CN104615620A - Map search type identification method and device and map search method and system - Google Patents
Map search type identification method and device and map search method and system Download PDFInfo
- Publication number
- CN104615620A CN104615620A CN201410289451.1A CN201410289451A CN104615620A CN 104615620 A CN104615620 A CN 104615620A CN 201410289451 A CN201410289451 A CN 201410289451A CN 104615620 A CN104615620 A CN 104615620A
- Authority
- CN
- China
- Prior art keywords
- search results
- query string
- search
- bar
- type
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a map search type identification method and device and a map search method and system. The map search type identification method comprises the steps of obtaining a search result by conducting map search for query strings; extracting feature information of a preset feature type of the search result and generating a feature vector containing the feature information; determining a search type of each query string as the probability of extensive retrieval according to the feature vector and an extensive retrieval prediction model, wherein the extensive retrieval prediction model is obtained by training the feature information of the preset feature type corresponding to each query string in a preset query string set. By means of the scheme of the map search type identification method and device and the map search method and system, identification for a map search type can be rapidly achieved without depending on a certain word list, the maintenance cost is low, and the precision is high.
Description
Technical field
The present invention relates to field of map search, particularly a kind of map search kind identification method, a kind of map search type identification device and a kind of map search method, a kind of map search system.
Background technology
Both provide function of search in map products numerous at present, its Query (query string) submitted to user, the relevant POI data of search also returns to user after sorted.Every number in map is stated to be a POI (Point of Interest, point of interest), each POI comprises title, classification (Category), longitude, latitude, importance degree (POIRank, represent the importance degree of POI data, gained is calculated based on the significance level of POI under affiliated classification or rank, the calculating of the lower POIRank of different classification may be different, in such as hospital's classification, the POIRank of Grade A hospital can higher than the hospital of other ranks, in cuisines classification user comment on quantity can and POIRank directly related) etc. much information.Type belonging to Query in map search, is generally divided into precise search and general retrieval two class.Precise search refers to that the Query that in map search, user submits to is searching for certain concrete POI data point.General retrieval refers to that the Query that user in map search submits to is class name, chain trade name, suffix word etc., the POI data of a certain class that what namely user wanted to search for is and certain POI data point of non-specific.The Search Results of general retrieval Query generally spatially distributes and compares dispersion, has certain general character in addition in classification.Typical general retrieval Query can be such as cuisines, hotel, university etc., and when user is at Beijing's search cuisines, map search can provide all data about cuisines in Beijing and Pagination Display.In map search technical finesse, general retrieval and precise search can be distinguished to some extent in search strategy, ordering strategy etc., and when such as user uses mobile phone searching cuisines, the data preferentially providing the cuisines near user position can have better Consumer's Experience.Therefore, be that general retrieval or precise search identify to the retrieval type of search, the identification of especially general retrieval becomes an important component part in map search.
In map search, basic general retrieval can based on vocabulary, and typical general term list is as item name vocabulary, chain store's vocabulary, label vocabulary, suffix vocabulary etc.If the Query of user's input belongs to certain known general term list, then think that the search-type of current Query is general retrieval.But based on the general retrieval of vocabulary, having certain limitation: on the one hand, have the general retrieval Query of various Inexact Newton methods intention in Query during actual user searches, vocabulary is difficult to cover all general retrieval vocabulary; In addition on the one hand, if general term list is excessive, relevant maintenance workload can strengthen, and may occur certain noise in vocabulary simultaneously.
Summary of the invention
Based on this, be necessary for above-mentioned problems of the prior art, a kind of map search kind identification method, a kind of map search type identification device and a kind of map search method, a kind of map search system are provided, its identification that can realize map search type rapidly without the need to relying on specific vocabulary, maintenance cost is low, and precision is high.
For achieving the above object, the embodiment of the present invention by the following technical solutions:
A kind of map search kind identification method, comprises step:
Obtain Search Results query string being carried out to map search acquisition;
Extract the characteristic information of the default characteristic type of described Search Results, generate the proper vector comprising described characteristic information;
Determine that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set.
A kind of map search type identification device, comprising:
Data obtaining module, for obtaining Search Results query string being carried out to map search acquisition;
Feature vector generation module, for extracting the characteristic information of the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Probability determination module, for determining that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set.
A kind of map search method, comprises step:
Receive map search instruction, described map search instruction comprises query string;
Search for according to described map search instruction, obtain Search Results;
Extract the characteristic information of the default characteristic type of described Search Results, obtain the proper vector of described Search Results;
Determine that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, determine the search-type of described query string according to the search-type of the described query string probability that is general retrieval, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set;
Determine the Graphics Processing mode of described Search Results according to the search-type of described query string, and according to this Graphics Processing mode, described Search Results is processed.
A kind of map search system, comprising:
Search module, for receiving map search instruction, described map search instruction comprises query string, searches for according to described map search instruction, obtains Search Results;
Feature vector generation module, for extracting the characteristic information of the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Search-type determination module, for determining that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, determine the search-type of described query string according to the search-type of the described query string probability that is general retrieval, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set;
Graphics Processing module, for determining the Graphics Processing mode of described Search Results according to this search-type, and processes described Search Results according to this Graphics Processing mode.
