CN108345609A

CN108345609A - A kind of method and apparatus of processing POI information

Info

Publication number: CN108345609A
Application number: CN201710054812.8A
Authority: CN
Inventors: 卢俊之; 汤沛; 季成晖; 吴坤; 孟凡超
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2017-01-24
Filing date: 2017-01-24
Publication date: 2018-07-31

Abstract

The invention discloses a kind of method and apparatus of processing POI information, belong to field of computer technology.The method includes：It obtains the first POI information and meets the second POI information of preset proximity condition with first POI information, the word for including to the first place name in first POI information carries out function division, and the word for including to the second place name in second POI information carries out function division, according to the matching degree of word with the same function in first place name and second place name, determine first POI information and second POI information whether be same physical entity POI information.Using the present invention, the accuracy of processing POI information can be improved.

Description

A kind of method and apparatus of processing POI information

Technical field

The present invention relates to field of computer technology, more particularly to a kind of method and apparatus of processing POI information.

Background technology

The side that the server of network map can be obtained by technical staff's field exploring, from partner or network crawls Formula gets POI (Point of Information, information point) information of each physical entity.Wherein, physical entity can be Sight spot, market, station etc. have the geographical place of fixed position attribute, and in GIS-Geographic Information System, a POI can be one House, a retail shop, a mailbox, a bus station etc., POI information may include the place name of physical entity, classification, The much informations such as location, coordinate (longitude and latitude), neighbouring retail shop of restaurant.Server can according to POI information in network map phase The position setting each physical entity corresponding ground map logo answered, in this way, user can directly determine physics by ground map logo The physical location of entity.And the different POI information that the server of network map is got from different approaches may be directed toward it is same Physical entity, therefore, before generating ground map logo, server needs to carry out homogeneity judgement to different POI information, will determine that POI information for same physical entity merges.Specifically, server can have been obtained first when carrying out homogeneity judgement Possible two POI information are repeated, then judge whether two POI information are directed toward same physical entity by the way that text is similar.It is right In the information that two POI information include, server can be selected according to preset rules using the letter in some POI information Breath, and for the information that only one POI information includes, then directly the information can be added in the POI information after merging.

For example, technical staff's field exploring is likely to be obtained a POI information, the place entitled " roads AB masses' holiday therein Hotel ", while server may get a POI information from partner, place therein is entitled " roads AB masses hotel ", and Actually the two POI information are the POI information of same physical entity, it is contemplated that the text similarity of two POI information is higher, So can be determined that above-mentioned two POI information is directed toward same physical entity, further, the POI of " roads AB masses' holiday inn " In information, including 4 kinds of place name, address, coordinate, neighbouring businessman information, in the POI information in " roads AB masses hotel ", including field 4 kinds of institute's name, address, classification, coordinate information, for all including in " place name ", " address " and " coordinate " these two POI information Information, server can select use " roads AB masses' holiday inn " in information, for " neighbouring businessman ", " classification " these Only there are one the information that POI information includes, which can be added in the POI information after merging.

In the implementation of the present invention, the inventor finds that the existing technology has at least the following problems：

Due to the diversity of naming method, some situations can not accurately judge two POI using the similar mode of text Whether information is directed toward same physical entity, for example, for a：" AC masses' holiday leisure hotel's hot spring ", b：" AC masses' holidays stop The two POI information of not busy Hotspring Hotel ", the POI information of the practical hot springs for certain hotel subordinate of a, the practical POI for certain hotel of b Information, however according to the similar judgement of text, the similarity of a and b are very high, then will be considered that the POI that a and b is same physical entity believes Breath, so the accuracy of processing POI information is relatively low.

Invention content

In order to solve problems in the prior art, an embodiment of the present invention provides a kind of method and apparatus of processing POI information. The technical solution is as follows：

In a first aspect, a kind of method of processing POI information is provided, the method includes：

The 2nd POI for obtaining the first POI information and meeting preset proximity condition with first POI information believes Breath；

The word for including to the first place name in first POI information carries out function division, and to the 2nd POI The word that the second place name in information includes carries out function division；

According to the matching degree of word with the same function in first place name and second place name, institute is determined State the first POI information and second POI information whether be same physical entity POI information.

Second aspect, provides a kind of device of processing POI information, and described device includes：

Data obtaining module, for obtaining the first POI information and meeting the preset degree of approach with first POI information Second POI information of condition；

Function division module, the word for including to the first place name in first POI information carry out function and draw Point, and the word for including to the second place name in second POI information carries out function division；

First determining module, for according to word with the same function in first place name and second place name The matching degree of language, determine first POI information and second POI information whether be same physical entity POI information.

The advantageous effect that technical solution provided in an embodiment of the present invention is brought is：

In the embodiment of the present invention, for server after getting a pending POI information, can choose has repetition with it A possible POI information carries out homogeneity judgement with it, then carries out function division, Jin Erke to the place of POI information name To be matched to the word with function of the same race in the name of place, in this way, can have according in the place name in two POI information The matching degree for having the word of identical function, judges whether two POI information are directed toward same physical entity.Wherein, two POI are believed The word of congenerous calculates matching degree in breath, and matching degree result can more accurately illustrate whether two POI information are directed toward together One physical entity, so as to improve the accuracy of processing POI information.

Description of the drawings

To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for For those of ordinary skill in the art, without creative efforts, other are can also be obtained according to these attached drawings Attached drawing.

Fig. 1 is a kind of method flow diagram of processing POI information provided in an embodiment of the present invention；

Fig. 2 is a kind of flow diagram carrying out homogeneity judgement to POI information provided in an embodiment of the present invention；

Fig. 3 is a kind of principle schematic of processing POI information provided in an embodiment of the present invention；

Fig. 4 is a kind of apparatus structure schematic diagram of processing POI information provided in an embodiment of the present invention；

Fig. 5 is a kind of apparatus structure schematic diagram of processing POI information provided in an embodiment of the present invention；

Fig. 6 is a kind of apparatus structure schematic diagram of processing POI information provided in an embodiment of the present invention；

Fig. 7 is a kind of apparatus structure schematic diagram of processing POI information provided in an embodiment of the present invention；

Fig. 8 is a kind of apparatus structure schematic diagram of processing POI information provided in an embodiment of the present invention；

Fig. 9 is a kind of apparatus structure schematic diagram of processing POI information provided in an embodiment of the present invention；

Figure 10 is a kind of structural schematic diagram of server provided in an embodiment of the present invention.

