CN109993568A

CN109993568A - The method and apparatus of information push

Info

Publication number: CN109993568A
Application number: CN201810005740.2A
Authority: CN
Inventors: 汪雷; 刘振杰
Original assignee: Beijing Jingdong Century Trading Co Ltd; Beijing Jingdong Shangke Information Technology Co Ltd
Current assignee: Beijing Jingdong Century Trading Co Ltd; Beijing Jingdong Shangke Information Technology Co Ltd
Priority date: 2018-01-03
Filing date: 2018-01-03
Publication date: 2019-07-09

Abstract

The embodiment of the present application discloses the method and apparatus of information push.One specific embodiment of this method includes: to obtain the article characteristics information for the multiple articles for belonging to same target item classification；Article characteristics information based on each article calculates the first similarity in multiple articles between any two article；Based on each first similarity, multiple articles are clustered to obtain multiple clustering clusters；The distance of cluster centre based on article to be assessed to each clustering cluster, determines target clustering cluster from each clustering cluster；According to the value and the second similarity of each article in target clustering cluster, the value of article to be assessed is determined；The value information of the value of the article to be assessed is used to indicate to user's push.The embodiment reduces the cost of labor of Item Value assessment, improves the accuracy of Item Value assessment, realizes the push rich in targetedly Item Information.

Description

The method and apparatus of information push

Technical field

The invention relates to field of computer technology, and in particular to Internet technical field more particularly to information push away The method and apparatus sent.

Background technique

Economic rapid development bring house property, jewelry, and other items transaction it is increasingly frequent, Item Value is (such as Price) the demand of assessment become more and more vigorous.Currently, the assessment of Item Value usually require to spend it is biggish manually at This, and take a long time, accuracy it is poor, it is difficult to meet the market demand.

By taking house property as an example, in the value assessment to house property, it usually needs real estate feature is collected by manual type, and Afterwards based on real estate feature acquired in manual analysis, to determine the value of house property, it is seen that it needs a large amount of work of artificial participation, Considerably increase cost of labor.Alternatively, in the value assessment of house property, it can also be by artificially selecting real estate feature, then The weight coefficient of each real estate feature is fitted based on the method for multiple linear regression to assess the value of house property, the house property The method of value assessment depends on artificial linear hypothesis, and accuracy is poor, therefore, it is difficult to needed for accurately pushing it to user The house property information wanted.

Summary of the invention

The embodiment of the present application proposes the method and apparatus of information push.

In a first aspect, the embodiment of the present application provides a kind of method of information push, this method comprises: obtain belong to it is same The article characteristics information of multiple articles of target item classification；Article characteristics information based on each article, calculates in multiple articles The first similarity between any two article；Based on each first similarity, multiple articles are clustered to obtain multiple clustering clusters；It is based on Article to be assessed to each clustering cluster cluster centre distance, target clustering cluster is determined from each clustering cluster, wherein to be assessed Article belongs to target item classification；According to the value and the second similarity of each article in target clustering cluster, object to be assessed is determined The value of product, wherein the second similarity is the similarity between each article in article to be assessed and target clustering cluster；To user Push is used to indicate the value information of the value of article to be assessed.

In some embodiments, article characteristics information includes the text information and use for describing the text feature of article In the feature vector of the data characteristics of description article；Article characteristics information based on each article calculates any two in multiple articles The first similarity between a article, comprising: calculate the text information of the first article and the text information of the second article in different dimensional The lower third similarity of degree, wherein the first article and the second article are any two article in multiple articles；Calculate the first object The vector distance of the feature vector of product and the feature vector of the second article under different dimensions is as the 4th similarity；By the first object The average value of the third similarity and the 4th similarity of product and the second article is determined as between first article and the second article First similarity.

In some embodiments, the distance of the cluster centre based on article to be assessed to each clustering cluster, from each clustering cluster Determine target clustering cluster, comprising: special according to the article of the article in the article characteristics information and each clustering cluster of article to be assessed Reference breath, calculates the distance between the cluster centre of the article to be assessed Yu each clustering cluster；It determines to cluster from each clustering cluster Center clustering cluster nearest at a distance from article to be assessed is as target clustering cluster.

In some embodiments, it according to the value and the second similarity of each article in target clustering cluster, determines to be assessed The value of article, comprising: calculate separately the second similarity between each article and article to be assessed in target clustering cluster；According to mesh The average value for marking the product of corresponding second similarity of each article and value in clustering cluster, determines the value of article to be assessed.

In some embodiments, according to the product of corresponding second similarity of each article and value in target clustering cluster Average value determines the value of article to be assessed, comprising: obtains the time difference at the first moment and the second moment, wherein when first Carving is article to be assessed at the time of need estimated value, the second moment be in target clustering cluster any article mark value when It carves；Using the second similarity corresponding with any article in target clustering cluster and the product of time difference as in target clustering cluster The value weight of the article；Each article in target clustering cluster is averaging in the value at the second moment and the product of value weight, Determine article to be assessed in the value at the first moment.

In some embodiments, the text information of the first article and the text information of the second article are calculated under different dimensions Third similarity, comprising: extract the first text from the text information of any dimension of the first article, from the second article should The second text is extracted in the text information of dimension；Calculate the text similarity of the first text and the second text, wherein text phase Third similarity like degree for the text information of the first article and the text information of the second article under the dimension.

In some embodiments, the feature vector of the first article and the feature vector of the second article are calculated under different dimensions Vector distance as the 4th similarity, comprising: calculate the feature vector and second article of the first article according to the following formula Vector distance under feature vector dimension in office:Wherein, D is the feature vector and the second object of the first article The 4th similarity under the feature vector of product dimension in office, x, y are respectively the feature vector and the second article of the first article Numerical value of the feature vector under the dimension.

