CN104615620B - Map search kind identification method and device, map search method and system - Google Patents
Map search kind identification method and device, map search method and system Download PDFInfo
- Publication number
- CN104615620B CN104615620B CN201410289451.1A CN201410289451A CN104615620B CN 104615620 B CN104615620 B CN 104615620B CN 201410289451 A CN201410289451 A CN 201410289451A CN 104615620 B CN104615620 B CN 104615620B
- Authority
- CN
- China
- Prior art keywords
- search result
- search
- query string
- type
- feature vector
- 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.)
- Active
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
A kind of map search kind identification method and device, a kind of map search method and system, which includes step：Obtain the search result that map search acquisition is carried out to query string；The characteristic information of the default characteristic type of described search result is extracted, the feature vector for including the characteristic information is generated；Determine that the search-type of the query string is the probability of general retrieval according to described eigenvector, general retrieval prediction model, the general retrieval preset model is trained acquisition by a pair characteristic information for default characteristic type corresponding with each query string in default inquiry set of strings.The present invention program can rapidly realize that maintenance cost is low to the identification of map search type without relying on specific vocabulary, and precision is high.
Description
Technical field
The present invention relates to field of map search, more particularly to 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
Function of search is both provided in map products numerous at present, to the Query (query string) that user submits, search
Relevant POI data simultaneously returns to user after sorted.Every data in map is known as a POI (Point of
Interest, point of interest), each POI includes title, classification (Category), longitude, latitude, (POIRank is indicated importance
The importance of POI data, calculated based on significance levels of the POI under affiliated classification or ranking obtained by, POIRank under different classifications
Calculating may be different, such as in hospital's classification Grade A hospital POIRank can be higher than other ranks hospital, cuisines
In classification user reviews quantity can and POIRank it is directly related) etc. much informations.The type belonging to Query in map search,
It is generally divided into two classes of precise search and general retrieval.Precise search refers to that the Query that user submits in map search is to be directed to some
The lookup of specific POI data point.It is class name, chain trade name, suffix that general retrieval, which refers to the Query that user submits in map search,
What word etc., i.e. user thought search is certain a kind of POI data rather than some specific POI data point.The search of general retrieval Query
As a result general to be spatially distributed more dispersed, in addition there is certain general character in classification.Typical general retrieval Query can be example
Such as cuisines, hotel, university, when user is when cuisines are searched for by Beijing, map search can provide all related cuisines in Beijing
Data and Pagination Display.In map search technical finesse, general retrieval and precise search are in the side such as search strategy, ordering strategy
Face can different from, such as user use mobile phone searching cuisines when, preferentially provide the data of the cuisines near user position
There can be better user experience.Therefore, it is that general retrieval or precise search are identified to the retrieval type of search, especially
The identification of general retrieval becomes an important component in map search.
In map search, basic general retrieval can be based on vocabulary, typical general term list such as item name
Vocabulary, chain store's vocabulary, label vocabulary, suffix vocabulary etc..If Query input by user belongs to some known general retrieval
Vocabulary, then it is assumed that the search-type of current Query is general retrieval.However the general retrieval based on vocabulary, there is certain limitation
Property：On the one hand, the general retrieval Query that various Inexact Newton methods are intended to, word are had in Query when actual user searches
Table is difficult to cover all general retrieval vocabulary；Still further aspect, if general term list is excessive, relevant maintenance workload can add
Greatly, it is possible that certain noise while in vocabulary.
Invention content
Based on this, it is necessary to for the above-mentioned prior art the problem of, a kind of map search type identification side is provided
Method, a kind of map search type identification device and a kind of map search method, a kind of map search system, it is special without relying on
Fixed vocabulary can rapidly realize that maintenance cost is low to the identification of map search type, and precision is high.
In order to achieve the above objectives, the embodiment of the present invention uses following technical scheme：
A kind of map search kind identification method, including step：
Obtain the search result that map search acquisition is carried out to query string；
Extract described search result default characteristic type characteristic information, generate comprising the characteristic information feature to
Amount；
Determine that the search-type of the query string is the general of general retrieval according to described eigenvector, general retrieval prediction model
Rate, the general retrieval preset model pass through a couple spy for default characteristic type corresponding with each query string in default inquiry set of strings
Reference breath is trained acquisition.
A kind of map search type identification device, including：
Data obtaining module, for obtaining the search result for carrying out map search acquisition to query string；
Feature vector generation module, the characteristic information of the default characteristic type for extracting described search result generate packet
Feature vector containing the characteristic information；
Probability determination module, the search for determining the query string according to described eigenvector, general retrieval prediction model
Type is the probability of general retrieval, and the general retrieval preset model passes through pair corresponding with the default each query string inquired in set of strings
The characteristic information of default characteristic type is trained acquisition.
A kind of map search method, including step：
Map search instruction is received, the map search instruction includes query string；
It is scanned for according to map search instruction, obtains search result；
The characteristic information for extracting the default characteristic type of described search result, obtains the feature vector of described search result；
Determine that the search-type of the query string is the general of general retrieval according to described eigenvector, general retrieval prediction model
Rate, according to the search-type of query string described in the determine the probability that the search-type of the query string is general retrieval, the general retrieval
Preset model is instructed by a pair characteristic information for default characteristic type corresponding with each query string in default inquiry set of strings
Practice and obtains；
Determine the display processing mode of described search result according to the search-type of the query string, and according to the display at
Reason mode handles described search result.