Specific implementation mode

To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention Formula is described in further detail.

An embodiment of the present invention provides applied to a kind of processing POI information method in GIS-Geographic Information System, this method fortune After the server of network map gets the POI information of pending physical entity, according to POI information in network map In the processing for adding corresponding ground map logo, herein, physical entity, which can be sight spot, market, station etc., has fixed position The executive agent in the geographical place of attribute, this method is server.Wherein, server can be the background service of network map Device.Server may include processor, memory, and processor can be used for carrying out the processing to POI information in following flows, Memory can be used for storing the data of the data and generation that are needed in following processing procedures.Server can also include transmission Component, input-output unit, transmission part can be used for handle POI information during data reception and transmission, input it is defeated Go out unit can be used for server side technical staff instruction input and data display.

Below in conjunction with specific implementation mode, process flow shown in FIG. 1 is described in detail, content can be as Under：

Step 101, it obtains the first POI information and meets the 2nd POI of preset proximity condition with the first POI information Information.

Wherein, in GIS-Geographic Information System, a POI can be a house, a retail shop, a mailbox, a public affairs Hand over the geographical place such as station, POI information may include the place name in geographical place, classification, address, coordinate (longitude and latitude), nearby The much informations such as retail shop of restaurant；Preset proximity condition can be place name similarity be more than certain similarity threshold and/ Or the distance between coordinate is less than certain distance threshold value, wherein if in view of threshold value is too low, may cause to meet preset connect The POI information of recency condition is excessive, and treating capacity is larger, if threshold value is excessively high, may omit a large amount of practical and the first POI information It is directed toward the POI information of same physical entity, so similarity the ratio of total word (i.e. identical word account for) threshold value can be arranged For 0.6-0.8, or set distance threshold to 100m-500m.

In force, server can be obtained by technical staff's field exploring, from partner or network crawl etc. it is more Kind approach mode gets the POI information of pending each physical entity, then can be to these pending POI information one by one It is handled.Server can first choose a pending POI information (i.e. the first POI information), then can be fixed The POI information for meeting preset proximity condition with the first POI information is filtered out in the standard POI information of physical entity, in turn First POI information and the POI information screened can be handled one by one, i.e., server can be in the POI screened The second POI information is obtained in information.

Step 102, the word for including to the first place name in the first POI information carries out function division, and to the 2nd POI The word that the second place name in information includes carries out function division.

In force, server, can be in the first POI information after obtaining the first POI information and the second POI information The first place name and the second POI information in the second place name handled, i.e., according to preset functional group respectively to first The word that the word and the second place name that place name includes include carries out function division.Specifically, server can be first to first Place name carries out semantic analysis, i.e. word segmentation processing obtains multiple words that the first place name includes, and then determines each word again Corresponding role, later, server can carry out structural analysis according to preset the first place of the structure criteria for classifying pair name, sentence Break the sentence ingredient for wherein each word in the name of the first place, so can be based on the corresponding role of above-mentioned each word with And affiliated sentence ingredient, function division is carried out to the word of the first place name.For example, for place name, " streets A B malls take Fill sales field (C distinguishes shop) ", it can segment to obtain several words such as " C distinguishes shop ", " streets A ", " B malls ", " clothes sales field ", Then role can be determined：" C distinguishes shop " is " region ", and " streets A " is " street ", " B malls " is " place ", " clothes sales field " For " business ".Then it is analyzed according to host-guest architecture, major part can be obtained：" B mall clothes sales field ", dependent part：" the streets A Road ", " areas C shop " can then obtain first layer " streets A ", the second layer " B malls ", third layer " sell by clothes according to step analysis ", the 4th floor " areas C shop " carries out boss's point analysis according to the relations of dependence, then it is principal point that can obtain " A malls ", and " clothes are sold " be sub- point, then unify by place name in word be divided into " core core ", " classification what ", " additional attach ", " other other " four functions, wherein the word with Core Feature can be the proper name for indicating physical entity uniqueness, number The words such as amount, orientation, the word with classification function can be the business for indicating physical entity, the word of generic, have The word of additional function can be the word to remark additionally to physical entity, and the word with other functions can be place Other words in name in addition to functions described above can obtain core for " streets A B mall clothes sales field (C distinguishes shop) " For：" B malls ", classification is：" clothes sales field ", adds and is：" C distinguishes shop ", other are：" streets A ".

Meanwhile server can also calculate each word corresponding weight in the name of place, specifically, server can be pre- Be first that a fixed weight each word is arranged, fixed weight can be technical staff according to each word to judging place The effect degree that corresponding physical entity can play, a fixed numbers of setting can also be by server according to a large amount of samples This (including positive sample：Include two POI information of the same physical entity of same words, anti-sample：I.e. not comprising same words With two POI information of physical entity) fixed numbers of each word that are voluntarily trained, later, server can be with Based on each word place name in corresponding role and affiliated sentence ingredient, the fixed weight of each word is adjusted It is whole, obtain final weights.This, which is sentenced, to the first place name illustrate for function division, to the place of the second place name Reason is similar, is no longer described in detail in the present embodiment.It should be noted that the word analysis process in the above-mentioned name to place is only It is a kind of feasible mode, server can also use other word analysis process, to obtain the corresponding function of each word.

Optionally, the processing that the function of step 102 divides can be based on sample training, and the algorithm of Model Matching is completed, Specific processing can be as follows：Pre-stored multiple training samples are obtained, training sample includes place name sample and place name Function of each word that sample includes in the name sample of the place, is based on multiple training samples, to preset initial algorithm Model is trained, and obtains function partitioning algorithm model, function partitioning algorithm model is based on, to first in the first POI information The word that place name includes carries out function division, and the word for including to the second place name in the second POI information carries out function It divides.

In force, a large amount of training sample can be supplied to server to carry out model training, training sample by technical staff Originally can be the function of each word in place name and place name.Meanwhile some can be arranged for determining word work(in technical staff The rule (i.e. preset initial algorithm model) of energy, such as：The words such as " street ", " community " can be regarded as other functions in the name of place Word, the words such as " clothes ", " food " can be regarded as the word of classification function in the name of place, and then server can be based on above-mentioned trained sample This, is trained preset initial algorithm model, obtains function partitioning algorithm model.In turn, server can be based on training Obtained function partitioning algorithm model respectively in the first POI information the first place name include word and the second POI information In the second place name include word carry out function division.It is noted that the thought of above-mentioned sample training can be used Into other relevant treatments in the present embodiment, for example, the processing for step 102, can be determined by way of sample training Go out the model divided for function, the input of the model is the place name of POI information, is exported as the work(of each word in the name of place Can, it can also determine the model for character labeling respectively, input is place name, exports as the role of each word, is used for The model of structural analysis, input are place name, export the sentence ingredient for each word, defeated for the model that function divides Enter the role for each word and sentence ingredient, export as the function of each word, similarly, the present embodiment to the whole of POI information at Reason, can be completed by a model, that is, it is two POI information to input, the POI information after exporting as merging, or be two A POI information is not the conclusion of the POI information of the same physical entity.

Step 103, according to the matching degree of word with the same function in the first place name and the second place name, the is determined One POI information and the second POI information whether be same physical entity POI information.

In force, after the word to the word of the first place name and the second place name carries out function division, server The word of word and the second place name that can be based on the first place of division result pair name carries out similarity mode, you can with determination The matching degree of word with the same function in first place name and the second place name.Specifically, determining different n, i, j values Corresponding X_(n,i,j), wherein X_(n,i,j)Be in the name of the first place in i-th of word with function n and the second place with work( The matching degree of j-th of word of energy n, according to the different corresponding X of n, i, j value_(n,i,j), determine the corresponding Q of different n values_n, Q_nIt is the matching degree of the set of words with function n in set of words and the second place name in the name of the first place with function n. Same principle can also determine the corresponding P of different n values_n, wherein P_nIt is the word collection with function n in the name of the second place Close the matching degree with the set of words with function n in the name of the first place.Later, server can be according to each of determining Matching degree between set of words with function n judges whether the first POI information and the second POI information are same physical entity POI information.

For example, having two word of a, b with function 1 in the name of the first place, have function 1 in the name of the second place has c, d two Word calculates the matching degree " X of a and c first_(1,a,c)", the matching degree " X of a and d_(1,a,d)", the matching degree " X of b and c_(1,b,c)", and Matching degree " the X of b and d_(1,b,d)”.Then according to aforementioned four matching degree, the word collection with function 1 of the first place name is calculated Close the matching degree Q with the set of words with function 1 in the name of the second place₁.Same principle can calculate separately c and a, b two The matching degree of word, the matching degree " X of two word of d and a, b_(1,c,a)”、“X_(1,c,b)”、“X_(1,d,a)”、“X_(1,d,b)", then according to above-mentioned Four matching degrees, calculate the second place name set of words with function 1 and the first place in the word collection with function 1 The matching degree P of conjunction₁.Later, 4 functions for being referred in step 102, it may be determined that go out Q₁、Q₂、Q₃、Q₄、P₁、P₂、P₃、P₄Eight A matching degree, then server can determine the corresponding matching degree Z of each function according to above-mentioned eight matching degrees_n, and then To the matching degree O of the first place name and the second place name, to judge whether the first POI information and the second POI information are same The POI information of physical entity.

Further, server is according to the matching degree between each of determining the set of words with function n, judges the It, can be according to each tool determined when whether one POI information and the second POI information are the POI information of same physical entity Matching degree between the set of words of functional n determines the matching degree of two places name, later by the matching degree of two places names with Matches criteria degree is compared, if the matching degree of two place names is more than matches criteria degree, it may be considered that the first POI information With the POI information that the second POI information is same physical entity.

For example, being based on the example above, the corresponding matching degree Z of four functions can be respectively obtained₁、Z₂、Z₃、Z₄, then obtain Matching degree O=(the Z of first place name and the second place name₁+Z₂+Z₃+Z₄)/4, then by O and preset matches criteria degree O₁It carries out Compare, if O is more than O₁, it may be considered that the first POI information and the POI information that the second POI information is same physical entity.

It should be noted that the technical staff of server side can provide a large amount of positive samples and anti-sample for server, one A positive sample can have been determined as two POI information of same physical entity, and an anti-sample can have determined that it is not same Two POI information of one physical entity, later server can be carried out by the way of machine learning based on above-mentioned positive/negative sample Training determines that, when two POI information are directed toward same physical entity, two POI information need the smallest match degree met, so After the smallest match degree can be set as matches criteria degree.Herein to the averaging of matching degree calculate, and with matches criteria degree The processing being compared is only to carry out a kind of feasible processing of homogeneity judgement, and this programme is also applied for other processing modes, this Place no longer carries out concrete example.

In addition, if the weight of each word is determined in step 102, during calculating the matching degree of two words, need Consider the weight of each word, for example, the weight of a is A, the weight of c is C, then the matching degree of a and c is " ACX_(1,a,c)”。 Have to be mentioned that, word match herein includes a variety of matched forms, can have same words matching, synonym matching to comply with one's wishes Think close match, abbreviation and full name matching, rectification of name and alias match, Chinese and English matching etc., the matching of fallibility word, such as " A figures are big Tall building " and " A states mansion ", " A loyalties road " and " Roads A " etc..Simultaneously, it is contemplated that the place name of " chain store " is general identical, and It is also possible in a small range while two physical entities, such as fast food quasi-linkage shop, gas station or business hall occurs, so During carrying out homogeneity judgement, the POI information to " chain store " class is needed individually to be considered.It is understood that In order to ensure that the accuracy of POI information homogeneity judgement, technical staff can carry out further based on the handling result of server Artificial judgment, can be specifically to some fallibility POI information carry out artificial judgment.It is appreciated that above-mentioned determination process also may be used With by server, by great amount of samples, the mode for establishing training pattern is determined.

Optionally, when whether determine two POI information is the POI information of same physical entity, it is also contemplated that different Matching degree between the word of function, correspondingly, the processing of step 103 can be as follows：According to the first place name and the second place The matching degree of word with the same function and the first place name and the word with different function in the name of the second place in name Matching degree, determine the first POI information and the second POI information whether be same physical entity POI information.

In force, it is determined that in the first place name and the second place name after the matching degree of word with the same function, The matching degree of the word with different function in the first place name and the second place name can also be determined, specifically, determining different The corresponding X of n, i, m, k value_(ni,mk), wherein X_(ni,mk)It is i-th of word and second with function n in the name of the first place The matching degree of k-th of word with function m in the name of place, according to the different corresponding X of n, i, m, k value_(ni,mk), determine not The corresponding Q with n and m values_(n,m), wherein Q_(n,m)It is the set of words and the second place name in the name of the first place with function n In with function m set of words matching degree, similarly, can also determine the second place name in function n set of words With the matching degree P of the set of words with function m in the name of the first place_(n,m).According to the word with function n in the name of the first place The matching degree of set and the set of words of each function in the name of the second place obtains the word with function n in the name of the first place Set and the second place name in functional set of words matching degree.Later, server can be every according to what is determined In set of words and the second place name in the name of a first place with function n functional set of words matching degree, really Fixed first POI information and the second POI information whether be same physical entity POI information.

For example, for 4 functions (1,2,3,4) in step 102, it can be according to Q_(1,1)、Q_(1,2)、Q_(1,3)、Q_(1,4)It determines Q₁', it can similarly determine Q₂'、Q₃'、Q₄'、P₁'、P₂'、P₃'、P₄', then server can be determined according to above-mentioned eight matching degrees Go out the corresponding matching degree O' of each function, and then obtain the matching degree of the first place name and the second place name, to judge first POI information and the second POI information whether be same physical entity POI information Z_n。

Further, server is according to the set of words each of determined in the name of the first place with function n and the Two places name in functional set of words matching degree, determine whether the first POI information and the second POI information are same object It, can be according to each of determining the set of words with function n and the word with function m when managing the POI information of entity The matching degree of set determines the matching degree of two place names, later carries out the matching degree of two place names with matches criteria degree Compare, if the matching degree of two place names is more than matches criteria degree, it may be considered that the first POI information and the second POI information For the POI information of same physical entity.For example, the corresponding matching degree Z of four functions can be respectively obtained₁'、Z₂'、Z₃'、Z₄', Then the matching degree O'=(Z of the first place name and the second place name are obtained₁'+Z₂'+Z₃'+Z₄')/4, then by O' and preset mark Quasi- matching degree O₁It is compared, if O' is more than O₁, it may be considered that the first POI information and the second POI information are that same physics is real The POI information of body.

It should be noted that for the set of words of different function, the set of words pair of some function can be preset Answer can the set of words of matching feature only matched with function 1,2,3 for example, for function 1, then can basis Q_(1,1)、Q_(1,2)、Q_(1,3)Determine Q₁'。

Optionally, can be that each function sets corresponding weights, it is true then in conjunction with the matching degree of weights and set of words The homogeneity of fixed two POI information, correspondingly, the processing of step 103 can be as follows：According to the first place name and the second place name In word with the same function matching degree, and each corresponding preset weights of function determine the first POI information and second POI information whether be same physical entity POI information.

In force, technical staff can in advance set each function different weights, can also be in different places name In, different weights are set for each function.In this way, server has in the first place name and the second place name is determined After the matching degree of the word of identical function, the corresponding preset weights of each function can be obtained, then in conjunction with weights and matching degree, Determine the first POI information and the second POI information whether be same physical entity POI information.Specifically, can in conjunction with weights and Matching degree obtains the matching degree of the first place name and the second place name, later by the matching degree and matches criteria degree of two place names It is compared, if the matching degree of two place names is more than matches criteria degree, it may be considered that the first POI information and the 2nd POI Information is the POI information of same physical entity.

For example, the corresponding matching degree Z of four functions can be respectively obtained₁、Z₂、Z₃、Z₄, and obtain corresponding weights A, B, C, then D can obtain the matching degree O=(AZ of the first place name and the second place name₁+BZ₂+CZ₃+DZ₄)/4, then by O and in advance If matches criteria degree O₁It is compared, if O is more than O₁, it may be considered that the first POI information and the second POI information are same The POI information of physical entity.

Optionally, when carrying out homogeneity judgement to two POI information, it is also contemplated that in two POI information The matching degree of address, correspondingly, the processing of step 103 can be as follows：The word for including to the first address in the first POI information Level division is carried out, determines the address level belonging to each word, and level is carried out to the second address in the second POI information and is drawn Point, the address level belonging to each word is determined, according to word with the same function in the first place name and the second place name Matching degree and identical address level in the first address word and two address word matching degree, determine the first POI Information and the second POI information whether be same physical entity POI information.

In force, server, can also be to the first POI information after obtaining the first POI information and the second POI information In the first address and the second POI information in the second address handled, i.e., according to preset address layer grade respectively to first Word that the word and the second address that address includes include carries out level division, specifically, address can be divided into country, The levels such as province, city, district, street, cell, floor, doorplate, for example, for the first address, " 11 building, the cities the A areas B C road D cells 301 ", the word of each address level can be obtained：" city：A ", " area：B ", " street：C ", " cell：D ", " floor：11 buildings ", doorplate： 301 ", then the word of identical address level can be matched one by one, obtain the word of the first address in identical address level The matching degree of language and two address word, in this way, can have phase according in obtained the first place name and the second place name The matching of the word and two address word of first address in the matching degree and identical address level of the word of congenerous Degree, determine the first POI information and the second POI information whether be same physical entity POI information.Specifically, can be according to phase With the first address in the level of address word and two address word the first place of matching degree pair name and the second place name Matching degree is adjusted, and the mode of adjustment can be by the word and two address word of the first address in identical address level Matching as weighting coefficient, the matching degree progress addition with the first place name and the second place name, it will be understood that the side of adjustment Formula is varied, and different situations can be selected with different adjustment modes, be not defined one by one herein, after adjusting later Two places name matching degree be compared with matches criteria degree, if adjustment after two places name matching degree be more than mark Quasi- matching degree, it may be considered that the first POI information and the POI information that the second POI information is same physical entity.

