Hot music mining method, electronic device and computer readable storage medium

Abstract

The present application discloses a hot music mining method, device, electronic device and computer-readable storage medium, the method comprising: determining a seed user among all users of a music platform, and constructing a first feature vector based on the operation behavior data of the seed user; determining the music to be mined based on the operation behavior data of the seed user, and constructing a second feature vector based on the basic information of the music to be mined; splicing the first feature vector and the second feature vector to obtain a target feature vector of the music to be mined; inputting the target feature vector into a trained neural network model to obtain a prediction result of potential hot music. The hot music mining method provided by the present application improves the accuracy of hot music mining.

Description

Technical Field

[0001] The present application relates to the technical field of hotspot mining, and more specifically, to a hotspot music mining method, an electronic device, and a computer-readable storage medium. Background Art

[0002] Among related technologies, hotspot mining technology is mainly used in video, text and other fields. It conducts statistics and monitoring based on the content itself, and can only passively monitor the changes in the popularity of related words. There is no way to migrate to the mining of content such as music works, nor can it mine content that is currently unpopular but will become popular in the future.

[0003] Therefore, how to improve the accuracy of mining hot music is a technical problem that technicians in this field need to solve. Summary of the invention

[0004] The purpose of this application is to provide a hot music mining method, an electronic device and a computer-readable storage medium, which improves the accuracy of hot music mining.

[0005] To achieve the above objectives, the present application provides a hot music mining method in a first aspect, comprising:

[0006] Determine a seed user among all users of the music platform, and construct a first feature vector based on the operation behavior data of the seed user;

[0007] Determine the music to be mined based on the operation behavior data of the seed user, and construct a second feature vector based on the basic information of the music to be mined;

[0008] Concatenating the first feature vector and the second feature vector to obtain a target feature vector of the music to be mined;

[0009] The target feature vector is input into the trained neural network model to obtain the prediction result of potential hot music.

[0010] To achieve the above-mentioned purpose, the second aspect of the present application provides an electronic device, including:

[0011] Memory for storing computer programs;

[0012] A processor is used to implement the steps of the hot music mining method as described above when executing the computer program.

[0013] To achieve the above-mentioned purpose, the third aspect of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the hot music mining method as mentioned above are implemented.

[0014] Through the above scheme, it can be known that a hot music mining method provided by the present application includes: determining a seed user among all users of the music platform, and constructing a first feature vector based on the operation behavior data of the seed user; determining the music to be mined based on the operation behavior data of the seed user, and constructing a second feature vector based on the basic information of the music to be mined; splicing the first feature vector and the second feature vector to obtain the target feature vector of the music to be mined; inputting the target feature vector into the trained neural network model to obtain the prediction result of the potential hot music. The hot music mining method provided by the present application first screens seed users in the music platform. Since the seed users have a strong appreciation ability or a large influence on music, the music to be mined determined based on the operation behavior data of the seed has a greater chance of becoming a potential hot music. Secondly, constructing a target feature vector based on the operation behavior data of the seed user on the music to be mined and the basic information of the music to be mined, and inputting it into the neural network model to obtain the potential hot music. Constructing the target feature vector based on the operation behavior data of the seed user ensures that the music has a certain degree of dissemination and detonation point, and constructing the target feature vector based on the basic information of the music to be mined ensures that the production quality of the music is high enough. The combination of the two can improve the accuracy of mining hot music. The present application also discloses an electronic device and a computer-readable storage medium, which can also achieve the above-mentioned technical effects.

[0015] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. The drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following specific implementation methods, they are used to explain the present disclosure, but do not constitute a limitation to the present disclosure. In the drawings:

[0017] Figure 1 An architecture diagram of a hot music mining system provided in an embodiment of the present application;

[0018] Figure 2 A flowchart of a hot music mining method provided in an embodiment of the present application;

[0019] Figure 3 A flowchart of another hot music mining method provided in an embodiment of the present application;

[0020] Figure 4 A flowchart of another hot music mining method provided in an embodiment of the present application;

[0021] Figure 5 A structural diagram of a hot music mining device provided in an embodiment of the present application;

[0022] Figure 6 A structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0024] In order to facilitate understanding of the hot music mining method provided by this application, the system used by it is introduced below. Figure 1 , which shows an architecture diagram of a hot music mining system provided by an embodiment of the present application, such as Figure 1 As shown, a server 10 is included.

