CN107273427A - Striding equipment network information search method and system based on data fusion - Google Patents
Striding equipment network information search method and system based on data fusion Download PDFInfo
- Publication number
- CN107273427A CN107273427A CN201710353743.0A CN201710353743A CN107273427A CN 107273427 A CN107273427 A CN 107273427A CN 201710353743 A CN201710353743 A CN 201710353743A CN 107273427 A CN107273427 A CN 107273427A
- Authority
- CN
- China
- Prior art keywords
- search
- user
- mrow
- equipment
- webpage
- 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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Abstract
The invention discloses a kind of striding equipment network information search method based on data fusion and system, the system includes:Data collection module, for recording and collecting behavioral data when user carries out web search on the first device;The behavioral data includes user the stay time on webpage;First equipment is the search equipment that user used last time;Historical data processing module, for the webpage for displaying for a user the query formulation of recommendation on the second device according to the data of data collection module and recommending;Second equipment is the search equipment that user is using;Data processing module is searched for, the rearrangement based on data fusion, bonding apparatus information is carried out for searching for the search result produced to user on the second device.The present invention supports user's striding equipment information search, and solve user the problem of repeat search, realizes the rearrangement of search result after striding equipment after striding equipment, it is possible to increase the search efficiency of user, improves the search experience of user.
Description
Technical field
The present invention relates to information retrieval technique, more particularly to a kind of striding equipment network information search based on data fusion
Method and system.
Background technology
The development of mobile Internet, promoted the terminal devices such as smart mobile phone, tablet personal computer functionally constantly carry
Rise;The continuous reduction of smart machine manufacturing cost, also promotes people to possess more different types of smart machines, such as intelligent hand
Mechanical, electrical brain, flat board, intelligent watch etc..Further convenient, fast, the interaction of user and distinct device with the mode of network insertion
Also further frequently, Internet user can often switch between distinct device in life and use for activity.
When particularly user is using the different equipment search network informations, due to network, equipment size, functions of the equipments, outer
The influence of the difference factor such as boundary's environment, can frequently result in the interruption of search activities, and is transferred in other equipment and continues search for living
It is dynamic (such as to be used in user family after the information that search computer tourism is gone on a journey, at outdoor due to forgetting previous search result
And continue search for relevant information using mobile phone;And for example user in library using the INFORMATION such as mobile phone searching paper, but by
It is limited in preview, download, rear in going back home to be continued search for using desktop computer).Nowadays user on this distinct device across
Equipment search is very universal, and special user is meeting the information need that some of is complex, the needs consuming time is more
, it is necessary to repeatedly be searched for when asking, its search activities often spans different search sessions, can also cross over no terminal
Equipment.
In striding equipment search, it is necessary to continue it in second equipment after the completion of user searches in first equipment
Preceding search mission, provides for search activities of the user after occurring equipment transfer and supports (as helped user to recall, provide
Related query formulation or search history), more preferable search service and experience, the search effect of lifting user can be provided the user with
Rate, the development of propulsion information retrieval technique.
The current technology interacted for user's progress striding equipment search and striding equipment has mainly used first in user
After equipment, when user begins to use second equipment, collected in user interface by accessing individual center, access browser
The functions such as folder, are the search history in its upper equipment of repetition according to the order of time.These technologies are mainly to aid in using
Family accesses previous webpage, content again, such as provides the web page listings accessed before striding equipment, accesses the thumbnail of webpage, or
With in a browser, by the synchronous search activities such as collection, bookmark after login account, there are some browsers to be proposed
Carry out the function of user data synchronization to aid in the striding equipment of user to search for by individual center.But these technologies are all introduced
Excessive user mutual, availability is not high, the complexity of user interactive and bears higher, especially in the user interface
The support of striding equipment search is not provided the user actively, and is only suitable for simple some web search across session lives
It is dynamic.
In addition, supporting the technology of striding equipment search, system all to ignore a problem, i.e. user and occur equipment turn at present
After shifting, in addition to the search activities before continuation, in addition it is also necessary to further scan for.And current simple duplicate customer is gone through
The technology and method of Records of the Historian record, it is impossible to judge whether the previous search activities of user are necessary to repeat for user again;User by
In after striding equipment transfer caused by the factors such as functions of the equipments, network environment, external environment, it is often necessary to further scan for,
So based on the previous search activities of user and interactive history, search when user occurring to search again for after striding equipment transfer
Sort result is particularly significant, it should with reference to the user data in previous session, the search result after striding equipment is arranged again
Sequence, it is to avoid the search that user repeats, lifts the search efficiency of user, improves the search experience of user.
The content of the invention
The technical problem to be solved in the present invention is to be based on data fusion there is provided one kind for defect of the prior art
Striding equipment network information search method and system.
The technical solution adopted for the present invention to solve the technical problems is:A kind of striding equipment network based on data fusion
Information search method and system,
Striding equipment network information search system based on data fusion, including:
Data collection module, for recording and collecting behavioral data when user carries out web search on the first device;
The behavioral data includes user the stay time on webpage;First equipment is the search equipment that user used last time;
Historical data processing module, is pushed away for being displayed for a user on the second device according to the data of data collection module
The query formulation recommended and the webpage recommended;Second equipment is the search equipment that user is using;
Data processing module is searched for, is carried out for searching for the search result produced to user on the second device based on number
According to fusion, the rearrangement of bonding apparatus information.
By such scheme, the user behavior data also includes user name, timestamp, the numbering of session, the equipment used
Type, the page type accessed, the URL addresses of accession page, the html source codes of accession page, use when accessing the page
Query formulation, accession page when occur event, IP address when user accesses the system, and user is in mobile terminal
In equipment access the system when on screen touch data;The event occurred during the accession page includes the activation page, closed
The page and jump page;The screen touch data includes touch-control direction, position of touch, touch-control speed and touch-control angle.