Second aspect, the embodiment of the present application provide a kind of device of information push, and device includes: acquiring unit, configuration For obtaining the article characteristics information for belonging to multiple articles of same target item classification；Computing unit is configured to based on each The article characteristics information of article calculates the first similarity in multiple articles between any two article；Cluster cell is configured to Based on each first similarity, multiple articles are clustered to obtain multiple clustering clusters；First determination unit is configured to based on to be assessed Article to each clustering cluster cluster centre distance, target clustering cluster is determined from each clustering cluster, wherein article category to be assessed In target item classification；Second determination unit is configured to similar according to the value of each article in target clustering cluster and second Degree, determines the value of article to be assessed, wherein the second similarity is between each article in article to be assessed and target clustering cluster Similarity；Push unit is configured to push the value information for being used to indicate the value of article to be assessed to user.

In some embodiments, article characteristics information includes the text information and use for describing the text feature of article In the feature vector of the data characteristics of description article；Computing unit includes: the first computing module, is configured to calculate the first article Text information and the second article third similarity of the text information under different dimensions, wherein the first article and the second object Product are any two article in multiple articles；Second computing module is configured to calculate the feature vector and the of the first article Vector distance of the feature vector of two articles under different dimensions is as the 4th similarity；First determining module, be configured to by The average value of the third similarity and the 4th similarity of first article and the second article is determined as first article and the second article Between the first similarity.

In some embodiments, the first determination unit is further configured to: being believed according to the article characteristics of article to be assessed The article characteristics information of article in breath and each clustering cluster, calculates between the article to be assessed and the cluster centre of each clustering cluster Distance；From determining cluster centre clustering cluster nearest at a distance from article to be assessed as target clustering cluster in each clustering cluster.

In some embodiments, the second determination unit includes: third computing module, is configured to calculate separately target cluster The second similarity between each article and article to be assessed in cluster；Second determining module is configured to according in target clustering cluster Corresponding second similarity of each article and value product average value, determine the value of article to be assessed.

The second determining module is further configured in some embodiments: obtaining the time at the first moment and the second moment Difference, wherein at the time of the first moment was that article to be assessed needs estimated value, the second moment was any object in target clustering cluster At the time of product mark value；The second similarity corresponding with any article in target clustering cluster and the product of time difference are made For the value weight of the article in target clustering cluster；To each article in target clustering cluster the second moment value and value weight Product be averaging, determine article to be assessed in the value at the first moment.

In some embodiments, the first computing module is further configured to: from the text of any dimension of the first article The first text is extracted in information, and the second text is extracted from the text information of the dimension of the second article；Calculate the first text and The text similarity of second text, wherein text similarity is the text information of the first article and the text envelope of the second article Cease the third similarity under the dimension.

In some embodiments, the second computing module is further configured to: calculating the first article according to the following formula Vector distance under the feature vector dimension in office of feature vector and the second article:Wherein, D is the first object The 4th similarity under the feature vector dimension in office of the feature vector of product and the second article, x, y are respectively the first article The numerical value of feature vector and the feature vector of the second article under the dimension.

The method and apparatus of information push provided by the embodiments of the present application, are primarily based on acquired target item classification The article characteristics information of multiple articles calculates the first similarity between any two article, calculated each based on institute later First similarity clusters each article for multiple clustering clusters, then by calculate article to be assessed to each clustering cluster cluster centre Distance determine target clustering cluster, finally according in target clustering cluster each article and article to be assessed the second similarity and respectively The value of article can determine the value of the article to be assessed, so as to be used to indicate the valence of article to be assessed to user's push The value information of value and it is greater than the article for presetting pre- threshold value with the similarity of the article to be assessed, reduces Item Value assessment Cost of labor improves the accuracy of Item Value assessment, realizes the push rich in targetedly Item Information.

Detailed description of the invention

By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other Feature, objects and advantages will become more apparent upon:

Fig. 1 is that this application can be applied to exemplary system architecture figures therein；

Fig. 2 is the flow chart according to one embodiment of the method for the information of the application push；

Fig. 3 is the schematic diagram according to an application scenarios of the method for the information of the application push；

Fig. 4 is the flow chart according to another embodiment of the method for the information of the application push；

Fig. 5 is the structural schematic diagram according to one embodiment of the device of the information of the application push；

Fig. 6 is adapted for the structural representation of the computer system for the terminal device or server of realizing the embodiment of the present application Figure.

Specific embodiment

The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that in order to Convenient for description, part relevant to related invention is illustrated only in attached drawing.

It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

Fig. 1 is shown can be using the example of the embodiment of the device for the method or information push that the information of the application pushes Property system architecture 100.

As shown in Figure 1, system architecture 100 may include terminal device 101,102,103, network 104 and server 105. Network 104 between terminal device 101,102,103 and server 105 to provide the medium of communication link.Network 104 can be with Including various connection types, such as wired, wireless communication link or fiber optic cables etc..

User can be used terminal device 101,102,103 and be interacted by network 104 with server 105, to receive or send out Send message etc..Various telecommunication customer end applications, such as the application of shopping class, net can be installed on terminal device 101,102,103 The application of page browsing device, searching class application, instant messaging tools, mailbox client, social platform software etc..

