A method of do something for the occasion in real time content matching and push
Technical field
The present invention relates to the methods of a kind of do something for the occasion in real time content matching and push.
Background technique
It is born at present from computer to the emergence of internet, then arrives the prosperity of mobile Internet, all be unable to do without digital content
And its development of technology.From local file to Web page, then arrive the extensive stream of the New Medias such as online audio-video, microblogging wechat
Row, these not only contain multi-media decoding and encoding technology, the development of multimedia communication technology further comprises content distribution techniques, people
The high development of machine interaction technique and internet large-scale application systems technology.Current most popular media platform, either
Google, Facebook or microblogging, wechat, above-mentioned technology all developed and be applied to ultimate attainment.
How an interaction simply content mistake natural, that meet users ' individualized requirement and displaying demand can be designed
Filter matches and pushes in real time application system, helps user more accurate in the epoch of information huge explosion and simply obtains useful
Content has become urgent problem, also provides the distribution mechanisms of content for internet of things equipment, and there are also users to be believed
While breath, one information of input is gone every time, these information are not also recorded and stored, i.e., user obtains each time
Information, server-side release information be all much do not had with last time it is associated, and much is all the unwanted information of user, lack
Few information filtering, so that user faces various mixed and disorderly information.
Summary of the invention
It is an object of the invention to overcome the defect of the prior art, a kind of do something for the occasion in real time content matching and push are provided
Method.
The present invention is implemented as follows: a kind of method of do something for the occasion in real time content matching and push comprising following steps:
(1) user terminal is that subscriber terminal equipment generates a unique identifier CID where it;
(2) user terminal and real time content Push Service module establish two-way real-time data channel, and CID is submitted to clothes
Business end, server-side are that the real-time data channel established generates a session unique identification SID, and maintains a binding relationship
{ SID, CID };
(3) current scene information X is sent to server-side by real-time data channel by user terminal;
(4) customer data base at server-side query service end checks whether the CID for having existed Yu obtaining in step 2
The user identifier of binding then executes step 5 if it does not exist, otherwise jumps to step 6;
(5) server-side inquires content contingency table in customer data base, if it exists content tab T0, T0 associated with CID
It is then that the user terminal generates content-preference model f (CID, X, T0)=T, server-side is according to this for a label vector
Preference pattern operator f matching related content is simultaneously pushed to user terminal;Content tab associated with CID if it does not exist, then from
The newest content of content source randomly chooses a content and returns to user terminal;
(6) it if there is the user UID bound with CID in customer data base, is generated in one for the user terminal
Hold preference pattern f (UID, X), server-side matches related content according to the preference pattern operator f and is pushed to user terminal;If with
There are the multiple user UID bound with CID in user data library, then based on the raw content-preference model f of the user group [UID]
([UID], X), server-side match related content according to the preference pattern operator f and are pushed to user terminal.
Further, user inputs voice data to terminal applies, and server-side identifies user by voiceprint identification module
UID, and it is bound with session SID and user terminal CID, server-side inquires customer data base, goes through according to current UID is corresponding
History characteristic calculates a content-preference model f (UID, X), wherein X be active user scene information, server-side according to
Preference pattern operator f matching related content is simultaneously pushed to user terminal.
Further, in user terminal and the real-time connection procedure of server-side, terminal applies timing is sent eventually to server-side
Scene information is held, the CID and UID of these scene informations and current sessions are associated and are stored in customer data base by server-side
The present invention is user or user terminal content construction preference pattern by machine learning, uses more in user
In the case of more understand user, can recommend to do something for the occasion content for user more accurately, in the age of information huge explosion, effectively help
User filtering falls useless or irrelevant information.Also, all the elements are all to be pushed to user terminal immediately, ensure that content and
Shi Xing.
Specific embodiment
Below in conjunction with the embodiment of the present invention, technical scheme in the embodiment of the invention is clearly and completely described,
Obviously, described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based in the present invention
Embodiment, every other embodiment obtained by those of ordinary skill in the art without making creative efforts, all
Belong to the scope of protection of the invention.
The embodiment of the present invention provides the method for a kind of do something for the occasion in real time content matching and push comprising following steps, specifically
Refer to following explanation.
(1) user terminal is that subscriber terminal equipment generation one unique identifier CID, CID can set where it for terminal
Standby sequence number, NIC address etc..
(2) user terminal and real time content Push Service module establish two-way real-time data channel, and CID is submitted to clothes
Business end, server-side are that the real-time data channel established generates a session unique identification SID, and maintains a binding relationship
{ SID, CID }.
(3) current scene information X is sent to server-side by real-time data channel by user terminal.
(4) customer data base at server-side query service end checks whether the CID for having existed Yu obtaining in step 2
The user identifier of binding then executes step 5 if it does not exist, otherwise jumps to step 6.
(5) server-side inquires content contingency table in customer data base, if it exists content tab T0, T0 associated with CID
It is then that the user terminal generates content-preference model f (CID, X, T0)=T, server-side is according to this for a label vector
Preference pattern operator f matching related content is simultaneously pushed to user terminal;Content tab associated with CID if it does not exist, then from
The newest content of content source randomly chooses a content and returns to user terminal.
(6) it if there is the user UID bound with CID in customer data base, is generated in one for the user terminal
Hold preference pattern f (UID, X), server-side matches related content according to the preference pattern operator f and is pushed to user terminal;If with
There are the multiple user UID bound with CID in user data library, then based on the raw content-preference model f of the user group [UID]
([UID], X), server-side match related content according to the preference pattern operator f and are pushed to user terminal.
User inputs voice data to terminal applies, and server-side identifies user UID by voiceprint identification module, and by its
It is bound with session SID and user terminal CID, server-side inquires customer data base, according to the corresponding history feature number of current UID
According to history feature data include but is not limited to the past geographical location information of user, historical reception content, by terminal applies pair
The feedback operation etc. that content carries out, calculates a content-preference model f (UID, X), and wherein X is the scene information of active user,
Server-side matches related content according to the preference pattern operator f and is pushed to user terminal;In addition, in user terminal and server-side
In real-time connection procedure, terminal applies timing sends terminal scene information, including but not limited to the position letter of terminal to server-side
The CID and UID of these scene informations and current sessions are associated and are stored in customer data base by breath, server-side.
It is user or user terminal content construction preference pattern by machine learning, in the case where user uses more
More understand user, content of doing something for the occasion can be recommended more accurately for user, in the age of information huge explosion, effectively helps user's mistake
Filter useless or irrelevant information.Also, all the elements are all to be pushed to user terminal immediately, ensure that the timeliness of content,
A kind of application mode and service also are provided for vast internet of things equipment, more wide distribution canal is also increased for content quotient
Road may advantageously facilitate more creative, high-quality content creation and publication, enrich also rich while internet content service
Rich people's lives.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.