: firstly, carrying out voice recognition on voice data requested by a user to obtain a recognition text of the voice data requested by the user; specifically, the acoustic feature is generally a spectrum feature of voice data, and is used to express timbre information of a user, and in the prior art, the acoustic feature of the user may be expressed by the following features, and the embodiment of the present invention is not particularly limited herein: PLP (Perceptual Linear Predictive coefficient) features, FilterBank features, MFCC (Mel-Frequency Cepstral Coefficients) features, and the like.
Subsequently, the existing speech recognition method is used to perform speech recognition on the user request speech data (the specific extraction method is the same as the prior art, and the embodiment is not specifically limited herein), so as to obtain a corresponding recognition text. Further, according to the recognition text of the user request voice data, extracting the semantic features of the user request voice data by utilizing the prior art; the semantic features are used for representing semantic information of the user request voice data, and specifically can be represented by word vectors or sentence vectors of the user request voice data recognition text; the sentence vectors may be obtained by adding Word vectors of each Word in the recognized text and then averaging, the Word vector extraction method is the same as that in the prior art, for example, a Word2Vec technology is adopted to extract the Word vectors of each Word in the recognized text, and this embodiment is not limited specifically herein.
The user attribute features corresponding to the user request voice data are attribute information used for describing the user's own special attribute, and can be the gender, age, occupation and the like of the user. Taking age as an example, the attribute of sexual age is children, young, middle-aged and old, or under 10 years old, 10 to 30 years old, and over 30 years old. When various user attribute features are extracted specifically, the acoustic feature extraction is directly carried out according to user request voice data, specifically, a pre-constructed user attribute feature is utilized to determine model determination, the model is input into the acoustic feature of each user request voice data, and the acoustic feature is output as an attribute feature value of a user; taking the attribute characteristics of the user as age as an example, the input of a preset model is voice request data of the user, and the output of the model is a specific age attribute value, namely young, middle-aged or old (under 10 years old, 10 to 30 years old or over 30 years old); the present embodiment is not particularly limited herein. The user attribute feature determination model is generally a neural network model in a deep learning method, such as a deep neural network model, a convolutional neural network model, and the like, and the specific construction method is the same as that in the prior art and is not described in detail herein; the present embodiment is not particularly limited herein.
The speaker characteristics corresponding to the user request speech data are speaker information used for characterizing the user, and in order to prevent the user from influencing speaker characteristic extraction due to different channels or environmental phonemes, the specific speaker characteristics can be obtained by any means in the prior art, which is not specifically limited in this embodiment. In the following embodiment, two speaker characteristics, namely a TV-vector characteristic and a CNN-driver characteristic, are extracted by respectively using the fully variable factor analysis method and the deep neural network in the prior art.
The Gaussian mixture-generic background model (GMM _ UBM) is a mainstream method for obtaining speaker characteristics. GMM _ UBM utilizes a Gaussian function to fit the characteristics of each person when speaking, and takes the problem of insufficient voice data quantity into consideration, a small amount of data of the speaker is self-adapted through a Universal Background Model (UBM) to obtain a target speaker model; a certain section of voice of a certain speaker can be represented by a mean value super vector M; the method for analyzing the all-variable factor comprises the steps of projecting the mean value super vector M to an all-variable space T, and obtaining the all-variable factor w representing information of a speaker, namely i-vector characteristics without distinguishing a speaker space, a channel space and a residual error space, wherein M is a UBM mean value vector; the i-vector feature includes not only speaker information but also channel information and residual information, and therefore channel compensation is required. Because the embodiment is an automatic recommendation process, namely an unsupervised process, a Principal Component Analysis (PCA) technology is adopted immediately, high-dimensional data is projected to low-dimensional data, and a projection which can represent the original high-dimensional data most in the sense of minimum mean square error is searched; and finally, obtaining the speaker characteristic vector TV _ ivctor under the full-variable factor analysis method.