Terminal device 101,102,103 can be with display screen and support the functions such as shopping online, web page browsing Various electronic equipments, including but not limited to smart phone, tablet computer, E-book reader, MP3 player (Moving Picture Experts Group Audio Layer III, dynamic image expert's compression standard audio level 3), MP4 (Moving Picture Experts Group Audio Layer IV, dynamic image expert's compression standard audio level 4) is broadcast Put device, pocket computer on knee and desktop computer etc..

Server 105 can be to provide the server of various services, for example, to terminal device 101,102,103 input to It assesses article and the background server supported is provided.Background server can be to the article characteristics information of the article to be assessed received Etc. data carry out the processing such as similarity calculation, and processing result (such as price of article to be assessed) is fed back into terminal device.

It should be noted that the method for the push of information provided by the embodiment of the present application is generally executed by server 105, phase The device of Ying Di, information push are generally positioned in server 105.

It should be understood that the number of terminal device, network and server in Fig. 1 is only schematical.According to realization need It wants, can have any number of terminal device, network and server.

With continued reference to Fig. 2, the process 200 of one embodiment of the method according to the push of the information of the application is shown.It should The method of information push, comprising the following steps:

Step 201, the article characteristics information for belonging to multiple articles of same target item classification is obtained.

In the present embodiment, the electronic equipment (such as server shown in FIG. 1) of the method operation of information push thereon The relevant information that multiple articles can be obtained from big data platform etc., then carries out feature from the relevant information of acquired article It extracts, to obtain the article characteristics information of each article.In general, article can be divided into according to its attribute such as house property, The different article classification such as the art work, antiques, each article acquired in above-mentioned electronic equipment can belong to same target item Classification.For example, the target item classification of above-mentioned each article can be house property classification, above-mentioned electronic equipment can obtain largely in advance House property relevant information, then therefrom extract floor, orientation of room, finishing situation, face where such as house of each house property The various articles characteristic informations such as product, the property right time limit.

Step 202, based on the article characteristics information of each article, the first phase in multiple articles between any two article is calculated Like degree.

In the present embodiment, the article characteristics information obtained based on step 201, above-mentioned electronic equipment (such as it is shown in FIG. 1 Server) any two article can be chosen from each article, then utilize the article characteristics information of two selected articles The similarity between the two articles is calculated, which is the first similarity.Therefore, above-mentioned electronic equipment can pass through by Multiple article every two articles acquired in step 201 are combined, and form an article pair, similar using above-mentioned calculating first The method of degree calculates the similarity of each article pair, to obtain the first similarity in multiple articles between every two article.

In general, similarity can be used for comparing the similitude between two articles, and similitude can generally pass through calculating The distance between characteristic information of article determines, if apart from small, then the similarity between two articles is big, if apart from big, Similarity between so two articles is small.It therefore, can be where house for two articles of such as house property A and house property B The various features information dimension such as floor, orientation of room, finishing situation, area, property right time limit is compared, come determine house property A and Similarity between house property B.Here the algorithm of similarity can include but is not limited to cosine similarity, Euclidean distance etc..

Step 203, it is based on each first similarity, multiple articles are clustered to obtain multiple clustering clusters.

In the present embodiment, above-mentioned electronic equipment can calculate in multiple articles acquired in it between every two article First similarity.Later, above-mentioned electronic equipment can be calculated according to calculated each first similarity using the cluster of machine learning Article similar in similarity in acquired multiple articles is divided into the same clustering cluster by method.In this way, above-mentioned electronic equipment can Acquired multiple articles are divided into multiple clustering clusters, wherein may include in each clustering cluster in multiple articles extremely A few article.It is different from traditional machine learning method, sample of the method independent of any mark in the present embodiment, machine The clustering algorithm of device study is also possible to unsupervised machine learning, realizes simple.

Cluster, can refer to that the set by physics or abstract object is divided into the process for the multiple classes being made of similar object. By clustering the set that cluster generated is one group of data object, these objects and the object in the same cluster are similar to each other, with it Object in his cluster is different.Correspondingly, similar to each other by multiple articles in a clustering cluster of cluster generation, and the clustering cluster In any one article and any one article in other clustering clusters it is (or can consider similarity very low) different from each other.

Step 204, the distance of the cluster centre based on article to be assessed to each clustering cluster, determines mesh from each clustering cluster Mark clustering cluster.

In the present embodiment, above-mentioned electronic equipment can be by wired connection mode or radio connection from user institute Terminal obtain the article characteristics information of article to be assessed He the article to be assessed, in the article to be assessed and step 201 Article classification belonging to multiple articles is identical, belongs to target item classification.Then, above-mentioned electronic equipment can be to each clustering cluster Included in article article characteristics information carry out cluster operation obtain the cluster centre of each clustering cluster.Last above-mentioned electronics is set The standby article characteristics information that can use article to be assessed determine the article to be assessed to each clustering cluster cluster centre away from From in order to which it can determine at least one target clustering cluster from each clustering cluster.It is pointed out that above-mentioned wireless connection Mode can include but is not limited to 3G/4G connection, WiFi connection, bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection and other currently known or exploitation in the future radio connections.

Step 205, according to the value and the second similarity of each article in target clustering cluster, the valence of article to be assessed is determined Value.

In the present embodiment, the target clustering cluster determined based on step 204, the available target of above-mentioned electronic equipment The article characteristics information of each article in clustering cluster.Then, each object in the characteristic information of article to be assessed and target clustering cluster is utilized The article characteristics information of product can calculate the similarity in the article to be assessed and target clustering cluster between each article, the phase It can be the second similarity like degree.Finally, above-mentioned electronic equipment can use each article in target clustering cluster value and The second similarity between each article and article to be assessed in target clustering cluster, calculates the value of the article to be assessed.

Step 206, the value information of the value of article to be assessed is used to indicate to user's push.

In the present embodiment, use can be generated in the value of the article to be assessed determined based on step 205, above-mentioned electronic equipment In the value information for the value for indicating the article to be assessed, and rear line pushes the value information.

In some optional implementations of the present embodiment, above-mentioned electronic equipment can also be obtained from target clustering cluster It is greater than the article of preset threshold with the article similarity to be assessed, acquired article is then pushed to user, so that User can also obtain article similar with article to be assessed while obtaining the value information of article to be assessed.It needs Bright, above-mentioned electronic equipment can use the value and/or article characteristics information and to be evaluated of each article in target clustering cluster The value and/or article characteristics information for estimating article calculate similar between each article and article to be assessed in target clustering cluster Degree.

As it can be seen that when user needs to assess the value of the article to be assessed of any article classification, user needs to input should be to The article characteristics information of article is assessed, in order to the article classification of the available multiple and article to be assessed of above-mentioned electronic equipment Identical article, and according to the first similarity between article gathers each article for multiple clustering clusters two-by-two in each article, then on It states electronic equipment and can calculate the article to be assessed and determine target clustering cluster at a distance from the cluster centre of each clustering cluster, last benefit The object in target clustering cluster is calculated with the article characteristics information of article each in target clustering cluster and the article to be assessed of user's input Second similarity of product and article to be assessed, so that electronic equipment can be according to each in each second similarity and target clustering cluster The value of article determines the value of the article to be assessed.Further, above-mentioned electronic equipment can also will be identified to be assessed The value information of article is pushed to user.

With continued reference to the schematic diagram that Fig. 3, Fig. 3 are according to the application scenarios of the method for the information of the present embodiment push. In the application scenarios of Fig. 3, for house property A to be assessed, user can input the various characteristic informations of house property A, such as house Place floor, orientation of room, finishing situation, area, property right time limit etc., in order to the spy of the available house property A of background server Reference breath；Later, the characteristic information of the available multiple house properties of background server, and believed based on the feature of acquired each house property Breath calculates the first similarity between any two house property；Then, it is obtained using each first similarity to what each house property was clustered Target clustering cluster is therefrom determined at a distance from the cluster centre of each clustering cluster to multiple clustering clusters, and according to house property A；Finally, The characteristic information that background server can use the house property A of user's input calculates each house property in house property A and target clustering cluster Second similarity can use the value of each house property in the second similarity and target clustering cluster in order to background server (such as Price) value (such as price) of determining house property A, when the user clicks when " house property valuation " button, as shown in figure 3, it can User is shown to the value (such as price) for the house property A that will be evaluated.Further, above-mentioned electronic equipment can also be pushed away to user Send house property 1, the house property 2 etc. for being greater than preset threshold with the similarity of house property A.

The method provided by the above embodiment of the application is primarily based on multiple articles of acquired target item classification Article characteristics information calculates the first similarity between any two article, later based on calculated each first similarity of institute Each article is clustered as multiple clustering clusters, the distance for then passing through cluster centre of the calculating article to be assessed to each clustering cluster determines Target clustering cluster, finally according to each article and the second similarity of article to be assessed and the value of each article in target clustering cluster It can determine the value of the article to be assessed, believe so as to be used to indicate the value of value of article to be assessed to user's push Breath and it is greater than the article for presetting pre- threshold value with the similarity of the article to be assessed, reduces the cost of labor of Item Value assessment, The accuracy for improving Item Value assessment realizes the push rich in targetedly Item Information.

With further reference to Fig. 4, it illustrates the processes 400 of another embodiment of the method for information push.The information pushes away The process 400 for the method sent, comprising the following steps:

Step 401, the article characteristics information for belonging to multiple articles of same target item classification is obtained.

In the present embodiment, the electronic equipment (such as server shown in FIG. 1) of the method operation of information push thereon The relevant information that multiple articles can be obtained from big data platform etc., then carries out feature from the relevant information of acquired article It extracts, to obtain the article characteristics information of each article.In general, article can be divided into according to its attribute such as house property, The different article classification such as the art work, antiques, each article acquired in above-mentioned electronic equipment can belong to same target item Classification.

Step 402, the text description information of the first article and the text description information of the second article are calculated in different dimensions Under third similarity.

In the present embodiment, the article characteristics information of the multiple articles obtained based on step 401, above-mentioned electronic equipment (example Server as shown in Figure 1) article characteristics information can be divided into text information and feature vector, wherein and text information is used In the text feature of description article, e.g., fitting case, the direction in house in house etc., feature vector can be used for describing article Data characteristics, such as the floor in the area in house, house.Then, above-mentioned electronic equipment can be from acquired each article Any two article is selected as the first article and the second article, and determines each dimension text of the first article and the second article Information.Finally, above-mentioned electronic equipment can calculate the text information of the first article and the text information of the second article in each dimension Under similarity, the similarity be third similarity.

In some optional implementations of the present embodiment, for the text information of any dimension, above-mentioned electronic equipment Can extract the first text in the text information of the dimension from the first article, and from the second article the dimension text envelope The second text is extracted in breath, then, calculates the text similarity of extracted the first text and the second text, the text is similar Third similarity of the degree for the text information of the first article and the text information of the second article under the dimension.As an example, right It can be the text information under orientation of room dimension in the text information of any dimension, the first article and the second article are respectively room 1 and house property 2 are produced, above-mentioned electronic equipment is from the first text extracted under orientation of room dimension in the text information of house property 1, from room The second text extracted under orientation of room dimension in 2 text information is produced, then by the first text extracted and second Text uses stringent matched method to be matched to obtain third similarity, it may be assumed that can match first as this text string (word Symbol string) and the text string of the second text it is whether completely the same, determine that the third similarity is 1, such as first if completely the same Text and the second text are " south "；Determine that the third similarity is 0 if not quite identical, such as the first text and the second text One's duty is by for " south " and " east ".

Step 403, vector of the feature vector of the feature vector and the second article that calculate the first article under different dimensions Distance is used as the 4th similarity.

In the present embodiment, the article characteristics information of the multiple articles obtained based on step 401, above-mentioned electronic equipment can be with The feature vector of each article is therefrom obtained, e.g., area, the property right in real estate time limit in house etc..Then, above-mentioned electronic equipment can be with Obtain the feature vector of above-mentioned first article and the second article different dimensions.Finally, above-mentioned electronic equipment can calculate the first object The similarity of the feature vector of product and the feature vector of the second article under each dimension, the similarity are the 4th similarity.

In some optional implementations of the present embodiment, for the feature vector of any dimension, above-mentioned electronic equipment Numerical value x, and the spy from the second article under the dimension can be extracted in the feature vector under the dimension from the first article Extract numerical value y in sign vector, then above-mentioned electronic equipment can use following formula calculate the feature vector of the first article with Vector distance under the feature vector of second article dimension in office, the vector distance are the feature vector and the of the first article 4th similarity D under the feature vector of two articles dimension in office:As an example, the feature of any dimension to Amount can be the floor space dimension of house property, and the first article and the second article are respectively house property 1 and house property 2, above-mentioned electronic equipment Can numerical value x and numerical value y be extracted respectively from the floor space dimension of the feature vector of house property 1 and house property 2 respectively, it is then sharp WithCalculate the 4th similarity D of house property 1 and house property 2 under floor space dimension, such as from house property 1 and room The numerical value that extracts of floor space dimension for producing 2 is respectively 100 and 80, thenI.e. The similarity of house property 1 and house property 2 under floor space dimension is 0.89.

Step 404, using the average value of the third similarity and the 4th similarity of the first article and the second article as first The first similarity between article and the second article.

In the present embodiment, it is based between step 402 and calculated first article of step 403 and the second article in each dimension Third similarity and the 4th similarity under degree, above-mentioned electronic equipment can be to calculated each third similarity and the two or four phase It sums like degree, then calculates the average value of this and value, which is the first article and the second article between First similarity.As it can be seen that using each article characteristics information of multiple articles acquired in this method processing electronic equipment, it can not It needs the text information to each article to quantize, but is directly clustered using text information and data characteristics, is not required to Additionally to increase artificial consumption, realize simple.

Step 405, it is based on each first similarity, multiple articles are clustered to obtain multiple clustering clusters.

In the present embodiment, above-mentioned electronic equipment can be calculated between every two article in each article acquired in it First similarity, above-mentioned electronic equipment can be incited somebody to action according to calculated each first similarity using the clustering algorithm of machine learning Article similar in similarity is divided into the same clustering cluster in acquired multiple articles.In this way, above-mentioned electronic equipment can incite somebody to action Multiple articles acquired in it are divided into multiple clustering clusters, wherein may include in each clustering cluster in multiple articles at least One article.

Step 406, the distance of the cluster centre based on article to be assessed to each clustering cluster, determines mesh from each clustering cluster Mark clustering cluster.

In the present embodiment, above-mentioned electronic equipment can obtain article to be assessed in advance and the article of the article to be assessed is special Reference breath.Later, each clustering cluster obtained based on step 405, above-mentioned electronic equipment can determine the cluster centre of each clustering cluster, Then according to the article characteristics information of article to be assessed calculate article to be assessed to each clustering cluster cluster centre distance.Most Afterwards, above-mentioned electronic equipment can therefrom determine the cluster centre clustering cluster nearest to the distance of article to be assessed, which determines Clustering cluster out is target clustering cluster.

Step 407, the second similarity between each article and article to be assessed in target clustering cluster is calculated separately.

In the present embodiment, the characteristic information of each article in the available target clustering cluster of above-mentioned electronic equipment, then Using the characteristic information of each article in target clustering cluster and the article characteristics information of article to be assessed, calculate in target clustering cluster Each article and article to be assessed between similarity, the similarity be the second similarity.

Step 408, the time difference at the first moment and the second moment is obtained.

In the present embodiment, above-mentioned electronic equipment can obtain the value of each article in target clustering cluster first, and each Article is corresponding with the second moment of the value, (for example, be 3,000,000 for the price of the house property 1 in target clustering cluster, the house property 1 The time that marked price/price is 3,000,000 is second moment).Then, the available article needs to be assessed of above-mentioned electronic equipment are commented First moment of assessment values, (for example, for house property A to be assessed, it is any in current time/future that user needs to estimate house property A The price at moment, the current time/future any moment were the first moment).Finally, above-mentioned electronic equipment can determine first The time difference at moment and the second moment.

Step 409, the second similarity corresponding with any article in target clustering cluster and the product of time difference are made For the value weight of the article in target clustering cluster.

In the present embodiment, for each article in target clustering cluster, above-mentioned electronic equipment it is available its with it is to be evaluated Estimate the second similarity and the time difference at its corresponding second moment and the first moment of article, then above-mentioned electronic equipment can To calculate the product of acquired the second similarity and time difference, which is the value power of each article in target clustering cluster Weight.Here it is possible to calculate the value weight of each article in target clustering cluster using following formula:

Weight=| Ta-Tb | × Similarity, wherein Weight is the value of any article in target clustering cluster Weight, Similarity are second similarity of the article and article to be assessed in target clustering cluster, and Ta was the first moment, and Tb is Second moment.