By such scheme, the historical data processing module also provides the user looking into search history in the first equipment
Inquiry formula and access webpage.
By such scheme, the historical data processing module displays for a user the query formulation and the webpage recommended of recommendation
Specific method is as follows:
The query formulation in search history and the calculated value of webpage are calculated according to following formula, according to calculated value from high to low
To being ranked up in the query formulation and webpage in user's search history, and according to the quantity of recommendation show recommend query formulation and
The webpage of recommendation:
Wherein, dwell represents stay time of the user on the Webpage, and λ is expression time importance parameter, λ's
General value is 0.1;Δ T represents the novelty that document is obtained, and this search time of Δ T=-last user accesses the document
Time;Wdevice is device type importance parameter, and Wdevice parameter value preferably chooses 0.8 recommendation effect more;If with
Device category is different after the striding equipment of family, and SD values are 0;If equipment is identical after user's striding equipment, SD values are 1;The device category bag
Include mobile device and non-mobile device.
By such scheme, the search data processing module provides the user the method for the search result of rearrangement such as
Under:
The calculating of initial sequence calculated value is carried out to the initial search result of search engine,
Wherein, Rel is the initial sequence calculated value of search-engine results, and wherein rank each document is in search result
Ranking (value that the document ranking such as searched is 1, rank is 1;2) value that the document ranking searched is 2, rank is.
Calculate the calculated value of document ordering in search result after the striding equipment search based on data fusion;
Scorefinal=Wrel*Rel-Wreaccess*scorereaccess
Wherein, ScorereaccessFor the calculating of the recommendation query formula in system home page based on data fusion and recommendation webpage
Value;ScorefinalFor the calculated value of document ordering in search result after the striding equipment search based on data fusion, WrelFor weighing apparatus text
Shelves correlation significance level parameter, value takes 0.9, WreaccessIt is to weigh ScorereaccessThe ginseng of weight in whole sequence
Number, was calculated by the device type in Score reaccess, residence time etc.;When general, when parameter value takes
When 0.5, the search results ranking best results of generation, therefore WrelParameter value preferably choose 0.9, WreaccessParameter value
Preferably choose 0.5.
According to ScorefinalCalculated value for user generate striding equipment after the search result resequenced.
Two algorithms are contained in striding equipment network information search method proposed by the present invention based on data fusion, this
Two algorithms can either meet user after striding equipment turns, and the search activities occurred in first equipment are proceeded
Demand, meets the demand that user further scans for again.
Striding equipment network information search method based on data fusion, comprises the following steps:
1) record and collect behavioral data when user carries out web search on the first device;The behavioral data bag
Include stay time of the user on webpage;First equipment is the search equipment that user used last time;
2) it is that user is being provided based on the recommendation net merged from the first device data when user uses the second equipment
Page and recommendation query formula, the demand always according to user provide the query formulation in the first equipment in search history and access webpage;
The specific method of the query formulation for providing the user recommendation and the webpage recommended is as follows:
The query formulation in search history and the calculated value of webpage are calculated according to following formula, according to calculated value from high to low
To being ranked up in the query formulation and webpage in user's search history, and according to the quantity of recommendation show recommend query formulation and
The webpage of recommendation:
Wherein, dwell represents stay time of the user on the Webpage, and λ is expression time importance parameter, λ's
General value is 0.1;Δ T represents the novelty that document is obtained, and this search time of Δ T=-last user accesses the document
Time;Wdevice is device type importance parameter, and Wdevice parameter value preferably chooses 0.8 recommendation effect more;If with
Device category is different after the striding equipment of family, and SD values are 0;If equipment is identical after user's striding equipment, SD values are 1;The device category bag
Include mobile device and non-mobile device;
2) user's input inquiry formula in the second equipment, starts new search;
3) complete after user's search, the search result based on distinct device data fusion is provided in result of page searching
There is provided the search result after sequence for rearrangement;
The method for the search result that the search data processing module provides the user rearrangement is as follows:
The calculating of initial sequence calculated value is carried out to the initial search result of search engine,
Wherein, Rel is the initial sequence calculated value of search-engine results, and wherein rank each document is in search result
Ranking (value that the document ranking such as searched is 1, rank is 1;2) value that the document ranking searched is 2, rank is.
Calculate the calculated value of document ordering in search result after the striding equipment search based on data fusion;
Scorefinal=Wrel*Rel-Wreaccess*Scorereaccess
Wherein, ScorereaccessFor the calculating of the recommendation query formula in system home page based on data fusion and recommendation webpage
Value;ScorefinalFor the calculated value of document ordering in search result after the striding equipment search based on data fusion, WrelFor weighing apparatus text
Shelves correlation significance level parameter, value takes 0.9, WreaccessIt is to weigh ScorereaccessThe ginseng of weight in whole sequence
Number, was calculated by the device type in Score reaccess, residence time etc.;When general, when parameter value takes
When 0.5, the search results ranking best results of generation, therefore WrelParameter value preferably choose 0.9, WreaccessParameter value
Preferably choose 0.5.
According to ScorefinalCalculated value for user generate striding equipment after the search result resequenced.
By such scheme, the step 1) in recommend query formulation and recommend webpage quantity be setting value.
By such scheme, the step 1) in the query formulation recommended and the webpage recommended be one-to-one.
The beneficial effect comprise that:Striding equipment network information search of the user between distinct device is supported, is solved
User after striding equipment repeat search, recover search, and the problem of search for more relevant informations;With reference to user across setting
Standby situation, the optimization based on algorithm, by fusion of the original searching results based on striding equipment contextual information, is searched after carrying out striding equipment
The rearrangement of hitch fruit, facilitates user to carry out seamless search between different devices, it is possible to increase the search efficiency of user,
Improve the search experience of user.
