CN109063069A - Song label determines method, apparatus, computer equipment and readable storage medium storing program for executing - Google Patents
Song label determines method, apparatus, computer equipment and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN109063069A CN109063069A CN201810812529.1A CN201810812529A CN109063069A CN 109063069 A CN109063069 A CN 109063069A CN 201810812529 A CN201810812529 A CN 201810812529A CN 109063069 A CN109063069 A CN 109063069A
- Authority
- CN
- China
- Prior art keywords
- target
- song
- information
- sample
- lyrics
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Electrophonic Musical Instruments (AREA)
Abstract
The present invention relates to a kind of song labels to determine method, which comprises the validity feature word information in the target lyrics is obtained according to the target lyrics of target song and default song tag database；The target cadence information of the target song is obtained according to the temporal information of the target song and the target lyrics；In conjunction with the validity feature word information, the target cadence information and the default song tag database, the target labels of the target song are obtained.The present invention program song label determines method and apparatus, can carry out label for labelling to song with fast automatic, effectively reduce cost of labor required for song label for labelling and time cost, in addition, Benchmark is unified, label body system unified standard.
Description
Technical field
The present invention relates to field of computer technology, determine that method, apparatus, computer are set more particularly to a kind of song label
Standby and readable storage medium storing program for executing.
Background technique
In order to classify to song, existing audio player would generally mark label to the song in library.
There are two types of types for general song label, are that song mark is corresponding one is the personnel by possessing musical expertise
Label, it is independently song mark label that another kind, which is by user,.
When marking label using expert, inefficiency and labor intensive；And when independently marking label using user, then can
Irregular due to user cognition field causes label system chaotic.
Summary of the invention
The purpose of the present invention is to provide a kind of song labels to determine method, apparatus computer equipment and readable storage medium
Matter can carry out label for labelling to song with fast automatic, effectively reduce cost of labor required for song label for labelling and time
Cost, in addition, Benchmark is unified, label body system unified standard.
The purpose of the present invention is achieved through the following technical solutions:
A kind of song label determines method, which comprises
Effective spy in the target lyrics is obtained according to the target lyrics of target song and default song tag database
Levy word information；
The target rhythm letter of the target song is obtained according to the temporal information of the target song and the target lyrics
Breath；
In conjunction with the validity feature word information, the target cadence information and the default song tag database, obtain
Take the target labels of the target song.
The default song tag database stores the sample song of multiple sample songs in one of the embodiments,
Word, sample cadence information, the corresponding standard label of each sample song, the corresponding sample song of each standard label sample
The sample characteristics word information of eigen word information and sample cadence information and the corresponding all sample songs of same standard label
Intersection and sample cadence information intersection.
It is described in one of the embodiments, to be obtained according to the target lyrics of target song and default song tag database
Multiple validity feature words in the target lyrics include:
The target lyrics of the target song are split into multiple group of words, the target extracted in the multiple group of words is real
Word information；
In conjunction with the sample characteristics word information of the target notional word information and the corresponding sample song of each standard label
The validity feature word information is obtained with sample cadence information.
Target notional word information described in the combination and the corresponding sample of each standard label in one of the embodiments,
The sample characteristics word information and sample cadence information of this song obtain the validity feature word information
In conjunction with the sample characteristics word information of the target notional word information and the corresponding sample song of each standard label
The information gain of each target notional word in the target notional word information is obtained with sample cadence information；
All target notional words in the target notional word information are sorted according to corresponding informance gain size, information is chosen and increases
The target notional word of the maximum predetermined number of benefit is as the validity feature word information.
Validity feature word information described in the combination, the target cadence information and institute in one of the embodiments,
Default song tag database is stated, the target labels for obtaining the target song include:
In conjunction with the validity feature word information, the target cadence information and the corresponding all samples of same standard label
The sample characteristics word information intersection and sample cadence information intersection of song calculate separately the target using default disaggregated model and sing
Song belongs to the probability of each standard label；
The corresponding standard label of the target song maximum probability is chosen as the target labels.
The target cadence information includes mean tempo rate information and rhythm pause duration in one of the embodiments,
Information.
It is described in one of the embodiments, that institute is obtained according to the temporal information of the target song and the target lyrics
The target cadence information for stating target song includes:
The described average of the target lyrics is obtained in conjunction with the time that the length and the target lyrics of the target lyrics occupy
Rhythm rate information；
The rhythm pause duration information is obtained in conjunction with lyrics pause duration in the target lyrics.
A kind of song label determining device, described device include:
Feature Words obtain module, for according to the target lyrics of target song and the acquisition of default song tag database
Validity feature word information in the target lyrics；
Rhythm obtains module, for obtaining the target according to the temporal information of the target song and the target lyrics
The target cadence information of song；
Label determining module, in conjunction with the validity feature word information, the target cadence information and described default
Song tag database obtains the target labels of the target song.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device realizes following steps when executing the computer program:
Effective spy in the target lyrics is obtained according to the target lyrics of target song and default song tag database
Levy word information；
The target rhythm letter of the target song is obtained according to the temporal information of the target song and the target lyrics
Breath；
In conjunction with the validity feature word information, the target cadence information and the default song tag database, obtain
Take the target labels of the target song.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
Following steps are realized when row:
Effective spy in the target lyrics is obtained according to the target lyrics of target song and default song tag database
Levy word information；
The target rhythm letter of the target song is obtained according to the temporal information of the target song and the target lyrics
Breath；
In conjunction with the validity feature word information, the target cadence information and the default song tag database, obtain
Take the target labels of the target song.
According to the scheme of aforementioned present invention, obtained according to the target lyrics of target song and default song tag database
Validity feature word information in the target lyrics can establish unified label and obtain in conjunction with default song tag database
Standard, label body system unified standard；The target song is obtained according to the temporal information of the target song and the target lyrics
Bent target cadence information；In conjunction with the validity feature word information, the target cadence information and the default song label
Database obtains the target labels of the target song, can carry out label for labelling to song with fast automatic, effectively reduce song
Cost of labor and time cost required for label for labelling.
Detailed description of the invention
Fig. 1 is the applied environment figure that song label determines method in one embodiment；
Fig. 2 is the flow diagram that song label determines method in one embodiment；
Fig. 3 is the flow diagram that song label determines method in one embodiment；
Fig. 4 is the flow diagram that song label determines method in another embodiment；
Fig. 5 is the structural block diagram of song label determining device in one embodiment；
Fig. 6 is the structural block diagram of song label determining device in one embodiment；
Fig. 7 is the structural block diagram of song label determining device in another embodiment；
Fig. 8 is the internal structure chart of computer equipment in one embodiment；
Fig. 9 is the internal structure chart of computer equipment in another embodiment.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention more comprehensible, with reference to the accompanying drawings and embodiments, to this
Invention is described in further detail.It should be appreciated that the specific embodiments described herein are only used to explain the present invention,
And the scope of protection of the present invention is not limited.
Song label provided by the present application determines method, can be applied in application environment as shown in Figure 1.Wherein, it takes
Business device obtains the validity feature word in the target lyrics according to the target lyrics and default song tag database of target song
Information；Server obtains the target rhythm of the target song according to the temporal information and the target lyrics of the target song
Information；Server in conjunction with the validity feature word information, the target cadence information and the default song tag database,
Obtain the target labels of the target song.Wherein, server can be formed with the either multiple servers of independent server
Server cluster realize.It will be understood by those skilled in the art that song label provided by the present application determines method, not only may be used
To be applied in application environment shown in FIG. 1, it can also apply but be not limited in various computers or server.
In one embodiment, as shown in Fig. 2, providing a kind of song label determines method, it is applied to Fig. 1 in this way
In server for be illustrated, comprising the following steps:
Step S102 is obtained in the target lyrics according to the target lyrics of target song and default song tag database
Validity feature word information；
Specifically, by being labeled with multiple sample songs of standard tag information, default song tag database is established, so
Default song tag database is combined to obtain target labels, the unifying datum in label determination process afterwards, label body system one is advised
Model.
Step S104 obtains the mesh of the target song according to the temporal information of the target song and the target lyrics
Mark cadence information；
Specifically, the target cadence information includes mean tempo rate information and rhythm pause duration information.
Step S106, in conjunction with the validity feature word information, the target cadence information and the default song label
Database obtains the target labels of the target song.
Above-mentioned song label determines in method, is obtained according to the target lyrics of target song and default song tag database
Validity feature word information in the target lyrics can establish unified label and obtain in conjunction with default song tag database
Standard, label body system unified standard；The target song is obtained according to the temporal information of the target song and the target lyrics
Bent target cadence information；In conjunction with the validity feature word information, the target cadence information and the default song label
Database obtains the target labels of the target song, can carry out label for labelling to song with fast automatic, effectively reduce song
Cost of labor and time cost required for label for labelling.
The default song tag database stores the sample song of multiple sample songs in one of the embodiments,
Word, sample cadence information, the corresponding standard label of each sample song, the corresponding sample song of each standard label sample
The sample characteristics word information of eigen word information and sample cadence information and the corresponding all sample songs of same standard label
Intersection and sample cadence information intersection.
In one of the embodiments, as shown in figure 3, the target lyrics and default song label according to target song
Database obtains multiple validity feature words in the target lyrics
The target lyrics of the target song are split into multiple group of words, extract the multiple group of words by step S1021
In target notional word information；
Specifically, the target lyrics are splitted by group of words one by one using segmenter, filters function word, modal particle, extract the lyrics
Interior notional word part, i.e. target notional word information.
Step S1022, in conjunction with the corresponding all sample songs of the target notional word information and each standard tag information
Bent sample characteristics word intersection obtains the validity feature word information.
In one of the embodiments, as shown in figure 4, target notional word information described in the combination and each standard mark
The sample characteristics word intersection of the corresponding all sample songs of label information obtains the validity feature word information
Step S10221, in conjunction with the sample of the target notional word information and the corresponding sample song of each standard label
Feature word information and sample cadence information obtain the validity feature word information.
Step S10222, by all target notional words in the target notional word information according to the big float of corresponding informance gain
Sequence chooses the target notional word of the maximum predetermined number of information gain as the validity feature word information.
Specifically, the feature selecting mode based on Information Gain Method is a kind of common hand for carrying out text feature selection
Method, Information Gain Method measures the information content of a system using the concept of entropy, and the information gain of some feature then refers to,
Under known conditions, the variation degree of the information content of whole system, attribute gain is an important indicator of feature selecting, it is defined
This feature how much information can be carried out for categorizing system synteny, and bring information is more, then illustrate that this feature is more important.
The comentropy of each target notional word in target notional word information is calculated first, and calculation formula is as follows:
In formula: wherein p (yi) chance event Y is represented as yiProbability, H (Y) is average information；Y refers to target notional word
Intersection, yiRefer to every i-th of target notional word.
Then the conditional entropy of each target notional word in target notional word information is calculated, the calculation formula of conditional entropy is as follows:
In formula: under conditions of H (Y | X) refers to known variables X, another expectation of variable Y entropy to X；P (y | x) refer to x thing
When part occurs, the probability of y event generation；X refers to standard label；Y is target notional word；X is the intersection of standard label；Y is feeling the pulse with the finger-tip
Mark the intersection of notional word；That is in the case that H (Y | X) is known standard label, the information content of target notional word.
Then information gain value is calculated according to conditional entropy and comentropy, calculation formula is as follows:
IG (Y, X)=H (Y)-H (Y | X) (3)
In formula: IG (Y, X) representative information yield value, information gain score have measured the information content of the lyrics feature, information
The amount the big, represents the importance of the lyrics feature.
Assuming that extracting four first sample songs, wherein the label of two songs is " Chinese feature ", two song labels are " feelings
Song ", the target notional word information chosen in target song include " in love ", " shedding tears " and " misty rain " these three words, it is assumed that this three
The number that a word occurs in four first songs is as shown in the table:
| (A) in love | Shed tears (B) | Misty rain (C) | Label |
| 1 | 1 | 0 | Love song |
| 0 | 1 | 1 | Love song |
| 1 | 0 | 0 | Chinese feature |
| 1 | 0 | 1 | Chinese feature |
The comentropy of the intersection of target notional word is calculated according to upper meter:
The conditional entropy of each Feature Words is calculated according to the formula of conditional entropy:
H (Y | B)==0-0-0-0=0
Then according to information gain formula, the information gain score of each Feature Words is calculated, the information of " shedding tears " is obtained
Gain is scored at 1, and the information gain of " in love " is scored at 0.31, and the information gain of " misty rain " is scored at 0.
" shedding tears " (1) > " in love " (0.31) > " misty rain " (0)
Select the target notional word of the maximum predetermined number of information gain as the validity feature word information, such as selection letter
The maximum two target notional words of gain are ceased as validity feature word information, then finally obtained validity feature word information just includes
" shedding tears " and " in love ".
The target cadence information includes mean tempo rate information and rhythm pause duration in one of the embodiments,
Information.
In one of the embodiments, as shown in figure 3, the temporal information and the target according to the target song
The target cadence information that the lyrics obtain the target song includes:
Step S1041 obtains the target lyrics in conjunction with the time of length and the target lyrics occupancy of the target lyrics
The mean tempo rate information；
Step S1042 obtains the rhythm pause duration information in conjunction with lyrics pause duration in the target lyrics.
It specifically, also include the temporal information of the lyrics, according to LRC due to not only possessing lyrics information in LRC lyrics file
The lyrics of format set the rhythm characteristic that following formula extracts song:
Avg(Si)=Duration (Si)/Len(Si) (4)
Avg (S in formulai) represent each word occupied time in current lyrics segment, that is, mean tempo quantity
Information, Duration (Si) then represent in the lyrics segment be occupy time, Len (Si) then represent the length of the lyrics.
Specifically, the length of the lyrics refers to the number of words of the lyrics, does not include the space between a lyrics.
By taking the following lyrics as an example:
[04:59.82] you always the too soft heart of the heart is too soft
[05:05.38] all oneself shoulders all problems
[05:09.93] is in love always simply to get along too difficult
[05:15.75] be not inadequate again without being you
[05:20.47] be not inadequate again without being you
[05:25.63] be not inadequate again without being you
[05:30.90] be not inadequate again without being you
[05:35.81]
By taking this section of lyrics " [04:59.82] you always the too soft heart of the heart is too soft " as an example, the time in bracket is this section of lyrics
At the beginning of T (Si), until T (S at the beginning of the next section of lyricsi+1) so the lyrics segment occupy time be then
Duration(Si)=T (Si+1)-T(Si)=305s-299s=6s in seconds, and Len (Si) it is then 9, then can obtain
The rhythm characteristic of the lyrics segment out:
Lyrics rhythm demarcation interval is carried out by section of 100ms, while can also be obtained by above-mentioned lyrics segment, [05:
35.81] there is the performance part of the not lyrics, belong to the transition stage of background or absolute music, the part is as song
Rhythm terminates Core feature word of the feature extraction as following categorizing songs models.The flat of every first song is extracted from LRC file
Equal rhythm rate and rhythm pause duration.
In one of the embodiments, as shown in figure 3, validity feature word information described in the combination, the target rhythm
Information and the default song tag database, the target labels for obtaining the target song include:
Step S1061, it is corresponding in conjunction with the validity feature word information, the target cadence information and same standard label
All sample songs sample characteristics word information intersection and sample cadence information intersection, calculated separately using default disaggregated model
The target song belongs to the probability of each standard label；
Step S1062 chooses the corresponding standard label of the target song maximum probability as the target labels.
Specifically, disaggregated model is using Naive Bayes Classification Model, in order to verify Naive Bayes Classification Model
Accuracy in computation, randomly select the song lyrics data that 5000 do not mark label and carry out test assessment, pass through song respectively
Label coverage, can the coverage of song of prediction label account for the accounting and accuracy of overall test song amount, that is,
The accuracy of song label for labelling is assessed, its coverage, the accuracy as the result is shown of final test are averagely respectively 80%
And 85%, it was demonstrated that by this method carry out song label for labelling coverage and accuracy a reasonable acceptable use water
In flat.
For example: assuming that the standard label and cadence information collection that store in default song tag database share vector table
Be shown as [" in love ", " 100ms ", " misty rain ", " 300ms "], and only " love song " and " R&B (and Rhythm and Blues, save
Play Bruce) " two tag along sorts, two class sample sizes i.e. validity feature word quantity are 8, are parsed with certain head song
The vector space model obtained for the feature [" in love ", " misty rain ", " 300ms "] come are as follows: x=[1,0,1,1] recycles Piao
Plain Bayesian model calculates classification prediction, it is assumed that the frequency statistics of sample data are as follows:
It is calculated using the model-naive Bayesian for being applied to text classification, while in order to avoid calculating probability
When, there is CiUnder do not occur feature Xi, and lead to this unreasonable situation of zero probability occur, it is flat to introduce Laplce
Sliding processing, carries out+1 to molecule, adds classification number N to denominator, then target labels are standard label C1Probability calculation formula are as follows:
In formula: i ∈ I, X indicate the intersection of validity feature word information x in target song；xiIndicate i-th of validity feature word,
The intersection of I expression Feature Words number；p(xi|C1) indicate that ith feature word appears in standard label C1Corresponding sample characteristics word
Probability in information intersection；A (i, 1) is indicated in standard label C1Ith feature word in corresponding sample characteristics word information intersection
The number of appearance；B1Expression appears in standard label C1In validity feature word total quantity；The classification number of N expression standard label.
It can be calculated in conjunction with upper table and formula (5):
Equally, then ith feature word is standard label C2Probability calculation formula are as follows:
In formula: p (X | C2) indicate that ith feature word appears in standard label C2In corresponding sample characteristics word information intersection
Probability；A (i, 2) is indicated in standard label C2The number that ith feature word occurs in corresponding sample characteristics word information intersection；
B2Expression appears in standard label C2In validity feature word total quantity；The classification number of N expression standard label.
It can be calculated in conjunction with upper table and formula (6):
Therefore, the probability that the target labels belong to " love song " is much larger than the probability for belonging to " R&B ", shows the target song
Bent target labels are " love song ".
In the remittance of wherein one embodiment, as shown in figure 5, providing a kind of song label determining device, described device packet
It includes:
Feature Words obtain module 102, for according to the target lyrics of target song and the acquisition of default song tag database
Validity feature word information in the target lyrics；
Rhythm obtains module 104, for according to the temporal information of the target song and target lyrics acquisition
The target cadence information of target song；
Label determining module 106, in conjunction with the validity feature word information, the target cadence information and described pre-
If song tag database obtains the target labels of the target song.
The default song tag database stores the sample song of multiple sample songs in one of the embodiments,
Word, sample cadence information, the corresponding standard label of each sample song, the corresponding sample song of each standard label sample
The sample characteristics word information of eigen word information and sample cadence information and the corresponding all sample songs of same standard label
Intersection and sample cadence information intersection.
In one of the embodiments, as shown in fig. 6, Feature Words acquisition module 102 includes:
Notional word acquiring unit 1021 extracts institute for the target lyrics of the target song to be split into multiple group of words
State the target notional word information in multiple group of words；
Feature Words determination unit 1022, in conjunction with the target notional word information and the corresponding sample of each standard label
The sample characteristics word information and sample cadence information of this song obtain the validity feature word information.
In one of the embodiments, as shown in fig. 7, the Feature Words determination unit 1022 includes:
Information gain computing unit 10221, for corresponding in conjunction with the target notional word information and each standard label
Sample song sample characteristics word information and sample cadence information obtain each target notional word in the target notional word information
Information gain；
Sequencing unit 10222, for by all target notional words in the target notional word information according to corresponding informance gain
Size sequence chooses the target notional word of the maximum predetermined number of information gain as the validity feature word information.
In one of the embodiments, as shown in fig. 6, the label determining module 106 includes:
Probability calculation unit 1061, in conjunction with the validity feature word information, the target cadence information and same
The sample characteristics word information intersection and sample cadence information intersection of the corresponding all sample songs of standard label, using default classification
Model calculates separately the probability that the target song belongs to each standard label；
Label selection unit 1062, for choosing the corresponding standard label of the target song maximum probability as the mesh
Mark label.
The target cadence information includes mean tempo rate information and rhythm pause duration in one of the embodiments,
Information.
In one of the embodiments, as shown in fig. 6, rhythm acquisition module 104 includes:
Rhythm rate calculation unit 1041, the time for length and target lyrics occupancy in conjunction with the target lyrics are obtained
Take the mean tempo rate information of the target lyrics；
Rhythm pause computing unit 1042 stops for obtaining the rhythm in conjunction with lyrics pause duration in the target lyrics
Immediately long message.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 8.The computer equipment include the processor connected by device bus, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating device, computer program and data
Library.The built-in storage provides environment for the operation of operating device and computer program in non-volatile memory medium.The calculating
The database of machine equipment is used to store song label and determines the data being related to.The network interface of the computer equipment is used for and outside
Terminal by network connection communication.To realize that a kind of song label determines method when the computer program is executed by processor.
In one embodiment, a kind of computer equipment is provided, which can be terminal, internal structure
Figure can be as shown in Figure 9.The computer equipment includes processor, the memory, network interface, display connected by system bus
Screen and input unit.Wherein, the processor of the computer equipment is for providing calculating and control ability.The computer equipment is deposited
Reservoir includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system and computer journey
Sequence.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The network interface of machine equipment is used to communicate with external terminal by network connection.When the computer program is executed by processor with
Realize that a kind of song label determines method.The display screen of the computer equipment can be liquid crystal display or electric ink is shown
Screen, the input unit of the computer equipment can be the touch layer covered on display screen, be also possible on computer equipment shell
Key, trace ball or the Trackpad of setting can also be external keyboard, Trackpad or mouse etc..
It will be understood by those skilled in the art that structure shown in Fig. 8-9, only part relevant to application scheme
The block diagram of structure, does not constitute the restriction for the computer equipment being applied thereon to application scheme, and specific computer is set
Standby may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, is stored in memory
Computer program, the processor perform the steps of the target lyrics according to target song and preset when executing computer program
Song tag database obtains the validity feature word information in the target lyrics；According to the temporal information of the target song and
The target lyrics obtain the target cadence information of the target song；In conjunction with the validity feature word information, the target section
Information and the default song tag database are played, the target labels of the target song are obtained.
The default song tag database stores when processor executes computer program in one of the embodiments,
The sample lyrics, sample cadence information, the corresponding standard label of each sample song, each standard mark of multiple sample songs
Sign the sample characteristics word information and sample cadence information and the corresponding all samples of same standard label of corresponding sample song
The sample characteristics word information intersection and sample cadence information intersection of song.
In one of the embodiments, processor execute computer program when the target lyrics according to target song and
It includes: by the target of the target song that default song tag database, which obtains multiple validity feature words in the target lyrics,
The lyrics split into multiple group of words, extract the target notional word information in the multiple group of words；In conjunction with the target notional word information
The sample characteristics word information and sample cadence information of sample song corresponding with each standard label obtain effective spy
Levy word information.
Target notional word information described in the combination and every when processor executes computer program in one of the embodiments,
The sample characteristics word information and sample cadence information of the corresponding sample song of one standard label obtain the validity feature word
Information includes: the sample characteristics word information in conjunction with the target notional word information and the corresponding sample song of each standard label
The information gain of each target notional word in the target notional word information is obtained with sample cadence information；The target notional word is believed
All target notional words in breath sort according to corresponding informance gain size, and the target for choosing the maximum predetermined number of information gain is real
Word is as the validity feature word information.
Validity feature word information described in combination when processor executes computer program in one of the embodiments,
The target cadence information and the default song tag database, the target labels for obtaining the target song include: knot
Close the sample of the validity feature word information, the target cadence information and the corresponding all sample songs of same standard label
Feature word information intersection and sample cadence information intersection, using default disaggregated model calculate separately the target song belong to it is each
The probability of the standard label；The corresponding standard label of the target song maximum probability is chosen as the target labels.
It includes mean tempo that processor, which executes target cadence information when computer program, in one of the embodiments,
Rate information and rhythm pause duration information.
Processor executes described according to the time of the target song letter when computer program in one of the embodiments,
Ceasing the target cadence information for obtaining the target song with the target lyrics includes: the length and mesh in conjunction with the target lyrics
The time that the mark lyrics occupy obtains the mean tempo rate information of the target lyrics；In conjunction with the lyrics in the target lyrics
Pause duration obtains the rhythm pause duration information.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor to be obtained according to the target lyrics and default song tag database of target song
Take the validity feature word information in the target lyrics；It is obtained according to the temporal information of the target song and the target lyrics
The target cadence information of the target song；In conjunction with the validity feature word information, the target cadence information and described pre-
If song tag database obtains the target labels of the target song.
Default song tag database storage when computer program is executed by processor in one of the embodiments,
There are the sample lyrics, sample cadence information, the corresponding standard label of each sample song, each standard of multiple sample songs
The sample characteristics word information and sample cadence information of the corresponding sample song of label and the corresponding all samples of same standard label
The sample characteristics word information intersection and sample cadence information intersection of this song.
The target lyrics according to target song when computer program is executed by processor in one of the embodiments,
Obtaining multiple validity feature words in the target lyrics with default song tag database includes: by the mesh of the target song
The mark lyrics split into multiple group of words, extract the target notional word information in the multiple group of words；Believe in conjunction with the target notional word
The sample characteristics word information and sample cadence information of breath and the corresponding sample song of each standard label obtain described effective
Feature word information.
When computer program is executed by processor in one of the embodiments, target notional word information described in the combination and
The sample characteristics word information and sample cadence information of the corresponding sample song of each standard label obtain the validity feature
Word information includes: to believe in conjunction with the sample characteristics word of the target notional word information and the corresponding sample song of each standard label
Breath and sample cadence information obtain the information gain of each target notional word in the target notional word information；By the target notional word
All target notional words in information sort according to corresponding informance gain size, choose the target of the maximum predetermined number of information gain
Notional word is as the validity feature word information.
Validity feature word described in the combination is believed when computer program is executed by processor in one of the embodiments,
Breath, the target cadence information and the default song tag database, the target labels for obtaining the target song include:
In conjunction with the sample of the validity feature word information, the target cadence information and the corresponding all sample songs of same standard label
Eigen word information intersection and sample cadence information intersection, calculate separately the target song using default disaggregated model and belong to often
The probability of one standard label；The corresponding standard label of the target song maximum probability is chosen as the target labels.
The target cadence information includes average section when computer program is executed by processor in one of the embodiments,
Play rate information and rhythm pause duration information.
The time according to target song when computer program is executed by processor in one of the embodiments,
Information and the target lyrics obtain the target song target cadence information include: in conjunction with the target lyrics length and
The time that the target lyrics occupy obtains the mean tempo rate information of the target lyrics；In conjunction with being sung in the target lyrics
Word pause duration obtains the rhythm pause duration information.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The embodiments described above only express several embodiments of the present invention, and the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to protection of the invention
Range.Therefore, the scope of protection of the patent of the invention shall be subject to the appended claims.
Claims (10)
1. a kind of song label determines method, which is characterized in that the described method includes:
The validity feature word in the target lyrics is obtained according to the target lyrics of target song and default song tag database
Information；
The target cadence information of the target song is obtained according to the temporal information of the target song and the target lyrics；
In conjunction with the validity feature word information, the target cadence information and the default song tag database, institute is obtained
State the target labels of target song.
2. song label according to claim 1 determines method, which is characterized in that the default song tag database storage
There are the sample lyrics, sample cadence information, the corresponding standard label of each sample song, each mark of multiple sample songs
The sample characteristics word information and sample cadence information of the corresponding sample song of fiducial mark label and same standard label are corresponding all
The sample characteristics word information intersection and sample cadence information intersection of sample song.
3. song label according to claim 2 determines method, which is characterized in that described to be sung according to the target of target song
Word and default song tag database obtain multiple validity feature words in the target lyrics
The target lyrics of the target song are split into multiple group of words, extract the target notional word letter in the multiple group of words
Breath；
In conjunction with the sample characteristics word information and sample of the target notional word information and the corresponding sample song of each standard label
Validity feature word information described in this rhythm acquisition of information.
4. song label according to claim 3 determines method, which is characterized in that target notional word information described in the combination
The sample characteristics word information and sample cadence information of sample song corresponding with each standard label obtain effective spy
Levying word information includes:
In conjunction with the sample characteristics word information and sample of the target notional word information and the corresponding sample song of each standard label
The information gain of each target notional word in target notional word information described in this rhythm acquisition of information；
All target notional words in the target notional word information are sorted according to corresponding informance gain size, choose information gain most
The target notional word of big predetermined number is as the validity feature word information.
5. song label according to claim 2 determines method, which is characterized in that the letter of validity feature word described in the combination
Breath, the target cadence information and the default song tag database, the target labels for obtaining the target song include:
In conjunction with the validity feature word information, the target cadence information and the corresponding all sample songs of same standard label
Sample characteristics word information intersection and sample cadence information intersection, the target song category is calculated separately using default disaggregated model
In the probability of each standard label；
The corresponding standard label of the target song maximum probability is chosen as the target labels.
6. song label according to claim 1 determines method, which is characterized in that the target cadence information includes average
Rhythm rate information and rhythm pause duration information.
7. song label according to claim 6 determines method, which is characterized in that it is described according to the target song when
Between information and the target lyrics obtain the target cadence information of the target song and include:
The mean tempo of the target lyrics is obtained in conjunction with the time that the length and the target lyrics of the target lyrics occupy
Rate information；
The rhythm pause duration information is obtained in conjunction with lyrics pause duration in the target lyrics.
8. a kind of song label determining device, which is characterized in that described device includes:
Feature Words obtain module, for obtaining the target according to the target lyrics of target song and default song tag database
Validity feature word information in the lyrics；
Rhythm obtains module, for obtaining the target song according to the temporal information of the target song and the target lyrics
Target cadence information；
Label determining module, in conjunction with the validity feature word information, the target cadence information and the default song
Tag database obtains the target labels of the target song.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810812529.1A CN109063069A (en) | 2018-07-23 | 2018-07-23 | Song label determines method, apparatus, computer equipment and readable storage medium storing program for executing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810812529.1A CN109063069A (en) | 2018-07-23 | 2018-07-23 | Song label determines method, apparatus, computer equipment and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109063069A true CN109063069A (en) | 2018-12-21 |
Family
ID=64836150
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810812529.1A Pending CN109063069A (en) | 2018-07-23 | 2018-07-23 | Song label determines method, apparatus, computer equipment and readable storage medium storing program for executing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109063069A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110633476A (en) * | 2019-09-27 | 2019-12-31 | 北京百度网讯科技有限公司 | Method and device for acquiring knowledge annotation information |
| CN110727840A (en) * | 2019-09-27 | 2020-01-24 | 浙江大搜车软件技术有限公司 | Vehicle inquiry tag pushing method and device, computer equipment and storage medium |
| CN111026908A (en) * | 2019-12-10 | 2020-04-17 | 腾讯科技（深圳）有限公司 | Song label determination method and device, computer equipment and storage medium |
| CN111859014A (en) * | 2020-06-29 | 2020-10-30 | 维沃移动通信有限公司 | Data labeling method and device |
| CN112163116A (en) * | 2020-09-28 | 2021-01-01 | 广州酷狗计算机科技有限公司 | Song classification method and device and computer readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101963972A (en) * | 2010-07-01 | 2011-02-02 | 深港产学研基地产业发展中心 | Method and system for extracting emotional keywords |
| CN102654859A (en) * | 2011-03-01 | 2012-09-05 | 北京彩云在线技术开发有限公司 | Method and system for recommending songs |
| CN103488782A (en) * | 2013-09-30 | 2014-01-01 | 华北电力大学 | Method for recognizing musical emotion through lyrics |
| CN103678274A (en) * | 2013-04-15 | 2014-03-26 | 南京邮电大学 | Feature extraction method for text categorization based on improved mutual information and entropy |
| CN105404674A (en) * | 2015-11-20 | 2016-03-16 | 焦点科技股份有限公司 | Knowledge-dependent webpage information extraction method |
| CN105868372A (en) * | 2016-03-31 | 2016-08-17 | 广州酷狗计算机科技有限公司 | Label distribution method and device |
-
2018
- 2018-07-23 CN CN201810812529.1A patent/CN109063069A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101963972A (en) * | 2010-07-01 | 2011-02-02 | 深港产学研基地产业发展中心 | Method and system for extracting emotional keywords |
| CN102654859A (en) * | 2011-03-01 | 2012-09-05 | 北京彩云在线技术开发有限公司 | Method and system for recommending songs |
| CN103678274A (en) * | 2013-04-15 | 2014-03-26 | 南京邮电大学 | Feature extraction method for text categorization based on improved mutual information and entropy |
| CN103488782A (en) * | 2013-09-30 | 2014-01-01 | 华北电力大学 | Method for recognizing musical emotion through lyrics |
| CN105404674A (en) * | 2015-11-20 | 2016-03-16 | 焦点科技股份有限公司 | Knowledge-dependent webpage information extraction method |
| CN105868372A (en) * | 2016-03-31 | 2016-08-17 | 广州酷狗计算机科技有限公司 | Label distribution method and device |
Non-Patent Citations (1)
| Title |
|---|
| 郑亚斌等: ""中文歌词的统计特征及其检索应用"", 《中文信息学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110633476A (en) * | 2019-09-27 | 2019-12-31 | 北京百度网讯科技有限公司 | Method and device for acquiring knowledge annotation information |
| CN110727840A (en) * | 2019-09-27 | 2020-01-24 | 浙江大搜车软件技术有限公司 | Vehicle inquiry tag pushing method and device, computer equipment and storage medium |
| CN110727840B (en) * | 2019-09-27 | 2022-07-05 | 浙江大搜车软件技术有限公司 | Vehicle inquiry tag pushing method and device, computer equipment and storage medium |
| CN110633476B (en) * | 2019-09-27 | 2024-04-05 | 北京百度网讯科技有限公司 | Method and device for acquiring knowledge annotation information |
| CN111026908A (en) * | 2019-12-10 | 2020-04-17 | 腾讯科技（深圳）有限公司 | Song label determination method and device, computer equipment and storage medium |
| CN111026908B (en) * | 2019-12-10 | 2023-09-08 | 腾讯科技（深圳）有限公司 | Song label determining method, device, computer equipment and storage medium |
| CN111859014A (en) * | 2020-06-29 | 2020-10-30 | 维沃移动通信有限公司 | Data labeling method and device |
| CN112163116A (en) * | 2020-09-28 | 2021-01-01 | 广州酷狗计算机科技有限公司 | Song classification method and device and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109063069A (en) | Song label determines method, apparatus, computer equipment and readable storage medium storing program for executing | |
| Benchimol et al. | Text mining methodologies with R: An application to central bank texts | |
| Lai et al. | Exploring the research in information technology implementation | |
| CN109002492B (en) | Performance point prediction method based on LightGBM | |
| CN110362823A (en) | The training method and device of text generation model are described | |
| CN108959271A (en) | Document creation method, device, computer equipment and readable storage medium storing program for executing are described | |
| CN107038173A (en) | Application query method and apparatus, similar application detection method and device | |
| CN107194430A (en) | A kind of screening sample method and device, electronic equipment | |
| CN110457677B (en) | Entity relationship identification method and device, storage medium and computer equipment | |
| CN107818491A (en) | Electronic installation, Products Show method and storage medium based on user's Internet data | |
| CN106874314A (en) | The method and apparatus of information recommendation | |
| CN110472049B (en) | Disease screening text classification method, computer device and readable storage medium | |
| CN110276382A (en) | Listener clustering method, apparatus and medium based on spectral clustering | |
| US20070136220A1 (en) | Apparatus for learning classification model and method and program thereof | |
| CN109858009A (en) | Device, method and its computer storage medium of control instruction are generated according to text | |
| CN106294128B (en) | A kind of automated testing method and device exporting report data | |
| CN105786898B (en) | A kind of construction method and device of domain body | |
| CN110111143A (en) | A kind of control method and control device for establishing mobile end subscriber portrait | |
| CN111159167B (en) | Labeling quality detection device and method | |
| CN109783740A (en) | Pay close attention to the sort method and device of the page | |
| CN109614982A (en) | Product analysis method, apparatus, computer equipment and storage medium | |
| CN108647714A (en) | Acquisition methods, terminal device and the medium of negative label weight | |
| CN114692889A (en) | Meta-feature training model for machine learning algorithm | |
| De Crombrugghe et al. | Statistical Demand Functions for Food in the USA and the Netherlands | |
| CN110134935A (en) | A kind of method, device and equipment for extracting font style characteristic |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181221 |