CN111767426B - Song recommendation method and device - Google Patents
Song recommendation method and device Download PDFInfo
- Publication number
- CN111767426B CN111767426B CN202010572038.1A CN202010572038A CN111767426B CN 111767426 B CN111767426 B CN 111767426B CN 202010572038 A CN202010572038 A CN 202010572038A CN 111767426 B CN111767426 B CN 111767426B
- Authority
- CN
- China
- Prior art keywords
- song
- user
- listening
- songs
- feature vector
- 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 49
- 239000013598 vector Substances 0.000 claims abstract description 314
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 84
- 230000006399 behavior Effects 0.000 claims description 84
- 238000000605 extraction Methods 0.000 claims description 29
- 238000004891 communication Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 13
- 238000012937 correction Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 7
- 238000010586 diagram Methods 0.000 description 9
- 238000012545 processing Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000002093 peripheral effect Effects 0.000 description 2
- 238000012896 Statistical algorithm Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/60—Information retrieval; Database structures therefor; File system structures therefor of audio data
- G06F16/63—Querying
- G06F16/635—Filtering based on additional data, e.g. user or group profiles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a song recommendation method and device. In the method, after receiving user attribute information of a user to be recommended sent by a terminal, a server searches a stored feature vector relation table, obtains a user feature vector corresponding to the user attribute information and a song feature vector of at least one song to be recommended except for a history listening song in a song listening log of the user to be recommended, calculates the user feature vector and the song feature vector of the at least one song to be recommended by adopting a preset recommendation algorithm, obtains association probability of the user to be recommended and each song to be recommended, and sends a preset number of songs to be recommended meeting preset association probability conditions to the terminal. The method can recommend the type of songs which are not listened to for the user, and user experience is improved.
Description
Technical Field
The application relates to the field of data processing, in particular to a song recommendation method and device.
Background
Intelligent devices such as intelligent sound boxes become part of life of people. At present, how to make the feedback of the intelligent sound box more personalized is one direction continuously pursued by the prior art.
The intelligent sound box can be a product of sound box upgrading, is a tool for home consumers to surf the internet by voice, such as song requesting, shopping on the internet, or weather forecast understanding, and can also control intelligent household equipment, such as opening a curtain, setting the temperature of a refrigerator, heating a water heater in advance, and the like.
The song recommendation system of the existing intelligent sound box generally only considers the accuracy of song recommendation and the diversity of different types of songs.
However, the existing song recommendation system may recommend the user with a type of song that the user frequently listens to, resulting in that the user listens to only one type of song, so the song recommendation of the existing song recommendation system has limitation in the type of song, and the user experience is reduced.
Disclosure of Invention
The embodiment of the application provides a song recommending method and device, solves the problems in the prior art, can recommend the type of songs which are not listened to for a user, and improves user experience.
In a first aspect, a song recommendation method is provided, which may include:
receiving user attribute information of a user to be recommended, which is sent by a terminal;
Searching a stored feature vector relation table, and acquiring a user feature vector corresponding to the user attribute information and a song feature vector of at least one song to be recommended except for a history listening song in a song listening log of the user to be recommended; the characteristic vector relation table comprises corresponding relations between user characteristic vectors corresponding to each piece of user attribute information and song characteristic vectors of different songs, association probabilities of corresponding relations and song attribute information corresponding to each piece of song characteristic vectors; the user characteristic vector is used for describing the listening behavior characteristics of a user listening to different songs; the song feature vector is used for describing the listened features of songs listened to by different users;
Calculating the user characteristic vector and the song characteristic vector of the at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended;
And sending the songs to be recommended which meet the preset association probability conditions to the terminal.
In an optional implementation, the obtaining process of the feature vector relation table specifically includes;
obtaining song listening logs of at least two users, wherein the song listening logs comprise user attribute information of each user, historical listening songs and listening behaviors of each user for listening to different songs;
extracting characteristics of listening behaviors of each user listening to different songs in the song listening log by adopting a characteristic extraction algorithm to obtain song characteristic vectors of each song listened to by different users;
acquiring at least one user listening to each song according to the songs listened to by each user;
Extracting features of listening behaviors of at least one user listening to each song in the song listening log by adopting the feature extraction algorithm to obtain user feature vectors of listening to different songs by each user;
And generating a feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vector, the song feature vector of the song listened to by each user and the corresponding song attribute information.
In an optional implementation, a feature extraction algorithm is adopted to perform feature extraction on a listening behavior of at least one user listening to each song in the song listening log, so as to obtain a user feature vector of each user listening to a different song, including:
Clustering listening behaviors of different users listening to each song in the song listening log by adopting a preset clustering algorithm to obtain a correlation degree value among the users;
acquiring initial user feature vectors of different songs listened to by each user;
Calculating the initial user feature vectors of each two users by adopting a preset similarity algorithm to obtain a calculated association degree value of each two users and an error of the calculated association degree value relative to the association degree value;
And carrying out iterative correction on the initial user feature vectors of each two target users according to the error, and obtaining the user feature vector of each user in each two users when the error meets a preset error threshold.
In an optional implementation, a feature extraction algorithm is adopted to perform feature extraction on listening behaviors of each user listening to different songs in the song listening log, so as to obtain song feature vectors of each song listened to by different users, including:
Clustering different songs listened to by each two users in the song listening log by adopting a preset clustering algorithm to obtain a correlation degree value between the songs;
Acquiring an initial song feature vector of each song to be listened to by different users;
Calculating the initial song feature vectors of every two songs by adopting a preset similarity algorithm to obtain a calculated association degree value of every two songs and an error relative to the association degree value;
and carrying out iterative correction on the initial song feature vectors of every two songs according to the error, and obtaining the song feature vector of each song in every two songs when the error meets a preset error threshold. In an optional implementation, generating the feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vector, the song feature vector of the song listened to by each user, and the corresponding song attribute information includes:
Obtaining a target song feature vector of each song according to the song feature vector of each song and the corresponding song attribute information, and obtaining a target user feature vector of each user according to the user attribute information and the corresponding user feature vector of each user;
Calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song;
A feature vector relationship table is generated.