In the embodiment, a deep learning method, namely, a Convolutional Neural Network (CNN) is adopted to extract CNN-driver characteristics, the network input is the frequency spectrum characteristics (such as FBank characteristics) of voice data of a speaker, the network output is the characteristic vector representing speaker information, the obtained characteristic vector is subjected to dimensionality reduction by adopting PCA, and the speaker characteristic vector CNN-driver under the deep learning method is obtained after principal components are extracted.
On the basis of the above specific embodiments of the present invention, a content recommendation method is provided, where the interactive recommendation feature further includes a feature of whether to use recommended content.
It should be noted that, when the current target user uses the intelligent device for the first time to make a voice request, since the intelligent device has not requested the voice data recommendation content for the current user, the recommendation content characteristic of whether the user uses the intelligent device is null. In addition, after the current target user finishes recommending content for at least one time, the user category of the current target user in the recommendation knowledge base can be dynamically adjusted according to the recommendation content of whether the current target user uses the intelligent device or not, that is, the distance between the current user request voice data and the cluster center of each type of user request voice data in the recommendation knowledge base is recalculated according to the recommendation content of whether the current user uses the intelligent device or not, so as to obtain a new cluster category of the current user request voice data, and more appropriate content is recommendable to the current target user according to the category to which the newly calculated current user request voice data belongs.
On the basis of the above specific embodiment of the present invention, a content recommendation method is provided, which further includes: and adding the user category of the target user, the interactive recommendation characteristics extracted from the sample voice request data of the target user, the corresponding recommendation content of the target user and the corresponding relation among the three into the recommendation knowledge base.
Specifically, after the content recommendation of the current target user is completed, the voice data requested by the current user, the interactive recommendation characteristics and the cluster type to which the interactive recommendation characteristics belong can be added into the recommendation knowledge base, and the recommendation knowledge base is updated; alternatively, the recommendation knowledge base is updated periodically, such as once a month or week.
Fig. 2 is a schematic diagram illustrating an overall framework of a content recommendation apparatus according to an embodiment of the present invention. In general, the apparatus includes the following modules.
The obtaining module a1 is configured to obtain voice request data of a target user, and extract a corresponding interactive recommendation feature in the voice request data.
In this embodiment, the obtaining module a1 in the apparatus first obtains voice request data newly sent by a current target user, and obtains at least one interactive recommendation feature from the voice request data, so that the recommending module a2 obtains corresponding recommended content from a pre-constructed recommendation knowledge base corresponding to the current target user according to the extracted interactive recommendation feature, and pushes the recommended content to the current target user.
The recommending module A2 is used for acquiring corresponding recommended content from a pre-constructed recommending knowledge base corresponding to the target user according to the interactive recommending feature and sending the recommended content to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users.
The recommendation knowledge base is constructed by collecting interaction information between a large number of other historical users and intelligent equipment executed by the scheme in advance, wherein the interaction information comprises user IDs, user voice data, interaction recommendation characteristics corresponding to the user voice data and contents which can only be recommended by the equipment for each historical user correspondingly. Specifically, the recommendation knowledge base is constructed in advance according to a large number of voice data of interaction between the historical users and the intelligent device, and the voice data requested by each historical user, the corresponding interaction recommendation characteristics of the voice data, and the cluster type of the voice data of each historical user are stored. And different clustering categories generated in the recommendation knowledge base respectively form a corresponding relation model, and the corresponding relation model is established based on interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user, wherein the interactive recommendation features extracted from sample voice request data of each historical user and the recommendation contents corresponding to each historical user are included in the corresponding user category, each historical user belonging to the user category, the interactive recommendation features extracted from sample voice request data of each historical user and the recommendation contents corresponding to each historical user.
It is further noted that the interaction recommendation feature can include at least one of: acoustic information, semantic information, user attribute information, speaker information of the user and the use condition of the recommended content by the user.