A kind of map search system, including：
Search module, for receiving map search instruction, the map search instruction includes query string, according to described
Graph search instruction scans for, and obtains search result；
Feature vector generation module, the characteristic information of the default characteristic type for extracting described search result generate packet
Feature vector containing the characteristic information；
Search-type determining module, for determining the query string according to described eigenvector, general retrieval prediction model
Search-type is the probability of general retrieval, according to query string described in the determine the probability that the search-type of the query string is general retrieval
Search-type, the general retrieval preset model pass through a pair default feature class corresponding with each query string in default inquiry set of strings
The characteristic information of type is trained acquisition；
Display processing module, the display processing mode for determining described search result according to the search-type, and according to
The display processing mode handles described search result.
It is to obtain the search that map search is carried out to inquiring according to the scheme of embodiment present invention as described above
As a result after, extract the characteristic information of the default characteristic type in the search result, generate the feature comprising these characteristic informations to
Amount, and based on this feature vector, the general retrieval prediction model established come the probability to the search-type of query string for general retrieval
Prediction determination is carried out, identification to map search type, maintenance cost can be rapidly realized without relying on specific vocabulary
It is low, can retrieval Query general to type have certain recognition capability, precision is high.
Description of the drawings
Fig. 1 is the flow diagram of the map search kind identification method embodiment of the present invention；
Fig. 2 is the flow diagram that general retrieval prediction model is established in an example；
Fig. 3 is the structural schematic diagram of the map search identification device embodiment of the present invention；
Fig. 4 is the flow diagram of the map search embodiment of the method for the present invention；
Fig. 5 is the structural schematic diagram of the map search system embodiment of the present invention；
Fig. 6 is the module map for the computer system 1000 that can realize the embodiment of the present invention.
Specific implementation mode
Below in conjunction with preferred embodiment therein, illustrated in greater detail is carried out to the present invention program.
The flow diagram of the map search kind identification method embodiment of the present invention is shown in Fig. 1.As shown in Figure 1,
Map search kind identification method in the present embodiment includes step：
Step S101：Obtain the search result that map search acquisition is carried out to query string；
Step S102：The characteristic information of the default characteristic type of described search result is extracted, generates and believes comprising the feature
The feature vector of breath；
Step S103：Determine that the search-type of the query string is general according to described eigenvector, general retrieval prediction model
The probability of retrieval, the general retrieval preset model pass through a pair default feature corresponding with each query string in default inquiry set of strings
The characteristic information of type is trained acquisition.
It is to obtain the search that map search is carried out to inquiring according to the scheme of embodiment present invention as described above
As a result after, extract the characteristic information of the default characteristic type in the search result, generate the feature comprising these characteristic informations to
Amount, and based on this feature vector, the general retrieval prediction model established come the probability to the search-type of query string for general retrieval
Prediction determination is carried out, identification to map search type, maintenance cost can be rapidly realized without relying on specific vocabulary
It is low, and precision is high.
The flow diagram that general retrieval prediction model is established in a specific example is shown in Fig. 2, as shown in Fig. 2, building
Founding the general process for retrieving prediction model may include：
Step S201：Structure inquires set of strings, includes general retrieval and inquisition string and non-general retrieval in the inquiry set of strings
Query string；
Step S202：Map search is carried out to the query string in the inquiry set of strings respectively, is obtained and each query string pair
The search result answered；
Step S203：The characteristic information of the default characteristic type of the corresponding search result of each query string is extracted respectively, respectively
Generate the feature vector for including the corresponding characteristic information of each query string；
Step S204：Logistic regression study processing is carried out to the feature vector of each query string, obtains the general retrieval prediction
Model.
It, can be between above-mentioned steps S203 and step S204 as shown in Fig. 2, when establishing general retrieval prediction model
Including step：
Step S2034：Pair feature vector corresponding with each query string is normalized.
Correspondingly, between above-mentioned steps S102 and step S103, can also include：
Step S1023：The feature vector obtained in step S102 is normalized.
After each feature vector is normalized, each feature vector can be enable to indicate in same area
Between, simplify subsequent processes, improves treatment effeciency.
Wherein, in above-mentioned map search type identification and during establish general retrieval prediction model, extraction search knot
When the characteristic information of the default characteristic type of fruit, above-mentioned default characteristic type can be based on actual needs and be set, wherein
Can be the arbitrary combination in following each information in one specific example：Inquire string length, the number of section of query string, preceding N items
Whether maximum classification in described search result is general retrieval, the maximum classification accounting of N search result, first search result
Text score, the difference of first search result and the text score of Article 2 search result, preceding N search result be averaged
Text score, the text score of first search result are searched with the difference of the average text score of preceding N search result, preceding N items
Whether the standard variance of the text score of hitch fruit, query string occur completely in the title of preceding N search result, first search
The POIRank of hitch fruit, the POIRank of preceding N search result, the POIRank of first search result and preceding N search result
The difference of average POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering convergence factor.
The map search kind identification method of the present invention is described in detail below in conjunction with one of specific example.
It is that the process of establishing of general retrieval prediction model is combined to illustrate in following the description.
Before specifically carrying out map search type identification, need to establish general retrieval prediction model, in the present invention program,
By analyzing the search result set of Query, the general relevant feature of retrieval is extracted, a large amount of flag datas is then based on, uses
The method of machine learning is trained, to obtain general retrieval prediction model.
When establishing general retrieval prediction model, it is necessary first to build a big Query set, which wraps in combining
Containing typical general retrieval Query and non-general retrieval Query.General retrieval sample Query can come from existing general term list,
Artificial mark is can come from, non-general retrieval (i.e. precise search) Query can come from general POI data point title, also may be used
With from artificial mark.Wherein, artificial mark refers to the search result manually based on current Query discriminates whether it is general retrieval
Query is simultaneously marked that (such as general retrieval is labeled as 1,0) non-general retrieval is labeled as.
Then it by above-mentioned marked sample Query, is submitted to map search system and scans for, from map search system
Backstage can obtain the essential attribute of every result, and the default characteristic type of general retrieval can be further extracted based on these attributes
Correlated characteristic.In the specific example of the present invention, with relevant some features of general retrieval and its correlation computations mode and
Feature description can be as shown in following table table 1.
