CN107977374A - Bent storehouse optimization method and device - Google Patents
Bent storehouse optimization method and device Download PDFInfo
- Publication number
- CN107977374A CN107977374A CN201610921779.XA CN201610921779A CN107977374A CN 107977374 A CN107977374 A CN 107977374A CN 201610921779 A CN201610921779 A CN 201610921779A CN 107977374 A CN107977374 A CN 107977374A
- Authority
- CN
- China
- Prior art keywords
- song
- user
- current
- current song
- mrow
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/60—Information retrieval; Database structures therefor; File system structures therefor of audio data
- G06F16/61—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/60—Information retrieval; Database structures therefor; File system structures therefor of audio data
- G06F16/68—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/686—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, title or artist information, time, location or usage information, user ratings
Abstract
The present invention relates to bent storehouse optimization method and device.The described method includes：Song behavioural information is listened for user, polled data information is determined according to preset code of points；According to polled data information, using Wilson's confidential interval model, to obtain the scope of every song confidential interval and the lower limit of Wilson's confidential interval that song behavioural information is listened for user；If the lower limit of the Wilson's confidential interval of current song is more than or equal to the first threshold of pre-set march storehouse optimization, current song is denoted as singing, and current song is chosen to library to optimize library.Wilson's confidential interval model during bent storehouse optimization is applied to by the embodiment of the present invention, realize the automation optimization of Qu Ku, improve the efficiency screened and sung, further, the diversification of song behavioural information is listened based on user, and data volume it is big the characteristics of so that the library of optimization is more accurate, meets the demand of user.
Description
Technical field
The present invention relates to Audiotechnica field, and specifically, the present invention relates to bent storehouse optimization method and device.
Background technology
The existing optimization method to library is mainly based upon the good song of artificial screening.The method for having screened song is specific
For：Audited by human-edited's audition, according to each editor's hobby of oneself, filter out and sing.
First method is specially：Editor carries out song screening according to song temperature.Screening master is carried out based on song temperature
If consider the factor such as playback volume, issuing date.The shortcomings that carrying out song screening based on song temperature is the good song number filtered out
It is also not comprehensive enough according to not accurate enough.For example, since the renewal of library is slow, sung as new, and the song that unexpected winner is pleasing to the ear, often
It is ignored.
Second method is specially：It is so different based on everyone hobby by manually into edlin, easily by individual
Subjectivity like being incorporated into the screening in song storehouse.In addition, the inefficiency screened completely by human-edited, does not also introduce
Eliminative mechanism so that the data volume of library is excessively huge, it is necessary to expend too many man power and material.Further, filter out
Good song data it is not accurate enough, it is also not comprehensive enough.Often this editor thinks pleasing to the ear song, another editor is considered as not
It is pleasing to the ear.
At present, for digital music service provider, bent storehouse optimization is mainly artificial optimization Qu Ku.But artificial optimization
Qu Ku not only needs to consume substantial amounts of human cost, but also inefficiency, and the accuracy rate of identification is low, it is difficult to the good song of optimization
The quality in storehouse carries out quality monitoring, therefore, it is necessary to a kind of automatic bent storehouse optimization method, realizes the Automatic Optimal to Qu Ku.
The content of the invention
The embodiment of the present invention is to provide bent storehouse optimization method and device, according to polled data information, is put using Wilson's
Interval model is believed, to obtain scope and Wilson's confidence area that the every song confidential interval for singing behavioural information is listened for user
Between lower limit；Compare the lower limit and the first of the optimization of pre-set march storehouse of the Wilson's confidential interval of current song
Threshold value, if the lower limit of the Wilson's confidential interval of current song is more than or equal to the first threshold of pre-set march storehouse optimization
Value, then be denoted as singing by current song, and current song is chosen to library to be optimized to library, it is achieved thereby that
The automation optimization of Qu Ku, improves the efficiency screened and sung.
In a first aspect, an embodiment of the present invention provides bent storehouse optimization method, the described method includes：
The daily record that the user in the scheduled time plays song is extracted, listens an old song form as letter to get user from the daily record
Breath；
Song behavioural information is listened for the user, polled data information is determined according to preset code of points；
According to the polled data information, using Wilson's confidential interval model, an old song form is listened for the user to obtain
For the scope of every song confidential interval and the lower limit of Wilson's confidential interval of information；
Compare the lower limit and the first threshold of pre-set march storehouse optimization of the Wilson's confidential interval of current song
Value, if the lower limit of the Wilson's confidential interval of current song is more than or equal to the first threshold of pre-set march storehouse optimization
Value, then be denoted as singing, and current song is chosen to library to optimize the library by current song.
Preferably, the user listens an old song form to specifically include following at least one operation：
User clicks on the operation of current song, user downloads the operation of current song, user search for current song operation,
User performs current song the point operation of red heart, user uploads the operation of current song, user collect current song operation,
User completely plays the operation of current song and user plays the operation that midway switches current song.
Preferably, the formula tool of the lower limit of the corresponding Wilson's confidential interval of the Wilson's confidential interval model is calculated
Body is：
Wherein,To listen the ratio of the affirmative vote of song behavioural information and ballot sum for the user, n is for described
User listens the ballot sum of song behavioural information, Z1-α/2To listen a certain confidence level of correspondence of song behavioural information for the user
Z statistics.
Preferably, the Z statistics that a certain confidence level of correspondence of song behavioural information is listened for the user are 13.
Preferably, the preset code of points specifically includes following at least one：
It is the scoring to current song made to listen an old song form for the user, and the operation of song is included for user
The scoring to current song made.
Preferably, it is that the scoring to current song made specifically includes following at least one to listen an old song form for the user
Kind：
It is positive integer 1 that user, which completely listens to the scoring that current song is denoted as to current song, and user repeatedly listens to current song
Scorings of the Qu Jizuo to current song is positive integer 3, and it is negative integer that user, which skips the scoring that current song is denoted as to current song,
1, it is negative integer 3 that user, which repeatedly skips the scoring that current song is denoted as to current song, and the multiple single of user circulates current song
It is positive integer 3 to be denoted as the scoring to current song.
Preferably, the scoring to current song made for song being included for user specifically includes following at least one
Kind：
The scoring that the operation that user performs collection current song is denoted as to current song is positive integer 1, and user record is liked
The scoring that the operation of current song is denoted as to current song is positive integer 2, and the operation that user performs comment current song is denoted as pair
The scoring of current song is positive integer 1, and the operation that user record is liked singing the singer of current song is denoted as to current song
Score as positive integer 1.
Preferably, if the user, which listens an old song form to go out two songs for presentation of information, has same song name and same singer,
The high song that currently scores then is chosen to library to be optimized to the library.
Preferably, it is less than pre-set march storehouse in the lower limit for judging the Wilson's confidential interval of current song
After the first threshold of optimization, the method further includes：Pass through the second threshold pair of pre-set march storehouse double optimization
Current song carry out programmed screening, if the lower limit of the Wilson's confidential interval of current song be more than or equal to it is pre-set into
The second threshold of row song storehouse optimization, then choose current song to treating selected library to be optimized to the library,
Wherein, the first threshold of the bent storehouse optimization is more than the second threshold of the bent storehouse optimization.
Second aspect, an embodiment of the present invention provides bent storehouse optimization method device, described device includes：
It is data obtaining module to listen an old song form, the daily record that the user in the scheduled time plays song is extracted, with from the daily record
In get user and listen song behavioural information；
Polled data information determination module, listens song behavioural information for the user, is determined according to preset code of points
Polled data information；
Model application module, according to the polled data information, using Wilson's confidential interval model, institute is directed to obtain
State the scope of every song confidential interval and the lower limit of Wilson's confidential interval that user listens song behavioural information；
Comparison module, the lower limit for comparing the Wilson's confidential interval of current song optimize with pre-set march storehouse
First threshold, if the lower limit of the Wilson's confidential interval of current song be more than or equal to pre-set march storehouse optimization
First threshold, then be denoted as singing by current song, and current song is chosen to library to be optimized to the library.
An embodiment of the present invention provides bent storehouse optimization method, extracts the daily record that the user in the scheduled time plays song, with
User is got from daily record and listens song behavioural information；Song behavioural information is listened for user, determines to throw according to preset code of points
Ticket data information；According to polled data information, using Wilson's confidential interval model, song behavioural information is listened for user to obtain
Scope per song confidential interval and Wilson's confidential interval lower limit；Compare the Wilson's confidence area of current song
Between the optimization of lower limit and pre-set march storehouse first threshold, if the lower limit of the Wilson's confidential interval of current song
Value is more than or equal to the first threshold of pre-set march storehouse optimization, then is denoted as singing by current song, and by current song
Choose to library to be optimized to library, it is achieved thereby that the automation optimization of Qu Ku, improves the effect screened and sung
Rate；Further, the diversification of song behavioural information is listened based on user, and the characteristics of data volume is big so that the library of optimization
More precisely, the demand of user is met.The embodiment of the present invention is according to polled data information, using Wilson's confidential interval model,
To obtain the scope of every song confidential interval and the lower limit of Wilson's confidential interval that song behavioural information is listened for user,
It is achieved thereby that the automation optimization of Qu Ku.
Brief description of the drawings
Fig. 1 is bent storehouse optimization method flow chart provided in an embodiment of the present invention；
Fig. 2 is bent storehouse optimization apparatus structure schematic diagram provided in an embodiment of the present invention.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, the technical solution in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
Part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
All other embodiments obtained without making creative work, belong to the scope of protection of the invention.
For ease of the understanding to the embodiment of the present invention, it is further explained below in conjunction with attached drawing with specific embodiment
It is bright.
In technical solution provided by the present invention, according to polled data information, using Wilson's confidential interval model, to obtain
Take the scope of every song confidential interval and the lower limit of Wilson's confidential interval that song behavioural information is listened for user；Compare
The lower limit of the Wilson's confidential interval of current song and the first threshold of pre-set march storehouse optimization, if current song
Wilson's confidential interval lower limit be more than or equal to pre-set march storehouse optimization first threshold, then by current song
It is denoted as singing, and current song is chosen to library to be optimized to library, it is achieved thereby that the automation of Qu Ku is excellent
Change, improve the efficiency screened and sung.
The technical solution that the invention will now be described in detail with reference to the accompanying drawings.
Song storehouse provided in an embodiment of the present invention optimization method flow chart, as shown in Figure 1, bent storehouse optimization method includes following step
Suddenly：
S101：Extract the scheduled time in user play song daily record, using got from daily record user listen an old song form as
Information.
In a specific embodiment of the present invention, user listens an old song form to specifically include following at least one operation：User clicks on
The operation of current song is downloaded in the operation of current song, user, user searches for the operation of current song, user holds current song
The operation of row red heart, user upload the operation of current song, user collects the operation of current song, user completely plays currently
The operation of song and user play the operation that midway switches current song.By aforesaid operations, user can be obtained and listen an old song form
For information.
Except through aforesaid operations, user can be obtained and listened outside song behavioural information, in another specific implementation of the present invention
In example, user listens song behavioural information also to specifically include following at least one information：User listen the path of song, user listen song the time with
And user listens song duration.
S102：Song behavioural information is listened for user, polled data information is determined according to preset code of points.
In a specific embodiment of the present invention, preset code of points includes：An old song form is listened to go out use for presentation of information user
Family completely listens song to be denoted as voting for the song, listens an old song form to go out user for presentation of information user and plays midway switching song note
The song of opposing is voted against.
S103：According to polled data information, using Wilson's confidential interval model, an old song form is listened as letter for user to obtain
The scope of every song confidential interval and the lower limit of Wilson's confidential interval of breath.
In a specific embodiment of the present invention, calculate under the corresponding Wilson's confidential interval of Wilson's confidential interval model
The formula of limit value is specially：
Wherein,To listen the ratio of the affirmative vote of song behavioural information and ballot sum for user, n is to listen song for user
The ballot sum of behavioural information, Z1-α/2To listen the Z statistics of a certain confidence level of correspondence of song behavioural information for user.
It should be noted that in one particular embodiment of the present invention, the correspondence of song behavioural information is listened for user
The Z statistics of one confidence level are 13.
S104：Compare the of the lower limit of the Wilson's confidential interval of current song and the optimization of pre-set march storehouse
One threshold value, if the lower limit of the Wilson's confidential interval of current song is more than or equal to the first of pre-set march storehouse optimization
Threshold value, then be denoted as singing by current song, and current song is chosen to library to be optimized to library.
It should be noted that in one particular embodiment of the present invention, when the correspondence that song behavioural information is listened for user
The Z statistics of a certain confidence level be 13 when, at this time, corresponding to for the Z statistics of the confidence level it is pre-set into
The first threshold of row song storehouse optimization is specially 0.6.
In concrete application scene, compare the lower limit of the Wilson's confidential interval of current song and pre-set progress
The first threshold 0.6 of bent storehouse optimization, if the lower limit of the Wilson's confidential interval of current song be more than or equal to it is pre-set into
The first threshold 0.6 of row song storehouse optimization, then be denoted as singing, and current song is chosen to library to come to song by current song
Storehouse optimizes.
, will " scope and Wilson's confidence area per song confidential interval in the specific embodiment of the invention described above
Between lower limit " as weighing the standard sung.
In the another specific embodiment of the present invention, will the scoring score value of song " often " as weighing the standard sung,
So so that scoring to song more standardized, using user's operation, the Experience Degree of user is improved.
Specifically, if user, which listens an old song form to go out two songs for presentation of information, has same song name and same singer,
Selection currently scores high song to library to be optimized to library.
In the another specific embodiment of the present invention, preset code of points specifically includes following at least one：For with
It is the scoring to current song made that an old song form is listened at family, and for user include song make to current song
Scoring.
Further, it is that the scoring to current song made specifically includes following at least one to listen an old song form for user：
It is positive integer 1 that user, which completely listens to the scoring that current song is denoted as to current song, and user repeatedly listens to current song and is denoted as pair
The scoring of current song is positive integer 3, and it is negative integer 1, Yong Huduo that user, which skips the scoring that current song is denoted as to current song,
Secondary to skip the scoring that current song is denoted as to current song be negative integer 3, and the multiple single circulation current song of user is denoted as to working as
The scoring of preceding song is positive integer 3.
Further, for user include song the scoring to current song made specifically include such as down toward
Few one kind：The scoring that the operation that user performs collection current song is denoted as to current song is positive integer 1, and user record is liked working as
The scoring that the operation of preceding song is denoted as to current song is positive integer 2, and the operation that user performs comment current song is denoted as to working as
The scoring of preceding song is positive integer 1, and the operation that user record is liked singing the singer of current song is denoted as commenting current song
It is divided into positive integer 1.
In practical application scene, there is same song name when user listens an old song form to go out current two song for presentation of information
When, if the scoring score value of First song is 12, the scoring score value of the second song is 15, then chooses the second song to song
Qu Ku is to optimize library.
Further, when user listen an old song form for presentation of information to go out current two song be the song of same singer when, if together
The scoring score value of the First song of one singer is 8, and the scoring score value of the second song of same singer is 12, then will be same
The second song of singer is chosen to library to be optimized to library.
Further, there is same song name when user listens an old song form to go out current two song for presentation of information, and is same
During the song of one singer, if First song is live version, the scoring score value of First song is 12, and the second song is
Lossless tonequality version, the scoring score value of the second song is 15, then chooses the second song to library with to library
Optimize.
It should be noted that in a specific embodiment of the present invention, in order to ensure the huge data volume of library, further
Introducing remaining and evaluate whether it is the good standard " carry out postsearch screening " sung, so that the library of optimization is more accurate,
And it is more in line with the demand of user.
Specifically, it is less than pre-set march in the lower limit for judging the Wilson's confidential interval of current song
After the first threshold of storehouse optimization, the method further includes：Pass through the second threshold of pre-set march storehouse double optimization
Programmed screening is carried out to current song, if the lower limit of the Wilson's confidential interval of current song is more than or equal to pre-set
The second threshold of march storehouse optimization, then choose current song to treating selected library to be optimized to library, its
In, the first threshold of bent storehouse optimization is more than the second threshold of bent storehouse optimization.By carrying out double optimization to bent storehouse, singing in antiphonal style is realized
The automatic Iterative mechanism of Qu Ku.
In a specific embodiment of the present invention, multiple automatic Iterative can carry out library according to the demand of user, it is more
Secondary screening, until reaching the demand of user, details are not described herein.
In conclusion song storehouse provided in an embodiment of the present invention optimization method, according to polled data information, puts using Wilson's
Interval model is believed, to obtain scope and Wilson's confidence area that the every song confidential interval for singing behavioural information is listened for user
Between lower limit；Compare the lower limit and the first of the optimization of pre-set march storehouse of the Wilson's confidential interval of current song
Threshold value, if the lower limit of the Wilson's confidential interval of current song is more than or equal to the first threshold of pre-set march storehouse optimization
Value, then be denoted as singing by current song, and current song is chosen to library to be optimized to library, it is achieved thereby that
The automation optimization of Qu Ku, improves the efficiency screened and sung.Further, the diversification of song behavioural information is listened based on user,
And data volume it is big the characteristics of so that the library of optimization is more accurate, meets the demand of user.
As shown in Fig. 2, it is the internal structure block diagram of bent storehouse optimization device provided in an embodiment of the present invention；As shown in Fig. 2, this
The bent storehouse optimization device that inventive embodiments are provided, including：An old song form is listened to be determined for data obtaining module 201, polled data information
Module 202, model application module 203 and comparison module 204.
Specifically, it is data obtaining module to listen an old song form, extract the scheduled time in user play song daily record, with from
User is got in daily record and listens song behavioural information.
Wherein, it is that the user that data obtaining module is got listens an old song form to specifically include following at least one behaviour to listen an old song form
Make：User clicks on the operation of current song, user downloads the operation of current song, user searches for the operation of current song, user
The operation of point red heart is performed to current song, user uploads the operation of current song, user collects the operation of current song, user
The complete operation for playing current song and user play the operation that midway switches current song.
Polled data information determination module, listens song behavioural information for user, determines to vote according to preset code of points
Data message.
In a specific embodiment of the present invention, preset code of points includes：An old song form is listened to go out use for presentation of information user
Family completely listens song to be denoted as voting for the song, listens an old song form to go out user for presentation of information user and plays midway switching song note
The song of opposing is voted against.
Model application module, according to polled data information, using Wilson's confidential interval model, is listened with obtaining for user
Sing the scope of every song confidential interval and the lower limit of Wilson's confidential interval of behavioural information.
It should be noted that model application module calculates the corresponding Wilson's confidential interval of Wilson's confidential interval model
The formula of lower limit is specially：
Wherein,To listen the ratio of the affirmative vote of song behavioural information and ballot sum for user, n is to listen song for user
The ballot sum of behavioural information, Z1-α/2To listen the Z statistics of a certain confidence level of correspondence of song behavioural information for user.
Further, in a specific embodiment of the invention, a certain confidence of correspondence of song behavioural information is listened for user
Horizontal Z statistics are 13.
Comparison module, the lower limit for comparing the Wilson's confidential interval of current song optimize with pre-set march storehouse
First threshold, if the lower limit of the Wilson's confidential interval of current song be more than or equal to pre-set march storehouse optimization
First threshold, then be denoted as singing by current song, and current song is chosen to library to be optimized to library.
Further, in order to ensure the accuracy of the library of optimization, it is that can cover use to listen an old song form according to user
The information that family listens song to be accustomed to and listen song to like is used to optimize library, obtains the library of optimization.
But since the data volume of the library of optimization is excessively huge, it is unfavorable for user and selects oneself liking
Song, therefore bent storehouse optimization device provided in an embodiment of the present invention further includes screening module (not marking in fig. 2).Pass through sieve
Modeling block, can carry out further automatic screening, to obtain the accurately song for meeting user demand to the library of optimization
Qu Ku.
In a specific embodiment of the present invention, screening module is used for, if user listens an old song form to go out two songs for presentation of information
With same song name and same singer, then the high song that currently scores is chosen to library to be optimized to library.
Specifically, the preset code of points that polled data information determination module is applied specifically includes following at least one
Kind：It is the scoring to current song made to listen an old song form for user, and pair made of song is included for user
The scoring of current song.
Further, it is that the scoring to current song made specifically includes following at least one to listen an old song form for user：
It is positive integer 1 that user, which completely listens to the scoring that current song is denoted as to current song, and user repeatedly listens to current song and is denoted as pair
The scoring of current song is positive integer 3, and it is negative integer 1, Yong Huduo that user, which skips the scoring that current song is denoted as to current song,
Secondary to skip the scoring that current song is denoted as to current song be negative integer 3, and the multiple single circulation current song of user is denoted as to working as
The scoring of preceding song is positive integer 3.
Further, the scoring to current song made that song is included for user specifically includes as follows at least
It is a kind of：The scoring that the operation that user performs collection current song is denoted as to current song is positive integer 1, and user record likes current
The scoring that the operation of song is denoted as to current song is positive integer 2, and the operation that user performs comment current song is denoted as to current
The scoring of song is positive integer 1, and user record likes the operation for singing the singer of current song to be denoted as the scoring to current song
For positive integer 1.
In the another specific embodiment of the present invention, the Wilson's confidential interval of current song is judged in comparison module
Lower limit is less than after the first threshold of pre-set march storehouse optimization, and screening module is additionally operable to, by pre-set
The second threshold of march storehouse double optimization carries out programmed screening to current song, if the Wilson's confidential interval of current song
Lower limit be more than or equal to the second threshold of pre-set march storehouse optimization, then current song is chosen to treating selected song
Qu Ku to be optimized to library, wherein, the first threshold of bent storehouse optimization is more than the second threshold of bent storehouse optimization.
In technical scheme, according to polled data information, using Wilson's confidential interval model, it is directed to obtaining
User listens the scope of every song confidential interval and the lower limit of Wilson's confidential interval of song behavioural information；Relatively more current song
The lower limit of bent Wilson's confidential interval and the first threshold of pre-set march storehouse optimization, if the Weir of current song
The lower limit of inferior confidential interval is more than or equal to the first threshold of pre-set march storehouse optimization, then is denoted as current song
Song, and current song is chosen to library to be optimized to library, it is achieved thereby that the automation optimization of Qu Ku, is improved
Screen the efficiency sung.
Above-described embodiment, has carried out the purpose of the present invention, technical solution and beneficial effect further
Describe in detail, it should be understood that the foregoing is merely the embodiment of the present invention, be not intended to limit the present invention
Protection domain, within the spirit and principles of the invention, any modification, equivalent substitution, improvement and etc. done, should all include
Within protection scope of the present invention.
Claims (10)
1. bent storehouse optimization method, it is characterised in that including：
The daily record that the user in the scheduled time plays song is extracted, listens song behavioural information to get user from the daily record；
Song behavioural information is listened for the user, polled data information is determined according to preset code of points；
According to the polled data information, using Wilson's confidential interval model, an old song form is listened as letter for the user to obtain
The scope of every song confidential interval and the lower limit of Wilson's confidential interval of breath；
Compare the lower limit and the first threshold of pre-set march storehouse optimization of the Wilson's confidential interval of current song, if
The lower limit of the Wilson's confidential interval of current song is more than or equal to the first threshold of pre-set march storehouse optimization, then will
Current song is denoted as singing, and current song is chosen to library to be optimized to the library.
2. according to the method described in claim 1, it is characterized in that, the user listens an old song form to specifically include following at least one
Operation：
User clicks on the operation of current song, user downloads the operation of current song, user searches for the operation of current song, user
The operation of point red heart is performed to current song, user uploads the operation of current song, user collects the operation of current song, user
The complete operation for playing current song and user play the operation that midway switches current song.
3. according to the method described in claim 1, it is characterized in that, calculate the corresponding Weir of the Wilson's confidential interval model
The formula of the lower limit of inferior confidential interval is specially：
<mrow>
<mfrac>
<mrow>
<mover>
<mi>p</mi>
<mo>^</mo>
</mover>
<mo>+</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>2</mn>
<mi>n</mi>
</mrow>
</mfrac>
<msub>
<msup>
<mi>Z</mi>
<mn>2</mn>
</msup>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msub>
<mo>&PlusMinus;</mo>
<msub>
<mi>Z</mi>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msub>
<msqrt>
<mrow>
<mfrac>
<mrow>
<mover>
<mi>p</mi>
<mo>^</mo>
</mover>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mover>
<mi>p</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
</mrow>
<mi>n</mi>
</mfrac>
<mo>+</mo>
<mfrac>
<mrow>
<msub>
<msup>
<mi>Z</mi>
<mn>2</mn>
</msup>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msub>
</mrow>
<mrow>
<mn>4</mn>
<msup>
<mi>n</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
</mrow>
</msqrt>
</mrow>
<mrow>
<mn>1</mn>
<mo>+</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<msub>
<msup>
<mi>Z</mi>
<mn>2</mn>
</msup>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
Wherein,To listen the ratio of the affirmative vote of song behavioural information and ballot sum for the user, n is for the user
Listen the ballot sum of song behavioural information, Z1-α/2To listen the Z of a certain confidence level of correspondence of song behavioural information to unite for the user
Metering.
4. according to the method described in claim 3, it is characterized in that, listen the correspondence of song behavioural information is a certain to put for the user
The horizontal Z statistics of letter are 13.
5. according to the method described in claim 3, it is characterized in that, the preset code of points specifically includes following at least one
Kind：
It is the scoring to current song made to listen an old song form for the user, and includes for user the operation of song and make
The scoring to current song.
6. according to the method described in claim 5, it is characterized in that, it is to make to current song to listen an old song form for the user
Scoring specifically include following at least one：
It is positive integer 1 that user, which completely listens to the scoring that current song is denoted as to current song, and user repeatedly listens to current song note
The scoring made to current song is positive integer 3, and it is negative integer 1 that user, which skips the scoring that current song is denoted as to current song, is used
It is negative integer 3 that the scoring that current song is denoted as to current song is repeatedly skipped at family, and the multiple single circulation current song of user is denoted as
Scoring to current song is positive integer 3.
7. according to the method described in claim 5, it is characterized in that, for user include song make to current song
Bent scoring specifically includes following at least one：
The scoring that the operation that user performs collection current song is denoted as to current song is positive integer 1, and user record likes current
The scoring that the operation of song is denoted as to current song is positive integer 2, and the operation that user performs comment current song is denoted as to current
The scoring of song is positive integer 1, and user record likes the operation for singing the singer of current song to be denoted as the scoring to current song
For positive integer 1.
8. the method according to claim 6 or 7, it is characterised in that if the user listens an old song form to go out two head for presentation of information
Song has same song name and same singer, then chooses the high song that currently scores to library to be carried out to the library
Optimization.
9. according to the method described in claim 1, it is characterized in that, in the case where judging the Wilson's confidential interval of current song
Limit value is less than after the first threshold of pre-set march storehouse optimization, and the method further includes：By it is pre-set into
The second threshold of row song storehouse double optimization carries out programmed screening to current song, if the Wilson's confidential interval of current song
Lower limit is more than or equal to the second threshold of pre-set march storehouse optimization, then chooses current song to treating selected song
Storehouse to be optimized to the library, wherein, the first threshold of song storehouse optimization is more than the second threshold of the bent storehouse optimization
Value.
10. bent storehouse optimization method device, it is characterised in that including：
It is data obtaining module to listen an old song form, the daily record that the user in the scheduled time plays song is extracted, to be obtained from the daily record
Get user and listen song behavioural information；
Polled data information determination module, listens song behavioural information for the user, determines to vote according to preset code of points
Data message；
Model application module, according to the polled data information, using Wilson's confidential interval model, the use is directed to obtain
Listen the scope of every song confidential interval and the lower limit of Wilson's confidential interval of song behavioural information in family；
Comparison module, compares the of lower limit and pre-set march the storehouse optimization of the Wilson's confidential interval of current song
One threshold value, if the lower limit of the Wilson's confidential interval of current song is more than or equal to the first of pre-set march storehouse optimization
Threshold value, then be denoted as singing by current song, and current song is chosen to library to be optimized to the library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610921779.XA CN107977374A (en) | 2016-10-21 | 2016-10-21 | Bent storehouse optimization method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610921779.XA CN107977374A (en) | 2016-10-21 | 2016-10-21 | Bent storehouse optimization method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107977374A true CN107977374A (en) | 2018-05-01 |
Family
ID=62004600
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610921779.XA Pending CN107977374A (en) | 2016-10-21 | 2016-10-21 | Bent storehouse optimization method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107977374A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109784666A (en) * | 2018-12-20 | 2019-05-21 | 国网北京市电力公司 | The detection method and device of equipment quality |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101719167A (en) * | 2010-01-15 | 2010-06-02 | 北京暴风网际科技有限公司 | Interactive movie searching method |
| CN103514172A (en) * | 2012-06-20 | 2014-01-15 | 同程网络科技股份有限公司 | Method for downloading words for setting search engine keywords |
| CN105163183A (en) * | 2015-08-28 | 2015-12-16 | 北京奇艺世纪科技有限公司 | Screening method and device for video pictures |
| 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 |
-
2016
- 2016-10-21 CN CN201610921779.XA patent/CN107977374A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101719167A (en) * | 2010-01-15 | 2010-06-02 | 北京暴风网际科技有限公司 | Interactive movie searching method |
| CN103514172A (en) * | 2012-06-20 | 2014-01-15 | 同程网络科技股份有限公司 | Method for downloading words for setting search engine keywords |
| CN105163183A (en) * | 2015-08-28 | 2015-12-16 | 北京奇艺世纪科技有限公司 | Screening method and device for video pictures |
| 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 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109784666A (en) * | 2018-12-20 | 2019-05-21 | 国网北京市电力公司 | The detection method and device of equipment quality |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107820138B (en) | Video broadcasting method, device, terminal and storage medium | |
| CN102467939B (en) | Song audio frequency cutting apparatus and method thereof | |
| CN108509534A (en) | Individualized music commending system based on deep learning and its implementation | |
| CN104123325B (en) | The recommendation method and recommendation server of multimedia file | |
| CN102752540B (en) | A kind of automated cataloging method based on face recognition technology | |
| CN106448630A (en) | Method and device for generating digital music file of song | |
| CN101504834B (en) | Humming type rhythm identification method based on hidden Markov model | |
| Wall | Finding an alternative: Music programming in US college radio | |
| CN101667422B (en) | Method and device for adjusting mode of song accompaniment | |
| CN107977373A (en) | A kind of recommendation method of song | |
| CN104991900A (en) | Method and apparatus for pushing music data | |
| CN107885745A (en) | A kind of song recommendations method and device | |
| CN109271550A (en) | A kind of music personalization classification recommended method based on deep learning | |
| CN101996627A (en) | Speech processing apparatus, speech processing method and program | |
| CN106228978A (en) | A kind of audio recognition method | |
| CN106791983A (en) | A kind of intelligent television user behavior analysis method and system | |
| CN109376265A (en) | Song recommendations list generation method, medium, device and calculating equipment | |
| CN107967280A (en) | A kind of method and system of label recommendations song | |
| CN101000681A (en) | Programme test system and control method thereof | |
| CN110175291A (en) | Hand trip recommended method, storage medium, equipment and system based on similarity calculation | |
| CN109214446A (en) | Potentiality good performance personnel kind identification method, system, terminal and computer readable storage medium | |
| CN107566856A (en) | The method and system of pushed information are provided | |
| CN109657144A (en) | Methods of marking, device, storage medium and the terminal device of works | |
| CN103955490A (en) | Audio playing method and audio playing equipment | |
| CN106168958A (en) | A kind of recommendation method of audio-frequency information and server |
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 |
Application publication date: 20180501 |
|
| RJ01 | Rejection of invention patent application after publication |