Specifically, the recommending module a2 obtains corresponding recommended content from a pre-constructed recommendation knowledge base corresponding to the current target user according to the extracted interactive recommendation features, and pushes the recommended content to the current target user, which specifically includes: the method comprises the steps of firstly determining a cluster category of a current target user in a recommendation knowledge base according to interactive recommendation characteristics of voice request data of the current user, and then recommending at least one piece of recommended content corresponding to other user requests under the cluster category to the current target user.
According to the device, a recommendation knowledge base is built by collecting a large amount of voice interaction information of historical users and intelligent equipment in advance, the user request category corresponding to the current target user request in the recommendation knowledge base is found according to the interaction recommendation characteristics contained in the current target user request voice data, content recommendation is carried out without the need of pre-registering an account number by the current target user, the device is convenient for the user to use, meanwhile, various characteristic parameters such as acoustics, semantics, user attributes and speaker information in the user request voice data are combined during recommendation, and the recommendation accuracy is further improved.
On the basis of the above embodiment of the present invention, a content recommendation apparatus is provided, in which the correspondence model includes the following four information and association relations therebetween: the method comprises the following steps of obtaining a user category, historical users belonging to the user category, interactive recommendation features extracted from sample voice request data of the historical users and recommendation contents corresponding to the historical users.
The different clustering categories generated in the recommendation knowledge base respectively form a corresponding relation model, and the corresponding relation model is established based on interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user, wherein the corresponding relation model comprises the corresponding user category, each historical user belonging to the user category, the interactive recommendation features extracted from the sample voice request data of each historical user and the recommendation contents corresponding to each historical user.
On the basis of the above-described embodiments of the present invention, there is provided a content recommendation apparatus, wherein the recommendation knowledge base is constructed by the following steps: acquiring interactive recommendation characteristics from sample voice request data of each historical user, clustering the interactive recommendation characteristics of each historical user, and correspondingly acquiring each user category and each historical user belonging to each user category according to clustering results; at least one user category and each corresponding historical user belonging to the user category, interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user form a corresponding relation model.
And the interactive recommendation features extracted from the sample voice request data of the historical users and the recommendation contents corresponding to the historical users form a corresponding relation model.
The different clustering categories generated in the recommendation knowledge base respectively form a corresponding relation model, and the corresponding relation model is established based on interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user, wherein the corresponding relation model comprises the corresponding user category, each historical user belonging to the user category, the interactive recommendation features extracted from the sample voice request data of each historical user and the recommendation contents corresponding to each historical user.
Specifically, the recommendation knowledge base is constructed by the following steps.
Collecting a large amount of sample data, wherein the sample data is interactive information of each historical user and intelligent equipment; the interactive information comprises request voice data of each historical user when interacting with the intelligent equipment each time, and the intelligent equipment recommends contents to each historical user. The sample data can be directly collected from the intelligent equipment or the server of the intelligent equipment according to the ID number of the intelligent equipment used by the user.
And step two, extracting interaction recommendation characteristics of the request voice data when each piece of sample data interacts with the intelligent equipment according to a large amount of collected sample data. The interaction recommendation feature can include a combination of at least one or more of: acoustic features, semantic features, user attribute features, user speaker features of the user request voice data, and recommended content of whether the user uses the intelligent device. The method for extracting the acoustic feature, the semantic feature and the user attribute feature of the user request voice data may be any extraction method in the prior art, and this embodiment is not specifically limited herein; whether the user uses the recommended content of the intelligent device can be directly determined according to the specific operation of the user on the recommended content in the user intelligent device.
And step three, clustering the collected voice data requested by each user according to the interactive recommendation characteristics of the voice data requested by the user. The recommended features obtained in step 2 are in a vector form, and a vector clustering method (such as a K-means method, which is not specifically limited in this embodiment) in the prior art may be adopted to cluster the recommended features of the voice data of each user, and divide the user request voice data of each user into a plurality of clustering categories, where the categories are user categories.