In an optional implementation, before the calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song, the method further includes:
Determining songs with preset listening behaviors as first-class songs and determining songs without the preset listening behaviors as second-class songs;
And calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song, wherein the method comprises the following steps:
If the song corresponding to the target song feature vector is the first type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the intermediate association probability of each user and each song in the first type song;
after the intermediate association probability and the preset probability are subjected to difference value operation, the association probability of each user and each song in the first type of songs is obtained;
And if the song corresponding to the target song feature vector is the second type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song in the second type song.
In an optional implementation, the preset listening behavior is a listening behavior of which a user listening duration is not greater than a preset time period.
In a second aspect, a song recommendation apparatus is provided, the apparatus may include: the device comprises a receiving unit, an acquisition unit, an operation unit and a sending unit;
the receiving unit is used for receiving user attribute information of the user to be recommended, which is sent by the terminal;
The obtaining unit is used for searching a stored characteristic vector relation table, and obtaining a user characteristic vector corresponding to the user attribute information and a song characteristic vector of at least one song to be recommended except for a historical listening song in a song listening log of the user to be recommended; the characteristic vector relation table comprises corresponding relations between user characteristic vectors corresponding to each piece of user attribute information and song characteristic vectors of different songs, association probabilities of corresponding relations and song attribute information corresponding to each piece of song characteristic vectors; the user characteristic vector is used for describing the listening behavior characteristics of a user listening to different songs; the song feature vector is used for describing the listened features of songs listened to by different users;
The computing unit is used for computing the user characteristic vector and the song characteristic vector of the at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended;
and the sending unit is used for sending the songs to be recommended which meet the preset association probability conditions to the terminal.
In an alternative implementation, the apparatus further includes: an extraction unit and a generation unit;
The acquisition unit is further used for acquiring song listening logs of at least two users, wherein the song listening logs comprise user attribute information of each user, historical listening songs and listening behaviors of each user for listening to different songs;
the extracting unit is used for extracting the characteristics of the listening behaviors of each user listening to different songs in the song listening log by adopting a characteristic extracting algorithm to obtain song characteristic vectors of each song listened to by different users;
The acquisition unit is further used for acquiring at least one user who listens to each song according to the songs listened to by each user;
The extracting unit is further used for extracting features of listening behaviors of at least one user listening to each song in the song listening log by adopting the feature extracting algorithm, so as to obtain user feature vectors of each user listening to different songs;
the generating unit is configured to generate a feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vectors, the song feature vectors of songs listened to by each user, and the corresponding song attribute information.
In an optional implementation, the extracting unit is specifically configured to use a preset clustering algorithm to cluster listening behaviors of different users listening to each song in the song listening log, so as to obtain a correlation degree value between the users;
acquiring initial user feature vectors of different songs listened to by each user;
Calculating the initial user feature vectors of each two users by adopting a preset similarity algorithm to obtain a calculated association degree value of each two users and an error of the calculated association degree value relative to the association degree value;
And carrying out iterative correction on the initial user feature vectors of each two target users according to the error, and obtaining the user feature vector of each user in each two users when the error meets a preset error threshold.
In an optional implementation, the extracting unit is further specifically configured to use a preset clustering algorithm to cluster different songs listened to by each two users in the song listening log, so as to obtain a correlation degree value between songs;
Acquiring an initial song feature vector of each song to be listened to by different users;
Calculating the initial song feature vectors of every two songs by adopting a preset similarity algorithm to obtain a calculated association degree value of every two songs and an error relative to the association degree value;
And carrying out iterative correction on the initial song feature vectors of every two songs according to the error, and obtaining the song feature vector of each song in every two songs when the error meets a preset error threshold.
In an optional implementation, the generating unit is specifically configured to obtain, according to the song feature vector of each song and the corresponding song attribute information, a target song feature vector of each song, and obtain, according to the user attribute information and the corresponding user feature vector of each user, a target user feature vector of each user;
Calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song;
A feature vector relationship table is generated.
In an alternative implementation, the apparatus further comprises a determining unit;
The determining unit is used for determining songs with preset listening behaviors as first-class songs and determining songs without the preset listening behaviors as second-class songs;
The obtaining unit is further configured to, if the song corresponding to the target song feature vector is the first type song, calculate the target song feature vector and the target user feature vector by using the preset similarity algorithm, so as to obtain an intermediate association probability between each user and each song in the first type song;
after the intermediate association probability and the preset probability are subjected to difference value operation, the association probability of each user and each song in the first type of songs is obtained;
And if the song corresponding to the target song feature vector is the second type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song in the second type song.
In an optional implementation, the preset listening behavior is a listening behavior of which a user listening duration is not greater than a preset time period.
In a third aspect, an electronic device is provided, the electronic device comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory are in communication with each other via the communication bus;
A memory for storing a computer program;
a processor for implementing the method steps of any one of the above first aspects when executing a program stored on a memory.
In a fourth aspect, a computer-readable storage medium is provided, in which a computer program is stored which, when being executed by a processor, carries out the method steps of any of the first aspects.
