CN108563648A

CN108563648A - data display method and device, storage medium and electronic device

Info

Publication number: CN108563648A
Application number: CN201711236981.XA
Authority: CN
Inventors: 罗波罗; 吴述雷; 骆奕先; 陈旭彪; 沈丽莉
Original assignee: Tencent Technology Shanghai Co Ltd
Current assignee: Tencent Technology Shanghai Co Ltd
Priority date: 2017-11-29
Filing date: 2017-11-29
Publication date: 2018-09-21
Anticipated expiration: 2037-11-29
Also published as: CN108563648B

Abstract

The invention discloses a kind of data display method and device, storage medium and electronic devices.Wherein, this method includes：Obtain data update instruction, wherein data update instruction is for updating data object shown in the first client；Response data more new command obtains the historical operation record of the first client, wherein historical operation, which records, includes：According to the interactive operation of data object executed included in the first client, and the history object feature vector generated；Partial history characteristics of objects vector during historical operation is recorded imports push model successively, obtains prediction characteristics of objects vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；Updated target object, and the display target object in the first client are determined according to prediction characteristics of objects vector.The present invention solves the poor technical problem of the display real-time present in available data display methods.

Description

Data display method and device, storage medium and electronic device

Technical field

The present invention relates to computer realms, in particular to a kind of data display method and device, storage medium and electricity Sub-device.

Background technology

It is browsed in order to facilitate user, intelligent data display methods is provided in many terminal applies.Wherein, above-mentioned number It is often the behavioral data dependent on the user account counted in terminal applies according to display methods, such as counts whithin a period of time Pageview, operating quantity of the user account etc..According to the statistical result got, to determine the browsing custom of the user account, To achieve the effect that the hobby for the user account shows individualized content.As statistical result indicates the user account to body The pageview for educating class data is maximum, then can continue as the user account and provide display sports type data.

However, it is above-mentioned data to display are determined according to data statistics result during, in order to ensure display Accuracy, for determining that the statistic processes of data to display usually takes relatively long, for terminal applies, this will significantly Influence the real-time that data are shown.

For above-mentioned problem, currently no effective solution has been proposed.

Invention content

A kind of data display method and device of offer of the embodiment of the present invention, storage medium and electronic device, at least to solve The poor technical problem of display real-time present in available data display methods.

One side according to the ... of the embodiment of the present invention provides a kind of data display method, including：Data update is obtained to refer to It enables, wherein above-mentioned data update instruction is for updating data object shown in the first client；Respond above-mentioned data update Instruction obtains the historical operation record of above-mentioned first client, wherein above-mentioned historical operation, which records, includes：According to above-mentioned first The interactive operation of data object executed included in client, and the history object feature vector generated；By above-mentioned history The above-mentioned history object feature vector in part in operation note imports push model successively, obtains prediction characteristics of objects vector, In, above-mentioned push model trains to obtain using the feature vector of multiple ordered sample objects by machine；According to above-mentioned prediction pair As feature vector determines updated target object, and above-mentioned target object is shown in above-mentioned first client.

Another aspect according to the ... of the embodiment of the present invention additionally provides a kind of data presentation device, including：First obtains list Member, for obtaining data update instruction, wherein above-mentioned data update instruction is for updating data shown in the first client Object；Response unit obtains the historical operation record of above-mentioned first client for responding above-mentioned data update instruction, wherein Above-mentioned historical operation records：According to the interactive operation of data object executed included in above-mentioned first client, And the history object feature vector generated；First processing units, for the above-mentioned history in part in recording above-mentioned historical operation Characteristics of objects vector imports push model successively, obtains prediction characteristics of objects vector, wherein above-mentioned push model has using multiple The feature vector of sequence sample object trains to obtain by machine；Display unit, for true according to above-mentioned prediction characteristics of objects vector Fixed updated target object, and show above-mentioned target object in above-mentioned first client.

Another aspect according to the ... of the embodiment of the present invention, additionally provides a kind of storage medium, and above-mentioned storage medium includes storage Program, wherein above procedure run when execute above-mentioned data display method.

Another aspect according to the ... of the embodiment of the present invention, additionally provides a kind of electronic device, including memory, processor and deposits The computer program that can be run on above-mentioned memory and on above-mentioned processor is stored up, above-mentioned processor passes through above computer journey Sequence executes above-mentioned data display method.

In embodiments of the present invention, after obtaining data update instruction, wherein data update instruction is for updating first Shown data object in client；Response data more new command obtains the historical operation record of the first client, wherein goes through History operation note includes：According to the interactive operation of data object executed included in the first client, and what is generated goes through History characteristics of objects vector；Partial history characteristics of objects vector during historical operation is recorded imports push model successively, obtains pre- Survey characteristics of objects vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；Root It is predicted that characteristics of objects vector determines updated target object, and the display target object in the first client, pass through basis History feature vector is poured into push model and obtained by the history object feature vector in the historical operation record of the first client successively To prediction characteristics of objects vector, updated target object is determined according to prediction characteristics of objects vector, has reached and has not needed frequency The purpose for reducing the operand of data in the case of numerous update push model, more can to realize updated target object Meet user demand, reduces data and show the technique effect of required duration, and then solve present in available data display methods The poor technical problem of display real-time.

Description of the drawings

Attached drawing described herein is used to provide further understanding of the present invention, and is constituted part of this application, this hair Bright illustrative embodiments and their description are not constituted improper limitations of the present invention for explaining the present invention.In the accompanying drawings：

Fig. 1 is a kind of application environment schematic diagram of optional data display method according to the ... of the embodiment of the present invention；

Fig. 2 is a kind of flow diagram of optional data display method according to the ... of the embodiment of the present invention；

Fig. 3 is a kind of schematic diagram of optional characteristics of objects vector according to the ... of the embodiment of the present invention；

Fig. 4 is a kind of optional schematic diagram for obtaining prediction characteristics of objects vector according to the ... of the embodiment of the present invention；

Fig. 5 is another optional schematic diagram for obtaining prediction characteristics of objects vector according to the ... of the embodiment of the present invention；

Fig. 6 is another optional schematic diagram for obtaining sample object feature vector according to the ... of the embodiment of the present invention；

Fig. 7 is the schematic diagram of another optional push model according to the ... of the embodiment of the present invention；

Fig. 8 is a kind of structural schematic diagram of optional data presentation device according to the ... of the embodiment of the present invention；

Fig. 9 is a kind of structural schematic diagram of optional electronic device according to the ... of the embodiment of the present invention.

Specific implementation mode

In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people The every other embodiment that member is obtained without making creative work should all belong to the model that the present invention protects It encloses.

It should be noted that term " first " in description and claims of this specification and above-mentioned attached drawing, " Two " etc. be for distinguishing similar object, without being used to describe specific sequence or precedence.It should be appreciated that using in this way Data can be interchanged in the appropriate case, so as to the embodiment of the present invention described herein can in addition to illustrating herein or Sequence other than those of description is implemented.In addition, term " comprising " and " having " and their any deformation, it is intended that cover It includes to be not necessarily limited to for example, containing the process of series of steps or unit, method, system, product or equipment to cover non-exclusive Those of clearly list step or unit, but may include not listing clearly or for these processes, method, product Or the other steps or unit that equipment is intrinsic.

In embodiments of the present invention, a kind of embodiment of above-mentioned data display method is provided.As a kind of optional reality Mode is applied, which can be, but not limited to be applied in application environment as shown in Figure 1, be installed in terminal 102 There are the first client, terminal 102 to be connect with server 106 by network 104, the first client obtained in terminal 102 is used for The data update instruction of the shown data object of update, and data update instruction is sent to server 106, server 106 response datas more new command obtains the historical operation record of the first client, wherein historical operation, which records, includes：According to The interactive operation of data object executed included in first client, and the history object feature vector generated；Server 106 historical operation is recorded in partial history characteristics of objects vector import successively push model, obtain prediction characteristics of objects to Amount, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；106 basis of server Prediction characteristics of objects vector determines updated target object, and display target object in the first client on the terminal 102.

It should be noted that in above-mentioned example, the update operation of data is enterprising in the corresponding server of the first client Capable, the display of data is carried out in the terminal where the first client, and in actual application scenarios, data are more New can also be to be carried out in the terminal where the first client, such as used account in being locally stored for the first client Number historical operation record etc., the above is only a kind of example, do not do any restriction herein.

In the present embodiment, data update instruction is obtained, wherein data update instruction is for updating institute in the first client The data object of display；Response data more new command obtains the historical operation record of the first client, wherein historical operation records Include：According to the interactive operation of data object executed included in the first client, and the history object feature generated Vector；Partial history characteristics of objects vector during historical operation is recorded imports push model successively, obtains prediction characteristics of objects Vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；According to prediction object Feature vector determines updated target object, and the display target object in the first client, by according to the first client Historical operation record in history object feature vector, by history feature vector successively pour into push model obtain prediction object Feature vector determines updated target object according to prediction characteristics of objects vector, has reached and has not needed frequent updating push The purpose for reducing the operand of data in the case of model can more meet user's need to realize updated target object It asks, reduces the technique effect that data show required duration, and then the display solved present in available data display methods is real-time The poor technical problem of property.

Optionally, in the present embodiment, above-mentioned terminal can include but is not limited at least one of：Mobile phone, tablet electricity Brain, laptop, desktop PC, DTV and the hardware device for running recommendation display application.Above-mentioned network can be with Including but not limited at least one of：Wide area network, Metropolitan Area Network (MAN), LAN.It is above-mentioned only a kind of example, the present embodiment to this not Do any restriction.

One side according to the ... of the embodiment of the present invention provides a kind of data display method, as shown in Fig. 2, this method includes：

S202 obtains data update instruction, wherein data update instruction is for updating number shown in the first client According to object；

S204, response data more new command obtain the historical operation record of the first client, wherein in historical operation record Including：According to the interactive operation of data object executed included in the first client, and the history object feature generated to Amount；

S206, the partial history characteristics of objects vector during historical operation is recorded import push model, are predicted successively Characteristics of objects vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；

S208 determines updated target object according to prediction characteristics of objects vector, and shows mesh in the first client Mark object.

Optionally, in the present embodiment, above-mentioned data display method can be, but not limited to during being shown applied to data, Wherein, above-mentioned data display method can include but is not limited to be applied to news application, social networking application and other have browsing In the application of function, the interested target object of user is shown according to operation browsing record to realize.Above-mentioned data update refers to Order can include but is not limited to：The refreshing instruction that refresh operation is generated is executed to browser interface, executes what register was given birth to Operational order, the operational order, etc. that corresponding content comment on and other interactive operations are generated.On for example, it is assumed that Data display method is stated using data as news content, the first client obtains data update instruction, and responds the data update and refer to Enable historical operation record (such as news category of user's browsing, browsing time, browsing content, comment, the point for obtaining the first client Praise), partial history characteristics of objects vector (such as vector of the news category of user's browsing, browsing during historical operation is recorded Content vector etc.) push model is imported successively, prediction characteristics of objects vector is obtained, update is determined according to prediction characteristics of objects vector Target object (such as predicting the interested news category of user, news content) afterwards, and the display target in the first client Object.Above-mentioned is only a kind of example, and any restriction is not done to this in the present embodiment.

Wherein, the updated target object determined in above-mentioned data display method may include a target pair As can also be multiple target objects.It is illustrated, is determined so that data are in news content as an example by above-mentioned data display method Updated target object is specific news content, can be news item content, can also be a plurality of news content, specifically Can be configured according to practical experience, do not do any restriction herein.

It should be noted that after obtaining data update instruction, wherein data update instruction is for updating the first client Shown data object on end；Response data more new command obtains the historical operation record of the first client, wherein history is grasped It notes down and includes：According to the interactive operation of data object executed included in the first client, and the history pair generated As feature vector；Partial history characteristics of objects vector during historical operation is recorded imports push model successively, obtains prediction pair As feature vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；According to pre- It surveys characteristics of objects vector and determines updated target object, and the display target object in the first client, by according to first Client historical operation record in history object feature vector, by history feature vector successively pour into push model obtain it is pre- Characteristics of objects vector is surveyed, updated target object is determined according to prediction characteristics of objects vector, has been reached need not frequently more The purpose for reducing the operand of data in the case of new push model, can more meet to realize updated target object User demand reduces data and shows the technique effect of required duration, and then solves aobvious present in available data display methods Show the poor technical problem of real-time.

Optionally, in the present embodiment, determine that updated target object includes but not according to prediction characteristics of objects vector It is limited to：It obtains and the matched data-object library of the first client, wherein store the first client in data-object library to be shown Total data object feature vector；The feature vector of total data object and prediction object in comparison data library of object successively Feature vector；Target object is determined according to the result of comparison.

Optionally, in the present embodiment, determine that target object includes but not limited to according to the result of comparison：From total data In the feature vector of object, the feature vector for being less than first threshold with prediction characteristics of objects vector distance is obtained；It will be away from It is determined as target object from data object corresponding less than the feature vector of first threshold.

In specific application scenarios, as shown in figure 3, after getting prediction characteristics of objects vectorial (white dot), it will Predict the feature vector (black dot) of characteristics of objects vector and the first client total data object to display successively into Row compares, and in rounding vector space and prediction characteristics of objects vector distance is less than one or more data objects of first threshold Characteristics of objects vector, the distance of two neighboring vector can use following Euclidean distance formula：

Wherein, x_iIndicate i-th dimension data in prediction characteristics of objects vector x, y_iIndicate the feature vector in data-object library I-th dimension data in y, d (x, y) is for indicating prediction the distance between characteristics of objects vector x and a feature vector y.

Optionally, in the present embodiment, the partial history characteristics of objects vector in historical operation being recorded imports successively to be pushed away Model is sent, obtaining prediction characteristics of objects vector includes but not limited to：It determines in historical operation record for importing push model Partial history characteristics of objects vector, wherein partial history characteristics of objects vector includes：Historical operation record in, executed and Execution interval is less than N number of history object feature vector corresponding to continuous N number of interactive operation of second threshold, and execution interval is to hold Execution time interval of the moment away from current time of row interactive operation, N is natural number；By identified partial history characteristics of objects Vector imports push model successively, obtains prediction characteristics of objects vector.

In specific application scenarios, the partial history characteristics of objects vector during historical operation is recorded pours into push successively Module, it is usually current from distance generally according to the time sequencing of the historical operation record of history object feature vector institute object Time is most as far as the sequence nearest apart from current time.It imports successively in push model sequentially in time, obtains prediction object Feature vector is illustrated by taking the news that user reads as an example, as shown in figure 4, inputting user into pushing module reads news Historical operation record, wherein user reads news, and there are three historical operations records, i.e. N=3, and wherein three history are grasped The execution interval noted down is less than second threshold, and the chronological order of wherein news sample is followed successively by the news 1 of user's reading, The news 2 that user reads, the news 3 that user reads.Pass through the history object feature vector recorded according to three historical operations Push model is poured into successively, after the history object feature vector for the news 3 for reading user is input to pushing module, is obtained and is used The predicted characteristics vector for the next chapter news that family will be read.

Optionally, in the present embodiment, identified partial history characteristics of objects vector is imported to push model successively, is obtained Include to prediction characteristics of objects vector：N number of history object feature vector is imported to push model successively, obtains N number of prediction object Feature vector；It is vectorial according to N-1 prediction characteristics of objects before executing moment discarding, n-th prediction characteristics of objects is vectorial, as The prediction characteristics of objects vector for determining target object of push model output.By data be user read information for into Row explanation, wherein managing operation history are sample to be entered, wherein N=3, by history object feature vector (user The information of reading) pour into push model successively sequentially in time, it pushes in model for the first time and in second of operation result Prediction result is dropped, and only retains hidden state (being used to indicate the mode of operation operated before).By second of operation output Hidden state and third time history object feature vector input model obtain final prediction characteristics of objects vector.

It should be noted that in the present embodiment, above-mentioned history object feature vector can be, but not limited to：Obtaining data Before more new command, the first client is generated according to the corresponding historical operation record of history interactive operation of executed.Wherein, on Historical operation record to be stated to can be, but not limited to record for offline historical operation, corresponding push model then obtains for off-line training, Above-mentioned historical operation record can also but be not limited to online historical operation record, and corresponding push model then obtains for on-line training It arrives.Wherein, history object feature vector can be, but not limited to that (such as history was read new to the object in history interactive operation Hear) feature extraction is carried out, to obtain the corresponding history object feature vector of the object.

Still illustrated by taking the news predicted characteristics vector acquisition process that above-mentioned user reads as an example, as shown in figure 5, its In, the corresponding history object feature vector of news 1,2,3 (i.e. historical operation records) read there are 3 users, i.e. N=3, The chronological order of middle news sample is followed successively by the news 1 of user's reading, the news 2 that user reads, the news that user reads 3, the corresponding history object feature vector of news that 3 users read is imported into push model sequentially in time, obtains 3 in advance Characteristics of objects vector is surveyed, the news corresponding time read according to user abandons preceding 2 historical forecast objects by push model The prediction characteristics of objects vector of output, retains the hidden state that the news 2 that user reads passes through push model output, user is read The hidden state for the news 2 that the history object feature vector of the news 3 of reading is read with user imports push model, will push model The result exported is vectorial by the prediction characteristics of objects of the next chapter news of reading as prediction user.

Optionally, in the present embodiment, before obtaining data update instruction, further include：Obtain multiple ordered samples pair The feature vector of elephant；Using the feature vector of multiple ordered sample objects, train to obtain push model by machine.Specific In application scenarios, there can be different push models for different sample databases, push model is typically to have by multiple The feature vector of sequence sample object, for example there is the sample object of time sequencing, by machine train to obtain push model (such as Trained neural network).

It should be noted that in the present embodiment, obtaining the feature vector of push model for being trained by machine, being more The feature vector of a ordered sample object.That is, in the training process, sample object will successively be imported according to predetermined order Push model is trained, so that the obtained push model of training depends on input sequence so that input sequence is to instruction Practice result and form influence, and then realizes after obtaining history object included in historical operation record, it can be according to its sequence It predicts and is associated with larger prediction object with the data object of last time execution interactive operation, push model is improved to reach The accuracy of prediction.

Optionally, in the present embodiment, the feature vector for obtaining multiple ordered sample objects includes but not limited to：It obtains successively Take multiple sample objects；Following operation is executed respectively to each sample object in multiple sample objects：It obtains in sample object Including data type, wherein data type includes at least one of：The title of sample object, the abstract of sample object, What data in sample object, sample object executed interactive operation executes moment, the interactive information of interactive operation, sample object Label；Sample object is respectively converted into corresponding subvector according to data type；It combines the subvector and obtains the sample The feature vector of object.

Optionally, in the present embodiment, above-mentioned that sample object is respectively converted into corresponding subvector according to data type It can be, but not limited to：Different data content in sample object is respectively converted into corresponding different subvector according to data type. Wherein, the feature vector of sample object can be, but not limited to obtain by being combined splicing to above-mentioned subvector.

Specifically in conjunction with shown in attached drawing 6, participle operation is carried out to the title of sample object, the abstract of sample object, is obtained pair It should segment, term vector conversion then is carried out to participle, obtain title subvector, abstract subvector；To the data in sample object Carry out keyword extraction operation, participle operation, term vector conversion then carried out to the keyword extracted, obtain content to Amount；The execution moment that interactive operation is executed to sample object converts, and obtains time subvector；To the interaction letter of interactive operation Breath is converted, and interactive information subvector is obtained；The label of sample object is converted, label information subvector is obtained.Into One step splices and combines above-mentioned subvector, to obtain the feature vector of the sample object.

Optionally, in the present embodiment, it using the feature vector of multiple ordered sample objects, is trained and is pushed away by machine The model is sent to include：The feature vector of multiple ordered sample objects is imported to push model successively, obtains multiple forecast sample objects Feature vector；Obtain multiple feature vector errors successively according to the feature vector of multiple forecast sample objects, wherein feature to Amount error is used to indicate the mistake between the feature vector of forecast sample object and the feature vector of the sample object of next importing Difference；Model is pushed according to feature vector error transfer factor, so that the feature vector error for the push model that training obtains is less than third Threshold value.

It is specifically illustrated in conjunction with shown in attached drawing 7, it is assumed that the feature vector of multiple ordered sample objects is followed successively by sample pair As A to sample object E.In the present embodiment, the feature vector of sample object A is imported into push model, obtains forecast sample pair As the feature vector of A；Then the feature vector of the sample object B of next importing is obtained, and compares the spy of above-mentioned sample object B The sign vector error a between the two with the feature vector of forecast sample object A, and so on, it is assumed that until getting sample pair As the corresponding error d of E.Above-mentioned error a to error d is compared, adjustment push model is controlled by minimizing error function, such as adjusts Model parameter in whole push model, so that the error that training obtains is gradually reduced, when error is less than the feelings of third threshold value Under condition, determines that push model training is completed, can be applied to determine target object to display during data are shown.

As a kind of optional technical solution, updated target object is determined according to prediction characteristics of objects vector, and Display target object includes in first client：

S1 is obtained and the matched data-object library of the first client, wherein the first client institute is stored in data-object library The feature vector of total data object to be shown；

S2, the feature vector of total data object and prediction characteristics of objects are vectorial in comparison data library of object successively；

S3 determines target object according to the result of comparison.

Optionally, determine that target object includes according to the result of comparison：

S1 is obtained from the feature vector of total data object and is less than the with prediction characteristics of objects vector distance The feature vector of one threshold value；

S2, the corresponding data object of feature vector that distance is less than to first threshold are determined as target object.

It is carried out by taking the characteristics of objects vector comparison process of prediction characteristics of objects vector and total data as shown in Figure 3 as an example Illustrate, it, will prediction characteristics of objects vector and the after getting prediction characteristics of objects vector (as shown in Figure 3 white dot) The feature vector (black dot as shown in Figure 3) of one client total data object to display is compared successively, Distance is less than one or more data objects of first threshold between prediction characteristics of objects vector in rounding vector space Characteristics of objects vector, the distance of two neighboring vector can use following Euclidean distance formula：

Through the embodiment of the present invention, successively in comparison data library of object the feature vector of total data object with prediction object Feature vector determines target object according to comparison result, realizes the purpose for reducing and obtaining target object data operand, improves Obtain the accuracy of target object.

As a kind of optional technical solution, the partial history characteristics of objects vector during historical operation is recorded imports successively Model is pushed, obtaining prediction characteristics of objects vector includes：

S1 determines the partial history characteristics of objects vector for importing push model in historical operation record, wherein part History object feature vector includes：In historical operation record, executed and execution interval are less than continuous N number of friendship of second threshold The corresponding N number of history object feature vector of interoperability, execution interval are to execute the execution moment of interactive operation away from current time Time interval, N is natural number；

S2, by identified partial history characteristics of objects vector successively import push model, obtain prediction characteristics of objects to Amount.

In specific application scenarios, illustrated by taking the news that user reads as an example, as shown in figure 4, to pushing module Middle input user reads the historical operation record of news, wherein user reads news, and there are three historical operations to record, i.e. N= 3, the execution interval of wherein three historical operations record is less than second threshold, the wherein chronological order of news sample successively For：The news 1 that user reads, the news 2 that user reads, the news 3 that user reads.By according to three historical operation records History object feature vector pour into push model successively, the news 3 for reading user history object feature vector input To after pushing module, the predicted characteristics vector for the next chapter news that user will read is obtained.

Through this embodiment, the partial history characteristics of objects in historical operation being recorded imports push model successively, obtains Predict characteristics of objects vector so that the accuracy higher of prediction characteristics of objects vector, and reduce acquisition prediction characteristics of objects The data operation quantity of vector.

As a kind of optional technical solution, identified partial history characteristics of objects vector is imported to push mould successively Type, obtaining prediction characteristics of objects vector includes：

N number of history object feature vector is imported push model by S1 successively, obtains N number of prediction characteristics of objects vector；

S2 abandons preceding N-1 prediction characteristics of objects vector, by n-th prediction characteristics of objects vector, work according to the moment is executed For the prediction characteristics of objects vector for determining target object of push model output.

Still illustrated by taking the news predicted characteristics vector acquisition process that above-mentioned user reads as an example, as shown in figure 5, its In, the corresponding history object feature vector of news 1,2,3 (i.e. historical operation records) read there are 3 users, i.e. N=3, The chronological order of middle news sample is followed successively by：The news 1 that user reads, the news 2 that user reads, the news that user reads 3, the corresponding history object feature vector of news that 3 users read is imported into push model sequentially in time, obtains 3 in advance Characteristics of objects vector is surveyed, the news corresponding time read according to user abandons preceding 2 historical forecast objects by push model The prediction characteristics of objects vector of output, retains the hidden state that the news 2 that user reads passes through push model output, user is read The hidden state for the news 2 that the history object feature vector of the news 3 of reading is read with user imports push model, will push model The result exported is vectorial by the prediction characteristics of objects of the next chapter news of reading as prediction user.

Through this embodiment, identified partial history characteristics of objects vector is imported to push model successively, and will be newest The output of history object feature vector prediction characteristics of objects vector, as by identified partial history characteristics of objects vector according to Secondary importing pushes model, obtains prediction characteristics of objects vector so that the accuracy higher of prediction characteristics of objects vector, and reduce Obtain the data operation quantity of prediction characteristics of objects vector.

Further include before obtaining data update instruction as a kind of optional technical solution：

S1 obtains the feature vector of multiple ordered sample objects；

S2 trains to obtain push model by machine using the feature vector of multiple ordered sample objects.

By embodiment provided by the present application, using the feature vector of multiple ordered sample objects, it is trained to carry out machine To push model.So that above-mentioned push model can go out predict object with input sequence Accurate Prediction, obtained in order to improve Shown target object accuracy.Further, training obtains the higher push model of accuracy, will also reduce to used The replacement number of model is pushed, and then ensures that the target object determined using push model is updated the real-time of display.

As a kind of optional technical solution, the feature vector for obtaining multiple ordered sample objects includes：

S1 obtains multiple sample objects successively；

S2 executes following operation respectively to each sample object in multiple sample objects：

S21 obtains data type included in sample object, wherein data type includes at least one of：Sample The title of object, the abstract of sample object, the data in sample object, sample object execute the execution moment of interactive operation, hand over The interactive information of interoperability, the label of sample object；

Sample object is respectively converted into corresponding subvector by S22 according to data type；

S23, combination subvector obtain the feature vector of sample object.

By embodiment provided by the present application, the data content by obtaining sample object is distinguished according to different types of data Corresponding subvector is converted to, the feature vector of sample object is obtained to realize combination splicing, so that the spy of sample object Sign vector can be with the data of multiple dimensions in Covering samples object, and are no longer only limitted to one dimension of statistic to sample object Degree to widen the coverage area of sample object, and then improves the coverage area of the obtained push model of training, realizes to unexpected winner The push of data object starts.

As a kind of optional technical solution, using the feature vector of multiple ordered sample objects, trained by machine Include to push model：

The feature vector of multiple ordered sample objects is imported push model, obtains multiple forecast sample objects by S1 successively Feature vector；

S2 obtains multiple feature vector errors successively according to the feature vector of multiple forecast sample objects, wherein feature to Amount error is used to indicate the mistake between the feature vector of forecast sample object and the feature vector of the sample object of next importing Difference；

S3 pushes model according to feature vector error transfer factor, so that the feature vector error for the push model that training obtains Less than third threshold value.

By embodiment provided by the present application, exported by by the feature vector of multiple ordered sample objects, and with it The feature vector of multiple forecast sample objects is compared, and the error obtained between vector carries to reach simplified training step The purpose of high model training efficiency.

It should be noted that for each method embodiment above-mentioned, for simple description, therefore it is all expressed as a series of Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the described action sequence because According to the present invention, certain steps can be performed in other orders or simultaneously.Secondly, those skilled in the art should also know It knows, embodiment described in this description belongs to preferred embodiment, and involved action and module are not necessarily of the invention It is necessary.

Through the above description of the embodiments, those skilled in the art can be understood that according to above-mentioned implementation The method of example can add the mode of required general hardware platform to realize by software, naturally it is also possible to by hardware, but it is very much In the case of the former be more preferably embodiment.Based on this understanding, technical scheme of the present invention is substantially in other words to existing The part that technology contributes can be expressed in the form of software products, which is stored in a storage In medium (such as ROM/RAM, magnetic disc, CD), including some instructions are used so that a station terminal equipment (can be mobile phone, calculate Machine, server or network equipment etc.) execute method described in each embodiment of the present invention.

According to embodiments of the present invention, a kind of data presentation device for implementing above-mentioned data display method is additionally provided, As shown in figure 8, the device includes：

1) first acquisition unit 802, for obtaining data update instruction, wherein data update instruction is for updating first Shown data object in client；

2) response unit 804 obtain the historical operation record of the first client for response data more new command, wherein Historical operation records：According to the interactive operation of data object executed included in the first client, and generate History object feature vector；

3) first processing units 806 import successively for the partial history characteristics of objects vector in recording historical operation Model is pushed, obtains prediction characteristics of objects vector, wherein push model is passed through using the feature vector of multiple ordered sample objects Machine trains to obtain；

4) display unit 808, for determining updated target object according to prediction characteristics of objects vector, and in the first visitor Display target object in the end of family.

Optionally, in the present embodiment, above-mentioned data presentation device can be, but not limited to during being shown applied to data, Wherein, above-mentioned data presentation device can include but is not limited to be applied to news application, social networking application and other have browsing In the application of function, the interested target object of user is shown according to operation browsing record to realize.Above-mentioned data update refers to Order can include but is not limited to：The refreshing instruction that refresh operation is generated is executed to browser interface, executes what register was given birth to Operational order, the operational order, etc. that corresponding content comment on and other interactive operations are generated.On for example, it is assumed that Data presentation device is stated using data as news content, the first client obtains data update instruction, and responds the data update and refer to Enable historical operation record (such as news category of user's browsing, browsing time, browsing content, comment, the point for obtaining the first client Praise), partial history characteristics of objects vector (such as vector of the news category of user's browsing, browsing during historical operation is recorded Content vector etc.) push model is imported successively, prediction characteristics of objects vector is obtained, update is determined according to prediction characteristics of objects vector Target object (such as predicting the interested news category of user, news content) afterwards, and the display target in the first client Object.Above-mentioned is only a kind of example, and any restriction is not done to this in the present embodiment.

Wherein, the updated target object determined in above-mentioned data presentation device may include a target pair As can also be multiple target objects.It is illustrated, is determined so that data are in news content as an example by above-mentioned data presentation device Updated target object is specific news content, can be news item content, can also be a plurality of news content, specifically Can be configured according to practical experience, do not do any restriction herein.

It should be noted that after obtaining data update instruction, wherein data update instruction is for updating the first client Shown data object on end；Response data more new command obtains the historical operation record of the first client, wherein history is grasped It notes down and includes：According to the interactive operation of data object executed included in the first client, and the history pair generated As feature vector；Partial history characteristics of objects vector during historical operation is recorded imports push model successively, obtains prediction pair As feature vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；According to pre- It surveys characteristics of objects vector and determines updated target object, and the display target object in the first client, by according to first Client historical operation record in history object feature vector, by history feature vector successively pour into push model obtain it is pre- Characteristics of objects vector is surveyed, updated target object is determined according to prediction characteristics of objects vector, has been reached need not frequently more The purpose for reducing the operand of data in the case of new push model, can more meet to realize updated target object User demand reduces data and shows the technique effect of required duration, and then solves and show real-time present in the relevant technologies Poor technical problem.

As a kind of optional technical solution, display unit 808 includes：

1) the first acquisition module, for obtaining and the matched data-object library of the first client, wherein in data-object library Store the feature vector of the first client total data object to display；

2) judgment module, the feature vector and prediction object for total data object in comparison data library of object successively are special Sign vector；

3) the first determining module, for determining target object according to the result of comparison.

Optionally, above-mentioned first determining module includes：

First acquisition submodule, for from the feature vector of total data object, obtaining and prediction characteristics of objects vector Distance is less than the feature vector of first threshold；

First determination sub-module, the corresponding data object of feature vector for distance to be less than to first threshold are determined as mesh Mark object.

As a kind of optional technical solution, first processing units 806 include：

1) the second determining module, it is special for importing the partial history object of push model in historical operation record for determining Sign vector, wherein partial history characteristics of objects vector includes：In historical operation record, executed and execution interval are less than the N number of history object feature vector corresponding to continuous N number of interactive operation of two threshold values, execution interval are to execute holding for interactive operation Time interval of the row moment away from current time, N are natural number；

2) first processing module is obtained for identified partial history characteristics of objects vector to be imported to push model successively To prediction characteristics of objects vector.

As a kind of optional technical solution, first processing module includes：

1) the first processing submodule obtains N number of pre- for N number of history object feature vector to be imported to push model successively Survey characteristics of objects vector；

2) second processing submodule, for N-1 prediction characteristics of objects vector before being abandoned according to the execution moment, by n-th Predict characteristics of objects vector, the prediction characteristics of objects vector for determining target object as push model output.

As a kind of optional technical solution, above-mentioned data presentation device further includes：

1) second acquisition unit, for before obtaining data update instruction, obtaining the feature of multiple ordered sample objects Vector；

2) second processing unit is trained by machine and is pushed away for the feature vector using multiple ordered sample objects Send model.

As a kind of optional technical solution, second acquisition unit includes：

1) the second acquisition module, for obtaining multiple sample objects successively；

2) Second processing module, for executing following operation respectively to each sample object in multiple sample objects：

S1 obtains data type included in sample object, wherein data type includes at least one of：Sample The title of object, the abstract of sample object, the data in sample object, sample object execute the execution moment of interactive operation, hand over The interactive information of interoperability, the label of sample object；

Sample object is respectively converted into corresponding subvector by S2 according to data type；

S3, combination subvector obtain the feature vector of sample object.

As a kind of optional technical solution, second processing unit includes：

1) third processing module is obtained for the feature vector of multiple ordered sample objects to be imported to push model successively The feature vector of multiple forecast sample objects；

2) third acquisition module, for obtaining multiple feature vectors successively according to the feature vector of multiple forecast sample objects Error, wherein feature vector error is used to indicate the feature vector of forecast sample object and the sample object of next importing Error between feature vector；

3) fourth processing module, for pushing model according to feature vector error transfer factor, so that the push mould that training obtains The feature vector error of type is less than third threshold value.

Another aspect according to the ... of the embodiment of the present invention additionally provides a kind of electricity for implementing above-mentioned data display method Sub-device, as shown in figure 9, the electronic device includes：One or more (one is only shown in figure) processors 902, memory 904, display 906, user interface 908, transmitting device 910.Wherein, memory 904 can be used for storing software program and mould Block, as in the embodiment of the present invention data display method and the corresponding program instruction/module of device, processor 902 pass through operation The software program and module being stored in memory 904 are realized above-mentioned to perform various functions application and data processing Data display method.Memory 904 may include high speed random access memory, can also include nonvolatile memory, such as one Or multiple magnetic storage devices, flash memory or other non-volatile solid state memories.In some instances, memory 904 can Further comprise that the memory remotely located relative to processor 902, these remote memories can pass through network connection to end End.The example of above-mentioned network includes but not limited to internet, intranet, LAN, mobile radio communication and combinations thereof.

Above-mentioned transmitting device 910 is used to receive via a network or transmission data.Above-mentioned network specific example It may include cable network and wireless network.In an example, transmitting device 910 includes a network adapter (Network Interface Controller, NIC), can be connected with other network equipments with router by cable so as to interconnection Net or LAN are communicated.In an example, transmitting device 910 is radio frequency (Radio Frequency, RF) module, For wirelessly being communicated with internet.

Wherein, specifically, memory 904 operation note of the interactive operation of object executed and is given birth to for storing data At characteristics of objects vector and the shown object information of object etc..Wherein, display 906 can be, but not limited to for showing Show target object.

Optionally, the specific example in the present embodiment can refer to the example described in above-described embodiment, the present embodiment Details are not described herein.

It will appreciated by the skilled person that structure shown in Fig. 9 is only to illustrate, electronic device can also be intelligence Mobile phone (such as Android phone, iOS mobile phones), tablet computer, applause computer and mobile internet device (Mobile Internet Devices, MID), the terminal devices such as PAD.Fig. 9 it does not cause to limit to the structure of above-mentioned electronic device.Example Such as, electronic device may also include than shown in Fig. 9 more either less components (such as network interface, display device) or With the configuration different from shown in Fig. 9.

One of ordinary skill in the art will appreciate that all or part of step in the various methods of above-described embodiment is can To be completed come command terminal device-dependent hardware by program, which can be stored in a computer readable storage medium In, storage medium may include：Flash disk, read-only memory (Read-Only Memory, ROM), random access device (Random Access Memory, RAM), disk or CD etc..

The another aspect of the embodiment of the present invention, additionally provides a kind of storage medium, optionally, in the present embodiment, above-mentioned Storage medium can be set to store the program code for executing following steps：

S1 obtains data update instruction, wherein data update instruction is for updating data shown in the first client Object；

S2, response data more new command obtain the historical operation record of the first client, wherein are wrapped in historical operation record It includes：According to the interactive operation of data object executed included in the first client, and the history object feature vector generated；

S3, the partial history characteristics of objects vector during historical operation is recorded import push model, obtain prediction pair successively As feature vector, wherein push model trains to obtain using the feature vector of multiple ordered sample objects by machine；

S4 determines updated target object, and the display target in the first client according to prediction characteristics of objects vector Object.

Optionally, in the present embodiment, above-mentioned storage medium can include but is not limited to：USB flash disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disc or The various media that can store program code such as CD.

The embodiments of the present invention are for illustration only, can not represent the quality of embodiment.

If the integrated unit in above-described embodiment is realized in the form of SFU software functional unit and as independent product Sale in use, can be stored in the storage medium that above computer can be read.Based on this understanding, skill of the invention Substantially all or part of the part that contributes to existing technology or the technical solution can be with soft in other words for art scheme The form of part product embodies, which is stored in a storage medium, including some instructions are used so that one Platform or multiple stage computers equipment (can be personal computer, server or network equipment etc.) execute each embodiment institute of the present invention State all or part of step of method.

In the above embodiment of the present invention, all emphasizes particularly on different fields to the description of each embodiment, do not have in some embodiment The part of detailed description may refer to the associated description of other embodiment.

In several embodiments provided herein, it should be understood that disclosed client, it can be by others side Formula is realized.Wherein, the apparatus embodiments described above are merely exemplary, for example, the unit division, only one Kind of division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component can combine or It is desirably integrated into another system, or some features can be ignored or not executed.Another point, it is shown or discussed it is mutual it Between coupling, direct-coupling or communication connection can be INDIRECT COUPLING or communication link by some interfaces, unit or module It connects, can be electrical or other forms.

The unit illustrated as separating component may or may not be physically separated, aobvious as unit The component shown may or may not be physical unit, you can be located at a place, or may be distributed over multiple In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme 's.

In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also It is that each unit physically exists alone, it can also be during two or more units be integrated in one unit.Above-mentioned integrated list The form that hardware had both may be used in member is realized, can also be realized in the form of SFU software functional unit.

The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered It is considered as protection scope of the present invention.

Claims

1. a kind of data display method, which is characterized in that including：

Obtain data update instruction, wherein the data update instruction is for updating data pair shown in the first client As；

Respond the historical operation record that the data update instruction obtains first client, wherein the historical operation note Record includes：According to the interactive operation of data object executed included in first client, and the history pair generated As feature vector；

The part history object feature vector during the historical operation is recorded imports push model successively, obtains prediction pair As feature vector, wherein the push model trains to obtain using the feature vector of multiple ordered sample objects by machine；

Updated target object is determined according to the prediction characteristics of objects vector, and in first client described in display Target object.

2. according to the method described in claim 1, it is characterized in that, described determined according to the prediction characteristics of objects vector updates Target object afterwards, and show that the target object includes in first client：

It obtains and the matched data-object library of the first client, wherein first visitor is stored in the data-object library The feature vector of family end total data object to display；

The feature vector of total data object described in the data-object library and prediction characteristics of objects vector are compared successively；

The target object is determined according to the result of comparison.

3. according to the method described in claim 2, it is characterized in that, the result according to comparison determines the target object packet It includes：

From the feature vector of the total data object, obtain with the prediction characteristics of objects vector distance less than the The feature vector of one threshold value；

The corresponding data object of feature vector that the distance is less than to the first threshold is determined as the target object.

4. according to the method described in claim 1, it is characterized in that, it is described by the historical operation record in part described in go through History characteristics of objects vector imports push model successively, obtains prediction characteristics of objects vector and includes：

Determine the part history object feature vector for importing the push model in the historical operation record, In, the part history object feature vector includes：In historical operation record, executed and execution interval are less than second N number of history object feature vector corresponding to continuous N number of interactive operation of threshold value, the execution interval are to execute the friendship Execution time interval of the moment away from current time of interoperability, the N are natural number；

The identified part history object feature vector is imported into the push model successively, it is special to obtain the prediction object Sign vector.

5. according to the method described in claim 4, it is characterized in that, it is described by the identified part history object feature to Amount imports the push model successively, and obtaining the prediction characteristics of objects vector includes：

N number of history object feature vector is imported into the push model successively, obtain N number of prediction characteristics of objects to Amount；

The described prediction characteristics of objects vectors of N-1 before being abandoned according to the execution moment, will prediction characteristics of objects described in n-th to Amount, the prediction characteristics of objects vector for determining the target object as the push model output.

6. according to the method described in claim 1, it is characterized in that, before acquisition data update instruction, further include：

Obtain the feature vector of the multiple ordered sample object；

Using the feature vector of the multiple ordered sample object, train to obtain the push model by machine.

7. according to the method described in claim 6, it is characterized in that, the feature for obtaining the multiple ordered sample object to Amount includes：

Multiple sample objects are obtained successively；

Following operation is executed respectively to each sample object in the multiple sample object：

Obtain data type included in the sample object, wherein the data type includes at least one of：It is described The title of sample object, the making a summary of the sample object, the data in the sample object, the sample object execute the friendship The execution moment of interoperability, the label of the interactive information of the interactive operation, the sample object；

The sample object is respectively converted into corresponding subvector according to the data type；

It combines the subvector and obtains the feature vector of the sample object.

8. according to the method described in claim 6, it is characterized in that, the feature using the multiple ordered sample object to Amount, train to obtain the push model by machine include：

The feature vector of the multiple ordered sample object is imported into the push model successively, obtains multiple forecast sample objects Feature vector；

Multiple feature vector errors are obtained successively according to the feature vector of the multiple forecast sample object, wherein the feature Vector error be used to indicate the feature vector of the forecast sample object and the sample object of next importing feature vector it Between error；

Model is pushed according to described in described eigenvector error transfer factor, so that the feature for the push model that training obtains Vector error is less than third threshold value.

9. a kind of data presentation device, which is characterized in that including：

First acquisition unit, for obtaining data update instruction, wherein the data update instruction is for updating the first client Upper shown data object；

Response unit obtains the historical operation record of first client for responding the data update instruction, wherein institute Stating historical operation record includes：According to the interactive operation of data object executed included in first client, and The history object feature vector of generation；

First processing units are imported for the part history object feature vector in recording the historical operation and are pushed away successively Model is sent, obtains prediction characteristics of objects vector, wherein the push model is logical using the feature vector of multiple ordered sample objects Machine is crossed to train to obtain；

Display unit, for determining updated target object according to the prediction characteristics of objects vector, and in first visitor The target object is shown in the end of family.

10. device according to claim 9, which is characterized in that the display unit includes：

First acquisition module, for obtaining and the matched data-object library of the first client, wherein the data-object library The feature vector of middle storage the first client total data object to display；

Judgment module, for comparing the feature vector of total data object described in the data-object library and the prediction successively Characteristics of objects vector；

First determining module, for determining the target object according to the result of comparison.

11. device according to claim 10, which is characterized in that first determining module includes：

First acquisition submodule, for from the feature vector of the total data object, obtaining and the prediction characteristics of objects Vectorial distance is less than the feature vector of first threshold；

First determination sub-module, the corresponding data object of feature vector for the distance to be less than to the first threshold determine For the target object.

12. device according to claim 9, which is characterized in that the first processing units include：

Second determining module, for determining the part history in the historical operation record for importing the push model Characteristics of objects vector, wherein the part history object feature vector includes：The historical operation record in, executed and Execution interval is less than N number of history object feature vector corresponding to continuous N number of interactive operation of second threshold, the execution Between be divided into execution time interval of the moment away from current time for executing the interactive operation, the N is natural number；

First processing module, for the identified part history object feature vector to be imported the push model successively, Obtain the prediction characteristics of objects vector.

13. device according to claim 12, which is characterized in that the first processing module includes：

First processing submodule obtains N number of for N number of history object feature vector to be imported the push model successively The prediction characteristics of objects vector；

Second processing submodule, for N-1 before being abandoned according to the execution moment prediction characteristics of objects vectors, by N A prediction characteristics of objects vector, the prediction for determining the target object as the push model output Characteristics of objects vector.

14. device according to claim 9, which is characterized in that further include：

Second acquisition unit, for before acquisition data update instruction, obtaining the spy of the multiple ordered sample object Sign vector；

Second processing unit trains to obtain described for the feature vector using the multiple ordered sample object by machine Push model.

15. device according to claim 14, which is characterized in that the second acquisition unit includes：

Second acquisition module, for obtaining multiple sample objects successively；

Second processing module, for executing following operation respectively to each sample object in the multiple sample object：

It combines the subvector and obtains the feature vector of the sample object.

16. device according to claim 14, which is characterized in that the second processing unit includes：

Third processing module is obtained for the feature vector of the multiple ordered sample object to be imported the push model successively To the feature vector of multiple forecast sample objects；

Third acquisition module is missed for obtaining multiple feature vectors successively according to the feature vector of the multiple forecast sample object Difference, wherein described eigenvector error is used to indicate the sample of the feature vector and next importing of the forecast sample object Error between the feature vector of object；

Fourth processing module, for pushing model according to described in described eigenvector error transfer factor so that training obtain it is described The described eigenvector error for pushing model is less than third threshold value.

17. a kind of storage medium, which is characterized in that the storage medium includes the program of storage, wherein when described program is run Execute the method described in any one of claim 1 to 8.

18. a kind of electronic device, including memory, processor and it is stored on the memory and can transports on the processor Capable computer program, which is characterized in that the processor executes the claim 1 to 8 times by the computer program Method described in one.