CN104504124A - Method for presenting entity popularity through video searching and playing behaviors - Google Patents

Abstract

The invention discloses a method for presenting entity popularity through video searching and playing behaviors. The method for presenting the entity popularity through the video searching and playing behaviors includes that selecting a certain sample time period, and selecting a plurality of sample entities with the same type to serve as sample objects; selecting a plurality of typical indexes which may represent the sample object popularity; confirming a research time for each sample object, calculating the standard deviation of each typical index in the research time, and averaging a plurality of standard deviations of the same typical index so as to respectively average the standard deviations; using the average standard deviations of different typical indexes to analyze weighted values of different typical indexes through an AHP analytic hierarchy process; multiplying the single-day standard value of each index of a sample by the weighted value to obtain the actual popularity weight of the index on that day, adding all the actual popularity weights to obtain the popularity of the sample on that day, and calculating the popularity of each entity on each day to obtain the popularity ranking list of the entities. The method for presenting the entity popularity through the video searching and playing behaviors realizes the quantitative showing of the top ranking list of all the entities.

Description

The method of entity temperature is gone out by video search and broadcasting behavior expression

Technical field

The application relates to video playback field, concrete, relates to the method being gone out entity temperature in video playback field by video search and broadcasting behavior expression.

Background technology

Along with the development of network video technique, increasing people obtain required video by video search engine.So-called video search engine refers to the featured videos resource on searching for Internet on Online Video playback website resource, video search, searching for Internet, inputs the key word such as film, TV play, MV that will see and search in " search " frame.And in numerous Internet videos, how can find the entity received publicity, the entity that namely temperature is higher becomes the technical matters that prior art needs solution badly.

Summary of the invention

The object of the invention is to propose one and go out entity temperature by video search and broadcasting behavior expression.

For reaching this object, the present invention by the following technical solutions:

By video search and the method playing behavior expression entity temperature, comprise the steps:

Step 1, chooses certain sample time section, select in described sample time section, have the significant impact time to occur multiple similar sample entity as sample object;

Step 2, select multiple typical index that may characterize described sample object temperature, described typical index comprises the volumes of searches for sample object, the positive playback volume of described sample object, the trailer of described sample object and titbit playback volume, the related news information playback volume of described sample object;

Step 3, for each sample object, determine the time point of its maximum effect time in described sample time section, and then determine that each a bit of time is as the search time of described sample object before and after this time point, calculate the standard variance of each typical index in described search time, then by multiple standard variance equalizations of the same typical index of all sample object, thus the average variance of different typical index is obtained respectively;

Step 4, utilizes the average variance of different typical index, by AHP analytical hierarchy process, is divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index;

Step 5, for certain sample of identical entity, by the actual value in each index odd-numbered day, divided by described index in maximal value on the same day, and be multiplied by described weighted value, obtain the actual temperature weight of this index on the same day, calculate the actual temperature weight of all indexs respectively, and described actual temperature weight is added, obtain the temperature of described sample on the same day, calculate the temperature of the every day of each entity, obtain the temperature ranking list of entity.

Preferably, the various research objects of described sample involved by video field.Described sample is one in star, director, film, TV play.

Preferably, described sample is star, and each step is as follows:

Step 1, chooses certain sample time section, select in described sample time section, have the significant impact time to occur multiple stars as sample object;

Step 2, select multiple typical index that may characterize star's temperature, described typical index comprises the volumes of searches of star, star takes part in a performance the positive playback volume of collection of drama, star takes part in a performance the trailer of collection of drama and titbit playback volume, the related news information playback volume of star, and star participate in the playback volume of variety show;

Step 3, for each sample star, determine the time point of its maximum effect time in described sample time section, and then determine each a bit of time before and after this time point, as the search time of described sample object, calculate the standard variance of each typical index in described search time, then by multiple standard variance equalizations of the same typical index of all sample object, thus obtain the average variance of different typical index respectively;

Step 4, utilize the average variance of different typical index, by AHP analytical hierarchy process, be divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index, utilize the average variance of different typical index by AHP analytical hierarchy process, be divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index;

Step 5, is extended to all star's scopes by temperature computer capacity from described multiple sample star, and for star i, the actual value of its odd-numbered day 5 indexs is respectively A volumes of searches _i, A positive _i, A trailer _i, A information _i, A variety _i, the maximal value in each index of this day simultaneously all star is designated as A volumes of searches _max, A positive _max, A trailer _max, A information _max, A participates in variety _max, then the hot value of this day star i is,

, hot value value between 0-10 of each star, thus obtain the temperature ranking list of star.

Preferably, in step 3, the average variance Sc computation process of volumes of searches S is as follows:

Absolute volumes of searches: S _i, wherein, 1≤i≤21

Standardized search amount: ${\mathrm{Sa}}_i=\frac{S_i}{\max (S_1,S_2,...,S_{21})};$

Average normalized volumes of searches: ${\mathrm{Sa}}_{\mathrm{avg}}=\frac{{\mathrm{Sa}}_1+{\mathrm{Sa}}_2+...+{\mathrm{Sa}}_{21}}{21}$

The variance of standardized search amount:

$\mathrm{Sb}=\frac{{({\mathrm{Sa}}_1-{\mathrm{Sa}}_{\mathrm{avg}})}^2+{({\mathrm{Sa}}_2-{\mathrm{Sa}}_{\mathrm{avg}})}^2+...+{({\mathrm{Sa}}_{21}-{\mathrm{Sa}}_{\mathrm{avg}})}^2}{21};$

Then the mean of variance of the standardized search amount of described multiple star is obtained the average variance Sc of the entirety of volumes of searches index.

Preferably, described sample time Duan Weisan month, the described a bit of time was 10 days, and described multiple star is 15 stars.

The present invention can not only obtain the temperature trend of each entity self, can also obtain the highest entity ranking list of temperature every day, for the top ranking list of all entities, achieves and quantizes to show.

Accompanying drawing explanation

Fig. 1 is by video search and the method playing behavior expression entity temperature according to of the present invention;

Fig. 2 is certain star's temperature schematic diagram.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.Be understandable that, specific embodiment described herein is only for explaining the present invention, but not limitation of the invention.It also should be noted that, for convenience of description, illustrate only part related to the present invention in accompanying drawing but not entire infrastructure.

First the present invention defines concept related to the present invention.

Entity, refers to the various research objects involved by video field, comprises star, director, film and TV play.Temperature, refers to the degree that above-mentioned entity is concerned, both can be to applaud to like by people, unfavorable comments also can be scold continuous, but all reflects the concerned degree come into question.In a particular embodiment of the present invention, the representative using star as all kinds of entity, explains the computation process of temperature mark.

See Fig. 1, show according to of the present invention by video search and the method playing behavior expression entity temperature.

Step 1, chooses certain sample time section, such as three months, select in described sample time section, have the significant impact time to occur multiple similar sample entity as sample object;

Step 2, select multiple typical index that may characterize described sample object temperature, described typical index comprises the volumes of searches for sample object, the positive playback volume of described sample object, the trailer of described sample object and titbit playback volume, the related news information playback volume of described sample object;

Step 3, for each sample object, determine the time point of its maximum effect time in described sample time section, and then determine that each a bit of time is as the search time of described sample object before and after this time point, calculate the standard variance of each typical index in described search time, then by multiple standard variance equalizations of the same typical index of all sample object, thus the average variance of different typical index is obtained respectively;

Step 4, utilizes the average variance of different typical index, by AHP analytical hierarchy process, is divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index;

Step 5, for certain sample of identical entity, by the actual value in each index odd-numbered day, divided by described index in maximal value on the same day, and be multiplied by described weighted value, obtain the actual temperature weight of this index on the same day, calculate the actual temperature weight of all indexs respectively, and described actual temperature weight is added, obtain the temperature of described sample on the same day, calculate the temperature of the every day of each entity, obtain the temperature ranking list of entity.

Preferably, the various research objects of described sample involved by video field, such as, in star, director, film, TV play one.

Preferably, described sample is star, and now, each step of this method is as follows:

Step 1, chooses certain sample time section, such as three months, select in described sample time section, have the significant impact time to occur multiple stars as sample object, such as 15 stars;

Step 2, select multiple typical index that may characterize star's temperature, described typical index comprises the volumes of searches of star, star takes part in a performance the positive playback volume of collection of drama, star takes part in a performance the trailer of collection of drama and titbit playback volume, the related news information playback volume of star, and star participate in the playback volume of variety show;

Step 3, for each sample star, determine the time point of its maximum effect time in described sample time section, and then determine each a bit of time before and after this time point, such as 10 days, as the search time of described sample object, calculate the standard variance of each typical index in described search time, then by multiple standard variance equalizations of the same typical index of all sample object, thus the average variance of different typical index is obtained respectively;

See Fig. 2, for star Huang ripple, 20140516 whoring events exposures are for maximum effect event, and therefore select 20140506-20140526 within totally 21 days, to be its of section, for volumes of searches S, its average variance Sc computation process is as follows computing time:

Absolute volumes of searches: S _i, wherein, 1≤i≤21

Standardized search amount: ${\mathrm{Sa}}_i=\frac{S_i}{\max (S_1,S_2,...,S_{21})};$

Average normalized volumes of searches: ${\mathrm{Sa}}_{\mathrm{avg}}=\frac{{\mathrm{Sa}}_1+{\mathrm{Sa}}_2+...+{\mathrm{Sa}}_{21}}{21}$

The variance of standardized search amount:

$\mathrm{Sb}=\frac{{({\mathrm{Sa}}_1-{\mathrm{Sa}}_{\mathrm{avg}})}^2+{({\mathrm{Sa}}_2-{\mathrm{Sa}}_{\mathrm{avg}})}^2+...+{({\mathrm{Sa}}_{21}-{\mathrm{Sa}}_{\mathrm{avg}})}^2}{21};$

Then the mean of variance of the standardized search amount of these 15 sample stars is obtained the average variance Sc of the entirety of volumes of searches index.

In these 21 days, the average variance of volumes of searches is larger, illustrates that the change fluctuation of volumes of searches before and after time point is larger, also more important for the change of reflection temperature with regard to description standard volumes of searches.In like manner can obtain the average variance yields of the entirety of other each indexs.

Exemplaryly in step 3 pick 15 sample stars, the equal hand picking of each star has gone out its time point the hottest in three months (time point of different star is different), each star can calculate aforementioned 5 typical index (volumes of searches of star before and after its time point, star takes part in a performance the positive playback volume of collection of drama, star takes part in a performance the trailer of collection of drama and titbit playback volume, the related news information playback volume of star, and star participate in the playback volume of variety show) corresponding normalize variance value.

The target of step 3 is the variance yields drawing these 5 indexs, for step 4 is served.But certain error may be there is in these 5 variance yields of single star, therefore the present invention have chosen multiple star, using the final reference variance of the mean value of the normalize variance of the volumes of searches of multiple star as volumes of searches index, other 4 indexs also adopt the average mode of the corresponding index of multiple star to obtain final reference variance equally.See table 1, obtain the final Average normalized variance of these 5 indexs.

The final Average normalized variance of table 1. star typical index

	Positive	Trailer and titbit	Volumes of searches	Domestic News	Variety show
						Variance	0.0336	0.0434	0.0461	0.0315	0.0203

Step 4, utilizes the average variance of different typical index, by AHP analytical hierarchy process, is divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index;

Those skilled in the art can know, utilize Average normalized variance by AHP analytical hierarchy process, and the weighted value obtaining different typical index belongs to the conventional Calculation Method of AHP.

9 grades is a requirement of AHP analytical hierarchy process Paired comparison matrix, namely weighs the importance between each index with 1-9 and inverse thereof, 9 represents relatively most important, 1/9 represents relatively least important.As most important in A index, B index is least important, then the relative B of A is 9.

According to the average variance of 5 indexs that step 3 generates, can see that volumes of searches index variance is maximum, variety show variance is minimum, therefore can show that gear difference is

(0.0461-0.0203)/(9-1)＝0.003225

And then extract any two indices, as " trailer and titbit " and " positive ", the gear differed between its variance is

(0.0434-0.0336)/0.003225＝3

Therefore " trailer and titbit " is (3+1)=4 relative to the importance of " positive ", and " positive " is then 1/4 relative to the importance of " trailer and titbit ".

That is, by maximal value and minimum value poor divided by numbers of gear steps, the gear obtaining every grade is poor, then the difference of any two indices is poor divided by described gear, obtains importance each other.

Obtain through similar process, finally can obtain Paired comparison matrix see table 2:

Table 2. Paired comparison matrix

, by AHP analytical hierarchy process, utilize Paired comparison matrix to generate the weighted value of final each index, be respectively P _{volumes of searches}, P _positive, P _trailer, P _information, P _variety.

Step 5, is extended to all star's scopes by temperature computer capacity from described multiple sample star, and for star i, the actual value of its odd-numbered day 5 indexs is respectively A volumes of searches _i, A positive _i, A trailer _i, A information _i, A variety _i, the maximal value in each index of this day simultaneously all star is designated as A volumes of searches _max, A positive _max, A trailer _max, A information _max, A participates in variety _max, then the hot value of this day star i is,

, hot value value between 0-10 of each star.

The many days hot value like this for certain star can form temperature trend, do the arrangement of hot value inverted order simultaneously can obtain the highest star's ranking list of this day temperature to the full dose star of one day.

Therefore, the present invention can not only obtain the temperature trend of each entity self, can also obtain the highest entity ranking list of temperature every day, for the top ranking list of all entities, achieves and quantizes to show.

Above content is in conjunction with concrete preferred implementation further description made for the present invention; can not assert that the specific embodiment of the present invention is only limitted to this; for general technical staff of the technical field of the invention; without departing from the inventive concept of the premise; some simple deduction or replace can also be made, all should be considered as belonging to the present invention by submitted to claims determination protection domain.

Claims (6)

1., by video search and the method playing behavior expression entity temperature, comprise the steps:

Step 1, chooses certain sample time section, select in described sample time section, have the significant impact time to occur multiple similar sample entity as sample object;

Step 2, select multiple typical index that may characterize described sample object temperature, described typical index comprises the volumes of searches for sample object, the positive playback volume of described sample object, the trailer of described sample object and titbit playback volume, the related news information playback volume of described sample object;

Step 3, for each sample object, determine the time point of its maximum effect time in described sample time section, and then determine that each a bit of time is as the search time of described sample object before and after this time point, calculate the standard variance of each typical index in described search time, then by multiple standard variance equalizations of the same typical index of all sample object, thus the average variance of different typical index is obtained respectively;

Step 4, utilizes the average variance of different typical index, by AHP analytical hierarchy process, is divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index;

Step 5, for certain sample of identical entity, by the actual value in each index odd-numbered day, divided by described index in maximal value on the same day, and be multiplied by described weighted value, obtain the actual temperature weight of this index on the same day, calculate the actual temperature weight of all indexs respectively, and described actual temperature weight is added, obtain the temperature of described sample on the same day, calculate the temperature of the every day of each entity, obtain the temperature ranking list of entity.

2. according to claim 1 by video search and the method playing behavior expression entity temperature, it is characterized in that:

The various research objects of described sample involved by video field.

3. according to claim 2 by video search and the method playing behavior expression entity temperature, it is characterized in that:

Described sample is one in star, director, film, TV play.

4. according to claim 3 by video search and the method playing behavior expression entity temperature, it is characterized in that:

Described sample is star, and each step is as follows:

Step 1, chooses certain sample time section, select in described sample time section, have the significant impact time to occur multiple stars as sample object;

Step 2, select multiple typical index that may characterize star's temperature, described typical index comprises the volumes of searches of star, star takes part in a performance the positive playback volume of collection of drama, star takes part in a performance the trailer of collection of drama and titbit playback volume, the related news information playback volume of star, and star participate in the playback volume of variety show;

Step 3, for each sample star, determine the time point of its maximum effect time in described sample time section, and then determine each a bit of time before and after this time point, as the search time of described sample object, calculate the standard variance of each typical index in described search time, then by multiple standard variance equalizations of the same typical index of all sample object, thus obtain the average variance of different typical index respectively;

Step 4, utilize the average variance of different typical index, by AHP analytical hierarchy process, be divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index, utilize the average variance of different typical index by AHP analytical hierarchy process, be divided into 9 grades, and then structure Paired comparison matrix, and on described Paired comparison matrix basis, obtain the weighted value of different typical index;

Step 5, is extended to all star's scopes by temperature computer capacity from described multiple sample star, and for star i, the actual value of its odd-numbered day 5 indexs is respectively A volumes of searches _i, A positive _i, A trailer _i, A information _i, A variety _i, the maximal value in each index of this day simultaneously all star is designated as A volumes of searches _max, A positive _max, A trailer _max, A information _max, A participates in variety _max, then the hot value of this day star i is,

, hot value value between 0-10 of each star, thus obtain the temperature ranking list of star.

5. according to claim 4 by video search and the method playing behavior expression entity temperature, it is characterized in that:

In step 3, the average variance Sc computation process of volumes of searches S is as follows:

Absolute volumes of searches: S _i, wherein, 1≤i≤21

Standardized search amount: ${\mathrm{Sa}}_i=\frac{S_i}{\max (S_1,S_2,...,S_{21})};$

Average normalized volumes of searches: ${\mathrm{Sa}}_{\mathrm{avg}}=\frac{{\mathrm{Sa}}_1+{\mathrm{Sa}}_1+...+{\mathrm{Sa}}_{21}}{21}$

The variance of standardized search amount:

$\mathrm{Sb}=\frac{{({\mathrm{Sa}}_1-{\mathrm{Sa}}_{\mathrm{avg}})}^2+{({\mathrm{Sa}}_2-{\mathrm{Sa}}_{\mathrm{avg}})}^2+...+{({\mathrm{Sa}}_{21}-{\mathrm{Sa}}_{\mathrm{avg}})}^2}{21};$

Then the mean of variance of the standardized search amount of described multiple star is obtained the average variance Sc of the entirety of volumes of searches index.

6. according to claim 4, it is characterized in that:

Described sample time Duan Weisan month, the described a bit of time was 10 days, and described multiple star is 15 stars.