1 general retrieval correlated characteristic of table and its feature description
Any side for having at present and being likely to occur later may be used in the method for determination of feature in above-mentioned table 1
Formula carries out.All kinds of features shown in above-mentioned table 1 can be used all, in specific implementation according to the feelings of real data
Condition can remove Partial Feature therein, can also increase Partial Feature.
After extracting the above-mentioned correlated characteristic of mentioned above searching results, the needs based on application can give each sample
The general retrieval mark that mono- codomain of Query is 0/1, indicates 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.To which each sample Query can be expressed as the vector of a N*1, entirely
Training set can be expressed as the vector of M*N dimensions, and the number that wherein N is characterized, M is Query in training examples set
Quantity.
Then the method that Logistic Regression (logistic regression) can be used to each feature vector, to above-mentioned spy
Sign vector carries out logistic regression study processing, converts general retrieval forecasting problem to recurrence learning and forecasting problem.
Before being trained using Logistic Regression, first above-mentioned each feature vector can be normalized
Processing improves treatment effeciency so that each feature vector can indicate, in same section, to facilitate subsequent processes.
When being normalized, various possible modes may be used and carry out.The typical feature normalizing of one of which
The mode that linear function conversion may be used in change method carries out, and the formula of linear function conversion is as described below：
Y=(x-MinValue)/(MaxValue-MinValue)
In above-mentioned formula, y indicates that the value after x is normalized, MinValue indicate the minimum value of characteristic value,
MaxValue indicates the maximum value in characteristic value.In the present invention program, for the different features in each feature vector, respectively
It is normalized using above-mentioned formula, to map that in [0,1] section.
Another typical normalization mode is the normalization being distributed based on attribute section sample, and detailed process can be as follows
It is described：
The value range of single feature value, N number of section is divided by the value range in statistics training set；
The percentage of positive example in each section of the attribute (being general retrieval Query) is counted, i.e., attribute is in current interval
Interior Query belongs to the probability of general retrieval；
Given original property value, based on the section belonging to the value, you can obtain the probability value of one [0,1].
After carrying out above-mentioned normalized, you can carry out Logistic for the feature vector after normalized is obtained
Regression learning trainings, to obtain general retrieval prediction model.The machine learning side of specific Logistic Regression
Method may be used the various possible modes for having at present and being likely to occur later and carry out, including but not limited to decision tree, branch
It holds vector machine (SVM), artificial neural network (ANN), gradient and is incremented by decision tree (GBDT) etc..
For ease of understanding, a brief description is done to Logistic Regression below.
Learn scene as a bare metal, gives N number of training sample (x1,y1),(x2,y2),(x3,y3),...,
(xN,yN), wherein xi∈RnIt is a n-dimensional vector, for indicating value of i-th of sample in this n feature, and yi∈{-1,
+ 1 } it illustrates that this sample is positive sample (+1) or negative sample (- 1), can be indicated to be general with positive sample in the present invention program
Retrieval.Logic Regression Models are exactly to pass through a function for being logical function (Logistic Function)The feature vector xi of i-th of sample is linked up with the sample for the probability of positive sample：
Wherein, p (yi=+1 | xi, w) and indicate sample xiFor the probability of positive sample, wTxi=β0+β1xi1+β2xi2+...+βnxin, the form of w parameters is β0,β1,β2...,βn, parameter w does different weightings to the n dimension of x, so as to calculate wTxi,
Then the value is adjusted to 0 to 1, the as probability of positive sample by the logical function of S types.
The target of logistic regression seeks to find best w so that and the score of positive sample is all bigger, while negative sample
Score is all smaller.Logistic Regression training result models are w parameter sets, i.e. β0,β1,β2...,βn。
Logistic Regression are found by maximal possibility estimation (Maximum Likelihood Estimation) principle
Best w.In a specific example, the libraries LibLinear can be used to carry out the processing of Logistic Regression, it should
Using Trusted Region Newton (TRON) method in library, so as to obtain preferable logistic regression performance.
After obtaining above-mentioned general retrieval prediction model, you can in actual map search whether be general to search-type
Retrieval is predicted.
When prediction, after giving Query, the set based on the search result that search returns extracts above-mentioned correlated characteristic, and
The feature vector for including these features is generated, and obtained feature vector is normalized, then calls above-mentioned formula
(1) general retrieval prediction model using the feature vector after the above-mentioned normalized of current Query as input calculate pre-
It surveys, you can the search-type to obtain current Query belongs to the probability of general retrieval, you can identifies whether current Query is general inspection
Rope.
Based on the map search kind identification method with aforementioned present invention, the present invention also provides a kind of knowledges of map search type
Other device.The structural schematic diagram of the map search type identification device embodiment of the present invention is shown in Fig. 3.As shown in figure 3, this
Embodiment device includes：
Data obtaining module 301, for obtaining the search result for carrying out map search acquisition to query string；
Feature vector generation module 302, the characteristic information of the default characteristic type for extracting described search result generate
Include the feature vector of the characteristic information；
Probability determination module 303, for determining searching for the query string according to described eigenvector, general retrieval prediction model
Rope type is the probability of general retrieval, and the general retrieval preset model passes through pair corresponding with the default each query string inquired in set of strings
The characteristic information of default characteristic type be trained acquisition.
It is to obtain the search that map search is carried out to inquiring according to the scheme of embodiment present invention as described above
As a result after, extract the characteristic information of the default characteristic type in the search result, generate the feature comprising these characteristic informations to
Amount, and based on this feature vector, the general retrieval prediction model established come the probability to the search-type of query string for general retrieval
Prediction determination is carried out, identification to map search type, maintenance cost can be rapidly realized without relying on specific vocabulary
It is low, and precision is high.
As shown in figure 3, in a wherein specific example, map search type identification device of the invention can also wrap
It includes：
Model construction module 300, for generating the general retrieval prediction model.
The structural schematic diagram of a specific example of the model construction module 300 is also shown in Fig. 3.In conjunction with shown in Fig. 3,
In this example, which includes：
Query string collection modules 3001 include that general retrieval is looked into for building inquiry set of strings, in the inquiry set of strings
Ask string and non-general retrieval and inquisition string；
Model eigenvectors generation module 3002 searches the query string progress map in the inquiry set of strings for obtaining
The search result corresponding with each query string that rope obtains, extracts the default characteristic type of the corresponding search result of each query string respectively
Characteristic information, generate include the feature vector of the corresponding characteristic information of each query string respectively；
Model generation module 3004 carries out logistic regression study processing for the feature vector to each query string, obtains institute
State general retrieval prediction model.
Wherein, as shown in figure 3, in this example, which can also include the first normalized mould
Block 3003 is normalized for a pair feature vector corresponding with each query string；
At this point, above-mentioned model generation module 3004, carries out for the feature vector to each query string after normalized
Logistic regression study is handled.
Corresponding, in the specific example of the present invention, which can also include：
Second normalized module 3023, the feature vector for being generated to feature vector generation module 302 are returned
One change is handled；
At this point, above-mentioned probability determination module 303, is based on the spy after 3023 normalized of the second normalized module
Vector is levied to determine that the search-type of the query string is the probability of general retrieval.
After each feature vector is normalized, each feature vector can be enable to indicate in same area
Between, simplify subsequent processes, improves treatment effeciency.
Wherein, it is above-mentioned be normalized and logistic regression study processing mode can be with the ground of aforementioned present invention
It is identical in graph search kind identification method, any mode for having at present and being likely to occur later may be used and carry out.
In above-mentioned map search type identification and during establish general retrieval prediction model, the pre- of search result is extracted
If when the characteristic information of characteristic type, above-mentioned default characteristic type can be based on actual needs and be set, wherein a tool
Can be the arbitrary combination in following each information in body example：Inquiry string length, the number of section of query string are searched described in preceding N items
Maximum classification in hitch fruit whether be general retrieval, the maximum classification accounting of N search result, first search result text
Score, first search result and the difference of the text score of Article 2 search result, the average text of preceding N search result obtain
Point, difference with the average text score of preceding N search result of the text score of first search result, preceding N search result
The standard variance of text score, query string whether occur completely in the title of preceding N search result, first search result
POIRank, preceding N search result POIRank, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering.
Map search kind identification method based on aforementioned present invention, the present invention also provides a kind of map search methods.Fig. 4
In show the present invention map search embodiment of the method flow diagram.As shown in figure 4, the map search in the present embodiment
Method includes step：
Step S401：Map search instruction is received, the map search instruction includes query string；
Step S402：It is scanned for according to map search instruction, obtains search result；
Step S403：The characteristic information for extracting the default characteristic type of described search result, obtains described search result
Feature vector；
Step S404：Determine that the search-type of the query string is general according to described eigenvector, general retrieval prediction model
The probability of retrieval, according to the search-type of query string described in the determine the probability that the search-type of the query string is general retrieval, institute
General retrieval preset model is stated by a pair feature for default characteristic type corresponding with each query string in default inquiry set of strings to believe
Breath is trained acquisition；
Step S405：The display processing mode of described search result, and root are determined according to the search-type of the query string
Described search result is handled according to the display processing mode.
Map search method based on embodiment present invention as described above obtains map when carrying out map search
After the search result of search, also by extracting the characteristic information of the default characteristic type in the search result, it includes these to generate
The feature vector of characteristic information, and based on this feature vector, the general retrieval prediction model established come the searching class to query string
Type is that the probability of general retrieval carries out prediction determination, and is handled search result based on determining search-type, in determination
When general retrieval probability, it can rapidly realize that maintenance cost is low to the identification of map search type without relying on specific vocabulary,
And precision is high.
Above-mentioned general retrieval prediction model, may be used mode shown in Fig. 2 and establishes, can specifically include following steps：
Step S201：Structure inquires set of strings, includes general retrieval and inquisition string and non-general retrieval in the inquiry set of strings
Query string；
Step S202：Map search is carried out to the query string in the inquiry set of strings, is obtained corresponding with each query string
Search result；
Step S203：The characteristic information of the corresponding search result of each query string is extracted respectively, and it includes each inquiry to generate respectively
Go here and there the feature vector of corresponding characteristic information；
Step S204：Logistic regression study processing is carried out to the feature vector of each query string, obtains the general retrieval prediction
Model.
It, can be between above-mentioned steps S203 and step S204 as shown in Fig. 2, when establishing general retrieval prediction model
Including step：
Step S2034：Pair feature vector corresponding with each query string is normalized.
Correspondingly, between above-mentioned steps S403 and step S404, can also include：
Step S4034：The feature vector obtained in step S403 is normalized.
After each feature vector is normalized, each feature vector can be enable to indicate in same area
Between, simplify subsequent processes, improves treatment effeciency.
Wherein, it is above-mentioned be normalized and logistic regression study processing mode can be with the ground of aforementioned present invention
It is identical in graph search kind identification method, any mode for having at present and being likely to occur later may be used and carry out.
In above-mentioned map search type identification and during establish general retrieval prediction model, the pre- of search result is extracted
If when the characteristic information of characteristic type, above-mentioned default characteristic type can be based on actual needs and be set, wherein a tool
Can be the arbitrary combination in following each information in body example：Inquiry string length, the number of section of query string are searched described in preceding N items
Maximum classification in hitch fruit whether be general retrieval, the maximum classification accounting of N search result, first search result text
Score, first search result and the difference of the text score of Article 2 search result, the average text of preceding N search result obtain
Point, difference with the average text score of preceding N search result of the text score of first search result, preceding N search result
The standard variance of text score, query string whether occur completely in the title of preceding N search result, first search result
POIRank, preceding N search result POIRank, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering.
Based on map search method present invention as described above, carried out below in conjunction with one of concrete processing procedure detailed
It describes in detail bright.
When carrying out map search, user is referred to by inputting Query by clicking the buttons such as search and sending out map and search element
It enables.After receiving map search instruction, map search is carried out for the Query, obtains search result.
Then the characteristic information for extracting the features described above type in search result, obtains corresponding feature vector, and right
After this feature vector is normalized, use above-mentioned general retrieval prediction model to the search-type of the Query for general retrieval
Probability predicted, after obtaining general retrieval probability, if it is general retrieval probability be more than setting threshold value, it may be considered that the Query
Retrieval type be general retrieval, otherwise, it is believed that the retrieval type of the Query be precise search.
Then, after being handled search result based on identified retrieval type, then will treated search result into
Row display.For example, for retrieving type and be precise search, then directly accurate search result can shown.With
It retrieves for type is general retrieval, then it can be to adjusting in the sequence of retrieval, display mode.For example, in search cuisines, hotel
Whens equal, can preferentially provide the search result near user position, and can carry out being highlighted etc., it is more preferable to have
User experience.
Based on thought identical with above-mentioned map search method, the present invention also provides a kind of map search systems.It is searched in Fig. 5
The structural schematic diagram of the map search system embodiment of the present invention is gone out.As shown in figure 5, the system in the present embodiment includes：
Search module 501, for receiving map search instruction, the map search instruction includes query string, according to institute
It states map search instruction to scan for, obtains search result；
Feature vector generation module 502, the characteristic information of the default characteristic type for extracting described search result generate
Include the feature vector of the characteristic information；
Search-type determining module 503, for determining the query string according to described eigenvector, general retrieval prediction model
Search-type be general retrieval probability, according to the search-type of the query string be general retrieval determine the probability described in query string
Search-type, the general retrieval preset model passes through a pair default feature corresponding with the default each query string inquired in set of strings
The characteristic information of type is trained acquisition；
Display processing module 504, the display processing mode for determining described search result according to the search-type, and root
Described search result is handled according to the display processing mode.
Map search system based on embodiment present invention as described above obtains map when carrying out map search
After the search result of search, also by extracting the characteristic information of the default characteristic type in the search result, it includes these to generate
The feature vector of characteristic information, and based on this feature vector, the general retrieval prediction model established come the searching class to query string
Type is that the probability of general retrieval carries out prediction determination, and is handled search result based on determining search-type, in determination
When general retrieval probability, it can rapidly realize that maintenance cost is low to the identification of map search type without relying on specific vocabulary,
And precision is high.
As shown in figure 5, in a wherein specific example, map search system of the invention can also include：
Model construction module 500, for generating the general retrieval prediction model.
The structural schematic diagram of a specific example of the model construction module 500 is also shown in Fig. 5.In conjunction with shown in Fig. 5,
In this example, which includes：
Query string collection modules 5001 include that general retrieval is looked into for building inquiry set of strings, in the inquiry set of strings
Ask string and non-general retrieval and inquisition string；
Model eigenvectors generation module 5002 searches the query string progress map in the inquiry set of strings for obtaining
The search result corresponding with each query string that rope obtains, extracts the default characteristic type of the corresponding search result of each query string respectively
Characteristic information, generate include the feature vector of the corresponding characteristic information of each query string respectively；
Model generation module 5004 carries out logistic regression study processing for the feature vector to each query string, obtains institute
State general retrieval prediction model.
Wherein, as shown in figure 5, in this example, which can also include the first normalized mould
Block 5003 is normalized for a pair feature vector corresponding with each query string；
At this point, above-mentioned model generation module 5004, carries out for the feature vector to each query string after normalized
Logistic regression study is handled.
Corresponding, in the specific example of the present invention, which can also include：
Second normalized module 5023, the feature vector for being generated to feature vector generation module 502 are returned
One change is handled；
At this point, above-mentioned probability determination module 503, is based on the spy after 5023 normalized of the second normalized module
Vector is levied to determine that the search-type of the query string is the probability of general retrieval.
After each feature vector is normalized, each feature vector can be enable to indicate in same area
Between, simplify subsequent processes, improves treatment effeciency.
Wherein, it is above-mentioned be normalized and logistic regression study processing mode can be with the ground of aforementioned present invention
It is identical in graph search kind identification method, any mode for having at present and being likely to occur later may be used and carry out.
In above-mentioned map search type identification and during establish general retrieval prediction model, the pre- of search result is extracted
If when the characteristic information of characteristic type, above-mentioned default characteristic type can be based on actual needs and be set, wherein a tool
Can be the arbitrary combination in following each information in body example：Inquiry string length, the number of section of query string are searched described in preceding N items
Maximum classification in hitch fruit whether be general retrieval, the maximum classification accounting of N search result, first search result text
Score, first search result and the difference of the text score of Article 2 search result, the average text of preceding N search result obtain
Point, difference with the average text score of preceding N search result of the text score of first search result, preceding N search result
The standard variance of text score, query string whether occur completely in the title of preceding N search result, first search result
POIRank, preceding N search result POIRank, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering.
One of ordinary skill in the art will appreciate that realizing all or part in embodiments of the present invention method
Flow is relevant hardware can be instructed to complete by computer program, and the program can be stored in a computer can
It reads in storage medium, the program is when being executed, it may include such as the flow of the embodiment of above-mentioned each method.Wherein, described to deposit
Storage media can be magnetic disc, CD, read-only memory (Read-Only Memory, ROM) or random access memory
(Random Access Memory, RAM) etc..Therefore, according to embodiments of the present invention scheme, the present invention also provides a kind of packets
Storage medium containing computer-readable program may be implemented above-mentioned when the computer-readable program in the storage medium executes
The map search kind identification method or map search method of the present invention in any of which.
The method of embodiment present invention as described above, can in the form of software be installed in corresponding machinery equipment,
And it is searched by controlling relevant processing equipment to complete above-mentioned map search type identification or map in the running software
The process of rope.Correspondingly, above-mentioned map search type identification device or map search system can be respectively set to be mounted on
Can also be corresponding terminal device itself on corresponding terminal device, terminal device here can be computer, mobile phone,
The arbitrary terminal devices such as tablet computer, PDA (Personal Digital Assistant, personal digital assistant), vehicle-mounted computer.
Fig. 6 is the module map for the computer system 1000 that can realize the embodiment of the present invention.The computer system 1000
An only example for being suitable for the invention computer environment is not construed as proposing appointing to the use scope of the present invention
What is limited.Computer system 1000 can not be construed to need to rely on or the illustrative computer system 1000 with diagram
One or more of component combination.
Computer system 1000 shown in Fig. 6 is the example of a computer system suitable for the present invention.Have
Other frameworks of different sub-systems configuration can also use.Such as there are big well known desktop computer, notebook, individual digital to help
The similar devices such as reason, smart phone, tablet computer, portable media player can be adapted for some embodiments of the present invention.
But it is not limited to equipment enumerated above.
As shown in fig. 6, computer system 1000 includes processor 1010, memory 1020 and system bus 1022.Including
Various system components including memory 1020 and processor 1010 are connected on system bus 1022.Processor 1010 is one
For executing the hardware of computer program instructions by arithmetic sum logical operation basic in computer system.Memory 1020
It is one to be used to temporarily or permanently store calculation procedure or the physical equipment of data (for example, program state information).System is total
Line 1020 can be any one in the bus structures of following several types, including memory bus or storage control, outer
If bus and local bus.Processor 1010 and memory 1020 can be by system bus 1022 into row data communication.Wherein
Memory 1020 includes read-only memory (ROM) or flash memory (being all not shown in figure) and random access memory (RAM), RAM
Typically refer to the main memory for being loaded with operating system and application program.
Computer system 1000 generally comprises a storage device 1030.Storage device 1030 can from a variety of computers
It reads to select in medium, computer-readable medium refers to any available medium that can be accessed by computer system 1000,
Including mobile and fixed two media.For example, computer-readable medium includes but not limited to, flash memory (miniature SD
Card), CD-ROM, digital versatile disc (DVD) or other optical disc storages, cassette, tape, disk storage or other magnetic storages are set
Any other medium that is standby, or can be used for storing information needed and being accessed by computer system 1000.
Computer system 1000 can carry out logical connection with one or more network equipment in a network environment.Network is set
Standby can be PC, server, router, smart phone, tablet computer or other common network nodes.Department of computer science
System 1000 is connected by LAN (LAN) interface 1040 or mobile communication module 1050 with the network equipment.LAN (LAN)
Refer in finite region, such as family, school, computer laboratory or using the network media office building, interconnection composition
Computer network.WiFi and twisted-pair feeder wiring Ethernet are two kinds of technologies of most common structure LAN.WiFi is a kind of
It can make 1000 swapping data of computer system or be connected to the technology of wireless network by radio wave.Mobile communication module
1050 are answered and are made a phone call by radio communication diagram while capable of being moved in a wide geographic area.In addition to logical
Other than words, mobile communication module 1050 is also supported to carry out in 2G, 3G or the 4G cellular communication system for providing mobile data service
Internet access.
Computer system 1000 includes the imaging sensor 1060 for photograph, is taken a picture and is obtained using imaging sensor 1060
Optical imagery is converted into electronic signal by image in 2 D code.
It should be pointed out that other includes than the computer system of 1000 more or fewer subsystems of computer system
It can be suitably used for inventing.For example, computer system 1000 may include that can exchange the bluetooth module of data in short distance.
As detailed above, map search type identification side can be executed by being suitable for the invention computer system 1000
The specified operation of method or map search method.Computer system 1000 operates in computer-readable medium by processor 1010
In the form of software instruction execute these operations.These software instructions from storage device 1030 or can pass through LAN
Interface 1040 is read into from another equipment in memory 1020.The software instruction being stored in memory 1020 makes processor
1010 execute above-mentioned map search kind identification method or map search method.In addition, passing through hardware circuit or hardware
Circuit combination software instruction also can equally realize the present invention.Therefore, realize the present invention is not limited to any specific hardware circuit and
The combination of software.
Several embodiments of the invention above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
Cannot the limitation to the scope of the claims of the present invention therefore be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the guarantor of the present invention
Protect range.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.
Claims (18)
1. a kind of map search kind identification method, which is characterized in that including step：
Obtain the search result that map search acquisition is carried out to query string；
The characteristic information of the default characteristic type of described search result is extracted, the feature vector for including the characteristic information is generated；
Determine that the search-type of the query string is the probability of general retrieval, institute according to described eigenvector, general retrieval prediction model
General retrieval prediction model is stated by a pair feature for default characteristic type corresponding with each query string in default inquiry set of strings to believe
Breath is trained acquisition.
2. map search kind identification method according to claim 1, which is characterized in that obtain the general retrieval and predict mould
The mode of type includes：
Structure inquires set of strings, includes general retrieval and inquisition string and non-general retrieval and inquisition string in the inquiry set of strings；
Map search is carried out to each query string in the inquiry set of strings respectively, obtains search knot corresponding with each query string
Fruit；
The characteristic information of the default characteristic type of the corresponding search result of each query string is extracted respectively, and it includes each inquiry to generate respectively
Go here and there the feature vector of corresponding characteristic information；
Logistic regression study processing is carried out to the feature vector of each query string, obtains the general retrieval prediction model.
3. map search kind identification method according to claim 2, it is characterised in that：
After generating the feature vector comprising the characteristic information, determine that the search-type of the query string is the general of general retrieval
Further include step before rate：Feature vector comprising the characteristic information is normalized；
It is carried out after generating the feature vector comprising the corresponding characteristic information of each query string respectively, to the feature vector of each query string
Further include step before logistic regression study processing：Pair feature vector corresponding with each query string is normalized.
4. the map search kind identification method according to claims 1 to 3 any one, which is characterized in that described default
Characteristic type includes the arbitrary combination in following each information：Inquire string length, query string section number, preceding N described search
As a result maximum classification in whether be general retrieval, the maximum classification accounting of N search result, first search result text obtain
Divide, first search result is obtained with the difference of the text score of Article 2 search result, the average text of preceding N search result
Point, difference with the average text score of preceding N search result of the text score of first search result, preceding N search result
The standard variance of text score, query string whether occur completely in the title of preceding N search result, first search result
POIRank, preceding N search result POIRank, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering.
5. a kind of map search type identification device, which is characterized in that including：
Data obtaining module, for obtaining the search result for carrying out map search acquisition to query string；
Feature vector generation module, the characteristic information of the default characteristic type for extracting described search result, it includes institute to generate
State the feature vector of characteristic information；
Probability determination module, the search-type for determining the query string according to described eigenvector, general retrieval prediction model
For the probability of general retrieval, the general retrieval prediction model is preset by pair corresponding with each query string in default inquiry set of strings
The characteristic information of characteristic type is trained acquisition.
6. map search type identification device according to claim 5, which is characterized in that further include：
Model construction module, for generating the general retrieval prediction model.
7. map search type identification device according to claim 6, which is characterized in that the model construction module packet
It includes：
Query string collection modules include general retrieval and inquisition string in the inquiry set of strings and non-for building inquiry set of strings
General retrieval and inquisition string；
Model eigenvectors generation module respectively obtains the query string progress map search in the inquiry set of strings for obtaining
The search result corresponding with each query string obtained extracts the feature letter of the default characteristic type of the corresponding search result of each query string
Breath generates the feature vector for including the corresponding characteristic information of each query string respectively；
Model generation module carries out logistic regression study processing for the feature vector to each query string, obtains the general retrieval
Prediction model.
8. map search type identification device according to claim 7, it is characterised in that：
Further include the second normalized module, for the feature vector comprising the characteristic information to be normalized；
The probability determination module is true according to the feature vector of the characteristic information after the normalized, general retrieval prediction model
The search-type of the fixed query string is the probability of general retrieval；
The model construction module further includes the first normalized module, for pair feature vector corresponding with each query string into
Row normalized；The model generation module carries out logic for the feature vector to each query string after normalized
Recurrence learning processing.
9. the map search type identification device according to claim 5 to 8 any one, which is characterized in that described default
Characteristic type includes the arbitrary combination in following each information：Inquire string length, query string section number, preceding N described search
As a result maximum classification in whether be general retrieval, the maximum classification accounting of N search result, first search result text obtain
Divide, first search result is obtained with the difference of the text score of Article 2 search result, the average text of preceding N search result
Point, difference with the average text score of preceding N search result of the text score of first search result, preceding N search result
The standard variance of text score, query string whether occur completely in the title of preceding N search result, first search result
POIRank, preceding N search result POIRank, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering.
10. a kind of map search method, which is characterized in that including step：
Map search instruction is received, the map search instruction includes query string；
It is scanned for according to map search instruction, obtains search result；
The characteristic information for extracting the default characteristic type of described search result, obtains the feature vector of described search result；
Determine that the search-type of the query string is the probability of general retrieval, root according to described eigenvector, general retrieval prediction model
Search-type according to the query string is the search-type of query string described in the determine the probability of general retrieval, and mould is predicted in the general retrieval
Type is trained acquisition by a pair characteristic information for default characteristic type corresponding with each query string in default inquiry set of strings；
The display processing mode of described search result is determined according to the search-type of the query string, and according to the display processing side
Formula handles described search result.
11. map search method according to claim 10, which is characterized in that the general retrieval prediction model passes through following
Mode obtains：
Structure inquires set of strings, includes general retrieval and inquisition string and non-general retrieval and inquisition string in the inquiry set of strings；
Map search is carried out to the query string in the inquiry set of strings respectively, obtains search result corresponding with each query string；
The characteristic information of the corresponding search result of each query string is extracted respectively, is generated respectively comprising the corresponding feature letter of each query string
The feature vector of breath；
Logistic regression study processing is carried out to the feature vector of each query string, obtains the general retrieval prediction model.
12. map search method according to claim 11, it is characterised in that：
After generating the feature vector comprising the characteristic information, determine that the search-type of the query string is the general of general retrieval
Further include step before rate：Feature vector comprising the characteristic information is normalized；
After generating the feature vector comprising the corresponding characteristic information of each query string respectively, the feature vector of each query string is carried out
Further include step before logistic regression study processing：Pair feature vector corresponding with each query string is normalized.
13. the map search method according to claim 10 to 12 any one, which is characterized in that the default feature class
Type includes the arbitrary combination in following each information：It inquires in string length, the number of section of query string, preceding N described search result
Maximum classification whether be general retrieval, the maximum classification accounting of N articles of search result, the text score of first article of search result, the
The difference of one search result and the text score of Article 2 search result, the average text score of preceding N search result, first
The text score of search result is obtained with the difference of the average text score of preceding N search result, the text of preceding N search result
Point standard variance, query string whether occur completely in the title of preceding N search result, first search result
POIRank, the POIRank of preceding N search result, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N search result clustering.
14. a kind of map search system, which is characterized in that including：
Search module, for receiving map search instruction, the map search instruction includes query string, is searched according to the map
Suo Zhiling is scanned for, and obtains search result；
Feature vector generation module, the characteristic information of the default characteristic type for extracting described search result, it includes institute to generate
State the feature vector of characteristic information；
Search-type determining module, the search for determining the query string according to described eigenvector, general retrieval prediction model
Type is the probability of general retrieval, according to the search of query string described in the determine the probability that the search-type of the query string is general retrieval
Type, the general retrieval prediction model pass through pair default characteristic type corresponding with each query string in default inquiry set of strings
Characteristic information is trained acquisition；
Display processing module, the display processing mode for determining described search result according to the search-type, and it is aobvious according to this
Show that processing mode handles described search result.
15. map search system according to claim 14, which is characterized in that further include：
Model construction module, for generating the general retrieval prediction model.
16. map search system according to claim 15, which is characterized in that the model construction module includes：
Query string collection modules include general retrieval and inquisition string in the inquiry set of strings and non-for building inquiry set of strings
General retrieval and inquisition string；
Model eigenvectors generation module carries out map search acquisition for obtaining to the query string in the inquiry set of strings
Search result corresponding with each query string extracts the feature letter of the default characteristic type of the corresponding search result of each query string respectively
Breath generates the feature vector for including the corresponding characteristic information of each query string respectively；
Model generation module carries out logistic regression study processing for the feature vector to each query string, obtains the general retrieval
Prediction model.
17. map search system according to claim 16, it is characterised in that：
Further include the second normalized module, for the feature vector comprising the characteristic information to be normalized；
Described search determination type module predicts mould according to the feature vector of the characteristic information after the normalized, general retrieval
Type determines that the search-type of the query string is the probability of general retrieval；
The model construction module further includes the first normalized module, for pair feature vector corresponding with each query string into
Row normalized；The model generation module carries out logic for the feature vector to each query string after normalized
Recurrence learning processing.
18. the map search system according to claim 14 to 17 any one, which is characterized in that the default feature class
Type includes the arbitrary combination in following each information：It inquires in string length, the number of section of query string, preceding N described search result
Maximum classification whether be general retrieval, the maximum classification accounting of N articles of search result, the text score of first article of search result, the
The difference of one search result and the text score of Article 2 search result, the average text score of preceding N search result, first
The text score of search result is obtained with the difference of the average text score of preceding N search result, the text of preceding N search result
Point standard variance, query string whether occur completely in the title of preceding N search result, first search result
POIRank, the POIRank of preceding N search result, first POIRank of search result and being averaged for preceding N search result
The convergence factor of the difference of POIRank, the cluster number of clusters of preceding N search result, preceding N 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 CN104615620A (en) | 2015-05-13 |
| CN104615620B true 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) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106933947B (en) * | 2017-01-20 | 2018-12-04 | 北京三快在线科技有限公司 | A kind of searching method and device, electronic equipment |
| 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 |
| 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 |
| CN112989224B (en) * | 2021-03-25 | 2024-06-25 | 北京百度网讯科技有限公司 | Retrieval method, retrieval device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101712988B1 (en) * | 2010-09-10 | 2017-03-07 | 삼성전자주식회사 | Method and apparatus providing for internet service in mobile communication terminal |
-
2014
- 2014-06-24 CN CN201410289451.1A patent/CN104615620B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104615620A (en) | 2015-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109145219B (en) | Method and device for judging validity of interest points based on Internet text mining | |
| CN106897428B (en) | Text classification feature extraction method and text classification method and device | |
| CN107346326B (en) | Method and system for information retrieval | |
| WO2020156389A1 (en) | Information pushing method and device | |
| WO2021143267A1 (en) | Image detection-based fine-grained classification model processing method, and related devices | |
| CN110297935A (en) | Image search method, device, medium and electronic equipment | |
| CN113434716B (en) | Cross-modal information retrieval method and device | |
| CN104615620B (en) | Map search kind identification method and device, map search method and system | |
| US20120084226A1 (en) | Measuring or estimating user credibility | |
| CN113806582B (en) | Image retrieval method, image retrieval device, electronic equipment and storage medium | |
| CN108629358A (en) | The prediction technique and device of object type | |
| CN107315833A (en) | Method and apparatus of the retrieval with downloading based on application program | |
| Ben-Shimon et al. | An ensemble method for top-N recommendations from the SVD | |
| CN110008740A (en) | A kind of processing method of document access authority, device, medium and electronic equipment | |
| CN113486989A (en) | Knowledge graph-based object recognition method and device, readable medium and equipment | |
| CN107968743A (en) | The method and apparatus of pushed information | |
| CN110598122B (en) | Social group mining method, device, equipment and storage medium | |
| CN114490923A (en) | Training method, device and equipment for similar text matching model and storage medium | |
| CN112214677B (en) | Point of interest recommendation method and device, electronic equipment and storage medium | |
| CN112862021A (en) | Content labeling method and related device | |
| CN107766881B (en) | Way finding method and device based on basic classifier and storage device | |
| CN107402886B (en) | Storehouse analysis method and relevant apparatus | |
| CN113704566B (en) | Identification number body identification method, storage medium and electronic equipment | |
| CN106651410A (en) | Application management method and application management device | |
| CN109446408A (en) | Retrieve method, apparatus, equipment and the computer readable storage medium of set of metadata of similar data |
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. |