For example, the first place name and the first ground in the matching degree O and identical address level in the second place can be obtained The matching degree α of the word and two address word of location is adjusted to obtain α O based on matching degree α to matching degree O, then by α O and Preset matches criteria degree O₁It is compared, if α O are more than O₁, it may be considered that the first POI information and the second POI information are same The POI information of one physical entity.It is understood that when the word for including in address matches, phase can also be used With modes such as word matching, synonym matching and the matchings of fallibility word.

Optionally, when carrying out homogeneity judgement to two POI information, it is also contemplated that in two POI information Coordinate distance, correspondingly, the processing of step 103 can be as follows：According in the coordinate and the second POI information in the first POI information Coordinate, determine the coordinate distance of the first POI information and the second POI information, according to the first place name and the second place name in have There are the matching degree of the word of identical function and the coordinate distance of the first POI information and the second POI information, determines that the first POI believes Breath and the second POI information whether be same physical entity POI information.

In force, server, can also be to the first POI information after obtaining the first POI information and the second POI information In coordinate and the second POI information in coordinate handled, that is, determine the coordinate of the first POI information and the second POI information away from It, later, then can be according to the first place name and second from (i.e. the distance between corresponding point of coordinate can be described as a point distance) The matching degree of word with the same function and above-mentioned coordinate distance in institute's name determine the first POI information and the 2nd POI letters Breath whether be same physical entity POI information.Specifically, can be according to the first place of coordinate distance pair name and the second place name Matching degree be adjusted, later by after adjustment two places name matching degree be compared with matches criteria degree, if tune The matching degree of two places name after whole is more than matches criteria degree, it may be considered that the first POI information and the second POI information are same The POI information of one physical entity.

For example, can obtain the first place name and the second place name matching degree O and coordinate distance D, based on coordinate away from It is adjusted to obtain DO to matching degree O from D, then by DO and preset matches criteria degree O₁It is compared, if DO is more than O₁, then It is considered that the first POI information and the POI information that the second POI information is same physical entity.It is noted that obtaining the After one POI information, the corresponding first area face of the first POI information can also be searched in pre-stored area surface information bank, Area surface is whole overlay areas of the corresponding physical entity of POI information, similarly, can also search the second POI information correspondence Second area face can determine the first POI information if looking only for second area face corresponding to the second POI information The shortest route length (can be described as point-to-plane distance) of the corresponding point of coordinate and second area face, and if finding the firstth area simultaneously Domain face and second area face can then determine that the shortest route length between first area face and second area face (can be described as face Identity distance from), can be in this way, when whether judge the first POI information and the second POI information is the POI information of same physical entity Consider above-mentioned coordinate distance and above-mentioned the shortest route length simultaneously.

Optionally, when carrying out homogeneity judgement to two POI information, it is also contemplated that two POI information coordinates Neighbouring road network information, correspondingly, the processing of step 103 can be as follows：According to the coordinate and the 2nd POI in the first POI information Coordinate in information and pre-stored road network information determine that the road that is separated by of the first POI information and the second POI information is believed Breath according to the matching degree of word with the same function in the first place name and the second place name, and is separated by road information, really Fixed first POI information and the second POI information whether be same physical entity POI information.It is noted that obtaining first After POI information, the corresponding first area face of the first POI information, area can also be searched in pre-stored area surface information bank Domain face is whole overlay areas of the corresponding physical entity of POI information, and it is corresponding similarly can also to search the second POI information Second area face can determine the seat of the first POI information if looking only for second area face corresponding to the second POI information The shortest route length in corresponding point and second area face is marked, and if finding first area face and second area face simultaneously, It can then determine the shortest route length between first area face and second area face, and then homogeneity judgement can carried out Above-mentioned the shortest route length is considered in the process.

In force, server can also obtain two POI letters after obtaining the first POI information and the second POI information Then coordinate in breath removes in pre-stored road network information library to obtain the road information between two coordinates (i.e. according to coordinate First POI information and the second POI information are separated by road information), specifically, can be by the company between two coordinate corresponding points The road that line is passed through, which can be regarded as, is separated by road, then determines that these are separated by the number of road, the category of roads of each road is such as public The length and width etc. of road, urban road, ride etc., each road is separated by road information, and then server can basis The matching degree of word with the same function in first place name and the second place name, and it is separated by road information, determine first POI information and the second POI information whether be same physical entity POI information.Specifically, can be according to being separated by road information pair The matching degree of first place name and the second place name is adjusted, later by the matching degree and standard of two places name after adjustment Matching degree is compared, if the matching degree of two places name after adjustment is more than matches criteria degree, it may be considered that the first POI Information and the POI information that the second POI information is same physical entity.For example, the first place name and the second place name can be obtained Matching degree O, and it is highway that be separated by the number of road, which be 2, category of roads, and then obtains the corresponding parameter t of road quality classification, Matching degree O is adjusted to obtain O/t based on road information is separated by², then by O/t²With preset matches criteria degree O₁Compared Compared with if O/t²More than O₁, it may be considered that the first POI information and the POI information that the second POI information is same physical entity.

It is noted that after obtaining the first POI information, can also be searched in pre-stored area surface information bank The corresponding first area face of first POI information, area surface are whole overlay areas of the corresponding physical entity of POI information, together Reason, can also search the corresponding second area face of the second POI information, if looking only for the secondth area corresponding to the second POI information Domain face can then determine that the shortest route in the corresponding point of the coordinate of the first POI information and second area face passed through is separated by Road, and if finding first area face and second area face simultaneously, can determine first area face and second area face it Between the shortest route passed through be separated by road.

Optionally, when carrying out homogeneity judgement to two POI information, it is also contemplated that two POI information are come Source information, correspondingly, the processing of step 103 can be as follows：There is identical function according in the first place name and the second place name The matching degree of word and the source-information of the first POI information and the source-information of the second POI information matching degree, determine First POI information and the second POI information whether be same physical entity POI information.

In force, server, can also be to two POI information after obtaining the first POI information and the second POI information Source-information matched (i.e. source match), specifically, can be compared to the source-information of two POI information, when When source-information matching degree is high, i.e., two POI information are the same source, then two POI information are the same physical entity The possibility of POI information is relatively low, on the contrary, if two POI information are separate sources, two POI information are the same physics The possibility of the POI information of entity is higher.In this way, server determine the first place name and the second place name in have it is identical After the matching degree of the word of function, the matching degree of the source-information of two POI information can be obtained, then determines the first POI letters Breath and the second POI information whether be same physical entity POI information.Specifically, can be according to the matching of two source-informations Spend to the first place name and the second place name matching degree be adjusted, later by after adjustment two places name matching degree with Matches criteria degree is compared, if the matching degree of two places name after adjustment is more than matches criteria degree, it may be considered that the One POI information and the POI information that the second POI information is same physical entity.

For example, the matching degree O of the first place name and the second place name and the source letter of two POI information can be obtained The matching degree β of breath is adjusted to obtain β O based on matching degree β to matching degree O, then by β O and preset matches criteria degree O₁It carries out Compare, if β O are more than O₁, it may be considered that the first POI information and the POI information that the second POI information is same physical entity.

Optionally, when carrying out homogeneity judgement to two POI information, it is also contemplated that the connection of two POI information It is information, correspondingly, the processing of step 103 can be as follows：There is identical function according in the first place name and the second place name The matching degree of word and the contact details of the contact details of the first POI information and the second POI information matching degree, determine First POI information and the second POI information whether be same physical entity POI information.

In force, server, can also be to two POI information after obtaining the first POI information and the second POI information Contact details matched (i.e. phone match), the contact details of the two are more similar, and matching degree is higher, then two POI information Probability for the POI information of same physical entity is higher, further, can be to home Tel when contact details are base number Code is by number is matched one by one after preceding, it will be understood that when former numbers mismatch in base number, then contact details Matching degree is relatively low, and two POI information are relatively low for the probability of the POI information of same physical entity, and rear several numbers in base number When word mismatches, then contact details may be the different extension sets of same physical entity, so two POI information are that same physics is real The probability of the POI information of body is higher；And contact details be phone number when, due to the flexibility and changeability of phone number, work as mobile phone When number difference, it is also possible to the phone number of different office workers in same physical entity, so when phone number is consistent, two A POI information is that the probability of the POI information of same physical entity is higher, and when phone number is inconsistent, to judging that two POI believe Breath whether be same physical entity POI information influence it is little.In this way, server is determining the first place name and second In institute's name after the matching degree of word with the same function, the matching degree of the contact details of two POI information can be obtained, then Determine the first POI information and the second POI information whether be same physical entity POI information.Specifically, can be joined according to two It is that the first place of matching degree pair name of information and the matching degree of the second place name are adjusted, later by two places after adjustment The matching degree of name is compared with matches criteria degree, if the matching degree of two places name after adjustment is more than matches criteria degree, It may be considered that the first POI information and the POI information that the second POI information is same physical entity.

For example, the matching degree O of the first place name and the second place name and the contact letter of two POI information can be obtained The matching degree χ of breath is adjusted to obtain χ O based on matching degree χ to matching degree O, then by χ O and preset matches criteria degree O₁It carries out Compare, if χ O are more than O₁, it may be considered that the first POI information and the POI information that the second POI information is same physical entity.

Based on above description, homogeneity judgement is carried out to the first POI information and the second POI information Fig. 2 shows a kind of Process flow, wherein while considering in two POI information：Place name in the corresponding set of words of 4 functions matching degree, The matching degree of classification, whether be " chain store " POI information, the matching degree of address, coordinate distance (including point distance, point face Apart from knead dough identity distance from), be separated by many factors such as road information, the matching degree of contact details, the matching degree of source-information.It can be with Understand, in the present solution, can a kind of only above-mentioned processing because usually carrying out homogeneity judgement, can also in conjunction with many factors it is common Carry out the processing of homogeneity judgement.In addition, the mode of model training may be used in server, completed by machine learning same Property judge, include a large amount of training samples in training set specifically, the training set of ten million scale can be constructed first, each training Sample can be two POI information for having been determined as same physical entity, and the pass in known two POI information between each information System, such server can establish the model for carrying out homogeneity judgement, which can be set with each information and need to meet Condition, if the matching degree of place name is more than A, the matching degree of address name is more than B, and coordinate distance is more than C etc., in turn, in determination After going out the relationship in two POI information between each information, it can judge whether two POI information are same object using above-mentioned model Manage the POI information of entity.

Optionally, the POI information for belonging to same physical entity can be merged processing, correspondingly, after step 103 Processing can be as follows：If the first POI information and the POI information that the second POI information is same physical entity, by the first POI Information and the second POI information merge.

In force, if it is determined that go out the first POI information and the second POI information for the POI information of same physical entity, then The same category information for including in first POI information and the second POI information can be merged, specifically, for two POI information Include certain category information, then can such as select more complete one of information according to preset rules, or selection source degree of belief compared with High one, for only there are one certain category information that POI information includes, then can directly carry out supplement addition.For example, first POI information include place name A1, address B1, coordinate C1, tetra- kinds of information of neighbouring businessman D, the second POI information include place name A2, Tetra- kinds of address B2, classification E, coordinate C2 information, it is known that the source degree of belief of the first POI information is higher than the source of the second POI information, Place name, address in second POI information, coordinate are more complete than the place name in the first POI information, address, coordinate, so If according to selection higher one principle of source degree of belief, merge after POI information include place name A1, address B1, Five kinds of coordinate C1, neighbouring businessman D, classification E information, if according to selection more complete one principle of information, after merging POI information includes five kinds of place name A2, address B2, coordinate C2, neighbouring businessman D, classification E information.Further, it is based on step 101 processing, it is known that the first POI information is pending POI information, the second POI information is to have determined that physical entity POI information, if not finding that same physics corresponding with the first POI information is real in all POI information for having determined that physical entity Body, then can filter out in POI information to be handled and meet the POI of preset proximity condition with the first POI information and believe Breath, and then the first POI information and the POI information screened can be handled one by one, i.e., server can filter out The second POI information is obtained in the POI information come, and then subsequent match merging treatment can be carried out, it finally, can also be according to merging POI information afterwards adds corresponding ground map logo on map.It is appreciated that if there is no the second POI information and the first POI Information is the POI information of same physical entity, then can first judge the accuracy of the first POI information, is then believed according to the first POI Breath adds corresponding ground map logo on map.