Brief description of the drawings
Below in conjunction with drawings and Examples, the invention will be further described, in accompanying drawing:
Fig. 1 is one embodiment of the striding equipment network information search method based on data fusion using the present invention
Flow chart;
Fig. 2 is one embodiment of the striding equipment network information search system based on data fusion using the present invention
Flow chart;
Fig. 3 is a specific embodiment of the striding equipment network information search system based on data fusion of the present invention
Schematic diagram.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to this hair
It is bright to be further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, and without
It is of the invention in limiting.
Including one key character of striding equipment search is the search sessions after striding equipment and the search sessions before striding equipment
Continuity in appearance, search need, especially because the complexity of search mission is high, before user possibly can not remember completely
Search procedure, therefore the present invention first against after user's striding equipment search sessions provide auxiliary, to support holding for task
It is continuous.
The inventive method is mainly by recording user data of the user in the equipment before striding equipment, and pass through this hair
" supporting that information reacquisition utilizes algorithm during striding equipment search " (algorithm 1) of bright proposition, preferably lives with the search before striding equipment
Displayed for a user in the related query formulation of dynamic height and the webpage accessed, the system home page proposed in the present invention.Display
The query formulation of height correlation and the webpage accessed are the calculating by algorithm, by the user interactive data of search activities before
Bring into the calculating of algorithm, be the webpages that user generates 5 query formulations recommended and 5 recommendations.Here the quantity recommended can
To be adjusted on backstage.Search history button is displayed for a user in homepage simultaneously, user, which can click on, checks that oneself owns
The query formulation history and the history of access webpage submitted.
Second object of the present invention is to support user after striding equipment search, to enter the heuristic search of row information.I.e.
User tentatively knows about before striding equipment to search mission, it is necessary to search further for after striding equipment.At this moment, Yong Huxu
That wants not just needs previous search history, accesses history, with greater need for it have submitted query formulation in second equipment after,
Search the information of more height correlations.
Based on this purpose, the present invention proposes a kind of " supporting search result rearrangement algorithm during striding equipment search " and (calculated
Method 2), this method is to combine the previous search activities of user, and when user's generation equipment is shifted and is searched again for, search is drawn
Hold up original search result to be resequenced, user can either be met in striding equipment search procedure to prior searches activity
Recover, can continue to explore new information again, meet the complicated information requirement of user.
This method mainly in conjunction with user striding equipment search before, inquiry when being scanned in first equipment
Operation note on content, result of page searching, the information of first equipment, search result is clicked on the key mouse on the page and touched
The stay time data on interaction data, and related web page are controlled, these data are brought into and " support to search during striding equipment search
In the calculating of hitch fruit rearrangement algorithm ", document re-ranking sequence is carried out to search result after striding equipment, realizes that search result is arranged
The optimization of sequence.
Based on the above method, the present invention propose support the kind of striding equipment network information search based on data fusion across
Device network information search method and system, support striding equipment search seamless connection problem of the user under multi-equipment environment.
Striding equipment network information search system based on data fusion, including:
Data collection module, for recording and collecting behavioral data when user carries out web search on the first device;
The behavioral data includes user the stay time on webpage;First equipment is the search equipment that user used last time;
Historical data processing module, is pushed away for being displayed for a user on the second device according to the data of data collection module
The query formulation recommended and the webpage recommended;Second equipment is the search equipment that user is using;
Data processing module is searched for, is carried out for searching for the search result produced to user on the second device based on number
According to fusion, the rearrangement of bonding apparatus information;
User behavior data in data collection module also includes user name, timestamp, the numbering of session, setting of using
Standby type, the page type accessed, the URL addresses of accession page, the html source codes of accession page, make when accessing the page
The event that occurs when query formulation, accession page, IP address when user accesses the system, and user is in movement
In end equipment access the system when on screen touch data;The event occurred during the accession page includes the activation page, closed
Close the page and jump page;The screen touch data includes touch-control direction, position of touch, touch-control speed and touch-control angle.
Historical data processing module displays for a user the query formulation of recommendation and the specific method of the webpage of recommendation is as follows:
The query formulation in search history and the calculated value of webpage are calculated according to following formula, according to calculated value from high to low
To being ranked up in the query formulation and webpage in user's search history, and according to the quantity of recommendation show recommend query formulation and
The webpage of recommendation:
Wherein, dwell represents stay time of the user on the Webpage, and λ is expression time importance parameter, λ's
General value is 0.1;Δ T represents the novelty that document is obtained, and this search time of Δ T=-last user accesses the document
Time;Wdevice is device type importance parameter, and Wdevice parameter value preferably chooses 0.8 recommendation effect more;If with
Device category is different after the striding equipment of family, and SD values are 0;If equipment is identical after user's striding equipment, SD values are 1;The device category bag
Include mobile device and non-mobile device.
The method for the search result that search data processing module provides the user rearrangement is as follows:
The calculating of initial sequence calculated value is carried out to the initial search result of search engine,
Wherein, Rel is the initial sequence calculated value of search-engine results, and wherein rank each document is in search result
Ranking (value that the document ranking such as searched is 1, rank is 1;2) value that the document ranking searched is 2, rank is.
Calculate the calculated value of document ordering in search result after the striding equipment search based on data fusion;
Scorefinal=Wrel*Rel-Wreaccess*Scorereaccess
Wherein, ScorereaccessFor the calculating of the recommendation query formula in system home page based on data fusion and recommendation webpage
Value;ScorefinalFor the calculated value of document ordering in search result after the striding equipment search based on data fusion, WrelFor weighing apparatus text
Shelves correlation significance level parameter, value takes 0.9, WreaccessIt is to weigh ScorereaccessThe ginseng of weight in whole sequence
Number, was calculated by the device type in Score reaccess, residence time etc.;When general, when parameter value takes
When 0.5, the search results ranking best results of generation, therefore WrelParameter value preferably choose 0.9, WreaccessParameter value
Preferably choose 0.5.
According to ScorefinalCalculated value for user generate striding equipment after the search result resequenced.
Two algorithms are contained in striding equipment network information search method proposed by the present invention based on data fusion, this
Two algorithms can either meet user after striding equipment turns, and the search activities occurred in first equipment are proceeded
Demand, meets the demand that user further scans for again.
Striding equipment network information search method based on data fusion, comprises the following steps:
1) record and collect behavioral data when user carries out web search on the first device;The behavioral data bag
Include stay time of the user on webpage;First equipment is the search equipment that user used last time;
2) it is that user is being provided based on the recommendation net merged from the first device data when user uses the second equipment
Page and recommendation query formula, the demand always according to user provide the query formulation in the first equipment in search history and access webpage;
The specific method of the query formulation for providing the user recommendation and the webpage recommended is as follows:
The query formulation in search history and the calculated value of webpage are calculated according to following formula, according to calculated value from high to low
To being ranked up in the query formulation and webpage in user's search history, and according to the quantity of recommendation show recommend query formulation and
The webpage of recommendation:
Wherein, dwell represents stay time of the user on the Webpage, and λ is expression time importance parameter, λ's
General value is 0.1;Δ T represents the novelty that document is obtained, and this search time of Δ T=-last user accesses the document
Time;Wdevice is device type importance parameter, and Wdevice parameter value preferably chooses 0.8 recommendation effect more;If with
Device category is different after the striding equipment of family, and SD values are 0;If equipment is identical after user's striding equipment, SD values are 1;The device category bag
Include mobile device and non-mobile device.
2) user's input inquiry formula in the second equipment, starts new search;
3) complete after user's search, the search result based on distinct device data fusion is provided in result of page searching
There is provided the search result after sequence for rearrangement;
The method for the search result that the search data processing module provides the user rearrangement is as follows:
The calculating of initial sequence calculated value is carried out to the initial search result of search engine,
Wherein, Rel is the initial sequence calculated value of search-engine results, and wherein rank each document is in search result
Ranking (value that the document ranking such as searched is 1, rank is 1;2) value that the document ranking searched is 2, rank is.
Calculate the calculated value of document ordering in search result after the striding equipment search based on data fusion;
Scorefinal=Wrel*Rel-Wreaccess*Scorereaccess
Wherein, ScorereaccessFor the calculating of the recommendation query formula in system home page based on data fusion and recommendation webpage
Value;ScorefinalFor the calculated value of document ordering in search result after the striding equipment search based on data fusion, WrelFor weighing apparatus text
Shelves correlation significance level parameter, value takes 0.9, WreaccessIt is to weigh ScorereaccessThe ginseng of weight in whole sequence
Number, was calculated by the device type in Score reaccess, residence time etc.;When general, when parameter value takes
When 0.5, the search results ranking best results of generation, therefore WrelParameter value preferably choose 0.9, WreaccessParameter value
Preferably choose 0.5.
According to ScorefinalCalculated value for user generate striding equipment after the search result resequenced.
Wherein step 1) in the query formulation recommended and the webpage quantity recommended be setting value, step 1) in the inquiry recommended
Formula is one-to-one with the webpage recommended.
The specific embodiment used of the inventive method and system is described below.
The invention provides a kind of method and system for supporting Internet user to carry out striding equipment network information search, Fig. 1
For according to the present invention the striding equipment network information search method based on data fusion one embodiment flow chart, mainly
Including:
Step S101, user use first equipment, such as desktop computer, notebook computer, smart mobile phone, tablet personal computer
Deng input inquiry formula progress web search.
Step S102, the striding equipment network information search system proposed by the present invention based on data fusion, can record and receive
Collection user carries out behavioral data during web search in first equipment.These data are all used for carrying out " supporting striding equipment to search
Information reacquires and utilizes algorithm during rope " (algorithm 1) and " supporting search result rearrangement algorithm when striding equipment is searched for " (calculation
Method 2) calculating.
Step S103, user use second equipment, such as desktop computer, notebook computer, smart mobile phone, tablet personal computer
Web search is carried out Deng, input inquiry formula, user has carried out striding equipment network letter due to being influenceed by extraneous factor here
Breath search.
Step S104, utilization " supporting that information reacquisition utilizes algorithm during striding equipment search " (algorithm 1) are in system home page
The webpage for displaying for a user the query formulation of recommendation and recommending, algorithm here will be described in detail in figure 3.
Step S105, utilization " supporting search result rearrangement algorithm during striding equipment search " (algorithm 2) are inputted in user
After query formulation, the rearrangement based on data fusion, bonding apparatus information is carried out to search result.
The striding equipment network information search method based on data fusion and system of the embodiment of the present invention, by user
Data when carrying out web search using distinct device are collected, analyze, handle, merged, and occur striding equipment turn in user
There is provided based on the recommendation query formula, recommendation webpage generated after distinct device data fusion, and search result after moving
Rearrangement.
One of the striding equipment network information search system based on data fusion according to the present invention is described in detail in Fig. 2
The flow chart of embodiment, mainly includes:
Step S201, user use equipment 1 (such as smart mobile phone) login system, and the address of system is:
Crosssearch.whu.edu.cn, the striding equipment network information search system proposed by the present invention based on data fusion is logical
User name is crossed, the data of user on different devices are associated, recommendation is produced.
After step S202, user are by user name and password login system, the striding equipment network information proposed by the present invention is searched
Cable system can be provided based on the recommendation webpage from distinct device data fusion, recommendation query formula, user for user in homepage
The query formulation in prior search history can also be accessed by search history and webpage is accessed.
And when new user accesses the system first, do not provided because system not yet records the history log of user, therefore
Recommend webpage and recommendation query formula.