According to the scheme of the embodiment of the present invention as above, its be obtain to inquire carry out map search Search Results after, extract the characteristic information of the default characteristic type in this Search Results, generate the proper vector comprising these characteristic informations, and based on this proper vector, the general retrieval forecast model set up carries out prediction to the probability that the search-type of query string is general retrieval and determines, its identification that can realize map search type rapidly without the need to relying on specific vocabulary, maintenance cost is low, certain recognition capability can be had to type general retrieval Query, precision is high.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of map search kind identification method embodiment of the present invention;
Fig. 2 is the schematic flow sheet setting up general retrieval forecast model in an example;
Fig. 3 is the structural representation of map search recognition device embodiment of the present invention;
Fig. 4 is the schematic flow sheet of map search embodiment of the method for the present invention;
Fig. 5 is the structural representation of map search system embodiment of the present invention;
Fig. 6 is the module map of a computer system 1000 that can realize the embodiment of the present invention.
Embodiment
Below in conjunction with preferred embodiment wherein, illustrated in greater detail is carried out to the present invention program.
The schematic flow sheet of map search kind identification method embodiment of the present invention has been shown in Fig. 1.As shown in Figure 1, the map search kind identification method in the present embodiment comprises step:
Step S101: obtain Search Results query string being carried out to map search acquisition;
Step S102: the characteristic information extracting the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Step S103: determine that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set.
According to the scheme of the embodiment of the present invention as above, its be obtain to inquire carry out map search Search Results after, extract the characteristic information of the default characteristic type in this Search Results, generate the proper vector comprising these characteristic informations, and prediction is carried out to the probability that the search-type of query string is general retrieval determine based on this proper vector, the general retrieval forecast model set up, its identification that can realize map search type rapidly without the need to relying on specific vocabulary, maintenance cost is low, and precision is high.
The schematic flow sheet setting up general retrieval forecast model in a concrete example has been shown in Fig. 2, and as shown in Figure 2, the process setting up general retrieval forecast model can comprise:
Step S201: build query string set, include general retrieval and inquisition string and non-general retrieval and inquisition string in described query string set;
Step S202: carry out map search to the query string in described query string set respectively, obtains the Search Results corresponding with each query string;
Step S203: the characteristic information extracting the default characteristic type of Search Results corresponding to each query string respectively, generates the proper vector comprising each query string characteristic of correspondence information respectively;
Step S204: logistic regression study process is carried out to the proper vector of each query string, obtains described general retrieval forecast model.
As shown in Figure 2, when setting up general retrieval forecast model, between above-mentioned steps S203 and step S204, can also step be comprised:
Step S2034: be normalized with each query string characteristic of correspondence vector.
Correspondingly, between above-mentioned steps S102 and step S103, can also comprise:
Step S1023: the proper vector obtained in step S102 is normalized.
After being normalized each proper vector, each proper vector can be made can to represent in same interval, simplify subsequent processes, improve treatment effeciency.
Wherein, in above-mentioned map search type identification and set up in the process of general retrieval forecast model, when extracting the characteristic information of the default characteristic type of Search Results, above-mentioned default characteristic type can set based on actual needs, wherein in a concrete example, can be the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
Below in conjunction with one of them concrete example, map search kind identification method of the present invention is described in detail.In the following description, be described in conjunction with the process of establishing of general retrieval forecast model.
Before specifically carrying out map search type identification, need to set up general retrieval forecast model, in the present invention program, by analyzing the search result set of Query, extract the feature that general retrieval is relevant, then based on a large amount of flag data, adopt the method for machine learning to train, thus obtain general retrieval forecast model.
When setting up general retrieval forecast model, the Query set first needing structure one large, this Query comprises typical general retrieval Query and non-general retrieval Query in combining.General retrieval sample Query can from existing general term list, also can from artificial mark, and non-general retrieval (i.e. precise search) Query can from general POI data point title, also can from artificial mark.Wherein, artificial mark refers to manually differentiate whether be that general retrieval Query is also marked (such as general retrieval is labeled as 1, and non-general retrieval is labeled as 0) based on the Search Results of current Query.
Then by the above-mentioned sample Query marked, be submitted to map search system and search for, the base attribute of every bar result can be obtained from map search system background, the correlated characteristic of the default characteristic type of general retrieval can be extracted based on these attributes further.In a concrete example of the present invention, some features relevant to general retrieval and correlation computations mode thereof and feature description can as shown in following table tables 1.
The general retrieval correlated characteristic of table 1 and feature description thereof
The determination mode of the feature in above-mentioned table 1, all can adopt any mode that is existing and that later may occur at present to carry out.All kinds of feature shown in above-mentioned table 1, when specific implementation, can all adopt, and according to the situation of real data, can remove Partial Feature wherein, also can increase Partial Feature.
After extracting the above-mentioned correlated characteristic of mentioned above searching results, based on the needs of application, can be the general retrieval mark of 0/1 to each sample Query codomain, represent whether current Query is general retrieval with this, in this example, mark value 1 is general retrieval, and mark value 0 is non-general retrieval.Thus each sample Query can be expressed as the vector of a N*1, whole training set just can be expressed as the vector of a M*N dimension, and wherein N is the number of feature, and M is the quantity of Query in training examples set.
Then can adopt the method for Logistic Regression (logistic regression) to each proper vector, logistic regression study process be carried out to above-mentioned proper vector, general retrieval forecasting problem is converted into recurrence learning and forecasting problem.
Before use Logistic Regression trains, first can be normalized above-mentioned each proper vector, to make each proper vector can represent in same interval, facilitate subsequent processes, improve treatment effeciency.
When being normalized, various possible mode can be adopted to carry out.The mode that wherein a kind of typical feature normalization method can adopt linear function to change is carried out, and the formula of linear function conversion is as described below:
y＝(x-MinValue)/(MaxValue-MinValue)
In above-mentioned formula, y represents the value after being normalized x, the minimum value of MinValue representation feature value, the maximal value in MaxValue representation feature value.In the present invention program, for the different feature in each proper vector, above-mentioned formula is adopted to be normalized respectively, to be mapped in [0,1] interval.
Another typical normalization mode is the normalization based on sample distribution between attribute area, and detailed process can be as described below:
In statistics training set, the span of single eigenwert, is divided into N number of interval by this span;
The number percent of positive example (being general retrieval Query) in each interval adding up this attribute, namely attribute Query in current interval belongs to the probability of general retrieval;
Given original property value, based on the interval belonging to this value, can obtain the probable value of [0,1].
After carrying out above-mentioned normalized, Logistic Regression learning training can be carried out for obtaining the proper vector after normalized, to obtain general retrieval forecast model.The machine learning method of concrete Logistic Regression can adopt various possible mode that is existing and that later may occur at present to carry out, and includes but not limited to that decision tree, support vector machine (SVM), artificial neural network (ANN), gradient increase progressively decision tree (GBDT) etc.
For ease of understanding, below one being done to Logistic Regression and briefly describing.
As a bare metal study scene, given N number of training sample (x
1, y
1), (x
2, y
2), (x
3, y
3) ..., (x
n, y
n), wherein x
i∈ R
nbe a n-dimensional vector, be used for the value of expression i-th sample in this n feature, and y
i{-1 ,+1} to illustrate this sample be positive sample (+1) or negative sample (-1) to ∈, can represent it is general retrieval in the present invention program with positive sample.Logic Regression Models is exactly the function being logical function (Logistic Function) by
the probability being positive sample by the proper vector xi of i-th sample and this sample links up:
Wherein, p (y
i=+1|x
i, w) represent sample x
ifor the probability of positive sample, w
tx
i=β
0+ β
1x
i1+ β
2x
i2+ ...+β
nx
in, the form of w parameter is β
0, β
1, β
2..., β
n, parameter w does different weightings to the n of an x dimension, thus can calculate w
tx
i, then this value is adjusted to 0 to 1 by the logical function of S type, is the probability of positive sample.
The target of logistic regression is just to locate best w, makes the mark of positive sample all larger, and the mark of negative sample is all smaller simultaneously.Logistic Regression training result model is w parameter sets, i.e. β
0, β
1, β
2..., β
n.Logistic Regression finds best w by maximal possibility estimation (Maximum LikelihoodEstimation) principle.In a concrete example, LibLinear storehouse can be used to carry out the process of Logistic Regression, what adopt in this storehouse is Trusted Region Newton (TRON) method, thus can obtain good logistic regression performance.
After obtaining above-mentioned general retrieval forecast model, can whether be that general retrieval is predicted to search-type in the map search of reality.
During prediction, after given Query, based on the set of searching for the Search Results returned, extract above-mentioned correlated characteristic, and generate the proper vector comprising these features, and the proper vector obtained is normalized, then the general retrieval forecast model of above-mentioned formula (1) is called, proper vector after the above-mentioned normalized of current Query is carried out computational prediction as input, the search-type that namely can obtain current Query belongs to the probability of general retrieval, and namely whether the current Query of identifiable design is general retrieval.
Based on the map search kind identification method with the invention described above, the present invention also provides a kind of map search type identification device.The structural representation of map search type identification device embodiment of the present invention has been shown in Fig. 3.As shown in Figure 3, the present embodiment device comprises:
Data obtaining module 301, for obtaining Search Results query string being carried out to map search acquisition;
Feature vector generation module 302, for extracting the characteristic information of the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Probability determination module 303, for determining that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set.
According to the scheme of the embodiment of the present invention as above, its be obtain to inquire carry out map search Search Results after, extract the characteristic information of the default characteristic type in this Search Results, generate the proper vector comprising these characteristic informations, and prediction is carried out to the probability that the search-type of query string is general retrieval determine based on this proper vector, the general retrieval forecast model set up, its identification that can realize map search type rapidly without the need to relying on specific vocabulary, maintenance cost is low, and precision is high.
As shown in Figure 3, wherein in a concrete example, map search type identification device of the present invention can also comprise:
Model construction module 300, for generating described general retrieval forecast model.
The structural representation of a concrete example of this model construction module 300 is also show in Fig. 3.Shown in composition graphs 3, in this example, this model construction module 300 comprises:
Query string collection modules 3001, for building query string set, includes general retrieval and inquisition string and non-general retrieval and inquisition string in described query string set;
Model eigenvectors generation module 3002, for obtaining the Search Results corresponding with each query string query string in described query string set being carried out to map search acquisition, extract the characteristic information of the default characteristic type of Search Results corresponding to each query string respectively, generate the proper vector comprising each query string characteristic of correspondence information respectively;
Model generation module 3004, for carrying out logistic regression study process to the proper vector of each query string, obtains described general retrieval forecast model.
Wherein, as shown in Figure 3, in this example, this model construction module 300 can also comprise the first normalized module 3003, for being normalized with each query string characteristic of correspondence vector;
Now, above-mentioned model generation module 3004, for carrying out logistic regression study process to the proper vector of each query string after normalized.
Corresponding, in a concrete example of the present invention, this map search type identification device can also comprise:
Second normalized module 3023, is normalized for the proper vector generated feature vector generation module 302;
Now, above-mentioned probability determination module 303, be based on the second normalized module 3023 normalized after proper vector determine that the search-type of described query string is the probability of general retrieval.
After being normalized each proper vector, each proper vector can be made can to represent in same interval, simplify subsequent processes, improve treatment effeciency.
Wherein, to be above-mentionedly normalized and the mode of logistic regression study process can identical with the map search kind identification method of the invention described above, any mode that is existing and that later may occur at present can be adopted to carry out.
In above-mentioned map search type identification and set up in the process of general retrieval forecast model, when extracting the characteristic information of the default characteristic type of Search Results, above-mentioned default characteristic type can set based on actual needs, wherein in a concrete example, can be the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
Based on the map search kind identification method of the invention described above, the present invention also provides a kind of map search method.The schematic flow sheet of map search embodiment of the method for the present invention has been shown in Fig. 4.As shown in Figure 4, the map search method in the present embodiment comprises step:
Step S401: receive map search instruction, described map search instruction comprises query string;
Step S402: search for according to described map search instruction, obtains Search Results;
Step S403: the characteristic information extracting the default characteristic type of described Search Results, obtains the proper vector of described Search Results;
Step S404: determine that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, determine the search-type of described query string according to the search-type of the described query string probability that is general retrieval, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set;
Step S405: the Graphics Processing mode determining described Search Results according to the search-type of described query string, and according to this Graphics Processing mode, described Search Results is processed.
Based on the map search method of the embodiment of the present invention as above, it is when carrying out map search, after obtaining the Search Results of map search, also by extracting the characteristic information of the default characteristic type in this Search Results, generate the proper vector comprising these characteristic informations, and based on this proper vector, the general retrieval forecast model set up carries out prediction to the probability that the search-type of query string is general retrieval and determines, and based on the search-type determined, Search Results is processed, it is when determining general retrieval probability, without the need to relying on the identification that specific vocabulary can realize map search type rapidly, maintenance cost is low, and precision is high.
Above-mentioned general retrieval forecast model, can adopt the mode shown in Fig. 2 to set up, specifically can comprise the steps:
Step S201: build query string set, include general retrieval and inquisition string and non-general retrieval and inquisition string in described query string set;
Step S202: carry out map search to the query string in described query string set, obtains the Search Results corresponding with each query string;
Step S203: the characteristic information extracting Search Results corresponding to each query string respectively, generates the proper vector comprising each query string characteristic of correspondence information respectively;
Step S204: logistic regression study process is carried out to the proper vector of each query string, obtains described general retrieval forecast model.
As shown in Figure 2, when setting up general retrieval forecast model, between above-mentioned steps S203 and step S204, can also step be comprised:
Step S2034: be normalized with each query string characteristic of correspondence vector.
Correspondingly, between above-mentioned steps S403 and step S404, can also comprise:
Step S4034: the proper vector obtained in step S403 is normalized.
After being normalized each proper vector, each proper vector can be made can to represent in same interval, simplify subsequent processes, improve treatment effeciency.
Wherein, to be above-mentionedly normalized and the mode of logistic regression study process can identical with the map search kind identification method of the invention described above, any mode that is existing and that later may occur at present can be adopted to carry out.
In above-mentioned map search type identification and set up in the process of general retrieval forecast model, when extracting the characteristic information of the default characteristic type of Search Results, above-mentioned default characteristic type can set based on actual needs, wherein in a concrete example, can be the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
Based on map search method of the present invention as above, be described in detail below in conjunction with one of them concrete processing procedure.
When carrying out map search, user, by input Query, waits button to send map by click search and searches plain instruction.After receiving this map search instruction, carry out map search for this Query, obtain Search Results.
Then the characteristic information of the above-mentioned characteristic type in Search Results is extracted, obtain characteristic of correspondence vector, and after this proper vector is normalized, above-mentioned general retrieval forecast model is adopted to predict the probability that the search-type of this Query is general retrieval, after obtaining general retrieval probability, if general retrieval probability is greater than the threshold value of setting, then can think that the retrieval type of this Query is general retrieval, otherwise, can think that the retrieval type of this Query is precise search.
Then, after Search Results being processed based on determined retrieval type, then the Search Results after process is shown.Such as, to retrieve type for precise search, then can directly will show by Search Results accurately.Be generally be retrieved as example to retrieve type, then can adjust in the sequence of retrieval, display mode.Such as, when searching for cuisines, hotel etc., preferentially can provide the Search Results near user position, and can highlighted display etc. be carried out, to have better Consumer's Experience.
Based on the thought identical with above-mentioned map search method, the present invention also provides a kind of map search system.The structural representation of map search system embodiment of the present invention has been found in Fig. 5.As shown in Figure 5, the system in the present embodiment comprises:
Search module 501, for receiving map search instruction, described map search instruction comprises query string, searches for according to described map search instruction, obtains Search Results;
Feature vector generation module 502, for extracting the characteristic information of the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Search-type determination module 503, for determining that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, determine the search-type of described query string according to the search-type of the described query string probability that is general retrieval, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set;
Graphics Processing module 504, for determining the Graphics Processing mode of described Search Results according to this search-type, and processes described Search Results according to this Graphics Processing mode.
Based on the map search system of the embodiment of the present invention as above, it is when carrying out map search, after obtaining the Search Results of map search, also by extracting the characteristic information of the default characteristic type in this Search Results, generate the proper vector comprising these characteristic informations, and based on this proper vector, the general retrieval forecast model set up carries out prediction to the probability that the search-type of query string is general retrieval and determines, and based on the search-type determined, Search Results is processed, it is when determining general retrieval probability, without the need to relying on the identification that specific vocabulary can realize map search type rapidly, maintenance cost is low, and precision is high.
As shown in Figure 5, wherein in a concrete example, map search system of the present invention can also comprise:
Model construction module 500, for generating described general retrieval forecast model.
The structural representation of a concrete example of this model construction module 500 is also show in Fig. 5.Shown in composition graphs 5, in this example, this model construction module 500 comprises:
Query string collection modules 5001, for building query string set, includes general retrieval and inquisition string and non-general retrieval and inquisition string in described query string set;
Model eigenvectors generation module 5002, for obtaining the Search Results corresponding with each query string query string in described query string set being carried out to map search acquisition, extract the characteristic information of the default characteristic type of Search Results corresponding to each query string respectively, generate the proper vector comprising each query string characteristic of correspondence information respectively;
Model generation module 5004, for carrying out logistic regression study process to the proper vector of each query string, obtains described general retrieval forecast model.
Wherein, as shown in Figure 5, in this example, this model construction module 500 can also comprise the first normalized module 5003, for being normalized with each query string characteristic of correspondence vector;
Now, above-mentioned model generation module 5004, for carrying out logistic regression study process to the proper vector of each query string after normalized.
Corresponding, in a concrete example of the present invention, this map search system can also comprise:
Second normalized module 5023, is normalized for the proper vector generated feature vector generation module 502;
Now, above-mentioned probability determination module 503, be based on the second normalized module 5023 normalized after proper vector determine that the search-type of described query string is the probability of general retrieval.
After being normalized each proper vector, each proper vector can be made can to represent in same interval, simplify subsequent processes, improve treatment effeciency.
Wherein, to be above-mentionedly normalized and the mode of logistic regression study process can identical with the map search kind identification method of the invention described above, any mode that is existing and that later may occur at present can be adopted to carry out.
In above-mentioned map search type identification and set up in the process of general retrieval forecast model, when extracting the characteristic information of the default characteristic type of Search Results, above-mentioned default characteristic type can set based on actual needs, wherein in a concrete example, can be the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
What one of ordinary skill in the art will appreciate that is, realize all or part of flow process in the invention described above embodiment method, that the hardware that can carry out instruction relevant by computer program has come, described program can be stored in a computer read/write memory medium, this program, when performing, can comprise the flow process of the embodiment as above-mentioned each side method.Wherein, described storage medium can be magnetic disc, CD, read-only store-memory body (Read-Only Memory, ROM) or random store-memory body (Random Access Memory, RAM) etc.Therefore, according to the invention described above embodiment scheme, the present invention also provides a kind of storage medium comprising computer-readable program, when the computer-readable program in this storage medium performs, the map search kind identification method of the present invention in above-mentioned any one mode or map search method can be realized.
The method of the embodiment of the present invention as above, can be installed in corresponding machinery and equipment in the form of software, and completes above-mentioned map search type identification or the process of map search when this running software by the treatment facility controlling to be correlated with.Correspondingly, above-mentioned map search type identification device or map search system, can be arrange respectively to be arranged on corresponding terminal device, also can be corresponding terminal device itself, here terminal device can be the terminal device arbitrarily such as computing machine, mobile phone, panel computer, PDA (Personal DigitalAssistant, personal digital assistant), vehicle-mounted computer.
Fig. 6 is the module map of a computer system 1000 that can realize the embodiment of the present invention.This computer system 1000 is an example being applicable to computer environment of the present invention, can not think to propose any restriction to usable range of the present invention.Computer system 1000 can not be interpreted as the combination needing the one or more parts depending on or have in illustrated exemplary computer system 1000.
Computer system 1000 shown in Fig. 6 is the examples being suitable for computer system of the present invention.Other framework with different sub-systems configuration also can use.The similar devices such as the desktop computer known by masses, notebook, personal digital assistant, smart phone, panel computer, portable electronic device are such as had to go for some embodiments of the present invention.But be not limited to above cited equipment.
As shown in Figure 6, computer system 1000 comprises processor 1010, storer 1020 and system bus 1022.The various system components comprising storer 1020 and processor 1010 are connected on system bus 1022.Processor 1010 is the hardware being used for being performed by arithmetic sum logical operation basic in computer system computer program instructions.Storer 1020 be one for storing the physical equipment of calculation procedure or data (such as, program state information) temporarily or permanently.System bus 1020 can be any one in the bus structure of following several types, comprises memory bus or memory controller, peripheral bus and local bus.Processor 1010 and storer 1020 can carry out data communication by system bus 1022.Wherein storer 1020 comprises ROM (read-only memory) (ROM) or flash memory (all not shown in figure), and random access memory (RAM), and RAM typically refers to the primary memory being loaded with operating system and application program.
Computer system 1000 generally comprises a memory device 1030.Memory device 1030 can be selected from multiple computer-readable medium, and computer-readable medium refers to any available medium can accessed by computer system 1000, that comprise movement and fixing two media.Such as, computer-readable medium includes but not limited to, flash memory (miniature SD card), CD-ROM, digital versatile disc (DVD) or other optical disc storage, tape cassete, tape, disk storage or other magnetic storage apparatus, or can be used for storing information needed and other medium any can accessed by computer system 1000.
Computer system 1000 can be carried out logic with one or more network equipment in a network environment and is connected.The network equipment can be PC, server, router, smart phone, panel computer or other common network node.Computer system 1000 is connected with the network equipment by LAN (Local Area Network) (LAN) interface 1040 or mobile communication module 1050.LAN (Local Area Network) (LAN) refers in limited area, such as family, school, computer laboratory or use the office building of the network media, the computer network of interconnected composition.WiFi and twisted-pair feeder wiring Ethernet are two kinds of technology of the most frequently used structure LAN (Local Area Network).WiFi is a kind of technology that can make computer system 1000 swapping data or be connected to wireless network by radiowave.Mobile communication module 1050 can be answered by radio communication diagram while movement and call in a wide geographic area.Except call, mobile communication module 1050 is also supported in the 2G providing mobile data service, carries out internet access in 3G or 4G cellular communication system.
Computer system 1000 comprises the imageing sensor 1060 for taking a picture, and uses imageing sensor 1060 to take a picture and obtains image in 2 D code, convert optical imagery to electronic signal.
It should be pointed out that other computer system comprising the subsystem more more or less than computer system 1000 also can be applicable to invention.Such as, computer system 1000 can comprise the bluetooth module that can exchange data in short distance.
As described in detail, be applicable to the assigned operation of computer system 1000 of the present invention energy place of execution graph search kind identification method or map search method above.The form of the software instruction that computer system 1000 is operated in computer-readable medium by processor 1010 performs these operations.These software instructions can be read into storer 1020 from memory device 1030 or by lan interfaces 1040 from another equipment.The software instruction be stored in storer 1020 makes processor 1010 perform above-mentioned map search kind identification method or map search method.In addition, also the present invention can be realized equally by hardware circuit or hardware circuit in conjunction with software instruction.Therefore, the combination that the present invention is not limited to any specific hardware circuit and software is realized.
The above embodiment only have expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but therefore can not be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.
Claims (18)
1. a map search kind identification method, is characterized in that, comprises step:
Obtain Search Results query string being carried out to map search acquisition;
Extract the characteristic information of the default characteristic type of described Search Results, generate the proper vector comprising described characteristic information;
Determine that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set.
2. map search kind identification method according to claim 1, is characterized in that, the mode obtaining described general retrieval forecast model comprises:
Build query string set, in described query string set, include general retrieval and inquisition string and non-general retrieval and inquisition string;
Respectively map search is carried out to each query string in described query string set, obtain the Search Results corresponding with each query string;
Extract the characteristic information of the default characteristic type of Search Results corresponding to each query string respectively, generate the proper vector comprising each query string characteristic of correspondence information respectively;
Logistic regression study process is carried out to the proper vector of each query string, obtains described general retrieval forecast model.
3. map search kind identification method according to claim 2, is characterized in that:
Generation comprise the proper vector of described characteristic information after, determine that the search-type of described query string is the probability of general retrieval before, also comprise step: the proper vector comprising described characteristic information is normalized;
Carry out logistic regression study process after generation comprises the proper vector of each query string characteristic of correspondence information respectively, to the proper vector of each query string before, also comprise step: be normalized with each query string characteristic of correspondence vector.
4. the map search kind identification method according to claims 1 to 3 any one, is characterized in that, described default characteristic type comprises the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
5. a map search type identification device, is characterized in that, comprising:
Data obtaining module, for obtaining Search Results query string being carried out to map search acquisition;
Feature vector generation module, for extracting the characteristic information of the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Probability determination module, for determining that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set.
6. map search type identification device according to claim 5, is characterized in that, also comprise:
Model construction module, for generating described general retrieval forecast model.
7. map search type identification device according to claim 6, it is characterized in that, described model construction module comprises:
Query string collection modules, for building query string set, includes general retrieval and inquisition string and non-general retrieval and inquisition string in described query string set;
Model eigenvectors generation module, for obtaining the Search Results corresponding with each query string respectively query string in described query string set being carried out to map search acquisition, extract the characteristic information of the default characteristic type of Search Results corresponding to each query string, generate the proper vector comprising each query string characteristic of correspondence information respectively;
Model generation module, for carrying out logistic regression study process to the proper vector of each query string, obtains described general retrieval forecast model.
8. map search type identification device according to claim 7, is characterized in that:
Also comprise the second normalized module, for being normalized the proper vector comprising described characteristic information; According to the proper vector of the described characteristic information after described normalized, general retrieval forecast model, described probability determination module determines that the search-type of described query string is the probability of general retrieval;
Described model construction module also comprises the first normalized module, for being normalized with each query string characteristic of correspondence vector; Described model generation module, for carrying out logistic regression study process to the proper vector of each query string after normalized.
9. the map search type identification device according to claim 5 to 8 any one, is characterized in that, described default characteristic type comprises the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
10. a map search method, is characterized in that, comprises step:
Receive map search instruction, described map search instruction comprises query string;
Search for according to described map search instruction, obtain Search Results;
Extract the characteristic information of the default characteristic type of described Search Results, obtain the proper vector of described Search Results;
Determine that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, determine the search-type of described query string according to the search-type of the described query string probability that is general retrieval, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set;
Determine the Graphics Processing mode of described Search Results according to the search-type of described query string, and according to this Graphics Processing mode, described Search Results is processed.
11. map search methods according to claim 10, is characterized in that, described general retrieval forecast model is obtained by following manner:
Build query string set, in described query string set, include general retrieval and inquisition string and non-general retrieval and inquisition string;
Respectively map search is carried out to the query string in described query string set, obtain the Search Results corresponding with each query string;
Extract the characteristic information of Search Results corresponding to each query string respectively, generate the proper vector comprising each query string characteristic of correspondence information respectively;
Logistic regression study process is carried out to the proper vector of each query string, obtains described general retrieval forecast model.
12. map search methods according to claim 11, is characterized in that:
Generation comprise the proper vector of described characteristic information after, determine that the search-type of described query string is the probability of general retrieval before, also comprise step: the proper vector comprising described characteristic information is normalized;
After generation comprises the proper vector of each query string characteristic of correspondence information respectively, before logistic regression study process is carried out to the proper vector of each query string, also comprise step: be normalized with each query string characteristic of correspondence vector.
13., according to claim 10 to the map search method described in 12 any one, is characterized in that, described default characteristic type comprises the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
14. 1 kinds of map search systems, is characterized in that, comprising:
Search module, for receiving map search instruction, described map search instruction comprises query string, searches for according to described map search instruction, obtains Search Results;
Feature vector generation module, for extracting the characteristic information of the default characteristic type of described Search Results, generates the proper vector comprising described characteristic information;
Search-type determination module, for determining that the search-type of described query string is the probability of general retrieval according to described proper vector, general retrieval forecast model, determine the search-type of described query string according to the search-type of the described query string probability that is general retrieval, described general retrieval preset model obtains by carrying out training to the characteristic information of the default characteristic type corresponding with each query string in default query string set;
Graphics Processing module, for determining the Graphics Processing mode of described Search Results according to this search-type, and processes described Search Results according to this Graphics Processing mode.
15. map search systems according to claim 14, is characterized in that, also comprise:
Model construction module, for generating described general retrieval forecast model.
16. map search systems according to claim 15, it is characterized in that, described model construction module comprises:
Query string collection modules, for building query string set, includes general retrieval and inquisition string and non-general retrieval and inquisition string in described query string set;
Model eigenvectors generation module, for obtaining the Search Results corresponding with each query string query string in described query string set being carried out to map search acquisition, extract the characteristic information of the default characteristic type of Search Results corresponding to each query string respectively, generate the proper vector comprising each query string characteristic of correspondence information respectively;
Model generation module, for carrying out logistic regression study process to the proper vector of each query string, obtains described general retrieval forecast model.
17. map search systems according to claim 16, is characterized in that:
Also comprise the second normalized module, for being normalized the proper vector comprising described characteristic information; According to the proper vector of the described characteristic information after described normalized, general retrieval forecast model, described search-type determination module determines that the search-type of described query string is the probability of general retrieval;
Described model construction module also comprises the first normalized module, for being normalized with each query string characteristic of correspondence vector; Described model generation module, for carrying out logistic regression study process to the proper vector of each query string after normalized.
18., according to claim 14 to the map search system described in 17 any one, is characterized in that, described default characteristic type comprises the combination in any in following each information: query string length, the number of the joint of query string, whether the maximum classification in Search Results described in front N bar is general retrieval, the maximum classification accounting of N bar Search Results, Article 1, the text score of Search Results, Article 1, the difference of the text score of Search Results and Article 2 Search Results, the average text score of front N bar Search Results, Article 1, the difference of the text score of Search Results and the average text score of front N bar Search Results, the standard variance of the text score of front N bar Search Results, whether query string occurs completely in the title of front N bar Search Results, Article 1, the POIRank of Search Results, the POIRank of front N bar Search Results, Article 1, the difference of the POIRank of Search Results and the mean P OIRank of front N bar Search Results, the cluster number of clusters of front N bar Search Results, the convergence factor of front N bar search result clustering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410289451.1A CN104615620B (en) | 2014-06-24 | 2014-06-24 | Map search kind identification method and device, map search method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410289451.1A CN104615620B (en) | 2014-06-24 | 2014-06-24 | Map search kind identification method and device, map search method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104615620A true CN104615620A (en) | 2015-05-13 |
| CN104615620B CN104615620B (en) | 2018-07-24 |
Family
ID=53150068
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410289451.1A Active CN104615620B (en) | 2014-06-24 | 2014-06-24 | Map search kind identification method and device, map search method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104615620B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018133648A1 (en) * | 2017-01-20 | 2018-07-26 | 北京三快在线科技有限公司 | Search method and apparatus, and non-temporary computer-readable storage medium |
| CN110516024A (en) * | 2019-08-30 | 2019-11-29 | 百度在线网络技术（北京）有限公司 | Map search result shows method, apparatus, equipment and storage medium |
| CN112528145A (en) * | 2020-12-11 | 2021-03-19 | 北京百度网讯科技有限公司 | Information recommendation method, device, equipment and readable storage medium |
| CN112989224A (en) * | 2021-03-25 | 2021-06-18 | 北京百度网讯科技有限公司 | Retrieval method, retrieval device, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120066234A1 (en) * | 2010-09-10 | 2012-03-15 | Korea University Research And Business Foundation | Method and apparatus for providing internet service in mobile communication terminal |
| CN102456054A (en) * | 2010-10-28 | 2012-05-16 | 腾讯科技（深圳）有限公司 | Searching method and system |
| CN102955798A (en) * | 2011-08-25 | 2013-03-06 | 腾讯科技（深圳）有限公司 | Search engine based search method and search server |
| CN103268348A (en) * | 2013-05-28 | 2013-08-28 | 中国科学院计算技术研究所 | Method for identifying user query intention |
-
2014
- 2014-06-24 CN CN201410289451.1A patent/CN104615620B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120066234A1 (en) * | 2010-09-10 | 2012-03-15 | Korea University Research And Business Foundation | Method and apparatus for providing internet service in mobile communication terminal |
| CN102456054A (en) * | 2010-10-28 | 2012-05-16 | 腾讯科技（深圳）有限公司 | Searching method and system |
| CN102955798A (en) * | 2011-08-25 | 2013-03-06 | 腾讯科技（深圳）有限公司 | Search engine based search method and search server |
| CN103268348A (en) * | 2013-05-28 | 2013-08-28 | 中国科学院计算技术研究所 | Method for identifying user query intention |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018133648A1 (en) * | 2017-01-20 | 2018-07-26 | 北京三快在线科技有限公司 | Search method and apparatus, and non-temporary computer-readable storage medium |
| CN110516024A (en) * | 2019-08-30 | 2019-11-29 | 百度在线网络技术（北京）有限公司 | Map search result shows method, apparatus, equipment and storage medium |
| CN110516024B (en) * | 2019-08-30 | 2022-05-20 | 百度在线网络技术（北京）有限公司 | Map search result display method, device, equipment and storage medium |
| CN112528145A (en) * | 2020-12-11 | 2021-03-19 | 北京百度网讯科技有限公司 | Information recommendation method, device, equipment and readable storage medium |
| CN112528145B (en) * | 2020-12-11 | 2023-09-05 | 北京百度网讯科技有限公司 | Information recommendation method, device, equipment and readable storage medium |
| CN112989224A (en) * | 2021-03-25 | 2021-06-18 | 北京百度网讯科技有限公司 | Retrieval method, retrieval device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104615620B (en) | 2018-07-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109145219B (en) | Method and device for judging validity of interest points based on Internet text mining | |
| CN107346326B (en) | Method and system for information retrieval | |
| US20200326197A1 (en) | Method, apparatus, computer device and storage medium for determining poi alias | |
| WO2015081720A1 (en) | Instant messaging (im) based information recommendation method, apparatus, and terminal | |
| CN112084413B (en) | Information recommendation method, device and storage medium | |
| CN102930048A (en) | Data abundance automatically found by semanteme and using reference and visual data | |
| CN104615620A (en) | Map search type identification method and device and map search method and system | |
| CN108334353B (en) | Skill development system and method | |
| CN112862021B (en) | Content labeling method and related device | |
| CN110674208B (en) | Method and device for determining position information of user | |
| CN114090792A (en) | Document relation extraction method based on comparison learning and related equipment thereof | |
| JP2020004217A (en) | Information display method, information display program and information display apparatus | |
| CN110990714B (en) | User behavior intention prediction method and device | |
| CN110598122B (en) | Social group mining method, device, equipment and storage medium | |
| CN111191107B (en) | System and method for recalling points of interest using annotation model | |
| CN110795547A (en) | Text recognition method and related product | |
| CN107766881B (en) | Way finding method and device based on basic classifier and storage device | |
| CN104050168B (en) | Information processing method, electronic equipment and dictionary server | |
| CN114297381A (en) | Text processing method, device, equipment and storage medium | |
| CN113515687B (en) | Logistics information acquisition method and device | |
| CN113449754B (en) | Label matching model training and displaying method, device, equipment and medium | |
| KR101363335B1 (en) | Apparatus and method for generating document categorization model | |
| EP4127957A1 (en) | Methods and systems for searching and retrieving information | |
| CN114329236A (en) | Data processing method and device | |
| CN112084764A (en) | Data detection method, device, storage medium and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210922 Address after: 518057 Tencent Building, No. 1 High-tech Zone, Nanshan District, Shenzhen City, Guangdong Province, 35 floors Patentee after: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd. Patentee after: TENCENT CLOUD COMPUTING (BEIJING) Co.,Ltd. Address before: 2, 518000, East 403 room, SEG science and Technology Park, Zhenxing Road, Shenzhen, Guangdong, Futian District Patentee before: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd. |