And step four, constructing a recommendation knowledge base according to the cluster type (user type) to which the voice data of each historical user request belongs. Specifically, each piece of historical user request voice data, corresponding interactive recommendation features and corresponding cluster types are stored in a recommendation knowledge base, and each cluster type (user type) and corresponding historical users belonging to the user type form a corresponding relation model together with the interactive recommendation features extracted from sample voice request data of the historical users and recommendation contents corresponding to the historical users.
On the basis of the foregoing specific embodiment of the present invention, there is provided a content recommendation apparatus, where the recommendation module a2 is configured to: calculating the distance between the interactive recommendation characteristics of the target user and the clustering center of the interactive recommendation characteristics of the sample voice request data of each historical user in the recommendation knowledge base; and recommending at least one recommended content corresponding to the user category to which the target user belongs in the recommendation knowledge base to the target user by taking the user category to which the history user with the minimum distance belongs in the recommendation knowledge base as the user category to which the target user belongs in the recommendation knowledge base.
On the basis of the foregoing specific embodiment of the present invention, there is provided a content recommendation apparatus, where the recommendation module a2 is further configured to: calculating the distance between the interactive recommendation characteristics of the target user and the clustering center of the interactive recommendation characteristics of the sample voice request data of each historical user in the recommendation knowledge base; and taking the user category to which the historical user with the minimum distance belongs in the recommendation knowledge base as the user category to which the target user belongs in the recommendation knowledge base, acquiring at least one sample voice request data with the highest similarity to the interactive recommendation feature of the target user from a corresponding relation model in which the user category is located in the recommendation knowledge base, and recommending the recommendation content corresponding to the sample voice request data to the target user.
On the basis of the foregoing specific embodiments of the present invention, a content recommendation apparatus is provided, where the interactive recommendation feature at least includes at least one of the following: acoustic characteristics of user request voice data, semantic characteristics of the user request voice data, user attribute characteristics and user speaker characteristics; correspondingly, the interactive recommendation feature is extracted through at least one of the following steps:
acquiring a frequency spectrum characteristic corresponding to user voice request data as an acoustic characteristic;
carrying out voice recognition on user voice request data to obtain a corresponding recognition text, and obtaining semantic features according to the recognition text;
inputting the acoustic features of the user into a pre-constructed user attribute feature determination model to obtain user attribute features; the user attribute features include at least one of: gender, age, and occupation;
and respectively extracting the user speaker feature vector TV-vector and the user speaker feature vector CNN-vector based on an all-variable factor analysis method and a pre-constructed convolutional neural network model according to the acoustic features.
The interactive recommendation characteristics comprise one or more combinations of acoustic characteristics, semantic characteristics, user attribute characteristics, user speaker characteristics and recommended contents of whether the user uses the intelligent equipment or not of the user request voice data. Whether the user uses the recommended content of the intelligent equipment can be directly determined according to the specific operation of the user on the recommended content in the user intelligent equipment; the method for extracting the acoustic feature, the semantic feature, and the user attribute feature of the user request voice data may be any extraction method in the prior art, and this embodiment is not particularly limited herein, and the following examples are given respectively.
The method for extracting the acoustic feature, the semantic feature, and the user attribute feature of the voice data requested by the user may be any extraction method in the prior art, and the above corresponding method embodiments have been illustrated, and the embodiment is not specifically limited herein.
On the basis of the above specific embodiments of the present invention, a content recommendation apparatus is provided, where the interactive recommendation feature further includes a feature of whether to use recommended content.
It should be noted that, when the current target user uses the intelligent device for the first time to make a voice request, since the intelligent device has not requested the voice data recommendation content for the current user, the recommendation content characteristic of whether the user uses the intelligent device is null. In addition, after the current target user finishes recommending content for at least one time, the user category of the current target user in the recommendation knowledge base can be dynamically adjusted according to the recommendation content of whether the current target user uses the intelligent device or not, that is, the distance between the current user request voice data and the cluster center of each type of user request voice data in the recommendation knowledge base is recalculated according to the recommendation content of whether the current user uses the intelligent device or not, so as to obtain a new cluster category of the current user request voice data, and more appropriate content is recommendable to the current target user according to the category to which the newly calculated current user request voice data belongs.