Step S203, user input query formula, start search.
After step S204, user's search, system can be provided in result of page searching based on distinct device data fusion
Search result is resequenced.The step that the method for sequence is below can be discussed in detail.
And when new user accesses the system first, do not provided because system not yet records the history log of user, therefore
Search result rearrangement based on data fusion.
After step S205, user search in first equipment, the striding equipment net proposed by the present invention based on data fusion
Network information search system can record the behavioral data of user in the background, for user's striding equipment search for after recommendation query formula,
Webpage, and the rearrangement of the search result based on data fusion is recommended to provide the related foundation that algorithm is calculated.
In step S205, the data of the striding equipment network information search system proposed by the present invention based on data fusion
Collection module can collect user using the system when various user data, including user name, timestamp, the numbering of session,
The device type used, the page type accessed, the URL addresses of accession page, the html source codes of accession page, access are somebody's turn to do
The event (such as the activation page, closing the page, jump page) that occurs when the query formulation that is used during the page, accession page, in webpage
On stay time, IP address when user accesses the system, and user is when mobile end equipment is complained to the higher authorities about an injustice and request fair settlement and asks the system
Screen touch data (including touch-control direction, position of touch, touch-control speed, touch-control angle on (such as smart mobile phone, tablet personal computer)
Degree) etc..
Step S206, after user data has been collected, the present invention propose it is a kind of based on data fusion striding equipment network letter
Searching method is ceased, this method is important to include two algorithms, be " to support that information reacquires utilization during striding equipment search respectively
Algorithm " (algorithm 1) and " supporting search result rearrangement algorithm when striding equipment is searched for " (algorithm 2).The two algorithms it is specific
Content can be described in detail in the following step.
Step S207, the user login system in equipment 2 (such as notebook computer), current search activities occur with set
In standby 1 different equipment.
Step S208, due to having been searched for before user, therefore system proposed by the present invention is in the Data Collection on backstage
The data of user have been have accumulated in module, and has generated and is melted based on data by the algorithm 1 and algorithm 2 in system algorithm module
The associated recommendation of conjunction.
User has been logged in after the system, can be provided in homepage the recommendation query formula from distinct device, recommend webpage,
Search history, recommendation query formula and recommendation webpage are the recommendations that the result calculated by algorithm 1 is generated, and are not simply to search
Rope history is repeated.
Start search after step S209, user input query formula.
After step S210, user's search, result of page searching is generated.By the algorithm 2 in system algorithm module, with reference to
Search history before user, resequences to search result.
The striding equipment network information search method based on data fusion and system of the embodiment of the present invention, can help to use
Previous search history is recovered at family, continues complex search activities, and provides the user in search result personalization
Recommend sequence, improve search efficiency and experience that user is searched under situation in striding equipment.
Step S2060, be described in detail here " support striding equipment search when information reacquire utilizes algorithm " (algorithm 1)
Specific composition.
In this algorithm, ScorereaccessFor the recommendation query formula in system home page based on data fusion and recommendation webpage
Calculated value, this value be for calculate user's striding equipment search in related important query formulation and webpage, final calculating
Value is used for being ranked up in the query formulation and webpage in user's search history, is that user preferably carries in system home page
For.Dwell represents stay time of the user on a Webpage.Parameter lambda=0.1.Δ T value subtracts equal to current time
Go the time of the access page.Algorithm debugging and user's test and appraisal by early stage, present invention discover that Wdevice parameter value takes
More preferably, therefore Wdevice parameter value preferably chooses 0.8 to recommendation effect when 0.8.If equipment is different after user's striding equipment
(being such as transferred to mobile phone from desktop computer), SD values are 0.If equipment is identical after user's striding equipment, SD values are 1.
Algorithm 1 in the striding equipment network information search method based on data fusion of the embodiment of the present invention can be preferred
Ground be user occur striding equipment transfer when, with more conforming to its continuation search need query formulation and webpage are provided for it, and
It is not to simply provide to fall the query formulation of row according to the time, more conforms to user's request.
Step S2061, be described in detail here " support striding equipment search when search result rearrangement algorithm " (algorithm
2) specific composition.
In this algorithm, several calculating process are contained.The system is ranked up to the initial results of search engine first,
Rel is the initial sequence calculated value of search-engine results, and wherein rank represents ranking of the different document in search result
(value that the document ranking such as searched is 1, rank is 1;2) value that the document ranking searched is 2, rank is.
ScorereaccessFor the recommendation query formula in system home page based on data fusion and the calculated value (i.e. algorithm 1) of recommendation webpage.
ScorefinalFor after the striding equipment search based on data fusion in search result document ordering calculated value, this
Calculated value is ultimately used to generate the search result resequenced after striding equipment for user.By early stage algorithm debugging and
User tests and assesses, it is a discovery of the invention that WrelParameter value take 0.9, WreaccessParameter value when taking 0.5, the search result row of generation
Sequence best results, therefore WrelParameter value preferably choose 0.9, WreaccessParameter value preferably choose 0.5.
Two algorithms are contained in striding equipment network information search method proposed by the present invention based on data fusion, this
Two algorithms can either meet user after striding equipment turns, and the search activities occurred in first equipment are proceeded
Demand, meets the demand that user further scans for again.
One of the striding equipment network information search system based on data fusion according to the present invention is described in detail in Fig. 3
The interface schematic diagram of specific embodiment.
Step S501, user are on smart mobile phone, input inquiry formula " machine learning " relevant search information.In the present invention
The striding equipment network information search system home page based on data fusion proposed has displayed for a user " recommending search ", " has recommended net
Page ", " recommending search " here, " recommendation webpage " are " to support that information is obtained again during striding equipment search according to proposed by the present invention
Take and utilize algorithm " (algorithm 1) realize, " recommend search ", " recommendation webpage " here shows 5 respectively, and this quantity can be with
It is adjusted in striding equipment network searching system proposed by the present invention.
