A smart on-demand switching method for local music and online music

Abstract

The present invention relates to the technical field of intelligent on-demand switching, and specifically discloses an intelligent on-demand switching method for local music and online music, wherein the method comprises the following steps: S1: obtaining music playing records within a monitoring period, counting the total number of times different music is played and the number of times music is played within different time intervals, counting the number of complete plays, calculating the favorite coefficient, and calculating the initial music value; S2: sorting the music in the music playing records according to the size of the initial music value, generating a playing sequence, and playing from the first place in the playing sequence; S3: obtaining target music, filtering out music to be selected from the target music, calculating the music heat of the music to be selected, and calculating the music value of the music to be selected, and putting the music to be selected with the largest music value into the playing sequence. The present invention realizes the reasonable arrangement of the song list according to the user's historical listening record, and prevents repeated types of songs from recurring.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent on-demand switching, and in particular to an intelligent on-demand switching method for local music and online music. Background Art

[0002] Smart on-demand switching is an advanced technology that allows users to seamlessly switch content based on personal preferences or network conditions during music or video playback. Smart on-demand switching technology has not only changed people's entertainment methods through its high interactivity and personalized settings, but also promoted the innovative development of related technologies. With the continuous advancement of technology, smart on-demand switching will provide a richer and more immersive experience in the future.

[0003] The intelligent music on-demand switching technology realizes fast and accurate recognition and response to user intentions through highly integrated AI algorithms and big data analysis. Through advanced voice recognition and natural language processing technology, the system can accurately understand the user's on-demand request. Once the user's on-demand instruction is received, the system will complete the song search, buffering and playback preparation in a very short time. And based on the user's historical listening records and preferences, the system can intelligently recommend music that suits the user's taste.

[0004] In the existing technology, the intelligent on-demand switching technology mainly relies on artificial intelligence and big data technology. By building an efficient database, the system can more accurately identify the user's intention and respond in real time. Although the intelligent on-demand switching system can make personalized recommendations based on the user's historical listening records and preferences, the accuracy of such recommendations still needs to be improved. At the same time, recommendations based on historical listening records and preferences are prone to limitations, resulting in repeated types of songs being played multiple times within a period of time. Therefore, it is necessary to reasonably arrange the playlist based on the user's historical listening records, while preventing repeated types of songs from recurring. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent on-demand switching method for local music and online music to solve the above technical problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A method for intelligent on-demand switching between local music and online music, comprising the following steps:

[0008] S1: Get the music playing records within the preset monitoring period T, and count the total number of times z the i-th music is played i , and count the number of times z the i-th music is played in the time interval [0, T / 2) and the time interval (T / 2, T] i_l and z i_r ;

[0009] Count the number of times the i-th music is played completely z i_all , calculate the favorite coefficient of the i-th music , calculate the initial music value , where µ1 and µ2 are both preset music value variation coefficients and µ1<µ2;

[0010] S2: Sort the music in the music playing record according to the size of the initial music value. The larger the music value, the higher the ranking. Obtain the music corresponding to the top M initial music values ​​as the playing sequence, where M is a preset playing sequence threshold, and start playing from the first place in the playing sequence.

[0011] S3: When the first music in the play sequence is finished, the music that does not belong to the play sequence in the music play record is used as the target music, and the target music whose music type is not an existing type is used as the selected music, and the existing type is the music type in the play sequence. The music heat R of the jth selected music is calculated. j , the music heat R j The specific calculation formula is:

[0012] ;

[0013] Among them, γ is the preset music heat adjustment coefficient, Z represents the total number of times all music plays are recorded, and Z j，type represents the number of times the type of the jth selected music has been played, w j represents the number of times the jth selected music is played completely, q j Represents the number of times the jth music is not played completely;

[0014] Calculate the music value YYZ of the jth selected music j * =YYZ j +R j , put the music with the largest music value into the last place in the play sequence.

[0015] As a further solution of the present invention: in S3, when the first music in the play sequence is finished playing, the first music is removed from the play sequence.

[0016] As a further solution of the present invention: In S1, the liking coefficient K of the i-th music is calculated i Another method includes:

[0017] When the playback progress of a certain music is less than 60%, the current playback of the music is marked as low-complete playback; when the playback progress of a certain music is greater than or equal to 60%, the current playback of the music is marked as high-complete playback;

[0018] Count the number of times the i-th music is marked as low complete play z i_low and high completion counts i_high , calculate the favorite coefficient of the i-th music .

[0019] As a further solution of the present invention: in S2, the playback sequence threshold M is less than the total number of music types.

[0020] As a further solution of the present invention: in S3, when the selected music with the largest music value is not unique, the selected music with high music popularity is put into the play queue.

[0021] As a further solution of the present invention: in S3, when the music type of the jth music to be selected is not unique, the music types of the j music to be selected are taken as special target types, the total number of times the music of a single special target type is played is counted, and the mean is calculated as Z j，type .

[0022] As a further solution of the present invention: if the playing progress of a certain music is less than 3%, the music will not be counted and subsequent operations will not be performed.

[0023] As a further solution of the present invention: for the music in the music playing record that has not been played within the monitoring period T, the corresponding initial music value is set to be equal to 1.

[0024] Beneficial effects of the present invention: In the present invention, firstly, the music playing records of the user within the monitoring period are obtained. By obtaining the music playing records of the user, the user's music type preference can be roughly inferred, so as to recommend songs of the same type to the user based on the user's preferred music type. The purpose of setting the monitoring period is to determine the music type that the user likes within the time period. The user may like different music types in different time periods. Therefore, it is necessary to determine a range for clarifying the type of music that the user likes.

[0025] Users' preference for music is a very abstract event that cannot be quantified. Therefore, it is necessary to obtain the types of music that users like through their music playback records and design a reasonable music playback order for users. This music playback order must meet the users' own preferences for music types and thus be more in line with their hobbies.

[0026] Users are more likely to play the music they like in its entirety. Therefore, the more times a piece of music is played in its entirety, the more the user likes the music, so the calculated liking coefficient is larger. The liking coefficient will affect the calculation of subsequent music values. The music value is positively correlated with the liking coefficient, which also reflects the degree of user liking for the music.

[0027] For the calculation of music value, the monitoring period is divided into two time intervals. Because the user's favorite music type is predicted based on the music playback record, the interval closer to the current time accounts for a larger proportion in the calculation, which can more accurately reflect the user's recent music preferences. The larger the music value of a piece of music, the more the user likes the music recently. The interval farther away from the current time accounts for a smaller proportion in the calculation. Because with the change of factors such as the environment and the user's mentality, the user's favorite music type may change greatly. Therefore, in order to comprehensively judge the changes in the user's favorite music type, it is necessary to make adjustments based on the music listened to by the user in different periods to ensure that the user's favorite type changes can be reasonably reflected.

[0028] Sort the music by its music value, select the top M music with the largest music value as the play sequence, and start playing from the first place in the play sequence. This allows users to hear their favorite music. After the play is completed, calculate the music popularity of all the music. The higher the music popularity, the faster the music value rises, the shorter the time to enter the play sequence, and the faster the user will hear the song.

[0029] When the music is finished playing, select the music with the largest music value and different music type from the music that does not belong to the playback sequence and add it to the playlist. This can avoid repeatedly playing the same type of music in a short period of time. Repeatedly playing the same type of music in a short period of time is likely to cause users to resist this type of music, causing users to have aesthetic fatigue of this type of music. Therefore, select the music with the largest music value and different music type from the music that does not belong to the playback sequence and add it to the playlist. This allows users to listen to more types of music, which not only helps users relieve aesthetic fatigue of the same type of music, but also allows users to listen to different types of music, enhance users' interest in other types of music, and thus make users interested in trying more types of music.

[0030] The calculation formula of music popularity shows that the music popularity of each piece of music is not only related to its own active and passive play times, but also to the number of times the type of music is played. In the formula, Z j，type Represents the number of times the type of the j-th music is played. Each music has a music type. The more times the music of this type is played, the higher the popularity of the music will be, so it will enter the loop faster, and the number of times the same type of music is passively played will also increase.

[0031] In summary, the present invention realizes the reasonable arrangement of the playlist according to the user's historical listening records, while preventing the repetitive appearance of repeated types of songs. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present invention will be further described below in conjunction with the accompanying drawings.

[0033] Figure 1 The present invention is a flowchart of a method for intelligent on-demand switching between local music and online music. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] See also Figure 1 As shown, the present invention is a method for intelligent on-demand switching between local music and online music, comprising the following steps:

[0036] S1: Get the music playing records within the preset monitoring period T, and count the total number of times z the i-th music is played i , and count the number of times z the i-th music is played in the time interval [0, T / 2) and the time interval (T / 2, T] i_l and z i_r ;

[0037] Count the number of times the i-th music is played completely z i_all , calculate the favorite coefficient of the i-th music , calculate the initial music value , where µ1 and µ2 are both preset music value variation coefficients and µ1<µ2;

[0038] S2: Sort the music in the music playing record according to the size of the initial music value. The larger the music value, the higher the ranking. Obtain the music corresponding to the top M initial music values ​​as the playing sequence, where M is a preset playing sequence threshold, and start playing from the first place in the playing sequence.

[0039] S3: When the first music in the play sequence is finished, the music that does not belong to the play sequence in the music play record is used as the target music, and the target music whose music type is not an existing type is used as the selected music, and the existing type is the music type in the play sequence. The music heat R of the jth selected music is calculated. j , the music heat R j The specific calculation formula is:

[0040] ;

[0041] Among them, γ is the preset music heat adjustment coefficient, Z represents the total number of times all music plays are recorded, and Z j，type represents the number of times the type of the jth selected music has been played, w j represents the number of times the jth selected music is played completely, q j Represents the number of times the jth music is not played completely;

[0042] Calculate the music value YYZ of the jth selected music j * =YYZ j +R j , put the music with the largest music value into the last place in the play sequence.

[0043] It should be noted that we first obtain the user's music playback records within the monitoring period. By obtaining the user's music playback records, we can roughly infer the user's music type preferences, and then recommend the same type of songs to the user based on the user's preferred music type. The purpose of setting the monitoring period is to determine the user's favorite music type within the time period. The user may like different music types in different time periods. Therefore, it is necessary to determine a range to clarify the type of music the user likes.

[0044] Users' preference for music is a very abstract event that cannot be quantified. Therefore, it is necessary to obtain the types of music that users like through their music playback records and design a reasonable music playback order for users. This music playback order must meet the users' own preferences for music types and thus be more in line with their hobbies.

[0045] Users are more likely to play the music they like in its entirety. Therefore, the more times a piece of music is played in its entirety, the more the user likes the music, so the calculated liking coefficient is larger. The liking coefficient will affect the calculation of subsequent music values. The music value is positively correlated with the liking coefficient, which also reflects the degree of user liking for the music.

[0046] For the calculation of music value, the monitoring period is divided into two time intervals. Because the user's favorite music type is predicted based on the music playback record, the interval closer to the current time accounts for a larger proportion in the calculation, which can more accurately reflect the user's recent music preferences. The larger the music value of a piece of music, the more the user likes the music recently. The interval farther away from the current time accounts for a smaller proportion in the calculation. Because with the change of factors such as the environment and the user's mentality, the user's favorite music type may change greatly. Therefore, in order to comprehensively judge the changes in the user's favorite music type, it is necessary to make adjustments based on the music listened to by the user in different periods to ensure that the user's favorite type changes can be reasonably reflected.

[0047] Sort the music by its music value, select the top M music with the largest music value as the play sequence, and start playing from the first place in the play sequence. This allows users to hear their favorite music. After the play is completed, calculate the music popularity of all the music. The higher the music popularity, the faster the music value rises, the shorter the time to enter the play sequence, and the faster the user will hear the song.

[0048] When the music is finished playing, select the music with the largest music value and different music type from the music that does not belong to the playback sequence and add it to the playlist. This can avoid repeatedly playing the same type of music in a short period of time. Repeatedly playing the same type of music in a short period of time is likely to cause users to resist this type of music, causing users to have aesthetic fatigue of this type of music. Therefore, select the music with the largest music value and different music type from the music that does not belong to the playback sequence and add it to the playlist. This allows users to listen to more types of music, which not only helps users relieve aesthetic fatigue of the same type of music, but also allows users to listen to different types of music, enhance users' interest in other types of music, and thus make users interested in trying more types of music.

[0049] The calculation formula of music popularity shows that the music popularity of each piece of music is not only related to its own active and passive play times, but also to the number of times the type of music is played. In the formula, Z j，type Represents the number of times the type of the j-th music is played. Each music has a music type. The more times the music of this type is played, the higher the popularity of the music will be, so it will enter the loop faster, and the number of times the same type of music is passively played will also increase.

[0050] And in the invention, it is necessary to classify the types of music. Since the standards for distinguishing music types are different, the final results may be different. Therefore, in order to make the final calculation results unified and avoid large errors, the classification of music types is based on the music software used by the user. The present invention does not make any requirements on the type of music.

[0051] In another preferred embodiment of the present invention, when the first music in the play sequence is finished playing, the first music is removed from the play sequence.

[0052] It is worth noting that when the first music in the playback sequence is finished, it needs to be removed from the playback sequence and the music value needs to be updated. If the user likes the music very much, it should be ensured that the music enters the playback sequence faster and is adjusted according to the music popularity of the first music in the playback sequence. Since it is calculated first and then removed, it will not cause the first music to enter the playback sequence again, thus solving the problem of repeated playback of the same music in a short period of time.

[0053] In another preferred embodiment of the present invention, the liking coefficient K of the i-th music is calculated. i Another method includes:

[0054] When the playback progress of a certain music is less than 60%, the current playback of the music is marked as low-complete playback; when the playback progress of a certain music is greater than or equal to 60%, the current playback of the music is marked as high-complete playback;

[0055] Count the number of times the i-th music is marked as low complete play z i_low and high completion counts i_high , calculate the favorite coefficient of the i-th music .

[0056] It is understandable that the user's liking for music can also be reflected through the progress of music playback. If a piece of music has a large number of high-complete playbacks, then it can be determined that the music is the user's favorite music. Therefore, when calculating the liking coefficient of the music, the more high-complete playbacks, the larger the calculated liking coefficient. Therefore, when calculating the music value, due to the larger liking coefficient, the larger the music value, the higher the probability of the music entering the playback sequence, which can allow the user's favorite music to be heard faster, thereby satisfying the user's demand for his or her favorite music type.

[0057] In another preferred embodiment of the present invention, the play sequence threshold M is less than the total number of music types.

[0058] It should be noted that if the value of the playback sequence threshold M needs to be greater than the total number of music types, the music entering the playback sequence will inevitably be of the same type, and the purpose of the present invention is to avoid repeated playback of music of the same type, so the playback sequence threshold M must be smaller than the total number of music types. If the value of the playback sequence threshold M is smaller than the total number of music types, in this case, the same type of music will appear in the playback sequence anyway, and the requirements of the present invention cannot be met in this case. Therefore, in order to prevent users from hearing the same type of music in a short period of time, it is necessary to ensure that the playback sequence threshold M is smaller than the total number of music types.

[0059] In another preferred embodiment of the present invention, when the selected music with the largest music value is not unique, the selected music with high music popularity is put into the play queue.

[0060] It should be noted that when the music to be selected with the largest music value is not unique, since only one music can be played in the same time period, only one music can be put into the playlist at a time. At this time, if there are multiple music to be selected with the same music value, the music to be selected with the largest music popularity is selected to enter the playlist. This is because the music with high music popularity will increase faster when the music value is updated next time. Therefore, when the music to be selected with the largest music value is not unique, the problem can be solved by putting the music to be selected with high music popularity into the play queue. Although there are currently multiple music to be selected with the same music value, it is speculated that after the current music is played, the value of the music with high music popularity will increase faster.

[0061] In another preferred embodiment of the present invention, when the music type of the jth music to be selected is not unique, the music types of the jth music to be selected are taken as special target types, the total number of times the music of a single special target type is played is counted, and the mean is calculated as Z. j，type .

[0062] It is understandable that one type of music can be divided into multiple types of music. If all statistics are taken, it will be unfair to some music of only one type of music, so the average value is taken to weaken the influence of multiple music types on the music recommended by the present invention.

[0063] For example, a type of music can be both light music and pop music. If all of them are counted during calculation, the final calculation result will have a large error, which is not conducive to accurately pushing the user's favorite music type.

[0064] In another preferred embodiment of the present invention, if the playing progress of a certain music is less than 3%, no statistics and subsequent operations will be performed on the music.

[0065] It is worth noting that if the playback progress of a piece of music is less than 3%, it is impossible to determine whether the user likes the music. The music may be played due to an accidental touch by the user, or the user may not like the music. Since it is impossible to judge the above two situations, in order to reduce the impact of uncertain factors and more accurately recommend music that users like, music with a playback progress of less than 3% will not be counted or subsequently operated.

[0066] There is also a certain difference between music and songs. Most music has a long prelude, and the climax of the music is in the middle of the entire playback content. If the music playback progress is too short, it is difficult for users to hear the climax of the music, and it is difficult to judge in this case.

[0067] In another preferred embodiment of the present invention, for the music in the music playing record that has not been played within the monitoring period T, the corresponding initial music value is set to be equal to 1.

[0068] It is worth noting that the music in the music library has not been played within the monitoring period, so the music value cannot be calculated using the formula. At the same time, since it has not been played within the monitoring period T, it may be because the user does not like the music very much recently. However, the music in the music library has been played at least once, so the corresponding music value is set equal to 1.

[0069] The music in the music library may not be played during the monitoring period because the music was released too early or there is too much music in the music library and the user has forgotten it. The user may like this old music very much, but it has not been played because of forgetfulness. As the number of music in the user's music library increases, the probability of passively playing old music will also decrease. Therefore, in order to give some opportunities to music released very early, the corresponding music value is set to 1. If the user is interested after playing it, it will be played manually, and the music value is set to 1. In the absence of user interference, the probability of being played again is low. Due to the lack of broadcast data, it is difficult to determine whether the user still likes the music, and the frequency of the music being played in a loop is reduced. Subsequently, adjustments are made based on the user's playback of the music, increasing the probability of old music being played.

[0070] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for intelligent on-demand switching between local music and online music, characterized in that: The following steps are involved: S1: Get the music playing records within the preset monitoring period T, and count the total number of times z the i-th music is played i , and count the number of times z the i-th music is played in the time interval [0, T / 2) and the time interval (T / 2, T] i_l and z i_r ; Count the number of times the i-th music is played completely z i_all , calculate the like coefficient of the i-th music , calculate the initial music value , where µ1 and µ2 are both preset music value variation coefficients and µ1<µ2; S2: Sort the music in the music playing record according to the size of the initial music value. The larger the music value, the higher the ranking. Obtain the music corresponding to the top M initial music values ​​as the playing sequence, where M is a preset playing sequence threshold, and start playing from the first place in the playing sequence. S3: When the first music in the play sequence is finished, the music that does not belong to the play sequence in the music play record is used as the target music, and the target music whose music type is not an existing type is used as the selected music, and the existing type is the music type in the play sequence. The music heat R of the jth selected music is calculated. j , the music heat R j The specific calculation formula is: ； Among them, γ is the preset music heat adjustment coefficient, Z represents the total number of times all music plays are recorded, and Z j，type represents the number of times the type of the jth selected music has been played, w j represents the number of times the jth selected music is played completely, q j Represents the number of times the jth music is not played completely; Calculate the music value YYZ of the jth selected music j * =YYZ j +R j , put the music with the largest music value into the last place in the play sequence.

2. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: In S3, when the first music in the play sequence is finished playing, the first music is removed from the play sequence.

3. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: In S1, the like coefficient K of the i-th music is calculated i Another method includes: When the playback progress of a certain music is less than 60%, the current playback of the music is marked as low-complete playback; when the playback progress of a certain music is greater than or equal to 60%, the current playback of the music is marked as high-complete playback; Count the number of times the i-th music is marked as low complete play z i_low and high completion counts i_high , calculate the favorite coefficient of the i-th music .

4. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: In S2, the playing sequence threshold M is less than the total number of music types.

5. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: In S3, when the selected music with the largest music value is not unique, the selected music with the largest music popularity is put into the play queue.

6. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: In S3, when the music type of the jth music to be selected is not unique, the music types of the jth music to be selected are taken as special target types, the total number of times the music of a single special target type is played is counted, and the mean is calculated as Z. j，type .

7. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: If the playing progress of a certain music is less than 3%, the music will not be counted and subsequent operations will not be performed.

8. The method for intelligent on-demand switching between local music and online music according to claim 1, characterized in that: For the music in the music playing record that has not been played within the monitoring period T, the corresponding initial music value is set to be equal to 1.

Images