On the basis of the foregoing specific embodiments of the present invention, there is provided a content recommendation apparatus, further including an update module, configured to: and adding the user category of the target user, the interactive recommendation characteristics extracted from the sample voice request data of the target user, the corresponding recommendation content of the target user and the corresponding relation among the three into the recommendation knowledge base.
Specifically, after the content recommendation of the current target user is completed, the voice data requested by the current user, the interactive recommendation characteristics and the cluster type to which the interactive recommendation characteristics belong can be added into the recommendation knowledge base, and the recommendation knowledge base is updated; alternatively, the recommendation knowledge base is updated periodically, such as once a month or week.
Based on the above specific embodiments, an electronic device is provided. Referring to fig. 3, the electronic device includes: a processor (processor)301, a memory (memory)302, and a bus 303;
the processor 301 and the memory 302 respectively complete communication with each other through a bus 303;
the processor 301 is configured to call the program instructions in the memory 302 to execute the content recommendation method provided by the above embodiment, for example, including: acquiring voice request data of a target user, and extracting corresponding interactive recommendation features in the voice request data; according to the interactive recommendation characteristics, acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users.
An embodiment of the present invention provides a non-transitory computer-readable storage medium, which stores computer instructions, where the computer instructions cause a computer to execute the content recommendation method provided in the foregoing embodiment, for example, including: acquiring voice request data of a target user, and extracting corresponding interactive recommendation features in the voice request data; according to the interactive recommendation characteristics, acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; and the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from the sample voice request data of the historical users and recommendation contents corresponding to the historical users.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The above-described embodiments of the information interaction device and the like are merely illustrative, where units illustrated as separate components may or may not be physically separate, and components displayed as units may or may not be physical units, may be located in one place, or may also be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute the various embodiments or some parts of the methods of the embodiments.
Finally, the method of the present application is only a preferred embodiment and is not intended to limit the scope of the embodiments of the present invention. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the embodiments of the present invention should be included in the protection scope of the embodiments of the present invention.
Claims (10)
1. A content recommendation method, comprising:
acquiring voice request data of a target user, and extracting corresponding interactive recommendation features in the voice request data, wherein the target user comprises a non-registered user;
according to the interactive recommendation characteristics, acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from sample voice request data of historical users corresponding to the user categories and recommendation contents corresponding to the historical users corresponding to the user categories;
the interactive recommendation feature includes a feature of whether to use recommended content, and further includes:
and dynamically adjusting the user category of the target user in a recommendation knowledge base according to whether the target user uses the recommended content.
2. The method according to claim 1, wherein the correspondence model includes the following four information and relationships between the information and the relationships: the method comprises the following steps of obtaining a user category, historical users belonging to the user category, interactive recommendation features extracted from sample voice request data of the historical users and recommendation contents corresponding to the historical users.
3. The method of claim 1, wherein the recommendation knowledge base is constructed by:
acquiring interactive recommendation characteristics from sample voice request data of each historical user, clustering the interactive recommendation characteristics of each historical user, and correspondingly acquiring each user category and each historical user belonging to each user category according to clustering results;
at least one user category and each corresponding historical user belonging to the user category, interactive recommendation features extracted from sample voice request data of each historical user and recommendation contents corresponding to each historical user form a corresponding relation model.
4. The method of claim 3, wherein according to the interactive recommendation feature, obtaining and sending corresponding recommended content to the target user from a pre-constructed recommendation knowledge base corresponding to the target user comprises:
calculating the distance between the interactive recommendation characteristics of the target user and the clustering center of the interactive recommendation characteristics of the sample voice request data of each historical user in the recommendation knowledge base; and recommending at least one recommended content corresponding to the user category to which the target user belongs in the recommendation knowledge base to the target user by taking the user category to which the history user with the minimum distance belongs in the recommendation knowledge base as the user category to which the target user belongs in the recommendation knowledge base.
5. The method of claim 3, wherein according to the interactive recommendation feature, obtaining and sending corresponding recommended content to the target user from a pre-constructed recommendation knowledge base corresponding to the target user comprises:
calculating the distance between the interactive recommendation characteristics of the target user and the clustering center of the interactive recommendation characteristics of the sample voice request data of each historical user in the recommendation knowledge base; and taking the user category to which the historical user with the minimum distance belongs in the recommendation knowledge base as the user category to which the target user belongs in the recommendation knowledge base, acquiring at least one sample voice request data with the highest similarity to the interactive recommendation feature of the target user from a corresponding relation model in which the user category is located in the recommendation knowledge base, and recommending the recommendation content corresponding to the sample voice request data to the target user.
6. The method of any of claims 1 to 5, wherein the interaction recommendation feature comprises at least one of: acoustic characteristics of user request voice data, semantic characteristics of the user request voice data, user attribute characteristics and user speaker characteristics; correspondingly, the interactive recommendation feature is extracted through at least one of the following steps:
acquiring a frequency spectrum characteristic corresponding to user voice request data as an acoustic characteristic;
carrying out voice recognition on user voice request data to obtain a corresponding recognition text, and obtaining semantic features according to the recognition text;
inputting the acoustic features of the user into a pre-constructed user attribute feature determination model to obtain user attribute features; the user attribute features include at least one of: gender, age, and occupation;
and respectively extracting the user speaker feature vector TV-vector and the user speaker feature vector CNN-vector based on an all-variable factor analysis method and a pre-constructed convolutional neural network model according to the acoustic features.
7. The method of claim 6, further comprising:
and adding the user category of the target user, the interactive recommendation characteristics extracted from the sample voice request data of the target user, the corresponding recommendation content of the target user and the corresponding relation among the three into the recommendation knowledge base.
8. A content recommendation apparatus characterized by comprising:
the system comprises an acquisition module, a recommendation module and a recommendation module, wherein the acquisition module is used for acquiring voice request data of a target user and extracting corresponding interactive recommendation characteristics in the voice request data, and the target user comprises a non-registered user;
the recommendation module is used for acquiring corresponding recommendation contents from a pre-constructed recommendation knowledge base corresponding to the target user according to the interactive recommendation characteristics and sending the recommendation contents to the target user; the recommendation knowledge base comprises a plurality of corresponding relation models; the corresponding relation models are respectively established on the basis of interactive recommendation features extracted from sample voice request data of historical users corresponding to the user categories and recommendation contents corresponding to the historical users corresponding to the user categories;
the interactive recommendation characteristics comprise the characteristics of whether the recommended content is used or not, and the recommendation module is further used for dynamically adjusting the user category of the target user in the recommendation knowledge base according to whether the target user uses the recommended content or not.
9. An electronic device, comprising:
at least one processor; and
at least one memory communicatively coupled to the processor, wherein:
the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1 to 7.
10. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711195319.4A CN107886949B (en) | 2017-11-24 | 2017-11-24 | Content recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711195319.4A CN107886949B (en) | 2017-11-24 | 2017-11-24 | Content recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107886949A CN107886949A (en) | 2018-04-06 |
| CN107886949B true CN107886949B (en) | 2021-04-30 |
Family
ID=61775110
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711195319.4A Active CN107886949B (en) | 2017-11-24 | 2017-11-24 | Content recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107886949B (en) |
Families Citing this family (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846429B (en) * | 2018-05-31 | 2023-04-07 | 清华大学 | Unsupervised learning-based network space resource automatic classification method and unsupervised learning-based network space resource automatic classification device |
| CN108833949A (en) * | 2018-06-25 | 2018-11-16 | 深圳市华慧品牌管理有限公司 | The video recommendation method and device of robot are accompanied for children |
| CN108962223A (en) * | 2018-06-25 | 2018-12-07 | 厦门快商通信息技术有限公司 | A kind of voice gender identification method, equipment and medium based on deep learning |
| CN108876548A (en) * | 2018-06-25 | 2018-11-23 | 北京金山云网络技术有限公司 | A kind of cloud Products Show method and device |
| CN109033281B (en) * | 2018-07-11 | 2019-12-13 | 国网技术学院 | Intelligent pushing system of knowledge resource library |
| CN109325096B (en) * | 2018-07-11 | 2019-12-06 | 国网技术学院 | Knowledge resource search system based on knowledge resource classification |
| CN108960934A (en) * | 2018-07-19 | 2018-12-07 | 苏州思必驰信息科技有限公司 | Information recommendation method and system during voice dialogue |
| CN109119069B (en) * | 2018-07-23 | 2020-08-14 | 深圳大学 | Specific crowd identification method, electronic device and computer readable storage medium |
| CN109165283B (en) * | 2018-08-20 | 2021-12-28 | 北京如布科技有限公司 | Resource recommendation method, device, equipment and storage medium |
| CN109408809A (en) * | 2018-09-25 | 2019-03-01 | 天津大学 | A kind of sentiment analysis method for automobile product comment based on term vector |
| CN109635209B (en) * | 2018-12-12 | 2021-03-12 | 广东小天才科技有限公司 | Learning content recommendation method and family education equipment |
| CN109686367B (en) * | 2018-12-17 | 2021-02-02 | 科大讯飞股份有限公司 | Earphone noise reduction method, device and equipment and readable storage medium |
| CN109545232A (en) * | 2019-01-21 | 2019-03-29 | 美的集团武汉制冷设备有限公司 | Information-pushing method, information push-delivery apparatus and interactive voice equipment |
| CN110223134B (en) * | 2019-04-28 | 2022-10-28 | 平安科技（深圳）有限公司 | Product recommendation method based on voice recognition and related equipment |
| CN110164415B (en) * | 2019-04-29 | 2024-06-14 | 腾讯科技（深圳）有限公司 | Recommendation method, device and medium based on voice recognition |
| CN110570837B (en) * | 2019-08-28 | 2022-03-11 | 卓尔智联(武汉)研究院有限公司 | Voice interaction method and device and storage medium |
| CN110992948B (en) * | 2019-11-18 | 2023-07-25 | 博泰车联网科技(上海)股份有限公司 | Restaurant reservation method and terminal based on multi-round voice interaction |
| CN111079001A (en) * | 2019-11-26 | 2020-04-28 | 贝壳技术有限公司 | Decoration recommendation information generation method and device, storage medium and electronic equipment |
| CN111261196A (en) * | 2020-01-17 | 2020-06-09 | 厦门快商通科技股份有限公司 | Age estimation method, device and equipment |
| EP4138358A4 (en) * | 2020-05-27 | 2023-09-20 | Baidu Online Network Technology (Beijing) Co., Ltd | Voice packet recommendation method, apparatus and device, and storage medium |
| CN111554300B (en) * | 2020-06-30 | 2021-04-13 | 腾讯科技（深圳）有限公司 | Audio data processing method, device, storage medium and equipment |
| CN111932296B (en) * | 2020-07-20 | 2024-05-28 | 中国建设银行股份有限公司 | Product recommendation method and device, server and storage medium |
| CN114125043A (en) * | 2020-09-01 | 2022-03-01 | 上海智臻智能网络科技股份有限公司 | Information pushing method and information pushing device |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105872792A (en) * | 2016-03-25 | 2016-08-17 | 乐视控股（北京）有限公司 | Voice-based service recommending method and device |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7949529B2 (en) * | 2005-08-29 | 2011-05-24 | Voicebox Technologies, Inc. | Mobile systems and methods of supporting natural language human-machine interactions |