Step 410, each article in target clustering cluster is averaging in the value at the second moment and the product of value weight, really Value of the fixed article to be assessed at the first moment.

In the present embodiment, the value weight of the article, above-mentioned electronics are set in the target clustering cluster determined based on step 409 The standby value that can continue to obtain each article in target clustering cluster.Then, for each article in target clustering cluster, above-mentioned electronics Equipment can calculate the product of the value and value weight of each article.Finally, multiplying to each article is corresponding in target clustering cluster Product value is averaging, which is value of the article to be assessed at the first moment.It is alternatively possible to be calculated using following formula Value of the article to be assessed at the first moment:

Wherein, Predicted price is value of the article to be assessed at the first moment, and L is the total of article in target clustering cluster Number, weight_kFor the value weight of k-th of article in target clustering cluster, Price_kIt is k-th of article in target clustering cluster The value at two moment.

Step 411, the value information of the value of article to be assessed is used to indicate to user's push.

In the present embodiment, the value based on the determining article to be assessed of step 410 at the first moment, above-mentioned electronic equipment The value information for being used to indicate the value of the article to be assessed can be generated, and rear line pushes the value information.

Figure 4, it is seen that compared with the corresponding embodiment of Fig. 2, the stream of the method for the information push in the present embodiment Journey 400 can introduce the value weight of time dependence, so that this method can not only assess article to be assessed current Value, the future value of article to be assessed can also be predicted.

With further reference to Fig. 5, as the realization to method shown in above-mentioned each figure, this application provides a kind of push of information One embodiment of device, the Installation practice is corresponding with embodiment of the method shown in Fig. 2, which specifically can be applied to In various electronic equipments.

As shown in figure 5, the device 500 that the information of the present embodiment pushes includes: acquiring unit 501, the cluster of computing unit 502 Unit 503, the first determination unit 504, the second determination unit 505 and push unit 506.Wherein, acquiring unit 501 is configured to Obtain the article characteristics information for belonging to multiple articles of same target item classification；Computing unit 502 is configured to based on each object The article characteristics information of product calculates the first similarity in multiple articles between any two article；Cluster cell 503 is configured to Based on each first similarity, multiple articles are clustered to obtain multiple clustering clusters；First determination unit 504 is configured to based on to be evaluated Estimate article to each clustering cluster cluster centre distance, target clustering cluster is determined from each clustering cluster, wherein article to be assessed Belong to target item classification；Second determination unit 505 is configured to the value and second according to each article in target clustering cluster Similarity determines the value of article to be assessed, wherein the second similarity is each article in article to be assessed and target clustering cluster Between similarity；Push unit 506 is configured to push the value information for being used to indicate the value of article to be assessed to user.

In some optional implementations of the present embodiment, article characteristics information includes special for describing the text of article The feature vector of the text information of sign and the data characteristics for describing article；Computing unit 502 includes: the first calculating mould Block, the text information for being configured to calculate the first article are similar to third of the text information of the second article under different dimensions Degree, wherein the first article and the second article are any two article in multiple articles；Second computing module is configured to count Vector distance of the feature vector of the feature vector and the second article of calculating the first article under different dimensions is as the 4th similarity； First determining module, the average value for being configured to the third similarity and the 4th similarity by the first article and the second article determine For the first similarity between first article and the second article.

In some optional implementations of the present embodiment, the first determination unit 504 is further configured to: according to The article characteristics information of article and the article characteristics information of the article in each clustering cluster are assessed, the article to be assessed is calculated and gathers with each The distance between the cluster centre of class cluster；Gather from determining that cluster centre is nearest at a distance from article to be assessed in each clustering cluster Class cluster is as target clustering cluster.

In some optional implementations of the present embodiment, the second determination unit 505 includes: third computing module, is matched It sets for calculating separately the second similarity between each article and article to be assessed in target clustering cluster；Second determining module, matches It sets for the average value according to the product of corresponding second similarity of each article and value in target clustering cluster, determines to be assessed The value of article.

In some optional implementations of the present embodiment, the second determining module is further configured to: obtaining first The time difference at moment and the second moment, wherein at the time of the first moment was that article to be assessed needs estimated value, the second moment At the time of marking value for article any in target clustering cluster；It is similar by corresponding with any article in target clustering cluster second Value weight of the product of degree and time difference as the article in target clustering cluster；To each article in target clustering cluster second The value at moment and the product for being worth weight are averaging, and determine article to be assessed in the value at the first moment.

In some optional implementations of the present embodiment, the first computing module is further configured to: from the first object The first text is extracted in the text information of any dimension of product, the second text is extracted from the text information of the dimension of the second article This；Calculate the text similarity of the first text and the second text, wherein text similarity be the first article text information with Third similarity of the text information of second article under the dimension.

In some optional implementations of the present embodiment, the second computing module is further configured to: according to as follows Formula calculates the vector distance under the feature vector of the first article and the feature vector dimension in office of the second article:Wherein, D is the 4 under the feature vector of the first article and the feature vector dimension in office of the second article Similarity, x, y are respectively the feature vector of the first article and numerical value of the feature vector under the dimension of the second article.

Below with reference to Fig. 6, it illustrates the terminal device/server computers for being suitable for being used to realize the embodiment of the present application The structural schematic diagram of system 600.Terminal device/server shown in Fig. 6 is only an example, should not be to the embodiment of the present application Function and use scope bring any restrictions.