Fig. 3 is a kind of schematic diagram of processing POI information disclosed in this programme, and first, server can be by " surveying on the spot The various ways such as spy ", " partner ", " network crawls " get pending POI information, then choose a pending POI letter It ceases (i.e. the first POI information), then in pending POI information or has determined that in the POI information of physical entity and select the 2nd POI Information carries out homogeneity judgement, if two POI information correspond to same object to the first POI information and the second POI information in turn Entity is managed, then merges two POI information, and create ground map logo in GIS-Geographic Information System according to the POI information after merging.

Based on the same technical idea, the embodiment of the present invention additionally provides a kind of device of processing POI information, such as Fig. 4 institutes Show, which includes：

Data obtaining module 401, for obtaining the first POI information and meeting preset connect with first POI information Second POI information of recency condition；

Function division module 402, the word for including to the first place name in first POI information carry out function It divides, and the word for including to the second place name in second POI information carries out function division；

First determining module 403, for there is identical function according in first place name and second place name Word matching degree, determine first POI information and second POI information whether be same physical entity POI letters Breath.

Optionally, as shown in figure 5, described device further includes：

Sample acquisition module 404, for obtaining pre-stored multiple training samples, the training sample includes place name The function for each word that sample and place name sample include；

Training module 405 is trained preset initial algorithm model, obtains for being based on the multiple training sample To function partitioning algorithm model；

The function division module 402, is used for：

Based on the function partitioning algorithm model, the word that includes to the first place name in first POI information into Row function divides, and the word for including to the second place name in second POI information carries out function division.

Optionally, first determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, Yi Jisuo The matching degree for stating the word with different function in the first place name and second place name, determines first POI information With second POI information whether be same physical entity POI information.

Optionally, first determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, and often The corresponding preset weights of a function, determine whether first POI information and second POI information are same physical entity POI information.

Optionally, as shown in fig. 6, described device further includes：

Level division module 406, the word for including to the first address in first POI information carry out level and draw Point, determine the address level belonging to each word, and level division is carried out to the second address in second POI information, really Address level belonging to fixed each word；

First determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, Yi Jixiang With the matching degree of the word of the first address and the two address word described in the level of address, the first POI letters are determined Breath and second POI information whether be same physical entity POI information.

Optionally, as shown in fig. 7, described device further includes：

Second determining module 407, for according in the coordinate and second POI information in first POI information Coordinate determines the coordinate distance of first POI information and second POI information；

First determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, Yi Jisuo The coordinate distance for stating the first POI information and second POI information determines first POI information and second POI information Whether be same physical entity POI information.

Optionally, as shown in figure 8, described device further includes：

Third determining module 408, for according in the coordinate and second POI information in first POI information Coordinate and pre-stored road network information determine that the road that is separated by of first POI information and second POI information is believed Breath；

First determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, Yi Jisuo State and be separated by road information, determine first POI information and second POI information whether be same physical entity POI letters Breath.

Optionally, first determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, Yi Jisuo The matching degree for stating the source-information of the first POI information and the source-information of second POI information determines the first POI letters Breath and second POI information whether be same physical entity POI information.

Optionally, first determining module 403, is used for：

According to the matching degree of word with the same function in first place name and second place name, Yi Jisuo The matching degree for stating the contact details of the first POI information and the contact details of second POI information determines the first POI letters Breath and second POI information whether be same physical entity POI information.

Optionally, as shown in figure 9, described device further includes：

Merging module 409, if being same physical entity for first POI information and second POI information POI information then merges first POI information and second POI information.

It should be noted that：Above-described embodiment provide processing POI information device when handling POI information, only more than The division progress of each function module is stated for example, in practical application, it can be as needed and by above-mentioned function distribution by difference Function module complete, i.e., the internal structure of device is divided into different function modules, with complete it is described above whole or Person's partial function.In addition, the embodiment of the method category of the device for the processing POI information that above-described embodiment provides and processing POI information In same design, specific implementation process refers to embodiment of the method, and which is not described herein again.

Figure 10 is the structural schematic diagram of server provided in an embodiment of the present invention.The server 1000 can be because of configuration or performance It is different and generate bigger difference, may include one or more central processing units (central processing Units, CPU) 1022 (for example, one or more processors) and memory 1032, one or more storage applications The storage medium 1030 (such as one or more mass memory units) of program 1042 or data 1044.Wherein, memory 1032 and storage medium 1030 can be of short duration storage or persistent storage.The program for being stored in storage medium 1030 may include one A or more than one module (diagram does not mark), each module may include to the series of instructions operation in server.More into One step, central processing unit 1022 could be provided as communicating with storage medium 1030, and storage medium is executed on server 1000 Series of instructions operation in 1030.

Server 1000 can also include one or more power supplys 1029, one or more wired or wireless nets Network interface 1050, one or more input/output interfaces 1058, one or more keyboards 1056, and/or, one or More than one operating system 1041, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM Etc..

Server 1000 may include have memory and one either more than one program one of them or one A procedure above is stored in memory, and be configured to by one either more than one processor execute it is one or one A procedure above includes the instruction for carrying out above-mentioned processing POI information.

One of ordinary skill in the art will appreciate that realizing that all or part of step of above-described embodiment can pass through hardware It completes, relevant hardware can also be instructed to complete by program, the program can be stored in a kind of computer-readable In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..