| US9493130B2 (en) * | 2011-04-22 | 2016-11-15 | Angel A. Penilla | Methods and systems for communicating content to connected vehicle users based detected tone/mood in voice input |
| KR20140092028A (en) * | 2013-01-15 | 2014-07-23 | 박종열 | Song recommendation system and terminal and song recommendation method using the same |
| CN104102819B (en) * | 2014-06-27 | 2017-12-19 | 北京奇艺世纪科技有限公司 | A kind of determination method and apparatus of user's natural quality |
| CN104361507A (en) * | 2014-11-20 | 2015-02-18 | 携程计算机技术（上海）有限公司 | Commodity recommending method and system |
| CN104991899B (en) * | 2015-06-02 | 2018-06-19 | 广州酷狗计算机科技有限公司 | The recognition methods of user property and device |
| US20170250930A1 (en) * | 2016-02-29 | 2017-08-31 | Outbrain Inc. | Interactive content recommendation personalization assistant |
| CN106128467A (en) * | 2016-06-06 | 2016-11-16 | 北京云知声信息技术有限公司 | Method of speech processing and device |
| CN106601259B (en) * | 2016-12-13 | 2021-04-06 | 北京奇虎科技有限公司 | Information recommendation method and device based on voiceprint search |
-
2017
- 2017-11-24 CN CN201711195319.4A patent/CN107886949B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105872792A (en) * | 2016-03-25 | 2016-08-17 | 乐视控股（北京）有限公司 | Voice-based service recommending method and device |
Non-Patent Citations (1)
| Title |
|---|
| 基于DNN算法的移动视频推荐策略;陈亮;《计算机学报》;20160229;第39卷(第8期);第1627-1637页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107886949A (en) | 2018-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107886949B (en) | Content recommendation method and device | |
| CN107507612B (en) | Voiceprint recognition method and device | |
| CN107978311B (en) | Voice data processing method and device and voice interaction equipment | |
| US11676575B2 (en) | On-device learning in a hybrid speech processing system | |
| US10623573B2 (en) | Personalized support routing based on paralinguistic information | |
| WO2018224034A1 (en) | Intelligent question answering method, server, terminal and storage medium | |
| CN107481720B (en) | Explicit voiceprint recognition method and device | |
| WO2022095380A1 (en) | Ai-based virtual interaction model generation method and apparatus, computer device and storage medium | |
| US10210867B1 (en) | Adjusting user experience based on paralinguistic information | |
| US10410628B2 (en) | Adjusting a ranking of information content of a software application based on feedback from a user | |
| CN110990685B (en) | Voiceprint-based voice searching method, voiceprint-based voice searching equipment, storage medium and storage device | |
| CN111625632A (en) | Question-answer pair recommendation method, device, equipment and storage medium | |
| CN111694940B (en) | User report generation method and terminal equipment | |
| RU2720359C1 (en) | Method and equipment for recognizing emotions in speech | |
| KR20180137932A (en) | Method and system for expansion to daily life language by using word vectorization technique based on social network content | |
| US11410643B2 (en) | Response generation for conversational computing interface | |
| US20170018268A1 (en) | Systems and methods for updating a language model based on user input | |
| CN113314119B (en) | Voice recognition intelligent household control method and device | |
| CN109582825B (en) | Method and apparatus for generating information | |
| KR101751898B1 (en) | Method for Translation Correction and Server for Providing Translation Correction Service | |
| CN112115248B (en) | Method and system for extracting dialogue strategy structure from dialogue corpus | |
| CN112470216B (en) | Voice application platform | |
| CN110099332B (en) | Audio environment display method and device | |
| CN111427444A (en) | Control method and device of intelligent device | |
| CN111556096B (en) | Information pushing method, device, medium and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220415 Address after: 230088 8th-10th floor, iFLYTEK building, 666 Wangjiang West Road, hi tech Zone, Hefei City, Anhui Province Patentee after: IFLYTEK INTELLIGENT SYSTEM Co.,Ltd. Address before: 230088 666 Wangjiang West Road, Hefei hi tech Development Zone, Anhui Patentee before: IFLYTEK Co.,Ltd. |