The patent application of the invention is divisional application. The patent number of the original case is 201811229455.5, the application date is 2018, 10 and 22, and the invention name is a big data bullet screen processing system and method.
Disclosure of Invention
The invention aims to provide a processing system for reasonably managing a big data barrage.
The big data bullet screen processing system comprises
The bullet screen input module is used for logging in a registered user and acquiring first information input by the registered user;
the rough filtering module is connected with the first big database, compares the first information collected by the bullet screen input module with the first big database, deletes the first information if the first information is consistent with the information in the first big database, and converts the first information into second information if the first information is inconsistent with the information in the first big database;
the user rating module is used for judging the percentage a of the number of the first information consistent with the information in the first big database and the total number of the first information sent by the registered user by the rough filtering module according to the first information input by the registered user;
the bullet screen server can set an upper limit b of the second information sending speed, and converts the second information sent by the registered users sending the designated percentage a into third information according to the upper limit of the second information sending speed;
and the streaming media server is used for synthesizing the streaming media and the third information of the barrage server and sending the synthesized information to the user terminal.
The big data bullet screen processing system comprises a bullet screen server and a second big database, wherein the second big database is connected with the bullet screen server, two or more synonym data corresponding to second information are arranged in the second big database, when the quantity of the second information of the registered users with the specified percentage a in unit time does not reach an upper limit b, the bullet screen server generates synonym data corresponding to the second information according to the second big database, so that the quantity of the sum of the second data and the synonym data in unit time is the same as the upper limit b, and the bullet screen server converts the synonym data and the second information corresponding to the second information into third information.
The big data bullet screen processing system comprises a bullet screen server, a third big database, a third information input module and a bullet screen input module, wherein the bullet screen server is connected with the third big database, at least one piece of third data corresponding to second information is arranged in the third big database, when the bullet screen server judges whether the second information acquired by the bullet screen input module is consistent with the second information corresponding to the third data in the third big database, if so, the bullet screen server converts the third data into the third information.
The big data bullet screen processing system comprises a bullet screen server, a fourth big database, a third database and a fourth database, wherein the bullet screen server is connected with the fourth big database, when the percentage a of the registered users is lower than a preset threshold value c, the bullet screen server judges whether second information input by the registered users is consistent with second information in the fourth big database, and if the percentage a of the registered users is consistent with the second information, the bullet screen server converts the second information into third information;
the bullet screen server outputs a threshold value c according to the following formula according to the total number d of the input bullet screens of the registered user, the number q of the input bullet screens in 10 minutes, the total online time f of the registered user, the current online time p of the user, the age h of the user and the gender j of the user:
wherein the unit of the total number d of the bullet screens is a strip;
the unit of the number q of the input bullet screens in 10 minutes is a strip;
the unit of the total online time f of the registered user is hour;
the unit of the online time p of the user is minutes;
the unit of the user age h is the year of the week;
the user gender j is respectively as follows: male 0.7, female 1.1.
According to the big data bullet screen processing system, when the percentage a of the registered users is lower than a preset threshold value c, the second information with the maximum sending times of the registered users is input into the first big database.
The processing method of the big data bullet screen processing system comprises the following steps:
s100, logging in a registered user and collecting first information input by the registered user;
s200, converting first information inconsistent with the first information in the first big database into second information;
s300, judging the percentage a of the number of first information consistent with the information in the first big database according to the first information input by the registered user and the total number of the first information sent by the registered user;
s400, setting an upper limit b of a second information sending speed, and converting second information sent by registered users sending specified percentage a into third information according to the upper limit of the second information sending speed;
and S500, synthesizing the streaming media and the third information of the bullet screen server and sending the synthesized information to the user terminal.
The big data barrage processing system is different from the prior art in that the big data barrage processing system can control the barrage which can be seen by the user at the root of the barrage through the mode, so that most of the barrages seen by the users are only barrages which are sent by a part of users screened by the user rating module and the barrage server, the right of the users who frequently send illegal information to be seen by more people is limited, and the probability of occurrence of the barrage of negative words such as abuse, rumor, sending no information and the like is avoided. Therefore, the method can better control the big data barrage, make the barrage more harmonious and increase the loyalty and viscosity of the users of the streaming media.
The big data bullet screen processing system of the present invention is further explained with reference to the attached drawings.
Detailed Description
As shown in FIG. 1, the big data bullet screen processing system of the present invention comprises
The bullet screen input module is used for logging in a registered user and acquiring first information input by the registered user;
the rough filtering module is connected with the first big database, compares the first information collected by the bullet screen input module with the first big database, deletes the first information if the first information is consistent with the information in the first big database, and converts the first information into second information if the first information is inconsistent with the information in the first big database;
the user rating module is used for judging the percentage a of the number of the first information consistent with the information in the first big database and the total number of the first information sent by the registered user by the rough filtering module according to the first information input by the registered user;
the bullet screen server can set an upper limit b of the second information sending speed, and converts the second information sent by the registered users sending the designated percentage a into third information according to the upper limit of the second information sending speed;
and the streaming media server is used for synthesizing the streaming media and the third information of the barrage server and sending the synthesized information to the user terminal.
The method can control the bullet screen which can be seen by the user at the root of the bullet screen, so that most of the bullet screens seen by the users are only the bullet screens which are sent by the users screened by the user rating module and the bullet screen server, the right of more people of the users who frequently send illegal information to see the bullet screen is limited, and the probability of occurrence of negative vocabulary' bullet screens such as abuse, rumor, non-information sending and the like is avoided. Therefore, the method can better control the big data barrage, make the barrage more harmonious and increase the loyalty and viscosity of the users of the streaming media.
For example, the first large database can be understood as a negative vocabulary database, the coarse filtering module filters first information collected by the bullet screen input module for the first time, so that direct negative vocabularies are filtered, the filtered second information can be safe vocabularies or positive vocabularies, the upper limit b of the speed of a bullet screen required to be sent by a user of the bullet screen server is set on the bullet screen server, and then the bullet screen sent by registered users with lower percentage a is sent according to the lower limit b, so that users with habitual network violence are filtered, the safety of the bullet screen is purified, and the impression of the bullet screen is improved.
The bullet screen server is set by a user or an administrator to set an upper limit of the second information sending speed, and converts the second information sent by the registered user with the lower sending percentage a into third information according to the smaller upper limit b of the second information sending speed. This allows for better screening of registered users.
For example,
mode(s)
|
Upper limit b (each second)
|
Specified percentage a
|
Less bullet screen mode
|
1.2
|
0~5%
|
Middle barrage mode
|
2.4
|
5.01%~10%
|
Multiple barrage mode
|
3.6
|
10.01%~20% |
Where the percentage a retains the percentile.
Wherein the above-mentioned modes may be set by a user, wherein if the percentage a of the registered users in each mode exceeds the upper limit b, the registered users are preferentially selected by the percentage a being less than or most than the upper limit b, so that the amount of the second data conforms to the upper limit b.
Preferably, the bullet screen server is connected with a second big database, the second big database is provided with two or more synonym data corresponding to the second information, when the number of the second information of the registered users in the specified percentage a in unit time does not reach an upper limit b, the bullet screen server generates the synonym data corresponding to the second information according to the second big database, so that the number of the sum of the second data and the synonym data in unit time is the same as the upper limit b, and the bullet screen server converts the synonym data and the second information corresponding to the second information into third information.
In practical applications, streaming media is a common network propagation path, so that popularity is high, and browsing volume is huge, wherein the number of barrages is very small from the very beginning to the large data volume, and in order to ensure that users who initially watch can have a good watching experience, a second big database is needed to send the barrages which are synonymous with the excellent barrages of a part of the registered users according to the excellent barrages of a specified percentage a, so that more people are driven to send the barrages, and the streaming media impression is improved.
When the number of the second information gradually increases and reaches the upper limit b in the unit time, the third information converted from the synonym data needs to be gradually reduced until no synonym data is converted into the third information.
Preferably, the bullet screen server is connected with a third big database, at least one third data corresponding to the second information is arranged in the third big database, when the bullet screen server determines whether the second information collected by the bullet screen input module is consistent with the second information corresponding to the third data in the third big database, if so, the bullet screen server converts the third data into the third information.
In the invention, the third data in the third big database can be defined as an error correction vocabulary, and when only the input second information is matched with the third big database, the third data corresponding to the second information is converted into the third information; therefore, the wrongly written characters, wrongly written characters and non-English alphabetic writing input by the user can be converted into the correct third data, so that the video author, the anchor and the audience can watch the video conveniently.
Wherein, several of the third big data may be:
second information
|
Third data
|
False oil
|
Refueling
|
jiayou
|
Refueling
|
jia oil
|
Refueling |
。
Preferably, the bullet screen server is connected with a fourth big database, when the percentage a of the registered users is lower than a preset threshold c, the bullet screen server determines whether the second information input by the registered users is consistent with the second information in the fourth big database, and if so, the bullet screen server converts the second information into third information;
the bullet screen server outputs a threshold value c according to the following formula according to the total number d of the input bullet screens of the registered user, the number q of the input bullet screens in 10 minutes, the total online time f of the registered user, the current online time p of the user, the age h of the user and the gender j of the user:
wherein the unit of the total number d of the bullet screens is a strip;
the unit of the number q of the input bullet screens in 10 minutes is a strip;
the unit of the total online time f of the registered user is hour;
the unit of the online time p of the user is minutes;
the unit of the user age h is the year of the week;
the user gender j is respectively as follows: male 0.7, female 1.1.
The invention presets a white list database as a fourth big database, when the percentage a of the registered users is lower than a preset threshold value c, the bullet screen sent by the registered users can not be shielded only by inputting the data in the white list database by the registered users, otherwise, the bullet screen is directly shielded, thereby the bullet screen can be better purified, and specifically, the invention can better purify the bullet screen according to the specific behaviors and characteristics of the users, such as: inputting the total number d of bullet screens, the total online time f of registered users, the online time p of the users, the age h of the users and the sex j of the users within 10 minutes, and outputting a threshold c according to the following formula to generate a personalized preset threshold c.
Wherein, when c exceeds 100%, c is 100%.
Preferably, when the percentage a of the registered users is lower than a preset threshold c, the second information with the maximum number of times of sending by the registered users at a time is input into the first big database.
According to the invention, the first big database has learning capacity, and the first big database is continuously updated by screening out the most frequently output words of the registered users with lower quality and lower than the preset threshold value c, so that the first big database can be updated following the change of the users.
The processing method of the big data bullet screen processing system comprises the following steps:
s100, logging in a registered user and collecting first information input by the registered user;
s200, converting first information inconsistent with the first information in the first big database into second information;
s300, judging the percentage a of the number of first information consistent with the information in the first big database according to the first information input by the registered user and the total number of the first information sent by the registered user;
s400, setting an upper limit b of a second information sending speed, and converting second information sent by registered users sending specified percentage a into third information according to the upper limit of the second information sending speed;
and S500, synthesizing the streaming media and the third information of the bullet screen server and sending the synthesized information to the user terminal.
The above-mentioned embodiments are merely illustrative of the preferred embodiments of the present invention, and do not limit the scope of the present invention, and various modifications and improvements of the technical solution of the present invention by those skilled in the art should fall within the protection scope defined by the claims of the present invention without departing from the spirit of the present invention.