The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all the present invention spirit and Within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.

Claims

1. a kind of method of processing POI information, which is characterized in that the method includes：

The 2nd POI for obtaining first information point POI information and meeting preset proximity condition with first POI information believes Breath；

The word for including to the first place name in first POI information carries out function division, and to second POI information In the second place name include word carry out function division；

According to the matching degree of word with the same function in first place name and second place name, described the is determined One POI information and second POI information whether be same physical entity POI information.

2. according to the method described in claim 1, it is characterized in that, the method further includes：

Pre-stored multiple training samples are obtained, the training sample includes place name sample and place name sample includes Each word the place name sample in function；

Based on the multiple training sample, preset initial algorithm model is trained, obtains function partitioning algorithm model；

The word that the first place name in first POI information includes carries out function division, and to the 2nd POI The word that the second place name in information includes carries out function division, including：

Based on the function partitioning algorithm model, the word for including to the first place name in first POI information carries out work( It can divide, and the word for including to the second place name in second POI information carries out function division.

3. according to the method described in claim 1, it is characterized in that, described according to first place name and second place The matching degree of word with the same function, determines whether first POI information and second POI information are same in name The POI information of physical entity, including：

According to the matching degree and described the of word with the same function in first place name and second place name The matching degree of the word with different function, determines first POI information and institute in one place name and second place name State the second POI information whether be same physical entity POI information.

4. according to the method described in claim 1, it is characterized in that, described according to first place name and second place The matching degree of word with the same function, determines whether first POI information and second POI information are same in name The POI information of physical entity, including：

According to the matching degree of word with the same function and each work(in first place name and second place name Can corresponding preset weights, determine first POI information and second POI information whether be same physical entity POI Information.

5. according to claim 1-4 any one of them methods, which is characterized in that the method further includes：

Level division is carried out to the word that the first address in first POI information includes, determines the ground belonging to each word Location level, and level division is carried out to the second address in second POI information, determine the address layer belonging to each word Grade；

The matching degree according to word with the same function in first place name and second place name determines institute State the first POI information and second POI information whether be same physical entity POI information, including：

According to the matching degree of word with the same function in first place name and second place name, and in the same manner The matching degree of the word of first address and the two address word described in the level of location, determine first POI information and Second POI information whether be same physical entity POI information.

6. according to claim 1-4 any one of them methods, which is characterized in that the method further includes：

According to the coordinate in the coordinate and second POI information in first POI information, first POI information is determined With the coordinate distance of second POI information；

According to the matching degree and described the of word with the same function in first place name and second place name The coordinate distance of one POI information and second POI information determines whether are first POI information and second POI information For the POI information of same physical entity.

7. according to claim 1-4 any one of them methods, which is characterized in that the method further includes：

According to the coordinate and pre-stored road network letter in the coordinate and second POI information in first POI information Breath, determine first POI information and second POI information is separated by road information；

According to the matching degree and the phase of word with the same function in first place name and second place name Every road information, determine first POI information and second POI information whether be same physical entity POI information.

8. according to claim 1-4 any one of them methods, which is characterized in that described according to first place name and described The matching degree of word with the same function, determines that first POI information and second POI information are in the name of second place The no POI information for same physical entity, including：

According to the matching degree and described the of word with the same function in first place name and second place name The matching degree of the source-information of one POI information and the source-information of second POI information, determine first POI information and Second POI information whether be same physical entity POI information.

9. according to claim 1-4 any one of them methods, which is characterized in that described according to first place name and described The matching degree of word with the same function, determines that first POI information and second POI information are in the name of second place The no POI information for same physical entity, including：

According to the matching degree and described the of word with the same function in first place name and second place name The matching degree of the contact details of the contact details of one POI information and second POI information, determine first POI information and Second POI information whether be same physical entity POI information.

10. according to claim 1-9 any one of them methods, which is characterized in that the method further includes：

If first POI information and the POI information that second POI information is same physical entity, by described first POI information and second POI information merge.

11. a kind of device of processing POI information, which is characterized in that described device includes：

Data obtaining module, for obtaining the first POI information and meeting preset proximity condition with first POI information The second POI information；

Function division module, the word for including to the first place name in first POI information carry out function division, and The word for including to the second place name in second POI information carries out function division；

First determining module, for according to word with the same function in first place name and second place name Matching degree, determine first POI information and second POI information whether be same physical entity POI information.

12. according to the devices described in claim 11, which is characterized in that described device further includes：

Sample acquisition module, for obtaining pre-stored multiple training samples, the training sample include place name sample and Function of each word that the place name sample includes in the name sample of the place；

Training module is trained preset initial algorithm model, obtains function and draw for being based on the multiple training sample Divide algorithm model；

The function division module, is used for：

13. according to the devices described in claim 11, which is characterized in that first determining module is used for：

14. according to the devices described in claim 11, which is characterized in that first determining module is used for：

15. according to claim 11-14 any one of them devices, which is characterized in that described device further includes：

Level division module, the word for including to the first address in first POI information carry out level division, determine Address level belonging to each word, and level division is carried out to the second address in second POI information, determine each word Address level belonging to language；

First determining module, is used for：

16. according to claim 11-14 any one of them devices, which is characterized in that described device further includes：

Second determining module is used for according to the coordinate in the coordinate and second POI information in first POI information, really The coordinate distance of fixed first POI information and second POI information；

First determining module, is used for：

17. according to claim 11-14 any one of them devices, which is characterized in that described device further includes：

Third determining module, for according to the coordinate in the coordinate and second POI information in first POI information, with And pre-stored road network information, determine first POI information and second POI information is separated by road information；

First determining module, is used for：

18. according to claim 11-14 any one of them devices, which is characterized in that first determining module is used for：

19. according to claim 11-14 any one of them devices, which is characterized in that first determining module is used for：

20. according to claim 11-19 any one of them devices, which is characterized in that described device further includes：

Merging module, if for the POI information that first POI information and second POI information are same physical entity, Then first POI information and second POI information are merged.