[0025] In the present application, the server 10 is used to execute the steps of the hot music mining method, including: determining a seed user among all users of the music platform; determining the music to be mined based on the operation behavior data of the seed user, and constructing a target feature vector of the music to be mined based on the operation behavior data and the basic information of the music to be mined; inputting the target feature vector into a trained neural network model to obtain a prediction result of potential hot music.

[0026] It is understandable that a neural network model can also be set up in the server 10 to predict potential hot music. The input of the neural network model is a target feature vector constructed based on the operational behavior data of the seed user on the music to be mined and the basic information of the music to be mined. The output of the neural network model can be the potential value corresponding to each music to be mined. The music to be mined with a potential value greater than the preset potential value is a potential hot music.

[0027] Furthermore, the hot music mining system may also include a number of clients 20 that establish communication connections with the server 10. The clients 20 may include fixed terminals such as PC (Chinese full name: Personal Computer, English full name: Personal Computer) and mobile terminals such as mobile phones, which are used to display potential hot music.

[0028] The embodiment of the present application discloses a hot music mining method, which improves the accuracy of hot music mining.

[0029] See also Figure 2 , a flowchart of a hot music mining method provided in an embodiment of the present application, such as Figure 2 As shown, including:

[0030] S101: determining a seed user among all users of the music platform, and constructing a first feature vector based on operation behavior data of the seed user;

[0031] The execution subject of this embodiment is the above-mentioned server, and the purpose is to screen potential hot music in the music platform. In this step, seed users are first determined among all users in the music platform. Since seed users have strong music appreciation ability or great influence, the music related to seed users has a greater chance of becoming potential hot music.

[0032] As a feasible implementation, the determination of seed users among all users of the music platform includes: determining the current popular music in the music platform and the unpopular time period corresponding to the current popular music, and determining the users who have operation behaviors on the current popular music in the unpopular time period as seed users. In this implementation, users with forward-looking operation behaviors, that is, users who successfully predict the current popular music, are regarded as high-quality users with advanced discovery vision, and high-quality users are used as seed users. The standard of the current popular music can be music with a current click-through rate greater than a preset click-through rate, and the forward-looking operation behaviors include playing it completely, playing it normally, commenting, collecting or sharing it in the unpopular time period of the current popular music, and the click-through rate of the current popular music in the unpopular time period is less than the preset click-through rate. In a specific implementation, the current popular music is analyzed in the music platform, and its corresponding unpopular time period is determined by going back in time, and the users who have operation behaviors on the current popular music in the unpopular time period are the above-mentioned high-quality users. As a preferred implementation, the number and duration of the above-mentioned forward-looking operation behaviors can be analyzed, and the number of forward-looking operation behaviors greater than the preset number, or the duration greater than the preset time, or the combination of the two users are determined as seed users.

[0033] As another feasible implementation, the step of determining seed users among all users of the music platform includes: determining high-quality music content in the music platform, and determining users who have operational behaviors on the high-quality music content as seed users. In this implementation, the content (User Generated Content, UGC) produced by ordinary users in the music platform is analyzed to screen high-quality music content. The high-quality music content here may include the number of high-quality comments produced by users, the number of high-quality playlist videos or podcasts released, etc. The users who produce the above-mentioned high-quality music content or the users who comment on, collect or analyze the above-mentioned music content are regarded as highly creative users, and the highly creative users are used as seed users.

[0034] As another feasible implementation, the step of determining seed users from all users of the music platform includes: determining users whose influence on the music platform is greater than a preset influence as seed users; wherein the influence includes the influence on the music platform and / or the influence on other platforms. In this implementation, the influence of all users on the music platform is analyzed, and users with greater influence are used as seed users. In a specific implementation, the influence of a user can be represented by the amount of attention, the number of works, or the interactive data of the user. The influence here can include the influence on the music platform, and can also include the influence on other platforms. That is, in this implementation, the influence of the user on the music platform can be analyzed, and the influence of the user on other platforms can also be analyzed.

[0035] Furthermore, this embodiment can also analyze the breadth, freshness, and unpopularity of users' music preferences, and accordingly screen out seed users who have the opportunity to predict popular music, which is not specifically limited in this embodiment.

[0036] On the basis of this step, as a preferred implementation mode, after determining the seed user among all users of the music platform, it also includes: constructing user characteristics of the seed user based on the user information and music preference portrait of the seed user; selecting non-seed training users from non-seed users, and constructing user characteristics of the non-seed training users based on the user information and music preference portrait of the non-seed training users; training a classification model using the user characteristics of the seed user and the user characteristics of the non-seed training users; inputting the user information and music preference portraits of non-seed users other than the non-seed training users into the trained classification model to predict extended seed users; and adding the extended seed users to the seed users.

[0037] In a specific implementation, user features are constructed based on the user's user information and music preference portrait. The user information may include age, gender, region, education, etc. The music preference portrait is used to describe the user's preferred music, and may include the user's preferred music style, preferred singer type, etc., which are not specifically limited here. The user features of the seed users are used as positive samples, and the user features of some non-seed users are used as negative samples to train the classification model. The classification model here may include a logistic regression model, etc., which are not specifically limited here. The trained classification model is used to predict the remaining non-seed users, and to screen out extended seed users who have comprehensive abilities but are not high in taste, creativity, and influence, and add them to the seed users. Furthermore, the attention relationship network and sharing and dissemination network of the seed users can also be analyzed, and the seed users can be expanded in the attention relationship network and sharing and dissemination network of the seed users.

[0038] Secondly, a first feature vector is constructed based on the operation behavior data of the seed user, and each operation behavior data includes an operation behavior, a seed user who performs the operation behavior, and the music being operated. As a feasible implementation method, the first feature vector is constructed based on the operation behavior data of the seed user, including: constructing a network topology based on the operation behavior data of the seed user; wherein the nodes in the network topology represent the seed user who performs the operation behavior or the music being operated, and the edges between the nodes represent the operation behavior; traversing the network topology according to the preset rules to obtain the music sequence to be mined, and performing feature representation on the music sequence to be mined to obtain the first feature vector. In a specific implementation, a network topology is constructed based on the operation behavior data of the seed user on the music, and the nodes of the network topology are the seed user and the music operated by the seed user, i.e., the music to be mined, and the edges between the nodes represent the operation behavior of the seed user on the music to be mined. Therefore, the network topology can be traversed by algorithms such as deepwalk or node2vec to determine the music sequence to be mined, which includes multiple identifiers of the music to be mined. Furthermore, the word2vec algorithm is used to perform feature representation on the music sequence to be mined to obtain the second feature vector.

[0039] S102: determining music to be mined based on the operation behavior data of the seed user, and constructing a second feature vector based on basic information of the music to be mined;

[0040] In this step, the music to be mined is determined based on the operation behavior data of the seed user, that is, the music related to the seed user is determined as the music to be mined, and the music to be mined can be determined based on the sequence of music to be mined obtained after traversing the network topology in the previous step. Convolutional neural networks (CNN, Convolutional Neural Networks), bidirectional encoder representation technology based on transformers (BERT, Bidirectional Encoder Representations from Transformers), graph neural networks (GNN, convolutional Neural Networks), etc. can be used to feature the basic information of the music to be mined, so as to construct a second feature vector. Here, the basic information of the music to be mined may include content knowledge graphs, audio, lyrics text, etc., and the content knowledge graph may include the lyricist, composer, singer, language, genre label, etc. of the music work, which is not specifically limited here.

[0041] S103: concatenating the first feature vector and the second feature vector to obtain a target feature vector of the music to be mined;

[0042] It should be noted that constructing the first feature vector based on the operational behavior data of seed users ensures that the music has a certain degree of dissemination and detonation point, and constructing the second feature vector based on the basic information of the music to be mined ensures that the production quality of the music is high enough. The target feature vector obtained by splicing the two can be used to evaluate whether the music is potential hot music.

[0043] S104: Input the target feature vector into the trained neural network model to obtain the prediction result of potential hot music.

[0044] In the specific implementation, the neural network model is used to predict potential hot music. In this step, the target feature vector constructed based on the operation behavior data of the seed user and the basic information of the music to be mined is input into the trained neural network model to obtain the prediction result of the potential hot music.

[0045] As a feasible implementation method, this step may include: inputting the target feature vector into the prediction result neural network model to obtain the potential value corresponding to each of the to-be-mined music; determining the to-be-mined music whose potential value is greater than the preset potential value as potential hotspot music. In a specific implementation, each row in the target feature vector corresponds to a piece of music to be mined, and the neural network model is used to predict the potential value of the to-be-mined music, that is, inputting the target feature vector into the neural network model can obtain the potential value corresponding to each piece of music to be mined, and determining the to-be-mined music whose potential value is greater than the preset potential value as potential hotspot music.

[0046] The hot music mining method provided in the embodiment of the present application first screens seed users in the music platform. Since seed users have strong appreciation ability or influence on music, the music to be mined determined based on the seed operation behavior data has a greater chance of becoming potential hot music. Secondly, a target feature vector is constructed based on the seed user's operation behavior data on the music to be mined and the basic information of the music to be mined, and it is input into the neural network model to obtain potential hot music. Constructing the target feature vector based on the seed user's operation behavior data ensures that the music has a certain degree of dissemination and detonation point, and constructing the target feature vector based on the basic information of the music to be mined ensures that the production quality of the music is high enough. The combination of the two can improve the accuracy of mining hot music.

[0047] The embodiment of the present application discloses a method for mining hot music. Compared with the previous embodiments, this embodiment further explains and optimizes the technical solution. Specifically:

[0048] See also Figure 3 , a flowchart of another hot music mining method provided by an embodiment of the present application, such as Figure 3 As shown, including:

[0049] S201: Determine a seed user among all users of the music platform;

[0050] S202: assigning a corresponding weight to each type of operation behavior data, and constructing a network topology based on the operation behavior data of the seed user and the weight corresponding to each type of operation behavior data;

[0051] In this embodiment, a corresponding weight can be assigned to each type of operation behavior data, for example, the weight of completed broadcast is greater than the weight of ordinary broadcast, and the weight of sharing is greater than the weight of commenting. The specific weight allocation standard can be flexibly set by those skilled in the art according to actual conditions. A network topology is constructed based on the operation behavior data of the seed user and the weight corresponding to each type of operation behavior data, that is, each edge in the network topology has a corresponding weight.

[0052] S203: traversing the network topology according to the weight corresponding to each type of operation behavior data to obtain a music sequence to be mined, and performing feature representation on the music sequence to be mined to obtain a first feature vector;

[0053] In this step, the music sequence to be mined is obtained by traversing based on the weight of each edge in the network topology, and the music sequence to be mined is represented by features to construct a first feature vector.

[0054] S204: determining the music to be mined based on the music sequence to be mined, and constructing a second feature vector based on basic information of the music to be mined;

[0055] S205: concatenating the first feature vector and the second feature vector to obtain a target feature vector of the music to be mined;

[0056] S206: Input the target feature vector into the trained neural network model to obtain the prediction result of potential hot music.

[0057] It can be seen that this embodiment assigns different weights to different types of operation behavior data. The weighted network topology can more accurately represent the operation behavior of the seed user. By traversing the weighted network topology, a more accurate target feature vector can be obtained, and the potential hot music predicted accordingly has a higher accuracy.

[0058] This embodiment introduces the training process of the neural network model, specifically:

[0059] See also Figure 4 , a flowchart of a training method for a neural network model provided in an embodiment of the present application, such as Figure 4 As shown, including:

[0060] S301: determining a seed user among all users of a music platform, and determining current popular music, current unpopular music, and an unpopular time period corresponding to the current popular music in the music platform;

[0061] In this embodiment, the current popular music in the music platform is used as positive samples, and a certain proportion of the current unpopular music in the music platform is randomly selected as negative samples to train the neural network model. The trained neural network model is used to predict potential popular music.

[0062] S302: constructing a training feature vector corresponding to the current popular music based on the operation behavior data of the seed user on the current popular music in the unpopular time period and the basic information of the current popular music;

[0063] S303: constructing a training feature vector corresponding to the current unpopular music based on the operation behavior data of the seed user on the current unpopular music and the basic information of the current unpopular music;

[0064] S304: Using the training feature vectors corresponding to the current popular music as positive samples and the training feature vectors corresponding to the current unpopular music as negative samples to train a neural network model, the trained neural network model is used to predict potential hot music based on the target feature vectors of the music to be mined.

[0065] In a specific implementation, a training feature vector corresponding to the current popular music, i.e., a training feature vector of a positive sample, is constructed based on the operational behavior data of the seed user on the current popular music during the unpopular time period and the basic information of the current popular music. A training feature vector corresponding to the current unpopular music, i.e., a training feature vector of a negative sample, is constructed based on the operational behavior data of the seed user on the current unpopular music and the basic information of the current unpopular music. The neural network model is trained using the training feature vectors corresponding to the positive samples and the negative samples, respectively, so that the trained neural network model can predict whether the music to be mined corresponding to the input target feature vector is potential hot music.

[0066] A hot music mining device provided in an embodiment of the present application is introduced below. The hot music mining device described below and the hot music mining method described above can be referenced to each other.

[0067] See also Figure 5 , a structural diagram of a hot music mining device provided in an embodiment of the present application, such as Figure 5 As shown, including:

[0068] A first construction module 100 is used to determine a seed user among all users of the music platform, and to construct a first feature vector based on the operation behavior data of the seed user;

[0069] A second construction module 200 is used to determine the music to be mined based on the operation behavior data of the seed user, and to construct a second feature vector based on the basic information of the music to be mined;

[0070] A concatenation module 300 is used to concatenate the first feature vector and the second feature vector to obtain a target feature vector of the music to be mined;

[0071] The input module 400 is used to input the target feature vector into the trained neural network model to obtain the prediction result of potential hot music.

[0072] The hot music mining device provided in the embodiment of the present application first screens seed users in the music platform. Since seed users have strong appreciation ability or influence on music, the music to be mined determined based on the seed user's operation behavior data has a greater chance of becoming potential hot music. Secondly, a target feature vector is constructed based on the seed user's operation behavior data on the music to be mined and the basic information of the music to be mined, and it is input into the neural network model to obtain potential hot music. Constructing the target feature vector based on the seed user's operation behavior data ensures that the music has a certain degree of dissemination and detonation point, and constructing the target feature vector based on the basic information of the music to be mined ensures that the production quality of the music is high enough. The combination of the two can improve the accuracy of mining hot music.

[0073] Based on the above embodiment, as a preferred implementation, the first construction module 100 is specifically a module for determining the current popular music in the music platform and the unpopular time period corresponding to the current popular music, determining the user who has an operation behavior on the current popular music in the unpopular time period as a seed user, and constructing the first feature vector based on the operation behavior data of the seed user;

[0074] Based on the above embodiment, as a preferred implementation, the first construction module 100 is specifically a module for determining high-quality music content in the music platform, determining users who have operation behaviors on the high-quality music content as seed users, and constructing a first feature vector based on the operation behavior data of the seed users;

[0075] Based on the above embodiments, as a preferred implementation mode, the first construction module 100 is specifically a module for determining users whose influence in the music platform is greater than a preset influence as seed users, and constructing a first feature vector based on the operation behavior data of the seed users; wherein the influence includes the influence in the music platform and / or the influence in other platforms.

[0076] Based on the above embodiment, as a preferred implementation, it also includes:

[0077] A third building module is used to build user characteristics of the seed user based on the user information and music preference portrait of the seed user;

[0078] A fourth construction module is used to select non-seed training users from non-seed users, and construct user features of the non-seed training users based on user information and music preference portraits of the non-seed training users;

[0079] A first training module, used for training a classification model using the user characteristics of the seed users and the user characteristics of the non-seed training users;

[0080] A prediction module, used for inputting user information and music preference portraits of non-seed users other than the non-seed training users into the trained classification model to predict the expanded seed users;

[0081] A adding module is used to add the extended seed user to the seed user.

[0082] Based on the above embodiment, as a preferred implementation, the first building module 100 includes:

[0083] A determination unit, used for determining a seed user among all users of the music platform;

[0084] A construction unit is used to construct a network topology based on the operation behavior data of the seed user; wherein each piece of the operation behavior data includes an operation behavior, a seed user who performs the operation behavior, and the operated music, and the nodes in the network topology represent the seed user who performs the operation behavior or the operated music, and the edges between the nodes represent the operation behavior;

[0085] The traversal unit is used to traverse the network topology according to preset rules to obtain the music sequence to be mined, and to perform feature representation on the music sequence to be mined to obtain a first feature vector.

[0086] Based on the above embodiment, as a preferred implementation, the second construction module 200 is specifically a module for determining the music to be mined based on the music sequence to be mined, and constructing a second feature vector based on the basic information of the music to be mined.

[0087] Based on the above embodiment, as a preferred implementation mode, the construction unit specifically assigns a corresponding weight to each type of operation behavior data, and constructs a network topology based on the operation behavior data of the seed user and the weight corresponding to each type of operation behavior data, and determines the unit of the music to be mined according to the network topology;

[0088] Correspondingly, the traversal unit is specifically a unit that traverses the network topology according to the weight corresponding to each type of operation behavior data to obtain the music sequence to be mined, and performs feature representation on the music sequence to be mined to obtain the first feature vector.

[0089] On the basis of the above embodiments, as a preferred implementation mode, the input module 400 is specifically a module that inputs the target feature vector into the trained neural network model to obtain the potential value corresponding to each of the music to be mined, and determines the music to be mined whose potential value is greater than the preset potential value as potential hot spot music.

[0090] Based on the above embodiment, as a preferred implementation, it also includes:

[0091] A determination module, used to determine a seed user among all users of the music platform, and to determine current popular music, current unpopular music, and an unpopular time period corresponding to the current popular music in the music platform;

[0092] A fifth construction module is used to construct a training feature vector corresponding to the current popular music based on the operation behavior data of the seed user on the current popular music in the unpopular time period and the basic information of the current popular music;

[0093] A sixth construction module, configured to construct a training feature vector corresponding to the current unpopular music based on the operation behavior data of the seed user on the current unpopular music and the basic information of the current unpopular music;

[0094] The second training module is used to train a neural network model using the training feature vectors corresponding to the current popular music as positive samples and the training feature vectors corresponding to the current unpopular music as negative samples. The trained neural network model is used to predict potential hot music based on the target feature vectors of the music to be mined.

[0095] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0096] The present application also provides an electronic device, see Figure 6 , a structural diagram of an electronic device 60 provided in an embodiment of the present application, such as Figure 6 As shown, a processor 61 and a memory 62 may be included.

[0097] Among them, the processor 61 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 61 can be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 61 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 61 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 61 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0098] The memory 62 may include one or more computer-readable storage media, which may be non-transitory. The memory 62 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In this embodiment, the memory 62 is at least used to store the following computer program 621, wherein, after the computer program is loaded and executed by the processor 61, it can implement the relevant steps in the hot music mining method performed by the server side disclosed in any of the aforementioned embodiments. In addition, the resources stored in the memory 62 may also include an operating system 622 and data 623, etc., and the storage method may be short-term storage or permanent storage. Among them, the operating system 622 may include Windows, Unix, Linux, etc.

[0099] In some embodiments, the electronic device 60 may further include a display screen 63 , an input / output interface 64 , a communication interface 65 , a sensor 66 , a power supply 67 , and a communication bus 68 .

[0100] certainly, Figure 6 The structure of the electronic device shown does not constitute a limitation on the electronic device in the embodiments of the present application. In actual applications, the electronic device may include Figure 6 More or fewer components than shown, or combinations of certain components.

[0101] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When the program instructions are executed by a processor, the steps of the hot music mining method executed by the electronic device in any of the above embodiments are implemented.

[0102] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of this application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of this application.

[0103] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

Claims

1. A hot music mining method, characterized in that: include: Determine a seed user among all users of the music platform, and construct a first feature vector based on the operation behavior data of the seed user; Determine the music to be mined based on the operation behavior data of the seed user, and construct a second feature vector based on the basic information of the music to be mined; Concatenating the first feature vector and the second feature vector to obtain a target feature vector of the music to be mined; Inputting the target feature vector into the trained neural network model to obtain the prediction result of potential hot music; The step of constructing a first feature vector based on the operation behavior data of the seed user includes: A network topology is constructed based on the operation behavior data of the seed user; wherein each piece of the operation behavior data includes an operation behavior, a seed user who performs the operation behavior, and the operated music, a node in the network topology represents a seed user who performs the operation behavior or the operated music, the operated music is the music to be mined, and an edge between nodes represents the operation behavior; The network topology is traversed according to preset rules to obtain a music sequence to be mined, and a feature representation is performed on the music sequence to be mined to obtain a first feature vector.

2. The hot music mining method according to claim 1, characterized in that: After determining the seed user among all users of the music platform, the method further includes: Constructing user characteristics of the seed user based on the user information and music preference portrait of the seed user; Selecting non-seed training users from non-seed users, and constructing user features of the non-seed training users based on user information and music preference portraits of the non-seed training users; Training a classification model using the user characteristics of the seed user and the user characteristics of the non-seed training user; Inputting user information and music preference profiles of non-seed users other than the non-seed training users into the trained classification model to predict expanded seed users; The extended seed user is added to the seed user.

3. The hot music mining method according to claim 1, characterized in that: Determining the seed user from all users of the music platform includes: Determine the current popular music in the music platform and the unpopular time period corresponding to the current popular music, and determine the users who have played the current popular music in the unpopular time period as seed users; and / or, determining high-quality music content in the music platform, and determining users who have operated the high-quality music content as seed users; And / or, users whose influence on the music platform is greater than a preset influence are determined as seed users; wherein the influence includes the influence on the music platform and / or the influence on other platforms.

4. The hot music mining method according to claim 1, characterized in that: The determining of the music to be mined based on the operation behavior data of the seed user comprises: The music to be mined is determined based on the sequence of music to be mined.

5. The hot music mining method according to claim 1, characterized in that: The constructing of a network topology based on the operation behavior data of the seed user includes: Assigning a corresponding weight to each type of operation behavior data, and constructing a network topology based on the operation behavior data of the seed user and the weight corresponding to each type of operation behavior data; Accordingly, traversing the network topology according to the preset rules to obtain the music sequence to be mined includes: The network topology is traversed according to the weight corresponding to each type of operation behavior data to obtain the music sequence to be mined.

6. The hot music mining method according to claim 1, characterized in that: The target feature vector is input into the trained neural network model to obtain the prediction results of potential hot music, including: Inputting the target feature vector into the trained neural network model to obtain the potential value corresponding to each of the music to be mined; The music to be mined whose potential value is greater than the preset potential value is determined as potential hot spot music.

7. A training method for a neural network model, characterized in that: include: Determine a seed user among all users of a music platform, and determine current popular music, current unpopular music, and an unpopular time period corresponding to the current popular music in the music platform; Constructing a training feature vector corresponding to the current popular music based on the operation behavior data of the seed user on the current popular music during the unpopular time period and the basic information of the current popular music; Constructing a training feature vector corresponding to the current unpopular music based on the operation behavior data of the seed user on the current unpopular music and the basic information of the current unpopular music; The training feature vector corresponding to the current popular music is used as a positive sample, and the training feature vector corresponding to the current unpopular music is used as a negative sample to train a neural network model. The trained neural network model is used to predict potential hot music based on the target feature vector of the music to be mined.

8. An electronic device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the hot music mining method as described in any one of claims 1 to 6 or the training method of the neural network model as described in claim 7 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the hot music mining method as described in any one of claims 1 to 6 or the training method of the neural network model as described in claim 7.

Images