In the song recommendation method of the embodiment of the invention, after receiving the user attribute information of the user to be recommended sent by the terminal, the server searches the stored feature vector relation table to obtain the user feature vector corresponding to the user attribute information and the song feature vector of at least one song to be recommended except the historical listening song in the song listening log of the user to be recommended; the feature vector relation table comprises the corresponding relation between the user feature vector corresponding to each piece of user attribute information and the song feature vectors of different songs, the association probability of the corresponding relation and the song attribute information corresponding to each piece of song feature vector; the user feature vector is used for describing the listening behavior features of the user listening to different songs; the song feature vector is used for describing the listened features of the song listened to by different users; calculating a user characteristic vector and a song characteristic vector of at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended; and sending the songs to be recommended which meet the preset association probability conditions to the terminal. The method can recommend the type of songs which are not listened to for the user, and user experience is improved.
Drawings
Fig. 1 is a schematic diagram of a song recommendation method and apparatus according to an embodiment of the present invention;
Fig. 2 is a schematic flow chart of a song recommendation method according to an embodiment of the present invention;
Fig. 3 is a schematic structural diagram of a song recommendation apparatus according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments, but not all embodiments of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The song recommendation method provided by the embodiment of the invention can be applied to a server, a terminal or a system formed by the server and the terminal (such as an intelligent sound box) as shown in fig. 1.
For example, the terminal in fig. 1 may be used to receive a recommended play voice command of a user, play a song to the user, store a relationship between voice features and user attribute information, that is, a relationship between voice features of different users and user attribute information of corresponding users, determine user attribute information according to current voice features of the users, and so on. The user attribute information may include user information such as user identification, age, occupation, and the like.
The server may be configured to store a song listening log for each user, where the song listening log may include user attribute information, historical listening songs, listening behaviors of listening to different songs, song listening time, and the like, and the listening behaviors may include a complete listening song, a collection behavior, a praise behavior, a loop play behavior, a cut song behavior, and the like.
And the method can be used for acquiring songs to be recommended according to the user attribute information determined by the terminal and sending the songs to be recommended to the terminal.
To ensure accuracy of the recommendation, the server may be an application server or a cloud server with a strong computing power; the Terminal may be a Mobile phone, a smart speaker, a notebook computer, a digital broadcast receiver, a Personal Digital Assistant (PDA), a tablet personal computer (PAD), a User Equipment (UE) with a strong computing power, a handheld device, a vehicle-mounted device, a wearable device, a computing device, or other processing device connected to a wireless modem, a Mobile Station (MS), a Mobile Terminal (Mobile Terminal), or the like. The terminal is capable of communicating with one or more core networks via a radio access network (Radio Access Network, RAN).
The song recommendation method provided by the application can recommend the song which is not in contact with the type of the listened song to for the current user from the surprise degree, wherein the song is the historic listened song of the similar user corresponding to the current user and is different from the historic listened song type of the current user. A similar user is one that has the same or similar song listening characteristics as the current user, such as a like to listen to the same song type, etc.
Because the current user and the similar user have similar song listening characteristics, different song types except the same song type as the current user in the song types listened by the similar user historically can be favored by the current user, and the songs of the different song types can have higher surprise degree relative to the current user, so that the songs of the different song types between the similar user and the current user can be recommended to the current user.
Wherein, surprise degree may refer to:
(1) Recommending songs which are not listened to in a listening list of the user history to the user, and not listening to songs of similar types;
(2) The user can complete listening to the recommended songs, or collect the songs, or perform secondary on-demand and the like;
Alternatively, surprise may also refer to the user ordering or listening to that type of song or beginning.
The preferred embodiments of the present application will be described below with reference to the accompanying drawings of the specification, it being understood that the preferred embodiments described herein are for illustration and explanation only, and not for limitation of the present application, and embodiments of the present application and features of the embodiments may be combined with each other without conflict.
Fig. 2 is a flowchart of a song recommendation method according to an embodiment of the present invention. As shown in fig. 2, the method may include:
Step 210, the terminal determines user attribute information of the user to be recommended according to the voice characteristics of the recommended play voice command of the user to be recommended.
Taking a terminal as an intelligent sound box as an example, the intelligent sound box receives a recommended playing voice instruction of a user to be recommended, and adopts a preset feature extraction algorithm to extract the voice feature of the recommended playing voice instruction.
The intelligent sound box searches the relation between the stored voice characteristics and the user attribute information, and obtains the user attribute information corresponding to the voice characteristics of the user to be recommended.
And 220, the terminal sends the user attribute information of the user to be recommended to the server.
Step 230, the server searches the stored feature vector relation table, and obtains the user feature vector corresponding to the user attribute information and the song feature vector of at least one song to be recommended except for the history listening song in the song listening log of the user to be recommended.
The feature vector relation table may include a correspondence between a user feature vector corresponding to each user attribute information and song feature vectors of different songs, association probabilities of the corresponding correspondence, and song attribute information corresponding to each song feature vector; the user feature vector may be used to describe the listening behavior features of a user listening to different songs; the song feature vector may be used to describe the listened to features of the song that are listened to by different users, the listened to features characterizing the listened to information of the song.
The listening behavior feature may include a complete listening behavior feature, a collection behavior feature, a praise behavior feature or a circulation playing behavior feature, a song cutting behavior feature, and the like.
The listened to features may include a listened to duration, a listened to number of times, a number of consecutive listened to times, a listened to time, etc.
Optionally, the obtaining process of the feature vector relation table specifically may include;
Firstly, song listening logs of at least two users are obtained, wherein the song listening logs can comprise user attribute information of each user, historical listening songs and listening behaviors of each user for listening to different songs; the listening behavior may include a behavior of completely listening to songs, a collection behavior, a praise behavior or a loop play behavior, a cut song behavior, etc.
Then, adopting a feature extraction algorithm to extract features of listening behaviors of each user listening to different songs in the song listening log, and obtaining song feature vectors of each song listened to by different users;
Specifically, a preset clustering algorithm, such as a word2vec algorithm, is adopted to cluster songs listened to by each two users in a song listening log, so as to obtain a correlation degree value between songs, namely, the correlation probability of continuous playing of each two songs;
randomly generating initial song feature vectors of each song listened to by at least two users, thereby acquiring initial song feature vectors of each song listened to by different users;
Calculating the initial song feature vectors of every two songs by adopting a preset similarity algorithm, such as a cosine similarity algorithm, an Euclidean distance algorithm and the like, so as to obtain a calculated association degree value of every two corresponding songs and an error of the calculated association degree value relative to the association degree value;
And carrying out iterative correction on the initial song feature vectors of every two songs according to the errors, and obtaining the song feature vector of each song in every two songs when the errors meet a preset error threshold.
For example, the steps may specifically include:
1, firstly, respectively randomly generating a random initial song characteristic vector from a song A and a song B;
2, adopting a statistical algorithm, forming a directed graph according to the random song feature vector, and simultaneously obtaining the association degree value of the song A and the song B;
3, when the first iteration starts, calculating cosine similarity of songs A and B by using the randomly generated initial song feature vector, taking the cosine similarity as a calculated association degree value of the two songs, and comparing the calculated association degree value to obtain an error between the two songs;
4, calculating the gradient of the error by utilizing an algorithm of counter propagation and gradient descent, and setting a fixed learning rate, so that the randomly generated initial song feature vector can be corrected to obtain the song feature vector;
and 5, the next iteration uses the song feature vector corrected by the previous iteration until the total error is not more than a preset error threshold value, and then the iteration is ended to obtain the current song feature vector corresponding to the songs A and B.
Further, the server may obtain at least one user who listens to each song according to the song that each user listens to; adopting a feature extraction algorithm to perform feature extraction on the listening behavior of at least one user listening to each song in the song listening log, and obtaining user feature vectors of different songs listened to by each user;
specifically, the server may use a preset clustering algorithm to cluster listening behaviors of different users listening to each song in the song listening log, so as to obtain a correlation degree value between the users;
acquiring initial user feature vectors of different songs listened to by each user;
Calculating the initial user feature vectors of each two users by adopting a preset similarity algorithm to obtain a calculated association degree value of each two corresponding users and an error relative to the association degree value;
and carrying out iterative correction on the initial user feature vectors of every two users according to the errors, and obtaining the user feature vector of each user in every two users when the errors meet a preset error threshold.
The server may then generate a feature vector relationship table based on the user attribute information of at least two users, the corresponding user feature vectors, the song feature vectors of songs listened to by each user, and the corresponding song attribute information.
The server can acquire the target song feature vector of each song according to the song feature vector of each song and the corresponding song attribute information, and acquire the target user feature vector of each user according to the user attribute information and the corresponding user feature vector of each user;
And calculating the target song feature vector and the target user feature vector by adopting a preset similarity algorithm to obtain the association probability of each user corresponding to each song, thereby generating a feature vector relation table.
Optionally, before obtaining the associated probability that each user corresponds to each song, the server may determine a song having a preset listening behavior as a first type of song, and determine a song not having a preset listening behavior as a second type of song; the preset listening behavior may be a listening behavior in which a user listening duration is not longer than a preset time period.
For example, based on a song listening log of the user, songs that the user has completely listened to, collected, praised or circularly played are taken as a second sample, and songs that the user has listened to within 30s are taken as a first sample.
At this time, if the song corresponding to the target song feature vector is a first type song, the server may use a preset similarity algorithm to calculate the target song feature vector and the target user feature vector to obtain an intermediate association probability between each user and each song in the first type song, and calculate the intermediate association probability and the preset probability to obtain an association probability between each user and each song in the first type song;
If the song corresponding to the target song feature vector is a second type song, the server may use a preset similarity algorithm to calculate the target song feature vector and the target user feature vector, so as to obtain the association probability of each user and each song in the second type song.
Step 240, the server calculates the user feature vector and the song feature vector of at least one song to be recommended by adopting a preset recommendation algorithm, so as to obtain the association probability of the user to be recommended and each song to be recommended.
The server may calculate the user feature vector and the song feature vector of each song to be recommended in at least one to obtain at least one association probability corresponding to the user to be recommended by adopting a cosine similarity algorithm.
Step 250, the server sends songs to be recommended which meet the preset association probability conditions to the terminal.
Optionally, the server may sort the obtained at least one association probability, and send the songs to be recommended that meet the preset association probability condition to the terminal, such as the intelligent sound box, so that the intelligent sound box plays the songs to be recommended.
The meeting of the preset association probability condition may be that the association probability is greater than a preset probability threshold, that is, the server sends the song to be recommended corresponding to the preset probability greater than the preset probability threshold to the terminal.
Or the preset association probability condition is satisfied, the preset number of preceding association probabilities in the association probabilities after the sorting from big to small are performed.
After receiving user attribute information of a user to be recommended sent by a terminal, a server in the song recommendation method of the embodiment of the invention determines the user information according to the voice characteristics of a voice command to be recommended by the user to be recommended and the relationship between the stored voice characteristics and the user attribute information; searching a stored feature vector relation table, and acquiring a user feature vector corresponding to the user attribute information and a song feature vector of at least one song to be recommended except for a history listening song in a song listening log of the user to be recommended; the feature vector relation table comprises the corresponding relation between the user feature vector corresponding to each piece of user attribute information and the song feature vectors of different songs, the association probability of the corresponding relation and the song attribute information corresponding to each piece of song feature vector; the user feature vector is used for describing the listening behavior features of the user listening to different songs; the song feature vector is used for describing the listened features of the song listened to by different users; calculating a user characteristic vector and a song characteristic vector of at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended; and sending the songs to be recommended which meet the preset association probability conditions to the terminal. The method can recommend the type of songs which are not listened to for the user, and user experience is improved.
Corresponding to the above method, the embodiment of the present invention further provides a song recommendation apparatus, as shown in fig. 3, where the song recommendation apparatus includes: a receiving unit 310, an acquiring unit 320, an operation unit 330, and a transmitting unit 340;
A receiving unit 310, configured to receive user attribute information of a user to be recommended sent by a terminal;
An obtaining unit 320, configured to search a stored feature vector relation table, and obtain a user feature vector corresponding to the user attribute information and a song feature vector of at least one song to be recommended except for a historical listening song in a song listening log of the user to be recommended; the characteristic vector relation table comprises corresponding relations between user characteristic vectors corresponding to each piece of user attribute information and song characteristic vectors of different songs, association probabilities of corresponding relations and song attribute information corresponding to each piece of song characteristic vectors; the user characteristic vector is used for describing the listening behavior characteristics of a user listening to different songs; the song feature vector is used for describing the listened features of songs listened to by different users;
The computing unit 330 is configured to perform a computing on the user feature vector and the song feature vector of the at least one song to be recommended by using a preset recommendation algorithm, so as to obtain a probability of association between the user to be recommended and each song to be recommended;
and the sending unit 340 is configured to send the song to be recommended that satisfies the preset association probability condition to the terminal.
In an alternative implementation, the apparatus further includes: an extraction unit 350 and a generation unit 360;
The obtaining unit 320 is further configured to obtain song listening logs of at least two users, where the song listening logs include user attribute information of each user, historical listening songs, and listening behaviors of each user listening to different songs;
The extracting unit 350 is configured to perform feature extraction on listening behaviors of each user listening to different songs in the song listening log by using a feature extraction algorithm, so as to obtain a song feature vector of each song listened to by different users;
The obtaining unit 320 is further configured to obtain, according to the songs listened to by each user, at least one user listening to each song;
The extracting unit 350 is further configured to perform feature extraction on a listening behavior of at least one user listening to each song in the song listening log by using the feature extraction algorithm, so as to obtain a user feature vector of each user listening to a different song;
The generating unit 360 is configured to generate a feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vectors, the song feature vectors of the songs listened to by each user, and the corresponding song attribute information.
In an optional implementation, the extracting unit 350 is specifically configured to use a preset clustering algorithm to cluster listening behaviors of different users listening to each song in the song listening log, so as to obtain a correlation degree value between the users;
acquiring initial user feature vectors of different songs listened to by each user;
Calculating the initial user feature vectors of each two users by adopting a preset similarity algorithm to obtain a calculated association degree value of each two users and an error of the calculated association degree value relative to the association degree value;
And carrying out iterative correction on the initial user feature vectors of each two target users according to the error, and obtaining the user feature vector of each user in each two users when the error meets a preset error threshold.
In an optional implementation, the extracting unit 350 is further specifically configured to use a preset clustering algorithm to cluster different songs listened to by each two users in the song listening log, so as to obtain a correlation degree value between songs;
Acquiring an initial song feature vector of each song to be listened to by different users;
Calculating the initial song feature vectors of every two songs by adopting a preset similarity algorithm to obtain a calculated association degree value of every two songs and an error relative to the association degree value;
And carrying out iterative correction on the initial song feature vectors of every two songs according to the error, and obtaining the song feature vector of each song in every two songs when the error meets a preset error threshold.
In an optional implementation, the generating unit 360 is specifically configured to obtain the target song feature vector of each song according to the song feature vector of each song and the corresponding song attribute information, and obtain the target user feature vector of each user according to the user attribute information and the corresponding user feature vector of each user;
Calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song;
A feature vector relationship table is generated.
In an alternative implementation, the apparatus further comprises a determining unit 370;
A determining unit 370 for determining songs having a preset listening behavior as a first type of songs and determining songs not having the preset listening behavior as a second type of songs;
The obtaining unit 320 is further configured to, if the song corresponding to the target song feature vector is the first type song, calculate the target song feature vector and the target user feature vector by using the preset similarity algorithm, so as to obtain an intermediate association probability between each user and each song in the first type song;
after the intermediate association probability and the preset probability are subjected to difference value operation, the association probability of each user and each song in the first type of songs is obtained;
And if the song corresponding to the target song feature vector is the second type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song in the second type song.
In an optional implementation, the preset listening behavior is a listening behavior of which a user listening duration is not greater than a preset time period.
The functions of each functional unit of the song recommendation apparatus provided in the foregoing embodiments of the present invention may be implemented by the foregoing method steps, so that the specific working process and beneficial effects of each unit in the song recommendation apparatus provided in the embodiments of the present invention are not repeated herein.
The embodiment of the present invention further provides an electronic device, as shown in fig. 4, including a processor 410, a communication interface 420, a memory 430, and a communication bus 440, where the processor 410, the communication interface 420, and the memory 430 complete communication with each other through the communication bus 440.
A memory 430 for storing a computer program;
The processor 410 is configured to execute the program stored in the memory 430, and implement the following steps:
receiving user attribute information of a user to be recommended, which is sent by a terminal;
Searching a stored feature vector relation table, and acquiring a user feature vector corresponding to the user attribute information and a song feature vector of at least one song to be recommended except for a history listening song in a song listening log of the user to be recommended; the characteristic vector relation table comprises corresponding relations between user characteristic vectors corresponding to each piece of user attribute information and song characteristic vectors of different songs, association probabilities of corresponding relations and song attribute information corresponding to each piece of song characteristic vectors; the user characteristic vector is used for describing the listening behavior characteristics of a user listening to different songs; the song feature vector is used for describing the listened features of songs listened to by different users;
Calculating the user characteristic vector and the song characteristic vector of the at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended;
And sending the songs to be recommended which meet the preset association probability conditions to the terminal.
In an optional implementation, the obtaining process of the feature vector relation table specifically includes;
obtaining song listening logs of at least two users, wherein the song listening logs comprise user attribute information of each user, historical listening songs and listening behaviors of each user for listening to different songs;
extracting characteristics of listening behaviors of each user listening to different songs in the song listening log by adopting a characteristic extraction algorithm to obtain song characteristic vectors of each song listened to by different users;
acquiring at least one user listening to each song according to the songs listened to by each user;
Extracting features of listening behaviors of at least one user listening to each song in the song listening log by adopting the feature extraction algorithm to obtain user feature vectors of listening to different songs by each user;
And generating a feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vector, the song feature vector of the song listened to by each user and the corresponding song attribute information.
In an optional implementation, a feature extraction algorithm is adopted to perform feature extraction on a listening behavior of at least one user listening to each song in the song listening log, so as to obtain a user feature vector of each user listening to a different song, including:
Clustering listening behaviors of different users listening to each song in the song listening log by adopting a preset clustering algorithm to obtain a correlation degree value among the users;
acquiring initial user feature vectors of different songs listened to by each user;
Calculating the initial user feature vectors of each two users by adopting a preset similarity algorithm to obtain a calculated association degree value of each two users and an error of the calculated association degree value relative to the association degree value;
And carrying out iterative correction on the initial user feature vectors of each two target users according to the error, and obtaining the user feature vector of each user in each two users when the error meets a preset error threshold.
In an optional implementation, a feature extraction algorithm is adopted to perform feature extraction on listening behaviors of each user listening to different songs in the song listening log, so as to obtain song feature vectors of each song listened to by different users, including:
Clustering different songs of each two users in the song listening log by adopting a preset clustering algorithm to obtain a correlation degree value between the songs;
Acquiring an initial song feature vector of each song to be listened to by different users;
Calculating the initial song feature vectors of every two songs by adopting a preset similarity algorithm to obtain a calculated association degree value of every two songs and an error relative to the association degree value;
And carrying out iterative correction on the initial song feature vectors of every two songs according to the error, and obtaining the song feature vector of each song in every two songs when the error meets a preset error threshold.
In an optional implementation, generating the feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vector, the song feature vector of the song listened to by each user, and the corresponding song attribute information includes:
Obtaining a target song feature vector of each song according to the song feature vector of each song and the corresponding song attribute information, and obtaining a target user feature vector of each user according to the user attribute information and the corresponding user feature vector of each user;
Calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song;
A feature vector relationship table is generated.
In an optional implementation, before the calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song, the method further includes:
Determining songs with preset listening behaviors as first-class songs and determining songs without the preset listening behaviors as second-class songs;
And calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song, wherein the method comprises the following steps:
If the song corresponding to the target song feature vector is the first type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the intermediate association probability of each user and each song in the first type song;
after the intermediate association probability and the preset probability are subjected to difference value operation, the association probability of each user and each song in the first type of songs is obtained;
And if the song corresponding to the target song feature vector is the second type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song in the second type song.
In an optional implementation, the preset listening behavior is a listening behavior of which a user listening duration is not greater than a preset time period.
The communication bus mentioned above may be a peripheral component interconnect standard (PERIPHERAL COMPONENT INTERCONNECT, PCI) bus or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, or the like. The communication bus may be classified as an address bus, a data bus, a control bus, or the like. For ease of illustration, the figures are shown with only one bold line, but not with only one bus or one type of bus.
The communication interface is used for communication between the electronic device and other devices.
The Memory may include random access Memory (Random Access Memory, RAM) or may include Non-Volatile Memory (NVM), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the aforementioned processor.
The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; but may also be a digital signal processor (DIGITAL SIGNAL Processing, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components.
Since the implementation manner and the beneficial effects of the solution to the problem of each device of the electronic apparatus in the foregoing embodiment may be implemented by referring to each step in the embodiment shown in fig. 2, the specific working process and the beneficial effects of the electronic apparatus provided by the embodiment of the present invention are not repeated herein.
In yet another embodiment of the present invention, a computer readable storage medium is provided, in which instructions are stored, which when run on a computer, cause the computer to perform the song recommendation method of any of the above embodiments.
In yet another embodiment of the present invention, there is also provided a computer program product containing instructions that, when run on a computer, cause the computer to perform the song recommendation method of any of the above embodiments.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, embodiments of 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, embodiments of the present application may take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
Embodiments of the present application are described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations 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.
While preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiment and all such alterations and modifications as fall within the scope of the embodiments of the application.
It will be apparent to those skilled in the art that various modifications and variations can be made in the embodiments of the present application without departing from the spirit or scope of the embodiments of the application. Thus, if such modifications and variations of the embodiments of the present application fall within the scope of the claims and the equivalents thereof, it is intended that such modifications and variations be included in the embodiments of the present application.
Claims (9)
1. A song recommendation method, the method comprising:
receiving user attribute information of a user to be recommended, which is sent by a terminal;
searching a stored feature vector relation table, and acquiring a user feature vector corresponding to the user attribute information and a song feature vector of at least one song to be recommended except for a history listening song in a song listening log of the user to be recommended;
The characteristic vector relation table comprises corresponding relations between user characteristic vectors corresponding to each piece of user attribute information and song characteristic vectors of different songs, association probabilities of corresponding relations and song attribute information corresponding to each piece of song characteristic vectors; the user characteristic vector is used for describing the listening behavior characteristics of a user listening to different songs; the song feature vector is used for describing the listened features of songs listened to by different users; the process for acquiring the feature vector relation table specifically comprises the following steps: obtaining song listening logs of at least two users, wherein the song listening logs comprise user attribute information of each user, historical listening songs and listening behaviors of each user for listening to different songs; extracting characteristics of listening behaviors of each user listening to different songs in the song listening log by adopting a characteristic extraction algorithm to obtain song characteristic vectors of each song listened to by different users; acquiring at least one user listening to each song according to the songs listened to by each user; extracting features of listening behaviors of at least one user listening to each song in the song listening log by adopting the feature extraction algorithm to obtain user feature vectors of listening to different songs by each user; generating a feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vectors, the song feature vectors of songs listened to by each user and the corresponding song attribute information; the song feature vectors of each song listened to by different users are determined according to the association degree between different songs listened to by each two users in the song listening log and the initial song feature vector of each song listened to by different users; the user characteristic vector of each user listening to different songs is determined according to the association degree between different users listening to each song in the song listening log and the initial user characteristic vector of each user listening to different songs;
Calculating the user characteristic vector and the song characteristic vector of the at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended;
And sending the songs to be recommended which meet the preset association probability conditions to the terminal.
2. The method of claim 1, wherein the feature extraction algorithm is used to perform feature extraction on the listening behavior of at least one user listening to each song in the song listening log to obtain a user feature vector for each user listening to a different song, comprising:
Clustering listening behaviors of different users listening to each song in the song listening log by adopting a preset clustering algorithm to obtain a correlation degree value among the users;
acquiring initial user feature vectors of different songs listened to by each user;
Calculating the initial user feature vectors of each two users by adopting a preset similarity algorithm to obtain a calculated association degree value of each two users and an error of the calculated association degree value relative to the association degree value;
And carrying out iterative correction on the initial user feature vectors of each two target users according to the error, and obtaining the user feature vector of each user in each two users when the error meets a preset error threshold.
3. The method of claim 1, wherein the feature extraction algorithm is used to perform feature extraction on listening behaviors of each user listening to different songs in the song listening log to obtain song feature vectors of each song listened to by different users, and the method comprises:
Clustering different songs listened to by each two users in the song listening log by adopting a preset clustering algorithm to obtain a correlation degree value between the songs;
Acquiring an initial song feature vector of each song to be listened to by different users;
Calculating the initial song feature vectors of every two songs by adopting a preset similarity algorithm to obtain a calculated association degree value of every two songs and an error relative to the association degree value;
And carrying out iterative correction on the initial song feature vectors of every two songs according to the error, and obtaining the song feature vector of each song in every two songs when the error meets a preset error threshold.
4. The method of claim 1, wherein generating the feature vector relationship table based on the user attribute information of the at least two users, the corresponding user feature vectors, the song feature vectors of the songs listened to by each user, and the corresponding song attribute information, comprises:
Obtaining a target song feature vector of each song according to the song feature vector of each song and the corresponding song attribute information, and obtaining a target user feature vector of each user according to the user attribute information and the corresponding user feature vector of each user;
Calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user corresponding to each song;
A feature vector relationship table is generated.
5. The method of claim 4, wherein the computing the target song feature vector and the target user feature vector using the preset similarity algorithm further comprises, before obtaining the probability of association of each user with each song:
Determining songs with preset listening behaviors as first-class songs and determining songs without the preset listening behaviors as second-class songs;
And calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song, wherein the method comprises the following steps:
If the song corresponding to the target song feature vector is the first type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the intermediate association probability of each user and each song in the first type song;
after the intermediate association probability and the preset probability are subjected to difference value operation, the association probability of each user and each song in the first type of songs is obtained;
And if the song corresponding to the target song feature vector is the second type song, calculating the target song feature vector and the target user feature vector by adopting the preset similarity algorithm to obtain the association probability of each user and each song in the second type song.
6. The method of claim 5, wherein the preset listening behavior is a listening behavior in which a user listening period is not greater than a preset period of time.
7. A song recommendation apparatus, the apparatus comprising: the device comprises a receiving unit, an acquisition unit, an operation unit and a sending unit;
the receiving unit is used for receiving user attribute information of the user to be recommended, which is sent by the terminal;
The obtaining unit is used for searching a stored characteristic vector relation table, and obtaining a user characteristic vector corresponding to the user attribute information and a song characteristic vector of at least one song to be recommended except for a historical listening song in a song listening log of the user to be recommended; the characteristic vector relation table comprises corresponding relations between user characteristic vectors corresponding to each piece of user attribute information and song characteristic vectors of different songs, association probabilities of corresponding relations and song attribute information corresponding to each piece of song characteristic vectors; the user characteristic vector is used for describing the listening behavior characteristics of a user listening to different songs; the song feature vector is used for describing the listened features of songs listened to by different users;
The computing unit is used for computing the user characteristic vector and the song characteristic vector of the at least one song to be recommended by adopting a preset recommendation algorithm to obtain the association probability of the user to be recommended and each song to be recommended;
The sending unit is used for sending songs to be recommended which meet the preset association probability conditions to the terminal;
The apparatus further comprises: an extraction unit and a generation unit;
The acquisition unit is further used for acquiring song listening logs of at least two users, wherein the song listening logs comprise user attribute information of each user, historical listening songs and listening behaviors of each user for listening to different songs;
The extracting unit is used for extracting the characteristics of the listening behaviors of each user listening to different songs in the song listening log by adopting a characteristic extracting algorithm to obtain song characteristic vectors of each song listened to by different users; the characteristic vector of the songs listened to by different users is determined according to the association degree between different songs listened to by each two users in the song listening log and the characteristic vector of the initial songs listened to by different users;
The acquisition unit is further used for acquiring at least one user who listens to each song according to the songs listened to by each user;
The extracting unit is further used for extracting features of listening behaviors of at least one user listening to each song in the song listening log by adopting the feature extracting algorithm, so as to obtain user feature vectors of each user listening to different songs; the user characteristic vector of each user listening to different songs is determined according to the association degree between different users listening to each song in the song listening log and the initial user characteristic vector of each user listening to different songs;
the generating unit is configured to generate a feature vector relation table according to the user attribute information of the at least two users, the corresponding user feature vectors, the song feature vectors of songs listened to by each user, and the corresponding song attribute information.
8. An electronic device, characterized in that the electronic device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are in communication with each other through the communication bus;
A memory for storing a computer program;
A processor for implementing the method steps of any of claims 1-6 when executing a program stored on a memory.
9. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored therein a computer program which, when executed by a processor, implements the method steps of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010572038.1A CN111767426B (en) | 2020-06-22 | 2020-06-22 | Song recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010572038.1A CN111767426B (en) | 2020-06-22 | 2020-06-22 | Song recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111767426A CN111767426A (en) | 2020-10-13 |
| CN111767426B true CN111767426B (en) | 2024-04-26 |
Family
ID=72721395
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010572038.1A Active CN111767426B (en) | 2020-06-22 | 2020-06-22 | Song recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111767426B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103970802A (en) * | 2013-02-05 | 2014-08-06 | 百度在线网络技术（北京）有限公司 | Song recommending method and device |
| CN106951527A (en) * | 2017-03-21 | 2017-07-14 | 北京邮电大学 | A kind of song recommendations method and device |
| CN108021568A (en) * | 2016-10-31 | 2018-05-11 | 北京酷我科技有限公司 | One kind song is single to recommend method and device |
| US10515399B1 (en) * | 2014-12-15 | 2019-12-24 | Amazon Technologies, Inc. | System, method and medium for utilizing current and historical device context information to determine content item recommendations |
| WO2020093559A1 (en) * | 2018-11-09 | 2020-05-14 | 平安科技（深圳）有限公司 | Music recommendation method and apparatus, and computer device |
-
2020
- 2020-06-22 CN CN202010572038.1A patent/CN111767426B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103970802A (en) * | 2013-02-05 | 2014-08-06 | 百度在线网络技术（北京）有限公司 | Song recommending method and device |
| US10515399B1 (en) * | 2014-12-15 | 2019-12-24 | Amazon Technologies, Inc. | System, method and medium for utilizing current and historical device context information to determine content item recommendations |
| CN108021568A (en) * | 2016-10-31 | 2018-05-11 | 北京酷我科技有限公司 | One kind song is single to recommend method and device |
| CN106951527A (en) * | 2017-03-21 | 2017-07-14 | 北京邮电大学 | A kind of song recommendations method and device |
| WO2020093559A1 (en) * | 2018-11-09 | 2020-05-14 | 平安科技（深圳）有限公司 | Music recommendation method and apparatus, and computer device |
Non-Patent Citations (2)
| Title |
|---|
| 基于会话记录的Word2Vec音乐推荐算法研究;周航帆;周莲英;;通信技术(第04期);全文 * |
| 基于歌曲标签聚类的协同过滤推荐算法的研究;赵宇峰;李新卫;;计算机应用与软件(第06期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111767426A (en) | 2020-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110457581B (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN107665710B (en) | Mobile terminal voice data processing method and device | |
| CN108694940B (en) | Voice recognition method and device and electronic equipment | |
| US8886635B2 (en) | Apparatus and method for recognizing content using audio signal | |
| CN106484777B (en) | Multimedia data processing method and device | |
| CN111209477B (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN107729578B (en) | Music recommendation method and device | |
| CN110570855A (en) | system, method and device for controlling intelligent household equipment through conversation mechanism | |
| CN106708829B (en) | Data recommendation method and recommendation system | |
| WO2017097118A1 (en) | Text classification processing method and apparatus | |
| CN104750718A (en) | Data information search method and data information search device | |
| CN111767426B (en) | Song recommendation method and device | |
| CN111951808B (en) | Voice interaction method, device, terminal equipment and medium | |
| CN108932262B (en) | Song recommendation method and device | |
| CN113643706B (en) | Speech recognition method, device, electronic equipment and storage medium | |
| CN111984867B (en) | Network resource determining method and device | |
| CN111723198B (en) | Text emotion recognition method, device and storage medium | |
| CN113383360B (en) | Content pushing method, device, server side and storage medium | |
| CN112631945A (en) | Test case generation method and device and storage medium | |
| CN109286408B (en) | Method for energy harvesting and energy receiver | |
| CN113051425A (en) | Method for acquiring audio representation extraction model and method for recommending audio | |
| CN111415191A (en) | User classification method and device, electronic equipment and storage medium | |
| CN111325575A (en) | Question information recommendation method and device, computer equipment and storage medium | |
| CN111898033A (en) | Content pushing method and device and electronic equipment | |
| CN111461118A (en) | Interest feature determination method, device, equipment and storage 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |