CN105447005A - Object push method and device - Google Patents

Abstract

The embodiment of the invention provides an object push method. The method comprises following steps: obtaining corresponding keywords of a client; obtaining M entities representing keywords, wherein M is an integer larger than or equal to 2; adjusting first probability of the ith entity of the M entities based on historical records of a user for the client in order to obtain second probability of the ith entity; sequencing target objects matching the M entities based on second probabilities of the M entities in order to obtain a sequencing result; and pushing the sequencing result to the client. The embodiment of the invention further provides an object push device. The technical scheme provided by the embodiment of the invention helps to improve reliability of object push.

Description

A kind of Object Push method and device

[technical field]

The present invention relates to internet, applications technology, particularly relate to a kind of Object Push method and device.

[background technology]

In current retrieval technique and recommended technology, often occur that a keyword can refer to the situation of multiple entity simultaneously.Such as, " apple " both can refer to a kind of food, also a kind of mobile phone can be referred to, in this case, according to the term of user's input, when pushing the destination object of retrieval to user, or, when initiatively recommending destination object to user, need first to determine to push the destination object matched with which entity, or need to determine preferentially to push the destination object matched with which entity.

In prior art, utilize the word in the contextual information of keyword, mate in entity storehouse, to obtain the similarity of word and each entity, then when showing the destination object matched with described keyword, the destination object that the entity that in preferential these destination objects of display, similarity is maximum matches.But, this method depends on the contextual information of keyword, each entity that keyword refers to cannot be distinguished effectively in keyword for shortage contextual information, therefore, the Reliability comparotive of the Object Push method of the entity referred to based on contextual information determination keyword in prior art is low.

[summary of the invention]

In view of this, embodiments provide a kind of Object Push method and device, the reliability improving Object Push can be realized.

Embodiments provide a kind of Object Push method, comprising:

Obtain the keyword that client is corresponding;

Obtain M the entity that described keyword refers to, M be greater than or equal to 2 integer;

According to the historical record of the user of the described client of use, the first probability of i-th entity in a described M entity is adjusted, to obtain the second probability of described i-th entity;

According to the second probability of a described M entity, the destination object matched with a described M entity is sorted, to obtain ranking results;

To ranking results described in described client push.

In said method, described method comprises:

Obtain the term that described client sends, using described term as keyword corresponding to described client; Or,

According to the historical record of the user of the described client of use, obtain the keyword that described client is corresponding.

In said method, the historical record of the described user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity, comprising:

Obtain the first probability of i-th entity in a described M entity; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M;

If use the user of described client to there is corresponding historical record, according to the first probability of described i-th entity, obtain the weighted value of described i-th entity in described historical record that described keyword refers to;

According to the weighted value of described i-th entity in described historical record, obtain the cumulative sum of the weighted value of a described M entity in described historical record;

The ratio of the cumulative sum of the weighted value of described i-th entity referred to according to described keyword in described historical record and the weighted value of a described M entity in described historical record, obtains the second probability of described i-th entity.

In said method, the first probability of i-th entity in the described M of a described acquisition entity, comprising:

Obtain the middle probability of i-th entity in a described M entity, described middle probability equals first probability of described i-th entity last time or equals default probability;

According to the middle probability of described i-th entity, obtain the weighted value of described i-th entity in the contextual information preset that described keyword refers to;

According to the weighted value of described i-th entity in described contextual information, obtain the cumulative sum of the weighted value of a described M entity in described contextual information;

The weighted value of described i-th entity referred to according to described keyword in the contextual information preset and the ratio of cumulative sum of the weighted value of a described M entity in described contextual information, obtain the first probability of described i-th entity.

In said method, described the first probability according to described i-th entity, obtains the weighted value of described i-th entity in described historical record that described keyword refers to, comprising:

The weighted value of described i-th entity in described historical record utilizing following formula to obtain described keyword to refer to:

$S\left(e_i^{\′}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\Σ}}}{s}_t\left(e_i^{\′}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P historical record, P be greater than or equal to 1 integer; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P historical record in t historical record, the value of t is the integer in 1 to P;

Wherein, described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _iweighted value in a described P historical record in t historical record:

$s_t\left(e_i^{\′}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\Σ}}}[W\left(n_{e_j}\right)\×p\left(e_j\right|n_{e_j})\×q^d(e_j,e_i^{\′})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with, R be greater than or equal to 2 integer; represent the entity e that described keyword k is relevant _jtitle weighted value; Q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance.

In said method, the historical record of the described user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity, comprising:

Obtain the first probability of i-th entity in a described M entity;

If use the user of described client there is no corresponding historical record, using second probability of the first probability of i-th entity in a described M entity as described i-th entity.

In said method, described the second probability according to a described M entity, sorts to the destination object matched with a described M entity, to obtain ranking results, comprising:

According to and the temperature information of destination object that matches of described i-th entity and the degree of correlation of destination object that matches with described i-th entity in the second probability of i-th entity at least one and a described M entity, the destination object that a described M entity matches is sorted, to obtain ranking results.

The embodiment of the present invention additionally provides a kind of Object Push device, comprising:

First acquiring unit, for obtaining keyword corresponding to client;

Second acquisition unit, for obtaining M the entity that described keyword refers to, M be greater than or equal to 2 integer;

Probability generation unit, for the historical record of the user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity;

Object order unit, for the second probability according to a described M entity, sorts to the destination object matched with a described M entity, to obtain ranking results;

Object Push unit, for ranking results described in described client push.

In said apparatus, described first acquiring unit specifically for:

Obtain the term that described client sends, using described term as keyword corresponding to described client; Or,

According to the historical record of the user of the described client of use, obtain the keyword that described client is corresponding.

In said apparatus, described probability generation unit specifically for:

Obtain the first probability of i-th entity in a described M entity; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M;

If use the user of described client to there is corresponding historical record, according to the first probability of described i-th entity, obtain the weighted value of described i-th entity in described historical record that described keyword refers to;

According to the weighted value of described i-th entity in described historical record, obtain the cumulative sum of the weighted value of a described M entity in described historical record;

The ratio of the cumulative sum of the weighted value of described i-th entity referred to according to described keyword in described historical record and the weighted value of a described M entity in described historical record, obtains the second probability of described i-th entity.

In said apparatus, in the described M of a described acquisition entity, the first probability of i-th entity is specially:

Obtain the middle probability of i-th entity in a described M entity, described middle probability equals first probability of described i-th entity last time or equals default probability;

According to the middle probability of described i-th entity, obtain the weighted value of described i-th entity in the contextual information preset that described keyword refers to;

According to the weighted value of described i-th entity in described contextual information, obtain the cumulative sum of the weighted value of a described M entity in described contextual information;

The weighted value of described i-th entity referred to according to described keyword in the contextual information preset and the ratio of cumulative sum of the weighted value of a described M entity in described contextual information, obtain the first probability of described i-th entity.

In said apparatus, described the first probability according to described i-th entity, obtains the weighted value of described i-th entity in described historical record that described keyword refers to and is specially:

The weighted value of described i-th entity in described historical record utilizing following formula to obtain described keyword to refer to:

$S\left(e_i^{\&prime;}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\&Sigma;}}}{s}_t\left(e_i^{\&prime;}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P historical record, P be greater than or equal to 1 integer; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P historical record in t historical record, the value of t is the integer in 1 to P;

Wherein, described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _iweighted value in a described P historical record in t historical record:

$s_t\left(e_i^{\&prime;}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\&Sigma;}}}[W\left(n_{e_j}\right)\&times;p\left(e_j\right|n_{e_j})\&times;q^d(e_j,e_i^{\&prime;})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with, R be greater than or equal to 2 integer; represent the entity e that described keyword k is relevant _jtitle weighted value; Q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance.

In said apparatus, described probability generation unit specifically for:

Obtain the first probability of i-th entity in a described M entity;

If use the user of described client there is no corresponding historical record, using second probability of the first probability of i-th entity in a described M entity as described i-th entity.

In said apparatus, described object order unit specifically for:

According to and the temperature information of destination object that matches of described i-th entity and the degree of correlation of destination object that matches with described i-th entity in the second probability of i-th entity at least one and a described M entity, the destination object that a described M entity matches is sorted, to obtain ranking results.

As can be seen from the above technical solutions, the embodiment of the present invention has following beneficial effect:

In the technical scheme that the embodiment of the present invention provides, utilize the prior probability using the historical record of the user of client and keyword to refer to entity, determine the entity that keyword refers to, thus realize the destination object of destination object as preferential display of the entity referred to by keyword, and be pushed to client, compared with the Object Push method of the entity referred to based on contextual information determination keyword in prior art, when lacking contextual information, still the entity that in current scene, keyword refers to can simple and effectively be determined, and complete the propelling movement of the destination object of the entity that keyword refers to, therefore the reliability improving Object Push can be realized.

[accompanying drawing explanation]

In order to be illustrated more clearly in the technical scheme of the embodiment of the present invention, be briefly described to the accompanying drawing used required in embodiment below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 is the system of the technical scheme use that the embodiment of the present invention provides;

Fig. 2 is the schematic flow sheet of the Object Push method that the embodiment of the present invention provides;

Fig. 3 is the first exemplary plot of the knowledge mapping that the embodiment of the present invention provides;

Fig. 4 is the second exemplary plot of the knowledge mapping that the embodiment of the present invention provides;

Fig. 5 is the functional block diagram of the Object Push device that the embodiment of the present invention provides.

[embodiment]

Technical scheme for a better understanding of the present invention, is described in detail the embodiment of the present invention below in conjunction with accompanying drawing.

Should be clear and definite, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making other embodiments all obtained under creative work prerequisite, belong to the scope of protection of the invention.

The term used in embodiments of the present invention is only for the object describing specific embodiment, and not intended to be limiting the present invention." one ", " described " and " being somebody's turn to do " of the singulative used in the embodiment of the present invention and appended claims is also intended to comprise most form, unless context clearly represents other implications.

Should be appreciated that term "and/or" used herein is only a kind of incidence relation describing affiliated partner, can there are three kinds of relations in expression, and such as, A and/or B, can represent: individualism A, exists A and B simultaneously, these three kinds of situations of individualism B.In addition, character "/" herein, general expression forward-backward correlation is to the relation liking a kind of "or".

Although should be appreciated that may adopt in embodiments of the present invention term first, second to describe probability, these probability should not be limited to these terms.These terms are only used for probability to be distinguished from each other out.Such as, when not departing from embodiment of the present invention scope, the first probability also can be called as the second probability, and similarly, the second probability also can be called as the first probability.

Depend on linguistic context, word as used in this " if " can be construed as into " ... time " or " when ... time " or " in response to determining " or " in response to detection ".Similarly, depend on linguistic context, phrase " if determination " or " if detecting (the conditioned disjunction event of statement) " can be construed as " when determining " or " in response to determining " or " when detecting (the conditioned disjunction event of statement) " or " in response to detection (the conditioned disjunction event of statement) ".

As shown in Figure 1, primarily of client and server composition, the method and apparatus that the embodiment of the present invention provides realizes at server side the system that the technical scheme that the embodiment of the present invention provides uses, mainly to client push destination object.The function of the embodiment of the present invention to client does not change.

Be understandable that, described client can comprise the client in all terminals, and described terminal can comprise personal computer (PersonalComputer, PC), notebook computer, mobile phone or panel computer.

The embodiment of the present invention provides a kind of Object Push method, please refer to Fig. 2, the schematic flow sheet of its Object Push method provided for the embodiment of the present invention, and as shown in the figure, the method comprises the following steps:

S201, obtains the keyword that client is corresponding.

Concrete, in the embodiment of the present invention, the method that server obtains keyword corresponding to client can comprise following two kinds:

The first: server can obtain the term that described client sends, using this term as keyword corresponding to described client.Wherein, the term that described client sends is that user inputs in described client, retrieves according to this keyword in order to trigger server, to obtain the destination object that the entity that refers to this term matches.

The second: server can obtain the historical record of the user using described client in nearest a period of time, then according to this historical record, obtain the term using the user of described client to input within nearest a period of time, URL(uniform resource locator) (the UniformResourceLocator that user clicks, URL) keyword etc. in the heading message in, using these terms and/or keyword as keyword corresponding to described client, in order to realize server when user does not input term on the client, can realize initiatively recommending destination object to client.

S202, obtains M the entity that described keyword refers to, M be greater than or equal to 2 integer.

Concrete, for described keyword, the knowledge mapping that server by utilizing is preset obtains M the entity that this keyword refers to, M be greater than or equal to 2 integer.

Such as, please refer to Fig. 3, first exemplary plot of its knowledge mapping provided for the embodiment of the present invention, as shown in the figure, keyword is fried rice with eggs, utilize keyword " fried rice with eggs " to mate in knowledge mapping, obtain 2 entities that " fried rice with eggs " refers to, comprise entity " food/fried rice with eggs " and entity " music/fried rice with eggs ".

Again such as, please refer to Fig. 4, second exemplary plot of its knowledge mapping provided for the embodiment of the present invention, as shown in the figure, keyword is Beyond, utilize keyword " Beyond " to mate in knowledge mapping, obtain 2 entities that " Beyond " refers to, comprise entity " music/Beyond " and entity " film/Beyond ".

S203, according to the historical record of the user of the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity.

Concrete, server, according to the historical record of the user of the described client of use, adjusts the first probability of i-th entity in a described M entity, can comprise following two kinds with the method for the second probability obtaining described i-th entity:

The first: server needs the first probability obtaining i-th entity in a described M entity in advance; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M.Then, server judges to use the user of described client whether to there is corresponding historical record, if use the user of described client there is no corresponding historical record, server using the first probability of i-th entity in a described M entity directly as the second probability of described i-th entity.Be equivalent to, when lacking the historical record of user, first probability of server to i-th entity obtained carries out zero adjustment.

The second: first, server needs the first probability obtaining i-th entity in a described M entity in advance; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M.Then, server judges to use the user of described client whether to there is corresponding historical record, if use the user of described client to there is corresponding historical record, server, according to the first probability of described i-th entity, obtains the weighted value of described i-th entity in described historical record that described keyword refers to; Again according to the weighted value of described i-th entity in described historical record, obtain the cumulative sum of the weighted value of a described M entity in described historical record.Finally, the ratio of the cumulative sum of the weighted value of described i-th entity that server refers to according to described keyword in described historical record and the weighted value of a described M entity in described historical record, obtains the second probability of described i-th entity.

Wherein, the information such as term that this user inputs in described client and the URL that user clicks that uses the historical record of the user of described client to comprise.Here, use the historical record of user as contextual information, can avoid, because single heading message causes the situation of contextual information deficiency, the problem effectively cannot distinguishing each entity that keyword refers to brought, the accuracy distinguishing the entity that keyword refers to can being improved.

Wherein, the method that server obtains the first probability of i-th entity in a described M entity is similar to the method that server obtains the second probability of i-th entity, difference is that the first probability of i-th entity obtains according to the weighted value of i-th entity in the contextual information preset, for the user of each use client, first probability of i-th entity is identical, and the second probability of i-th entity obtains according to the weighted value of i-th entity in the historical record of user using active client, for the user of each use client, second probability of i-th entity is different, because use the historical record of the user of client different.

Illustrate, the method that server obtains the first probability of i-th entity in a described M entity can comprise: first, server obtains the middle probability of i-th entity in a described M entity, wherein, this middle probability can equal first probability of described i-th entity last time or equal default probability.Then, server, according to the middle probability of described i-th entity, obtains the weighted value of described i-th entity in the contextual information preset that described keyword refers to; Again according to the weighted value of described i-th entity in described contextual information, obtain the cumulative sum of the weighted value of a described M entity in described contextual information.Finally, the ratio of the cumulative sum of the weighted value of described i-th entity that server refers to according to described keyword in the contextual information preset and the weighted value of a described M entity in described contextual information, obtains the first probability of described i-th entity.

Such as, below with keyword k, it is example that this keyword k can refer to M entity, and the historical record according to the user using described client is described, first probability of i-th entity in a described M entity is adjusted, to obtain the method for the second probability of described i-th entity:

Following formula can be utilized to obtain the second probability of i-th entity:

$p\left(e_i^{\&prime;}\right|k)=S\left(e_i^{\&prime;}\right|k)/\underset{x=1}{\overset{M}{\mathrm{\&Sigma;}}}S\left(e_x\right|k)$

Wherein, p (e ' _i| k) represent i-th entity e ' in M entity _ithe second probability, S (e ' _i| k) represent described i-th entity e ' that keyword k refers to _iweighted value in described historical record.

Wherein, represent the cumulative sum of the weighted value of a described M entity in described historical record.

The weighted value of described i-th entity in described historical record that following formula can be utilized to obtain described keyword refer to:

$S\left(e_i^{\&prime;}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\&Sigma;}}}{s}_t\left(e_i^{\&prime;}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P historical record, P be greater than or equal to 1 integer; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P historical record in t historical record, the value of t is the integer in 1 to P.

Understandable, S (e ' _i| k) be the equal of described i-th entity e ' that keyword k refers to _ithe cumulative sum of the weighted value in each historical record.

Described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _iweighted value in a described P historical record in t historical record:

$s_t\left(e_i^{\&prime;}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\&Sigma;}}}[W\left(n_{e_j}\right)\&times;p\left(e_j\right|n_{e_j})\&times;q^d(e_j,e_i^{\&prime;})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with, R be greater than or equal to 2 integer; represent the entity e that described keyword k is relevant _jtitle weighted value, value can arrange in advance, if do not pre-set, be then defaulted as q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance, when obtaining this distance, can knowledge mapping be regarded as directionless collection of illustrative plates, then obtaining entity e _jcorresponding node and entity e ' _ireached at step number between corresponding node, can reach this at step number as distance, especially, as sporocarp e _jwith entity e ' _ifor same entity, then d (e _j, e ' _i)=0; If two nodes are unreachable, then d (e _j, e ' _i)=∞, now, q^d (e _j, e ' _i)=0.

Wherein, the entity that described keyword k is correlated with can comprise the entity that keyword k can refer to, the entity that other keywords except this keyword k of comprising in historical record can refer to can also be comprised, such as, the entity that can refer to other keywords of keyword k co-occurrence in heading message.Such as, historical record comprises " fried rice with eggs " and " Beyond ", the entity that " fried rice with eggs " is relevant can comprise " food/fried rice with eggs ", " music/fried rice with eggs ", " music/Beyond " and " film/Beyond ", wherein, " food/fried rice with eggs " and " music/fried rice with eggs " is the entity that keyword " fried rice with eggs " can refer to, and " music/Beyond " and " film/Beyond " is the entity that in historical record, keyword " Beyond " can refer to except keyword " fried rice with eggs ".

Again such as, below with keyword k, it is example that this keyword k can refer to M entity, and the method for the first probability obtaining i-th entity in M entity is described:

Following formula can be utilized to obtain the first probability of i-th entity:

$p\left(e_i^{\&prime;}\right|k)=S\left(e_i^{\&prime;}\right|k)/\underset{x=1}{\overset{M}{\mathrm{\&Sigma;}}}S\left(e_x\right|k)$

Wherein, p (e ' _i| k) represent i-th entity e ' in M entity _ithe first probability, S (e ' _i| k) represent described i-th entity e ' that keyword k refers to _iweighted value in the contextual information preset.

Wherein, represent the cumulative sum of the weighted value of a described M entity in the contextual information preset.

Wherein, described contextual information can comprise term near the heading message of keyword k place webpage, keyword k and the pushed information etc. relevant to keyword k.

The weighted value of described i-th entity in contextual information that following formula can be utilized to obtain described keyword k refer to:

$S\left(e_i^{\&prime;}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\&Sigma;}}}{s}_t\left(e_i^{\&prime;}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P contextual information; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P contextual information in t contextual information, the value of t is the integer in 1 to P.

Understandable, S (e ' _i| k) be the equal of described i-th entity e ' that keyword k refers to _ithe cumulative sum of the weighted value in each contextual information.

Described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _jweighted value in a described P contextual information in t contextual information:

$s_t\left(e_i^{\&prime;}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\&Sigma;}}}[W\left(n_{e_j}\right)\&times;p\left(e_j\right|n_{e_j})\&times;q^d(e_j,e_i^{\&prime;})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with; represent the entity e that described keyword k is relevant _jtitle weighted value, value can arrange in advance, if do not pre-set, be then defaulted as q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance, when obtaining this distance, can knowledge mapping be regarded as directionless collection of illustrative plates, then obtaining entity e _jcorresponding node and entity e ' _ireached at step number between corresponding node, can reach this at step number as distance, especially, as sporocarp e _jwith entity e ' _ifor same entity, then d (e _j, e ' _i)=0; If two nodes are unreachable, then d (e _j, e ' _i)=∞, now, q^d (e _j, e ' _i)=0.

Wherein, the keyword k title of entity of being correlated with refer to entity e _jprior probability can equal the first probability that this entity last time calculates, or, when first time calculates the first probability of this entity, the title of the entity that keyword k is correlated with refer to entity e _jprior probability equal default probability.

Such as, keyword k refers to entity e ' _iprior probability can be p (e ' _i| k)=1/M.

Understandable, iterative computation mode can be utilized, the entity e that the last time can be calculated _jthe first probability, as this computational entity e _jthe first probability time need the keyword k utilized to be correlated with the title of entity refer to entity e _jprior probability, until the numerical value of the square error of the first probability that several times calculate is less than default error threshold or iterations when reaching default iteration threshold, stop iterative computation.

Wherein, the entity that described keyword k is correlated with can comprise other keywords except this keyword k comprised in the entity and contextual information that keyword k can refer to can refer to entity.Such as, contextual information comprises " fried rice with eggs " and " Beyond ", the entity that " fried rice with eggs " is relevant can comprise " food/fried rice with eggs ", " music/fried rice with eggs ", " music/Beyond " and " film/Beyond ", wherein, " food/fried rice with eggs " and " music/fried rice with eggs " is the entity that keyword " fried rice with eggs " can refer to, and " music/Beyond " and " film/Beyond " is the entity that the keyword " Beyond " in historical record except keyword " fried rice with eggs " can refer to.

Understandable, obtain the first probability of entity in advance, namely keyword refers to the prior probability of this entity, can in retrieval scene or under recommending scene, when not or when lacking contextual information, can using the first probability of entity obtained in advance directly as the second probability needing during sequence to utilize, with realize when not or lack contextual information time still effectively can distinguish each entity that keyword refers to, the reliability of raising Object Push method.

S204, according to the second probability of a described M entity, sorts to the destination object matched with a described M entity, to obtain ranking results.

Concrete, server is after the second probability obtaining i-th entity in M entity, and the value due to i is the integer in 1 to M, so server can the second probability of each entity in M entity.

First, server, after the second probability obtaining each entity in M entity, according to each entity in M entity, can obtain the destination object matched with each entity.

Then, server, according to the second probability of at least one and this entity in the temperature information of the destination object matched with each entity and the degree of correlation of destination object that matches with each entity, obtains the weighted value of each destination object in all destination objects.

Such as, the product of the second probability of the entity that can mate with this destination object according to the temperature information of destination object, obtains the weighted value of this destination object.

Again such as, the product of the second probability of the entity that can mate with this destination object according to the degree of correlation of destination object, obtains the weighted value of this destination object.

Finally, server, according to the weighted value of each destination object, according to the order that weighted value is descending, sorts to all destination objects that a described M entity matches, to obtain ranking results.

S205, to ranking results described in described client push.

Concrete, this ranking results, after the ranking results obtaining the destination object matched with a described M entity, is sent to client by server, to make client can show this ranking results, thus realizes pushing destination object to user.

Embodiment one

For keyword " fried rice with eggs ", this keyword may refer to food, also likely refers to the MV works " fried rice with eggs " of Yu Chengqing.

Wherein, the heading message " Foods-how to cook fried rice with eggs " of webpage can be there is, also can there is term " Yu Cheng celebrates fried rice with eggs ".The heading message of relevant webpage can be utilized as contextual information.

For above-mentioned keyword, if only according to heading message, which be sometimes difficult to determine preferentially to provide to user the destination object matched with the entity referred to, if with reference to other contextual informations, just can determine referred to entity, as heading message be " fried rice with eggs " picture the destination object of being correlated with, based on the contextual information of this picture, what can obtain that heading message is that the picture of " fried rice with eggs " refers to easily is food, instead of music.As term " Yu Chengqing " and " fried rice with eggs " of user successively input, what the historical record according to this user can obtain that term " fried rice with eggs " refers to is music, instead of food.

As shown in Figure 3, the textual representation entity in figure in circle, the textual representation keyword in rectangle frame, the knowledge mapping shown in Fig. 3 can provide the topology of the connection between entity and entity, between entity and keyword.

As shown in the knowledge mapping on the left side in Fig. 3, in the context comprising keyword " Ha Lin ", " fried rice with eggs " and " only having for you " in information, knowledge mapping is located these keywords.Because " Ha Lin " and " only having for you " can be the support that " song/Yu Chengqing/fried rice with eggs " this entity provides weighted value, and " food/fried rice with eggs " this entity lacks the support of weighted value, therefore, can determine that keyword " fried rice with eggs " refers to entity " song/Yu Chengqing/fried rice with eggs ".

As shown in the knowledge mapping on the right in Fig. 3, in the contextual information comprising keyword " cuisines " and " fried rice with eggs ", knowledge mapping is located these keywords, because " cuisines " can be the support that entity " food/fried rice with eggs " provides weighted value, therefore according to this contextual information, can determine that keyword " fried rice with eggs " refers to entity " food/fried rice with eggs ".

As shown in Figure 3, if two entities that keyword " fried rice with eggs " refers to are owing to lacking contextual information, cause the support all lacking weighted value, can according to the first probability of each entity obtained in advance, namely keyword " fried rice with eggs " refers to the prior probability of each entity, determines that entity that keyword " fried rice with eggs " refers to is which entity in two entities.Such as, if there is prior probability p (food/fried rice with eggs | fried rice with eggs) > prior probability p (song/Yu Chengqing/fried rice with eggs | fried rice with eggs), determine that the entity that " fried rice with eggs " refers to is " food/fried rice with eggs ".

Embodiment two

As shown in Figure 4, be " fried rice with eggs " with keyword below, contextual information is heading message " fried rice with eggs Beyond " is example, is described in the method for the weighted value of this contextual information obtaining the entity that keyword " fried rice with eggs " refers to.

As shown in Figure 4, by mating in knowledge mapping, two entities that keyword " fried rice with eggs " can refer to can be obtained, i.e. " food/fried rice with eggs " and " music/fried rice with eggs ".In like manner, by mating in knowledge mapping, the related entities of keyword " fried rice with eggs " can be obtained, two entities that namely can refer to other keywords " Beyond " of keyword " fried rice with eggs " co-occurrence in contextual information, i.e. " music/Beyond " and " film/Beyond ".

Target obtains following weighted value:

S _t(food/fried rice with eggs | fried rice with eggs)

S _t(music/fried rice with eggs | fried rice with eggs)

S _t(music/Beyond | Beyond)

S _t(film/Beyond | Beyond)

Can pre-set:

W (fried rice with eggs)=1

W (Beyond)=1

Can obtain the first probability of precalculated entity, namely keyword refers to the prior probability of this entity, that is:

P (food/fried rice with eggs | fried rice with eggs)=0.5

P (music/fried rice with eggs | fried rice with eggs)=0.5

P (music/Beyond | Beyond)=0.5

P (film/Beyond | Beyond)=0.5

Pre-set decay factor q=0.6.

According to the knowledge mapping shown in Fig. 4, following distance can be obtained:

D (food/fried rice with eggs, music/Beyond)=∞

D (food/fried rice with eggs, film/Beyond)=∞

D (music/fried rice with eggs, music/Beyond)=2

D (music/fried rice with eggs, film/Beyond)=∞

D (food/fried rice with eggs, food/fried rice with eggs)=0

D (food/fried rice with eggs, music/fried rice with eggs)=∞

D (music/fried rice with eggs, music/fried rice with eggs)=0

D (music/Beyond, music/Beyond)=0

D (music/Beyond, film/Beyond)=∞

D (film/Beyond, film/Beyond)=0

Based on above-mentioned information, obtain the weighted value of entity " food/fried rice with eggs " in above-mentioned heading message " fried rice with eggs Beyond " that keyword " fried rice with eggs " refers to:

S _t(food/fried rice with eggs | fried rice with eggs)=W (fried rice with eggs) × p (food/fried rice with eggs | fried rice with eggs)

× q^d (food/fried rice with eggs, food/fried rice with eggs)

+ W (fried rice with eggs) × p (music/fried rice with eggs | fried rice with eggs) × q^d (music/fried rice with eggs, food/fried rice with eggs)

+ W (Beyond) × p (music/Beyond | Beyond)

× q^d (music/Beyond, food/fried rice with eggs)

+ W (Beyond) × p (film/Beyond | Beyond)

× q^d (film/Beyond, food/fried rice with eggs)=1 × 0.5 × 1+0+0+0

＝0.5

Obtain the weighted value of entity " music/fried rice with eggs " in above-mentioned heading message " fried rice with eggs Beyond " that keyword " fried rice with eggs " refers to:

S _t(music/fried rice with eggs | fried rice with eggs)=W (fried rice with eggs) × p (food/fried rice with eggs | fried rice with eggs)

× q^d (food/fried rice with eggs, music/fried rice with eggs)

+ W (fried rice with eggs) × p (music/fried rice with eggs | fried rice with eggs) × q^d (music/fried rice with eggs, music/fried rice with eggs)

+ W (Beyond) × p (music/Beyond | Beyond)

× q^d (music/Beyond, music/fried rice with eggs)

+ W (Beyond) × p (film/Beyond | Beyond)

× q^d (film/Beyond, music/fried rice with eggs)

＝0+1×0.5×1+1×0.5×0.6^2+0＝0.68

Obtain the weighted value of entity " music/Beyond " in above-mentioned heading message " fried rice with eggs Beyond " that keyword " fried rice with eggs " is relevant:

S _t(music/Beyond | Beyond)=W (fried rice with eggs) × p (food/fried rice with eggs | fried rice with eggs)

× q^d (food/fried rice with eggs, music/Beyond)

+ W (fried rice with eggs) × p (music/fried rice with eggs | fried rice with eggs) × q^d (music/fried rice with eggs, music/Beyond)

+ W (Beyond) × p (music/Beyond | Beyond)

× q^d (music/Beyond, music/Beyond)

+ W (Beyond) × p (film/Beyond | Beyond)

× q^d (film/Beyond, music/Beyond)

＝0+1×0.5×0.6^2+1×0.5×1+0＝0.68

Obtain the weighted value of entity " film/Beyond " in above-mentioned heading message " fried rice with eggs Beyond " that keyword " fried rice with eggs " is relevant:

S _t(film/Beyond | Beyond)=W (fried rice with eggs) × p (food/fried rice with eggs | fried rice with eggs)

× q^d (food/fried rice with eggs, film/Beyond)

+ W (fried rice with eggs) × p (music/fried rice with eggs | fried rice with eggs) × q^d (music/fried rice with eggs, film/Beyond)

+ W (Beyond) × p (music/Beyond | Beyond)

× q^d (music/Beyond, film/Beyond)

+ W (Beyond) × p (film/Beyond | Beyond)

× q^d (film/Beyond, film/Beyond)=0+0+0+1 × 0.5 × 1

＝0.5

The embodiment of the present invention provides the device embodiment realizing each step and method in said method embodiment further.

Please refer to Fig. 5, the functional block diagram of its Object Push device provided for the embodiment of the present invention.As shown in the figure, this device comprises:

First acquiring unit 501, for obtaining keyword corresponding to client;

Second acquisition unit 502, for obtaining M the entity that described keyword refers to, M be greater than or equal to 2 integer;

Probability generation unit 503, for the historical record of the user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity; Object order unit 504, for the second probability according to a described M entity, sorts to the destination object matched with a described M entity, to obtain ranking results;

Object Push unit 505, for ranking results described in described client push.

Preferably, described first acquiring unit 501 specifically for:

Obtain the term that described client sends, using described term as keyword corresponding to described client; Or,

According to the historical record of the user of the described client of use, obtain the keyword that described client is corresponding.

Preferably, described probability generation unit 503 specifically for:

Obtain the first probability of i-th entity in a described M entity; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M;

If use the user of described client to there is corresponding historical record, according to the first probability of described i-th entity, obtain the weighted value of described i-th entity in described historical record that described keyword refers to;

According to the weighted value of described i-th entity in described historical record, obtain the cumulative sum of the weighted value of a described M entity in described historical record;

The ratio of the cumulative sum of the weighted value of described i-th entity referred to according to described keyword in described historical record and the weighted value of a described M entity in described historical record, obtains the second probability of described i-th entity.

Preferably, in the described M of a described acquisition entity, the first probability of i-th entity is specially:

Obtain the middle probability of i-th entity in a described M entity, described middle probability equals first probability of described i-th entity last time or equals default probability;

According to the middle probability of described i-th entity, obtain the weighted value of described i-th entity in the contextual information preset that described keyword refers to;

According to the weighted value of described i-th entity in described contextual information, obtain the cumulative sum of the weighted value of a described M entity in described contextual information;

The weighted value of described i-th entity referred to according to described keyword in the contextual information preset and the ratio of cumulative sum of the weighted value of a described M entity in described contextual information, obtain the first probability of described i-th entity.

Preferably, described the first probability according to described i-th entity, obtains the weighted value of described i-th entity in described historical record that described keyword refers to and is specially:

The weighted value of described i-th entity in described historical record utilizing following formula to obtain described keyword to refer to:

$S\left(e_i^{\&prime;}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\&Sigma;}}}{s}_t\left(e_i^{\&prime;}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P historical record, P be greater than or equal to 1 integer; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P historical record in t historical record, the value of t is the integer in 1 to P;

Wherein, described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _iweighted value in a described P historical record in t historical record:

$s_t\left(e_i^{\&prime;}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\&Sigma;}}}[W\left(n_{e_j}\right)\&times;p\left(e_j\right|n_{e_j})\&times;q^d(e_j,e_i^{\&prime;})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with, R be greater than or equal to 2 integer; represent the entity e that described keyword k is relevant _jtitle weighted value; Q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance.

Preferably, described probability generate Unit 503 specifically for:

Obtain the first probability of i-th entity in a described M entity;

If use the user of described client there is no corresponding historical record, using second probability of the first probability of i-th entity in a described M entity as described i-th entity.

Preferably, described object order unit 504 specifically for:

According to and the temperature information of destination object that matches of described i-th entity and the degree of correlation of destination object that matches with described i-th entity in the second probability of i-th entity at least one and a described M entity, the destination object that a described M entity matches is sorted, to obtain ranking results.

Because each unit in the present embodiment can perform the method shown in Fig. 2, the part that the present embodiment is not described in detail, can with reference to the related description to Fig. 2.

The technical scheme of the embodiment of the present invention has following beneficial effect:

1, in the technical scheme that the embodiment of the present invention provides, utilize the prior probability using the historical record of the user of client and keyword to refer to entity, determine the entity that keyword refers to, thus realize the destination object of destination object as preferential display of the entity referred to by keyword, and be pushed to client, compared with the Object Push method of the entity referred to based on contextual information determination keyword in prior art, when lacking contextual information, still the entity that in current scene, keyword refers to can simple and effectively be determined, and complete the propelling movement of the destination object of the entity that keyword refers to, therefore the reliability improving Object Push can be realized.

2, the technical scheme that the embodiment of the present invention provides can according to the historical record of user, the Search Requirement of user is analyzed, thus determine the entity of the destination object coupling pushed, so just preferentially can push the destination object of this Entities Matching, make user can obtain required object in time, therefore the technical scheme that the embodiment of the present invention provides can improve the accuracy of the destination object of propelling movement, the running cost reducing retrieval and recommend, and improves recall precision and recommends efficiency.

Those skilled in the art can be well understood to, and for convenience and simplicity of description, the system of foregoing description, the specific works process of device and unit, with reference to the corresponding process in preceding method embodiment, can not repeat them here.

In several embodiment provided by the present invention, should be understood that, disclosed system, apparatus and method, can realize by another way.Such as, device embodiment described above is only schematic, such as, the division of described unit, is only a kind of logic function and divides, and actual can have other dividing mode when realizing, such as, multiple unit or assembly can in conjunction with or another system can be integrated into, or some features can be ignored, or do not perform.Another point, shown or discussed coupling each other or direct-coupling or communication connection can be by some interfaces, and the indirect coupling of device or unit or communication connection can be electrical, machinery or other form.

The described unit illustrated as separating component or can may not be and physically separates, and the parts as unit display can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of unit wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.

In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, also can be that the independent physics of unit exists, also can two or more unit in a unit integrated.Above-mentioned integrated unit both can adopt the form of hardware to realize, and the form that hardware also can be adopted to add SFU software functional unit realizes.

The above-mentioned integrated unit realized with the form of SFU software functional unit, can be stored in a computer read/write memory medium.Above-mentioned SFU software functional unit is stored in a storage medium, comprising some instructions in order to make a computer installation (can be personal computer, server, or network equipment etc.) or processor (Processor) perform the part steps of method described in each embodiment of the present invention.And aforesaid storage medium comprises: USB flash disk, portable hard drive, ROM (read-only memory) (Read-OnlyMemory, ROM), random access memory (RandomAccessMemory, RAM), magnetic disc or CD etc. various can be program code stored medium.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within the scope of protection of the invention.

Claims (14)

1. an Object Push method, is characterized in that, described method comprises:

Obtain the keyword that client is corresponding;

Obtain M the entity that described keyword refers to, M be greater than or equal to 2 integer;

According to the historical record of the user of the described client of use, the first probability of i-th entity in a described M entity is adjusted, to obtain the second probability of described i-th entity;

According to the second probability of a described M entity, the destination object matched with a described M entity is sorted, to obtain ranking results;

To ranking results described in described client push.

2. method according to claim 1, is characterized in that, described method comprises:

Obtain the term that described client sends, using described term as keyword corresponding to described client; Or,

According to the historical record of the user of the described client of use, obtain the keyword that described client is corresponding.

3. method according to claim 1, is characterized in that, the historical record of the described user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity, comprising:

Obtain the first probability of i-th entity in a described M entity; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M;

If use the user of described client to there is corresponding historical record, according to the first probability of described i-th entity, obtain the weighted value of described i-th entity in described historical record that described keyword refers to;

According to the weighted value of described i-th entity in described historical record, obtain the cumulative sum of the weighted value of a described M entity in described historical record;

The ratio of the cumulative sum of the weighted value of described i-th entity referred to according to described keyword in described historical record and the weighted value of a described M entity in described historical record, obtains the second probability of described i-th entity.

4. method according to claim 3, is characterized in that, the first probability of i-th entity in the described M of a described acquisition entity, comprising:

Obtain the middle probability of i-th entity in a described M entity, described middle probability equals first probability of described i-th entity last time or equals default probability;

According to the middle probability of described i-th entity, obtain the weighted value of described i-th entity in the contextual information preset that described keyword refers to;

According to the weighted value of described i-th entity in described contextual information, obtain the cumulative sum of the weighted value of a described M entity in described contextual information;

The weighted value of described i-th entity referred to according to described keyword in the contextual information preset and the ratio of cumulative sum of the weighted value of a described M entity in described contextual information, obtain the first probability of described i-th entity.

5. method according to claim 3, is characterized in that, described the first probability according to described i-th entity, obtains the weighted value of described i-th entity in described historical record that described keyword refers to, comprising:

The weighted value of described i-th entity in described historical record utilizing following formula to obtain described keyword to refer to:

$S\left(e_i^{\&prime;}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\&Sigma;}}}{s}_t\left(e_i^{\&prime;}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P historical record, P be greater than or equal to 1 integer; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P historical record in t historical record, the value of t is the integer in 1 to P;

Wherein, described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _iweighted value in a described P historical record in t historical record:

$s_t\left(e_i^{\&prime;}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\&Sigma;}}}[W\left(n_{e_j}\right)\&times;p\left(e_j\right|n_{e_j})\&times;q^d(e_j,e_i^{\&prime;})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with, R be greater than or equal to 2 integer; represent the entity e that described keyword k is relevant _jtitle weighted value; Q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance.

6. method according to claim 1, is characterized in that, the historical record of the described user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity, comprising:

Obtain the first probability of i-th entity in a described M entity;

If use the user of described client there is no corresponding historical record, using second probability of the first probability of i-th entity in a described M entity as described i-th entity.

7. method according to claim 1, is characterized in that, described the second probability according to a described M entity, sorts, to obtain ranking results, comprising the destination object matched with a described M entity:

According to and the temperature information of destination object that matches of described i-th entity and the degree of correlation of destination object that matches with described i-th entity in the second probability of i-th entity at least one and a described M entity, the destination object that a described M entity matches is sorted, to obtain ranking results.

8. an Object Push device, is characterized in that, described device comprises:

First acquiring unit, for obtaining keyword corresponding to client;

Second acquisition unit, for obtaining M the entity that described keyword refers to, M be greater than or equal to 2 integer;

Probability generation unit, for the historical record of the user according to the described client of use, adjusts the first probability of i-th entity in a described M entity, to obtain the second probability of described i-th entity;

Object order unit, for the second probability according to a described M entity, sorts to the destination object matched with a described M entity, to obtain ranking results;

Object Push unit, for ranking results described in described client push.

9. device according to claim 8, is characterized in that, described first acquiring unit specifically for:

Obtain the term that described client sends, using described term as keyword corresponding to described client; Or,

According to the historical record of the user of the described client of use, obtain the keyword that described client is corresponding.

10. device according to claim 8, is characterized in that, described probability generation unit specifically for:

Obtain the first probability of i-th entity in a described M entity; Wherein, the first probability of described i-th entity is the prior probability that described keyword refers to i-th entity, and the value of i is the integer in 1 to M;

If use the user of described client to there is corresponding historical record, according to the first probability of described i-th entity, obtain the weighted value of described i-th entity in described historical record that described keyword refers to;

According to the weighted value of described i-th entity in described historical record, obtain the cumulative sum of the weighted value of a described M entity in described historical record;

The ratio of the cumulative sum of the weighted value of described i-th entity referred to according to described keyword in described historical record and the weighted value of a described M entity in described historical record, obtains the second probability of described i-th entity.

11. devices according to claim 10, is characterized in that, in the described M of a described acquisition entity, the first probability of i-th entity is specially:

Obtain the middle probability of i-th entity in a described M entity, described middle probability equals first probability of described i-th entity last time or equals default probability;

According to the middle probability of described i-th entity, obtain the weighted value of described i-th entity in the contextual information preset that described keyword refers to;

According to the weighted value of described i-th entity in described contextual information, obtain the cumulative sum of the weighted value of a described M entity in described contextual information;

The weighted value of described i-th entity referred to according to described keyword in the contextual information preset and the ratio of cumulative sum of the weighted value of a described M entity in described contextual information, obtain the first probability of described i-th entity.

12. devices according to claim 10, is characterized in that, described the first probability according to described i-th entity, obtain the weighted value of described i-th entity in described historical record that described keyword refers to and are specially:

The weighted value of described i-th entity in described historical record utilizing following formula to obtain described keyword to refer to:

$S\left(e_i^{\&prime;}\right|k)=\underset{t=1}{\overset{P}{\mathrm{\&Sigma;}}}{s}_t\left(e_i^{\&prime;}\right|k)$

Wherein, S (e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in P historical record, P be greater than or equal to 1 integer; s _t(e ' _i| k) represent described i-th entity e ' that described keyword k refers to _iweighted value in a described P historical record in t historical record, the value of t is the integer in 1 to P;

Wherein, described i-th the entity e ' utilizing following formula to obtain described keyword k to refer to _iweighted value in a described P historical record in t historical record:

$s_t\left(e_i^{\&prime;}\right|k)=\underset{j=1}{\overset{R}{\mathrm{\&Sigma;}}}[W\left(n_{e_j}\right)\&times;p\left(e_j\right|n_{e_j})\&times;q^d(e_j,e_i^{\&prime;})]$

Wherein, R represents the total number of the title of the entity that described keyword k is correlated with, R be greater than or equal to 2 integer; represent the entity e that described keyword k is relevant _jtitle weighted value; Q represents decay factor, and 0 < q < 1; presentation-entity e _jthe first probability; D (e _j, e ' _i) represent in knowledge mapping, the title of the entity that keyword k is correlated with refer to entity e _jwith described i-th entity e ' _ibetween distance.

13. devices according to claim 8, is characterized in that, described probability generation unit specifically for:

Obtain the first probability of i-th entity in a described M entity;

If use the user of described client there is no corresponding historical record, using second probability of the first probability of i-th entity in a described M entity as described i-th entity.

14. devices according to claim 8, is characterized in that, described object order unit specifically for:

According to and the temperature information of destination object that matches of described i-th entity and the degree of correlation of destination object that matches with described i-th entity in the second probability of i-th entity at least one and a described M entity, the destination object that a described M entity matches is sorted, to obtain ranking results.