As shown in fig. 6, computer system 600 includes central processing unit (CPU) 601, it can be read-only according to being stored in Program in memory (ROM) 602 or be loaded into the program in random access storage device (RAM) 603 from storage section 608 and Execute various movements appropriate and processing.In RAM 603, also it is stored with system 600 and operates required various programs and data. CPU 601, ROM 602 and RAM 603 are connected with each other by bus 604.Input/output (I/O) interface 605 is also connected to always Line 604.

I/O interface 605 is connected to lower component: the importation 606 including keyboard, mouse etc.；It is penetrated including such as cathode The output par, c 607 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage section 608 including hard disk etc.； And the communications portion 609 of the network interface card including LAN card, modem etc..Communications portion 609 via such as because The network of spy's net executes communication process.Driver 610 is also connected to I/O interface 605 as needed.Detachable media 611, such as Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 610, in order to read from thereon Computer program be mounted into storage section 608 as needed.

Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description Software program.For example, embodiment of the disclosure includes a kind of computer program product comprising be carried on computer-readable medium On computer program, which includes the program code for method shown in execution flow chart.In such reality It applies in example, which can be downloaded and installed from network by communications portion 609, and/or from detachable media 611 are mounted.When the computer program is executed by central processing unit (CPU) 601, limited in execution the present processes Above-mentioned function.It should be noted that computer-readable medium described herein can be computer-readable signal media or Computer readable storage medium either the two any combination.Computer readable storage medium for example can be --- but Be not limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or any above combination. The more specific example of computer readable storage medium can include but is not limited to: have one or more conducting wires electrical connection, Portable computer diskette, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type may be programmed read-only deposit Reservoir (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory Part or above-mentioned any appropriate combination.In this application, computer readable storage medium, which can be, any include or stores The tangible medium of program, the program can be commanded execution system, device or device use or in connection.And In the application, computer-readable signal media may include in a base band or the data as the propagation of carrier wave a part are believed Number, wherein carrying computer-readable program code.The data-signal of this propagation can take various forms, including but not It is limited to electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be computer Any computer-readable medium other than readable storage medium storing program for executing, the computer-readable medium can send, propagate or transmit use In by the use of instruction execution system, device or device or program in connection.Include on computer-readable medium Program code can transmit with any suitable medium, including but not limited to: wireless, electric wire, optical cable, RF etc., Huo Zheshang Any appropriate combination stated.

The calculating of the operation for executing the application can be write with one or more programming languages or combinations thereof Machine program code, described program design language include object oriented program language-such as Java, Smalltalk, C+ +, it further include conventional procedural programming language-such as " C " language or similar programming language.Program code can Fully to execute, partly execute on the user computer on the user computer, be executed as an independent software package, Part executes on the remote computer or executes on a remote computer or server completely on the user computer for part. In situations involving remote computers, remote computer can pass through the network of any kind --- including local area network (LAN) Or wide area network (WAN)-is connected to subscriber computer, or, it may be connected to outer computer (such as utilize Internet service Provider is connected by internet).

Flow chart and block diagram in attached drawing are illustrated according to the system of the various embodiments of the application, method and computer journey The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation A part of one module, program segment or code of table, a part of the module, program segment or code include one or more use The executable instruction of the logic function as defined in realizing.It should also be noted that in some implementations as replacements, being marked in box The function of note can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are actually It can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it to infuse Meaning, the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart can be with holding The dedicated hardware based system of functions or operations as defined in row is realized, or can use specialized hardware and computer instruction Combination realize.

Being described in unit involved in the embodiment of the present application can be realized by way of software, can also be by hard The mode of part is realized.Described unit also can be set in the processor, for example, can be described as: a kind of processor packet Include acquiring unit, computing unit, cluster cell, the first determination unit and the second determination unit.Wherein, the title of these units exists The restriction to the unit itself is not constituted in the case of certain, " acquisition belongs to same for example, acquiring unit is also described as The unit of the article characteristics information of multiple articles of target item classification ".

As on the other hand, present invention also provides a kind of computer-readable medium, which be can be Included in device described in above-described embodiment；It is also possible to individualism, and without in the supplying device.Above-mentioned calculating Machine readable medium carries one or more program, when said one or multiple programs are executed by the device, so that should Device: the article characteristics information for belonging to multiple articles of same target item classification is obtained；Article characteristics letter based on each article Breath, calculates the first similarity in multiple articles between any two article；Based on each first similarity, multiple articles are clustered To multiple clustering clusters；The distance of cluster centre based on article to be assessed to each clustering cluster, determines target from each clustering cluster Clustering cluster, wherein article to be assessed belongs to target item classification；According to the value and the second phase of each article in target clustering cluster Like degree, determine the value of article to be assessed, wherein the second similarity be each article in article to be assessed and target clustering cluster it Between similarity；The value information of the value of the article to be assessed is used to indicate to user's push.

Above description is only the preferred embodiment of the application and the explanation to institute's application technology principle.Those skilled in the art Member is it should be appreciated that invention scope involved in the application, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic Scheme, while should also cover in the case where not departing from foregoing invention design, it is carried out by above-mentioned technical characteristic or its equivalent feature Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed herein Can technical characteristic replaced mutually and the technical solution that is formed.

Claims

1. a kind of method of information push, comprising:

Obtain the article characteristics information for belonging to multiple articles of same target item classification；

Based on the article characteristics information of each article, calculate in the multiple article first similar between any two article Degree；

Based on each first similarity, the multiple article is clustered to obtain multiple clustering clusters；

The distance of cluster centre based on article to be assessed to each clustering cluster determines that target is poly- from each clustering cluster Class cluster, wherein the article to be assessed belongs to the target item classification；

According to the value and the second similarity of each article in the target clustering cluster, the value of the article to be assessed is determined, Wherein, second similarity is the similarity between each article in the article to be assessed and the target clustering cluster；

The value information of the value of the article to be assessed is used to indicate to user's push.

2. according to the method described in claim 1, wherein, the article characteristics information includes the text feature for describing article Text information and the data characteristics for describing article feature vector；

The article characteristics information based on each article, calculates the first phase in the multiple article between any two article Like degree, comprising:

Third similarity of the text information of the text information and the second article that calculate the first article under different dimensions, wherein First article and second article are any two article in the multiple article；

Vector distance of the feature vector of the feature vector and second article that calculate first article under different dimensions As the 4th similarity；

The average value of the third similarity and the 4th similarity of first article and the second article is determined as first article and The first similarity between two articles.

3. described based on article to be assessed to the cluster centre of each clustering cluster according to the method described in claim 1, wherein Distance, determine target clustering cluster from each clustering cluster, comprising:

According to the article characteristics information of the article in the article characteristics information of the article to be assessed and each clustering cluster, calculate The distance between the cluster centre of the article to be assessed and each clustering cluster；

Described in determining cluster centre in each clustering cluster at a distance from the article to be assessed nearest clustering cluster be used as Target clustering cluster.

4. the method according to claim 1, wherein the valence according to each article in the target clustering cluster Value and the second similarity, determine the value of the article to be assessed, comprising:

Calculate separately the second similarity between each article and the article to be assessed in the target clustering cluster；

According to the average value of the product of corresponding second similarity of each article and value in the target clustering cluster, determine described in The value of article to be assessed.

It is corresponding second similar according to each article in the target clustering cluster 5. according to the method described in claim 4, wherein The average value of the product of degree and value, determines the value of the article to be assessed, comprising:

Obtain the time difference at the first moment and the second moment, wherein first moment is that the article needs to be assessed are commented At the time of assessment values, at the time of second moment is any article mark value in the target clustering cluster；

It will be with the product of corresponding second similarity of any article and the time difference in the target clustering cluster as institute State the value weight of the article in target clustering cluster；

Each article in the target clustering cluster is averaging in the value at the second moment and the product of the value weight, determines institute Article to be assessed is stated in the value at first moment.

6. according to the method described in claim 2, wherein, the text of the text information for calculating the first article and the second article Third similarity of the information under different dimensions, comprising:

The first text is extracted from the text information of any dimension of first article, from the dimension of second article The second text is extracted in text information；

Calculate the text similarity of first text and the second text, wherein text similarity is the text of the first article Third similarity of the text information of information and the second article under the dimension.

7. according to the method described in claim 2, wherein, the feature vector for calculating first article and second object Vector distance of the feature vector of product under different dimensions is as the 4th similarity, comprising:

The feature vector of first article and the feature vector dimension in office of second article are calculated according to the following formula Under vector distance:

Wherein, D is under the feature vector of the first article and the feature vector dimension in office of the second article 4th similarity, x, y are respectively the feature vector of the first article and numerical value of the feature vector under the dimension of the second article.

8. a kind of device of information push, comprising:

Acquiring unit is configured to obtain the article characteristics information for the multiple articles for belonging to same target item classification；

Computing unit is configured to calculate any two in the multiple article based on the article characteristics information of each article The first similarity between article；

Cluster cell is configured to cluster to obtain multiple clustering clusters to the multiple article based on each first similarity；

First determination unit is configured to the distance of the cluster centre based on article to be assessed to each clustering cluster, from each institute It states and determines target clustering cluster in clustering cluster, wherein the article to be assessed belongs to the target item classification；

Second determination unit is configured to value and the second similarity according to each article in the target clustering cluster, determines The value of the article to be assessed, wherein second similarity is in the article to be assessed and the target clustering cluster Similarity between each article；

Push unit is configured to push the value information for being used to indicate the value of the article to be assessed to user.

9. device according to claim 8, wherein the article characteristics information includes the text feature for describing article Text information and the data characteristics for describing article feature vector；

The computing unit includes:

First computing module is configured to calculate the text information of the first article and the text information of the second article in different dimensions Under third similarity, wherein first article and second article are any two article in the multiple article；

Second computing module is configured to calculate the feature vector of first article and the feature vector of second article exists Vector distance under different dimensions is as the 4th similarity；

First determining module is configured to the average value of the third similarity and the 4th similarity by the first article and the second article The first similarity being determined as between first article and the second article.

10. device according to claim 8, wherein first determination unit is further configured to:

11. device according to claim 8, wherein second determination unit includes:

Third computing module is configured to calculate separately between each article and the article to be assessed in the target clustering cluster Second similarity；

Second determining module is configured to according to corresponding second similarity of each article and value in the target clustering cluster The average value of product determines the value of the article to be assessed.

12. device according to claim 11, wherein second determining module is further configured to:

13. device according to claim 9, wherein first computing module is further configured to:

14. device according to claim 9, wherein second computing module is further configured to:

15. a kind of equipment, comprising:

One or more processors；

Storage device, for storing one or more programs,

When one or more of programs are executed by one or more of processors, so that one or more of processors are real The now method as described in any in claim 1-7.

16. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that described program is processed The method as described in any in claim 1-7 is realized when device executes.