After step S502, user input query formula, in result of page searching, it is shown that " supported by proposed by the present invention
The sequence for the search result that search result rearrangement algorithm when striding equipment is searched for " (algorithm 2) is realized.
Step S503, user are clicked on after first search result, browse search result, the information on the webpage can lead to
Step S205 records are crossed, and analysis calculating is carried out by step S206.
Striding equipment transfer occurs for step S504, user, is searched on notebook computer, and input inquiry formula " calculate by machine learning
Method " relevant search information.
Displayed for a user in the striding equipment network information search system home page proposed by the present invention based on data fusion
" recommending search ", " recommendation webpage ", user can also click on " search history " and check oneself all search record, including look into
Inquiry formula, the webpage accessed, search history here is arranged according to time inverted order.
In " recommending search ", " recommendation webpage ", the equipment letter that the query formulation and the webpage occur also has been displayed for a user
Breath, such as query formulation come from terminal console equipment (such as notebook computer, desktop computer), then use iconRepresent;Come
From mobile end equipment (such as smart mobile phone, tablet personal computer), then icon is usedRepresent.In webpage is recommended, the net also show
The search time of page, user is helped to recover its previous search mission.
In step S505, after user input query formula, in result of page searching, it is shown that by " branch proposed by the present invention
The sequence for the search result that search result rearrangement algorithm when holding striding equipment search " (algorithm 2) is realized.It is shown in figure
The sectional drawing of partial search results, and the search result of not all.
The search results ranking that the algorithm is calculated, the on the one hand recovery in view of user to prior searches task, such as
Search result link (first and Article 3 search result) lower section previously clicked in user, it is shown that search time, use
Device type and click on the relevant inquiring formula used during the webpage that family is used when clicking on.On the other hand user is also allowed for enter
The demand of one step relevant search information.
Therefore, in striding equipment network information search method and system proposed by the present invention based on data fusion, searching
During rope sort result, be not merely consider user may repeat click on webpage and document, but by " support across
Search result rearrangement algorithm when equipment is searched for " (algorithm 2), has considered friendship when user searches on different devices
Mutual data, search history, so as to support seamless search problem of the user between distinct device.
In summary, the invention provides a kind of striding equipment network information search method based on data fusion and system,
Support striding equipment network information search of the user between distinct device, solve user after striding equipment repeat search, obtain more
The problem of many relevant informations, the rearrangement of search result after striding equipment is realized, facilitate user to carry out between different devices
Seamless search, it is possible to increase the search efficiency of user, improves the search experience of user.
It should be appreciated that for those of ordinary skills, can according to the above description be improved or be become
Change, and all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Claims (8)
1. a kind of striding equipment network information search system based on data fusion, it is characterised in that including:
Data collection module, for recording and collecting behavioral data when user carries out web search on the first device;It is described
Behavioral data includes user the stay time on webpage;First equipment is the search equipment that user used last time;
Historical data processing module, for displaying for a user looking into for recommendation on the second device according to the data of data collection module
Inquiry formula and the webpage recommended;Second equipment is the search equipment that user is using;
Data processing module is searched for, on the second device searching for user the search result produced based on data melt
Conjunction, the rearrangement of bonding apparatus information.
2. striding equipment network information search system according to claim 1, it is characterised in that the user behavior data is also
Including user name, timestamp, the numbering of session, the device type used, access page type, the URL addresses of accession page,
The html source codes of accession page, access the event occurred when the query formulation used during the page, accession page, user access this
IP address during system, and user when mobile end equipment is complained to the higher authorities about an injustice and request fair settlement and asks the system on screen touch data;It is described to visit
Ask that the event occurred during the page includes the activation page, closes the page and jump page;The screen touch data includes touch-control side
To, position of touch, touch-control speed and touch-control angle.
3. striding equipment network information search system according to claim 1, it is characterised in that the historical data handles mould
Block, also provides the user the query formulation in the first equipment in search history and accesses webpage.
4. striding equipment network information search system according to claim 1, it is characterised in that the historical data handles mould
Block displays for a user the query formulation of recommendation and the specific method of the webpage of recommendation is as follows:
The query formulation in search history and the calculated value of webpage are calculated according to following formula, according to calculated value from high to low to user
It is ranked up in query formulation and webpage in search history, and the query formulation recommended and the net of recommendation is shown according to the quantity of recommendation
Page:
<mrow>
<msub>
<mi>Score</mi>
<mrow>
<mi>r</mi>
<mi>e</mi>
<mi>a</mi>
<mi>c</mi>
<mi>c</mi>
<mi>e</mi>
<mi>s</mi>
<mi>s</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>d</mi>
<mi>w</mi>
<mi>e</mi>
<mi>l</mi>
<mi>l</mi>
</mrow>
<mn>10.0</mn>
</mfrac>
<mo>&times;</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mi>&lambda;</mi>
<mo>&times;</mo>
<mi>&Delta;</mi>
<mi>T</mi>
</mrow>
</msup>
<mo>&times;</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>d</mi>
<mi>e</mi>
<mi>v</mi>
<mi>i</mi>
<mi>c</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mi>S</mi>
<mi>D</mi>
</mrow>
</msup>
<mo>&times;</mo>
<msup>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>d</mi>
<mi>e</mi>
<mi>v</mi>
<mi>i</mi>
<mi>c</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>S</mi>
<mi>D</mi>
</mrow>
</msup>
</mrow>
Wherein, dwell represents stay time of the user on the Webpage, and λ is expression time importance parameter, and Δ T is represented
The novelty that document is obtained, this search time of Δ T=-last user accesses the time of the document;Wdevice is device type
Importance parameter;If device category is different after user's striding equipment, SD values are 0;If equipment is identical after user's striding equipment, SD values are 1;
The device category includes mobile device and non-mobile device.
5. striding equipment network information search system according to claim 1, it is characterised in that the search data processing mould
The method that block provides the user the search result of rearrangement is as follows:
The calculating of initial sequence calculated value is carried out to the initial search result of search engine,
<mrow>
<mi>Re</mi>
<mi>l</mi>
<mo>&Proportional;</mo>
<mfrac>
<mn>1.0</mn>
<mrow>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mi>a</mi>
<mi>n</mi>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
</mrow>
Wherein, Rel is the initial sequence calculated value of search-engine results, row of the wherein rank each document in search result
Name;
Calculate the calculated value of document ordering in search result after the striding equipment search based on data fusion;
Scorefinal=Wrel*Rel-Wreaccess*Scorereaccess
Wherein, ScorereaccessFor the recommendation query formula in system home page based on data fusion and the calculated value of recommendation webpage;
ScorefinalFor the calculated value of document ordering in search result after the striding equipment search based on data fusion, WrelFor weighing apparatus document phase
Closing property significance level parameter, value takes 0.9, WreaccessIt is to weigh ScorereaccessThe parameter of weight in whole sequence, works as ginseng
When numerical value takes 0.5, the search results ranking best results of generation, therefore WrelParameter value preferably choose 0.9, Wreaccess's
Parameter value preferably chooses 0.5;
According to ScorefinalCalculated value for user generate striding equipment after the search result resequenced.
6. a kind of striding equipment network information search method based on data fusion, it is characterised in that comprise the following steps:
1) record and collect behavioral data when user carries out web search on the first device;The behavioral data includes user
Stay time on webpage;First equipment is the search equipment that user used last time;
2) it is that user is providing based on the recommendation webpage merged from the first device data and pushed away when user uses the second equipment
Query formulation is recommended, the demand always according to user provides the query formulation in the first equipment in search history and accesses webpage;
The specific method of the query formulation for providing the user recommendation and the webpage recommended is as follows:
The query formulation in search history and the calculated value of webpage are calculated according to following formula, according to calculated value from high to low to user
It is ranked up in query formulation and webpage in search history, and the query formulation recommended and the net of recommendation is shown according to the quantity of recommendation
Page:
<mrow>
<msub>
<mi>Score</mi>
<mrow>
<mi>r</mi>
<mi>e</mi>
<mi>a</mi>
<mi>c</mi>
<mi>c</mi>
<mi>e</mi>
<mi>s</mi>
<mi>s</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>d</mi>
<mi>w</mi>
<mi>e</mi>
<mi>l</mi>
<mi>l</mi>
</mrow>
<mn>10.0</mn>
</mfrac>
<mo>&times;</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mi>&lambda;</mi>
<mo>&times;</mo>
<mi>&Delta;</mi>
<mi>T</mi>
</mrow>
</msup>
<mo>&times;</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>d</mi>
<mi>e</mi>
<mi>v</mi>
<mi>i</mi>
<mi>c</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mi>S</mi>
<mi>D</mi>
</mrow>
</msup>
<mo>&times;</mo>
<msup>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>d</mi>
<mi>e</mi>
<mi>v</mi>
<mi>i</mi>
<mi>c</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>S</mi>
<mi>D</mi>
</mrow>
</msup>
</mrow>
Wherein, dwell represents stay time of the user on the Webpage, and λ is expression time importance parameter, and Δ T is represented
The novelty that document is obtained, this search time of Δ T=-last user accesses the time of the document;Wdevice is device type
Importance parameter;If device category is different after user's striding equipment, SD values are 0;If equipment is identical after user's striding equipment, SD values are 1;
The device category includes mobile device and non-mobile device;
2) user's input inquiry formula in the second equipment, starts new search;
3) complete after user's search, the search result based on distinct device data fusion is provided in result of page searching and arranged again
There is provided the search result after sequence for sequence;
The method for the search result that the search data processing module provides the user rearrangement is as follows:
The calculating of initial sequence calculated value is carried out to the initial search result of search engine,
<mrow>
<mi>Re</mi>
<mi>l</mi>
<mo>&Proportional;</mo>
<mfrac>
<mn>1.0</mn>
<mrow>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mi>a</mi>
<mi>n</mi>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
</mrow>
Wherein, Rel is the initial sequence calculated value of search-engine results, row of the wherein rank each document in search result
Name;
Calculate the calculated value of document ordering in search result after the striding equipment search based on data fusion;
Scorefinal=Wrel*Rel-Wreaccess*Scorereaccess
Wherein, ScorereaccessFor the recommendation query formula in system home page based on data fusion and the calculated value of recommendation webpage;
ScorefinalFor the calculated value of document ordering in search result after the striding equipment search based on data fusion, WrelFor weighing apparatus document phase
Closing property significance level parameter, WreaccessIt is to weigh ScorereaccessThe parameter of weight in whole sequence;
According to ScorefinalCalculated value for user generate striding equipment after the search result resequenced.
7. striding equipment network information search method according to claim 6, it is characterised in that the step 1) middle recommendation
Query formulation and the webpage quantity of recommendation are setting value.
8. striding equipment network information search method according to claim 6, it is characterised in that the step 1) middle recommendation
Query formulation is one-to-one with the webpage recommended.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710353743.0A CN107273427B (en) | 2017-05-18 | 2017-05-18 | Cross-device network information searching method and system based on data fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710353743.0A CN107273427B (en) | 2017-05-18 | 2017-05-18 | Cross-device network information searching method and system based on data fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107273427A true CN107273427A (en) | 2017-10-20 |
CN107273427B CN107273427B (en) | 2020-09-01 |
Family
ID=60064164
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710353743.0A Active CN107273427B (en) | 2017-05-18 | 2017-05-18 | Cross-device network information searching method and system based on data fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107273427B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108334536A (en) * | 2017-11-30 | 2018-07-27 | 中国电子科技集团公司电子科学研究院 | A kind of information recommendation method, equipment and storage medium |
CN113904827A (en) * | 2021-09-29 | 2022-01-07 | 恒安嘉新(北京)科技股份公司 | Method and device for identifying counterfeit website, computer equipment and medium |
CN115617600A (en) * | 2021-07-15 | 2023-01-17 | 北京特纳飞电子技术有限公司 | Collecting runtime information for debugging and analysis |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101681377A (en) * | 2007-05-23 | 2010-03-24 | 微软公司 | User-defined relevance ranking for search |
US20120150855A1 (en) * | 2010-12-13 | 2012-06-14 | Yahoo! Inc. | Cross-market model adaptation with pairwise preference data |
CN103412958A (en) * | 2013-08-30 | 2013-11-27 | 广州市动景计算机科技有限公司 | Display method and device for searching result |
CN103533530A (en) * | 2013-09-26 | 2014-01-22 | 林毅 | Cross-device user corresponding and user tracking methods and systems |
CN105324754A (en) * | 2013-06-03 | 2016-02-10 | 微软技术许可有限责任公司 | Task continuance across devices |
CN105359136A (en) * | 2013-06-04 | 2016-02-24 | 微软技术许可有限责任公司 | Responsive input architecture |
CN106663116A (en) * | 2014-11-19 | 2017-05-10 | 谷歌公司 | Method, systems, and media for presenting links to media content |
-
2017
- 2017-05-18 CN CN201710353743.0A patent/CN107273427B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101681377A (en) * | 2007-05-23 | 2010-03-24 | 微软公司 | User-defined relevance ranking for search |
US20120150855A1 (en) * | 2010-12-13 | 2012-06-14 | Yahoo! Inc. | Cross-market model adaptation with pairwise preference data |
CN105324754A (en) * | 2013-06-03 | 2016-02-10 | 微软技术许可有限责任公司 | Task continuance across devices |
CN105359136A (en) * | 2013-06-04 | 2016-02-24 | 微软技术许可有限责任公司 | Responsive input architecture |
CN103412958A (en) * | 2013-08-30 | 2013-11-27 | 广州市动景计算机科技有限公司 | Display method and device for searching result |
CN103533530A (en) * | 2013-09-26 | 2014-01-22 | 林毅 | Cross-device user corresponding and user tracking methods and systems |
CN106663116A (en) * | 2014-11-19 | 2017-05-10 | 谷歌公司 | Method, systems, and media for presenting links to media content |
Non-Patent Citations (1)
Title |
---|
吴丹等: "多设备环境下网络信息搜索行为研究综述", 《中国图书馆学报》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108334536A (en) * | 2017-11-30 | 2018-07-27 | 中国电子科技集团公司电子科学研究院 | A kind of information recommendation method, equipment and storage medium |
CN108334536B (en) * | 2017-11-30 | 2023-10-24 | 中国电子科技集团公司电子科学研究院 | Information recommendation method, device and storage medium |
CN115617600A (en) * | 2021-07-15 | 2023-01-17 | 北京特纳飞电子技术有限公司 | Collecting runtime information for debugging and analysis |
CN113904827A (en) * | 2021-09-29 | 2022-01-07 | 恒安嘉新(北京)科技股份公司 | Method and device for identifying counterfeit website, computer equipment and medium |
CN113904827B (en) * | 2021-09-29 | 2024-03-19 | 恒安嘉新(北京)科技股份公司 | Identification method and device for counterfeit website, computer equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN107273427B (en) | 2020-09-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102004794B (en) | Search engine system and implementation method thereof | |
Xie et al. | Efficient browsing of web search results on mobile devices based on block importance model | |
Karlson et al. | FaThumb: a facet-based interface for mobile search | |
KR101667344B1 (en) | Method and system for providing search results | |
CN102368262B (en) | Method and equipment for providing searching suggestions corresponding to query sequence | |
US7930287B2 (en) | Systems and methods for compound searching | |
US20100058202A1 (en) | Method system and program product for providing enabling an interactive and social search engine | |
CN103246678B (en) | A kind of web page content preview method and apparatus | |
CN107016020A (en) | The system and method for aiding in searching request using vertical suggestion | |
US20100169756A1 (en) | Automated bookmarking | |
CN1353838A (en) | Server-side WEB summary generation and presentation | |
CN102779136A (en) | Method and device for information search | |
CN107273427A (en) | Striding equipment network information search method and system based on data fusion | |
CN102937975B (en) | A kind of Webpage search equipment and method | |
CN103793495B (en) | Application message search method and system and application message acquisition methods and system | |
JPWO2009072174A1 (en) | Information search apparatus, information search method, and search processing program | |
JP2014081918A (en) | Method for providing recommendation result cooperated with search word automatic completion function | |
WO2003012687A1 (en) | Contents service system and method using image, and computer readable storage medium stored therein computer executable instructions to implement contents service method | |
US7975238B2 (en) | Identifying previously bookmarked hyperlinks in a received Web page in a World Wide Web network browser system for searching | |
EP1216448A2 (en) | A system and method for advanced network viewing | |
TW200426775A (en) | Extracting displayed numerical data from displayed documents received from communication networks, e.g. world wide web, and processing the extracted numerical data independent of the received document | |
CN102663070B (en) | Method and system for supporting browser application | |
JP2001331486A (en) | Website integrated retrieval method on communication and recording medium storing software programmed so as to perform the method | |
Heimonen et al. | Mobile findex: supporting mobile web search with automatic result categories | |
CN1922606B (en) | For dynamic keyword processing system and the method for user oriented internet navigation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |