CN103020206A - Knowledge-network-based search result focusing system and focusing method - Google Patents
Knowledge-network-based search result focusing system and focusing method Download PDFInfo
- Publication number
- CN103020206A CN103020206A CN2012105185537A CN201210518553A CN103020206A CN 103020206 A CN103020206 A CN 103020206A CN 2012105185537 A CN2012105185537 A CN 2012105185537A CN 201210518553 A CN201210518553 A CN 201210518553A CN 103020206 A CN103020206 A CN 103020206A
- Authority
- CN
- China
- Prior art keywords
- search results
- knowledge network
- focusing
- engine
- knowledge
- 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.)
- Pending
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a knowledge-network-based search result focusing system and a knowledge-network-based search result focusing method. The focusing system comprises a knowledge network library, a search result module, a mode engine and a focusing presentation module, wherein the mode engine is connected with the search result module, the knowledge network library and the focusing presentation module respectively; the search result module inputs search results of a universal search engine into the mode engine; the mode engine performs mode judgment and classification on the search results from a root node of the knowledge network library according to mode judgment functions in knowledge network nodes so as to finish focusing the search results; and the focusing presentation module outputs and displays the focused search results. According to the system and the method, a knowledge network is used for focusing the search results, so that the search results can be accurately classified, the number of obtained classes is definite, and the name of each class is clear and readable.
Description
Technical field
The present invention relates to a kind of Search Results focusing system and focus method, relate in particular to a kind of knowledge network that utilizes predetermined field as the focusing foundation, thereby improve focusing system and the focus method of Search Results focusing effect, belong to the web search technical field.
Background technology
Current, the data total amount in the internet is calculated with the hundreds of terabyte, and still is exponential increase.In order to help user's required information of quick obtaining from the data ocean that this extends endlessly, search engine is being brought into play irreplaceable effect.Internet information is extremely vast and numerous, and any one keyword all may search hundreds of even tens thousand of relevant webpage or links.And the Search Results that obtains by universal search engine often precision is low, be difficult to allow rapidly the user obtain required content.Therefore, be necessary Search Results is focused on classification, rearrange classification and ordination, so that rapidly lock onto target set of user.
In this technical field, at present general method is to utilize search result clustering, according to the condition of certain user's the unknown with search result clustering in some subsets, and provide the subset label to be beneficial to the user to lock rapidly.For example application number is in 200410091772.7 the Chinese patent application, discloses a kind of method of search result clustering, comprises the steps: that pre-recorded indexed document is with respect to one or more classifications of its certain or certain several keywords that comprise; , with respect to the classification that is included in certain or certain the several keywords in the searching request document in the described Search Results is divided into groups according to pre-recorded document.Described classification can be arbitrarily document classification mark or keyword.Each classification can arrange a weighted value.Document in the Search Results is placed in the classification set of the document with respect to searching keyword, and the rank of resulting each cluster classification can be calculated by the rank of its document that comprises.This clustering method is applicable to extensive DRS (for example internet search engine) to the cluster of Search Results, and by to cluster classification grading, can with have the cluster of higher level and wherein the document of higher level preferentially present to the user.
U.S. Yahoo has also proposed a kind of for search data and Search Results is grouped into the method and apparatus of the cluster that sorts according to relevance of searches in application number is 200780049318.7 Chinese patent application.Each cluster comprises one or more data types, such as image, the WEB page, local information, news, advertisement etc.In one embodiment, estimate search terms for the related notion of the classification of indicating the data source of searching for.Also can be by coming identification data source such as the position of client device, the contextual informations such as application of current operation.Search Results in each cluster sorts by correlativity, and each cluster is based on the gathering of the correlativity within this cluster and be given mark.Each cluster score can be revised based on one or more corresponding concepts and/or contextual information.Based on the modified mark cluster that sorts.The content that comprises advertisement also can be added to tabulation through ordering to show as another cluster.
In addition, be in the Chinese invention patent of ZL 200810226637.7 in the patent No., further propose a kind of method and device thereof of optimizing cluster search results, can not satisfy the problem of user's personalized search demand in order to solve Search Results that existing cluster seeking technology returns.Concrete technical scheme comprises: according to the cluster classification in the current cluster search results, search the historical weights corresponding with described cluster classification from user's historical search information of pre-save; According to lookup result and the current weight corresponding to described cluster classification of described historical weights, determine the as a result weights that described cluster classification is corresponding; According to described as a result weights, determine the priority when described cluster classification returns to the user.The cluster search results that returns by this technical scheme can satisfy user's personalized search demand, improves user's Experience Degree.
Because the existing limitation of clustering technique, the search result data amount that causes existing universal search engine to provide is large, precision is low.The user has to carry out artificial binary search, and work efficiency is low.
Summary of the invention
Technical matters to be solved by this invention is to provide a kind of Search Results focusing system and focus method based on knowledge network.The knowledge network in the predetermined field of this technical scheme utilization can effectively improve the focusing effect of Search Results as focusing on foundation.
For realizing above-mentioned goal of the invention, the present invention adopts following technical scheme:
A kind of Search Results focusing system based on knowledge network, comprise that knowledge network storehouse, Search Results module, pattern engine and focusing present module, wherein said modeling engine connects respectively described Search Results module, described knowledge network storehouse and described focusing and presents module;
Described Search Results module is inputted the Search Results of universal search engine in the described pattern engine, described pattern engine is from the root node in described knowledge network storehouse, according to the pattern discrimination function in the knowledge network node described Search Results is carried out pattern discrimination and sort out, finish the focusing work of described Search Results;
Described Search Results output display after described focusing presents module and will focus on.
Wherein more preferably, also comprise auxiliary conceptional tree in the described Search Results focusing system, described auxiliary conceptional tree is connected with described knowledge network storehouse.
Wherein more preferably, described knowledge network storehouse is made of a plurality of knowledge network nodes, and each knowledge network node has self exclusive pattern discrimination function and concept label.
Wherein more preferably, a plurality of knowledge network nodes in the described knowledge network storehouse adopt tree-shaped hierarchical structure distribution mode.
A kind of Search Results focus method based on knowledge network is realized based on above-mentioned Search Results focusing system, comprises the steps:
At first by user's input, cause the Search Results of universal search engine;
After described Search Results is loaded into the pattern engine, described pattern engine is from the root node in knowledge network storehouse, according to the pattern discrimination function in the knowledge network node described Search Results being carried out pattern discrimination sorts out, with described Search Results respectively merger enter in the corresponding knowledge network node, thereby finish the focusing work of described Search Results;
Described Search Results output display after focusing presents module and will focus on.
Wherein more preferably, in the deterministic process of described pattern engine, further utilize the auxiliary semantic logic that carries out of the concept word set of storing in the auxiliary conceptional tree to differentiate.
Wherein more preferably, record the original alignment order of described Search Results, the Search Results that each knowledge network Nodes focuses on is arranged according to original clooating sequence.
Wherein more preferably, the output that described focusing presents module is tree-shaped layering and increases progressively form, and shows the concept label of each knowledge network node.
The present invention utilizes knowledge network that Search Results is focused on, and can realize accurate classification, and classification quantity is definite, and the name that makes each classification is clear readable, avoided existing clustering method to generate the classification number at random, the classification semanteme is unclear, allows the user be difficult to locate fast the problem of required content.
Description of drawings
Fig. 1 is the one-piece construction synoptic diagram of Search Results focusing system provided by the present invention;
Fig. 2 is the tree-shaped hierarchical structure synoptic diagram in knowledge network storehouse;
Fig. 3 is an embodiment synoptic diagram in knowledge network storehouse shown in Figure 2.
Embodiment
Below in conjunction with the drawings and specific embodiments, the technical solution adopted in the present invention is described in further detail.
Fig. 1 has shown the one-piece construction of Search Results focusing system provided by the present invention.This Search Results focusing system presents module by auxiliary conceptional tree, knowledge network storehouse, Search Results module, pattern engine and focusing and forms.Wherein, modeling engine connects respectively Search Results module, knowledge network storehouse and focusing and presents module, and auxiliary conceptional tree connects the knowledge network storehouse.This Search Results module, focus on and to present the ripe algorithm that module etc. can adopt computer realm, realize with software or firmware mode.Auxiliary conceptional tree and knowledge network storehouse etc. can be realized in the nonvolatile memory mode.In this Search Results focusing system, at first by user's input, cause the Search Results of universal search engine.The Search Results module obtains Search Results from universal search engine, and with in this Search Results input pattern engine.The knowledge network storehouse is made of the knowledge network node that a plurality of concepts and logic discrimination merge, and each knowledge network node has concept label and concept discrimination model (being presented as the pattern discrimination function) thereof.Auxiliary conceptional tree is the storage set of concept word set, and the concept supporting function is played in the knowledge network storehouse.The pattern engine is the execution core of pattern discrimination, after the Search Results record loads, the pattern engine is from the root node in knowledge network storehouse, according to the pattern discrimination function g (x) in the knowledge network node Search Results record is carried out the pattern discrimination classification, thereby finish the focusing work of Search Results.Search Results output display after focusing presents module and will focus on, output is tree-shaped layering and increases progressively form, and shows the concept label (Tag) of each knowledge network node.
Fig. 2 has shown in the knowledge network storehouse, the tree-shaped hierarchical structure distribution mode of a plurality of knowledge network nodes.On the inner structure in knowledge network storehouse, the form that the present invention increases progressively with tree-shaped hierarchical structure represents the hierarchical relationship between knowledge concepts, and each knowledge network node is the fusion of the machine expression of concept label and knowledge, the i.e. fusion of Tag and g (x).Each knowledge network node is by the concept segmentation of pattern discrimination function g (x) realization next stage, in order to can realize quick focusing.For example in Fig. 2, i child node sub (1 arranged under the root node root, 1), sub (1,2) ... sub (1, i), and child node sub (1,1) have again under j Sun Jiedian sub (2,1), sub (2,2) ... sub (2, j), wherein i, j are positive integer.In each knowledge network node, all have himself exclusive pattern discrimination function g (x) and concept label (Tag).Wherein, the concept label has determined this basic fixed position of knowledge network node in the knowledge network storehouse, for example " China ", " U.S. " and " Australia " these three knowledge network nodes all have concept label " place name ", and " cloudy ", " fine day " and " rainy day ", these three knowledge network nodes all had concept label " weather " etc.Pattern discrimination function g (x) has determined the concept generic relation between certain knowledge network node and other knowledge network node.This pattern discrimination function g (x) is the conventional techniques means that pattern-recognition (Pattern Recognition) those skilled in the art can both grasp, specifically can consult Bian Zhaoqi etc. writes " pattern-recognition (the 2nd edition) " (publishing house of Tsing-Hua University publishes in January, 2000, ISBN:9787302010593), just do not describe in detail at this.
Based on above-mentioned Search Results focusing system, the technical characterstic of this Search Results focus method is to utilize the pattern discrimination function g (x) of knowledge network node, and the position that the differentiation Search Results is recorded in the knowledge concepts space is divided.In this Search Results focus method, at first by user's input, cause the Search Results of universal search engine.This Search Results need to record original clooating sequence.After the Search Results record is loaded into the pattern engine, the pattern engine begins according to the discrimination model g (x) the knowledge network node Search Results record to be carried out the pattern discrimination classification from the root node in knowledge network storehouse, Search Results is recorded respectively merger enter in the corresponding knowledge network node.Simultaneously, the record that focuses at each knowledge network Nodes is arranged according to original clooating sequence.In the deterministic process of pattern engine, further utilize the auxiliary semantic logic that carries out of the concept word set of storing in the auxiliary conceptional tree to differentiate.At last, focus on the Search Results output display after presenting module and will focusing on, export and be tree-shaped layering and increase progressively form, and show the concept label of each knowledge network node.
In knowledge network storehouse example shown in Figure 3, three child nodes are arranged: " world ", " continent " and " Hong Kong, Macao and Taiwan " under the knowledge network node " news ", three child nodes are arranged: " history ", " present age " and " modern times " under the knowledge network node " personage ", three child nodes are arranged under the knowledge network node " the local introduction ": " China ", " U.S. " and " Australia ", etc.Can also further segment out Sun Jiedian under each child node, the rest may be inferred.When having new data to arrive the knowledge network node, differentiate its subclass classification (introduce in case of necessity the concept word set of storing in the auxiliary conceptional tree and assist differentiation) by pattern discrimination function g (x), and put into corresponding child node.For example the Search Results module provides three Search Results: " US President Washington XXX ", " US President Barack Obama XXX " and " President Roosevelt XXX ", when these three Search Results arrive knowledge network node " news ", can judge that by pattern discrimination function g (x) they all belong to the world news subclass, therefore further put into " world " child node; When these three Search Results arrive knowledge network nodes " personage ", can judge that by pattern discrimination function g (x) they should put into respectively " history ", " present age " and " modern times " three child nodes.After all Search Results focus on, further generate the cluster result of tree-shaped hierarchical structure, each cluster title is taken from the title of this knowledge network node.
The foundation that the present invention directly utilizes the knowledge network in predetermined field to focus on as Search Results by the combination that concept label and the Concept Semantic of utilization knowledge network are differentiated, can realize that clear significant Search Results focuses on.Focus on the result and be tree-shaped hierarchical structure, clear.Focus on the semantic clear and definite of classification of set, make the user can lock rapidly focus.
More than Search Results focusing system and the focus method based on knowledge network of the present invention had been described in detail.For one of ordinary skill in the art, any apparent change of under the prerequisite that does not deviate from connotation of the present invention it being done all will consist of infringement of patent right of the present invention, will bear corresponding legal liabilities.
Claims (8)
1. Search Results focusing system based on knowledge network is characterized in that:
Described Search Results focusing system comprises that knowledge network storehouse, Search Results module, pattern engine and focusing present module, and wherein said modeling engine connects respectively described Search Results module, described knowledge network storehouse and described focusing and presents module;
Described Search Results module is inputted the Search Results of universal search engine in the described pattern engine, described pattern engine is from the root node in described knowledge network storehouse, according to the pattern discrimination function in the knowledge network node described Search Results is carried out pattern discrimination and sort out, finish the focusing work of described Search Results;
Described Search Results output display after described focusing presents module and will focus on.
2. Search Results focusing system as claimed in claim 1 is characterized in that:
Also comprise auxiliary conceptional tree in the described Search Results focusing system, described auxiliary conceptional tree is connected with described knowledge network storehouse.
3. Search Results focusing system as claimed in claim 1 is characterized in that:
Described knowledge network storehouse is made of a plurality of knowledge network nodes, and each knowledge network node has self exclusive pattern discrimination function and concept label.
4. Search Results focusing system as claimed in claim 1 is characterized in that:
A plurality of knowledge network nodes in the described knowledge network storehouse adopt tree-shaped hierarchical structure distribution mode.
5. the Search Results focus method based on knowledge network is realized based on Search Results focusing system claimed in claim 1, it is characterized in that:
At first by user's input, cause the Search Results of universal search engine;
After described Search Results is loaded into the pattern engine, described pattern engine is from the root node in knowledge network storehouse, according to the pattern discrimination function in the knowledge network node described Search Results being carried out pattern discrimination sorts out, with described Search Results respectively merger enter in the corresponding knowledge network node, thereby finish the focusing work of described Search Results;
Described Search Results output display after focusing presents module and will focus on.
6. Search Results focus method as claimed in claim 5 is characterized in that:
In the deterministic process of described pattern engine, further utilize the auxiliary semantic logic that carries out of the concept word set of storing in the auxiliary conceptional tree to differentiate.
7. Search Results focus method as claimed in claim 5 is characterized in that:
Record the original alignment order of described Search Results, the Search Results that each knowledge network Nodes focuses on is arranged according to original clooating sequence.
8. Search Results focus method as claimed in claim 5 is characterized in that:
The output that described focusing presents module is tree-shaped layering and increases progressively form, and shows the concept label of each knowledge network node.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012105185537A CN103020206A (en) | 2012-12-05 | 2012-12-05 | Knowledge-network-based search result focusing system and focusing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012105185537A CN103020206A (en) | 2012-12-05 | 2012-12-05 | Knowledge-network-based search result focusing system and focusing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103020206A true CN103020206A (en) | 2013-04-03 |
Family
ID=47968810
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012105185537A Pending CN103020206A (en) | 2012-12-05 | 2012-12-05 | Knowledge-network-based search result focusing system and focusing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103020206A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504082A (en) * | 2014-12-24 | 2015-04-08 | 北京德塔普博软件有限公司 | Path showing method and system for target knowledge node sets of multiple knowledge networks |
| WO2017005196A1 (en) * | 2015-07-08 | 2017-01-12 | 谭红 | Knowledge point structure-based search apparatus |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020099700A1 (en) * | 1999-12-14 | 2002-07-25 | Wen-Syan Li | Focused search engine and method |
| CN1955958A (en) * | 2005-10-26 | 2007-05-02 | 腾讯科技（深圳）有限公司 | Sort data storage and split catalog inquiry method based on catalog tree |
| CN101000607A (en) * | 2006-01-12 | 2007-07-18 | 国际商业机器公司 | Visual method and device for strenthenzing search result guide |
| CN101216826A (en) * | 2007-01-05 | 2008-07-09 | 鸿富锦精密工业（深圳）有限公司 | Information search system and method |
| US20080177731A1 (en) * | 2007-01-23 | 2008-07-24 | Justsystems Corporation | Data processing apparatus, data processing method and search apparatus |
| CN102332031A (en) * | 2011-10-18 | 2012-01-25 | 中国科学院自动化研究所 | Method for clustering retrieval results based on video collection hierarchical theme structure |
-
2012
- 2012-12-05 CN CN2012105185537A patent/CN103020206A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020099700A1 (en) * | 1999-12-14 | 2002-07-25 | Wen-Syan Li | Focused search engine and method |
| CN1955958A (en) * | 2005-10-26 | 2007-05-02 | 腾讯科技（深圳）有限公司 | Sort data storage and split catalog inquiry method based on catalog tree |
| CN101000607A (en) * | 2006-01-12 | 2007-07-18 | 国际商业机器公司 | Visual method and device for strenthenzing search result guide |
| CN101216826A (en) * | 2007-01-05 | 2008-07-09 | 鸿富锦精密工业（深圳）有限公司 | Information search system and method |
| US20080177731A1 (en) * | 2007-01-23 | 2008-07-24 | Justsystems Corporation | Data processing apparatus, data processing method and search apparatus |
| CN102332031A (en) * | 2011-10-18 | 2012-01-25 | 中国科学院自动化研究所 | Method for clustering retrieval results based on video collection hierarchical theme structure |
Non-Patent Citations (6)
| Title |
|---|
| CARLOS N. SILLA JR ET AL: "《A Survey of Hierarchical Classifcation Across Di®erent Application Domains》", 《DATA MINING AND KNOWLEDGE》 * |
| SHAILESH KUMAR ET AL: ""Hierarchical Fusion of Multiple Classifiers for Hyperspectral Data Analysis"", 《PATTERN ANALYSIS & APPLICATIONS》 * |
| YU-CHIANG FRANK WANG ET AL: ""Soft-Decision Hierarchical Classification Using SVM-Type Classifiers"", 《2008 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS》 * |
| 张健钦 等: "《应用知识决策树分类遥感影像识别潘阳湖区螺钉孳生地》", 《中国吸血虫病防治杂志2008年第20卷第1期》, vol. 20, no. 1, 29 February 2008 (2008-02-29) * |
| 王冬丽 等: ""基于后验概率的SVM决策树多类分类算法"", 《微型电脑应用》 * |
| 袁辉华: ""一种基于标记和本体的分布式P2P资源获取方法"", 《硅谷》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504082A (en) * | 2014-12-24 | 2015-04-08 | 北京德塔普博软件有限公司 | Path showing method and system for target knowledge node sets of multiple knowledge networks |
| WO2017005196A1 (en) * | 2015-07-08 | 2017-01-12 | 谭红 | Knowledge point structure-based search apparatus |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829104B (en) | Semantic similarity based pseudo-correlation feedback model information retrieval method and system | |
| CN105210064B (en) | Classifying resources using deep networks | |
| US9589208B2 (en) | Retrieval of similar images to a query image | |
| CN109918532A (en) | Image search method, device, equipment and computer readable storage medium | |
| CN112819023B (en) | Sample set acquisition method, device, computer equipment and storage medium | |
| JP2010165348A (en) | Method for performing annotation and computer program therefor | |
| CN102508859A (en) | Advertisement classification method and device based on webpage characteristic | |
| CN110059181A (en) | Short text stamp methods, system, device towards extensive classification system | |
| CN104573130A (en) | Entity resolution method based on group calculation and entity resolution device based on group calculation | |
| CN104199965A (en) | Semantic information retrieval method | |
| CN103412925A (en) | System and method for integrated searching of structured data and unstructured data | |
| CN102902826A (en) | Quick image retrieval method based on reference image indexes | |
| CN111177432A (en) | Large-scale image retrieval method based on hierarchical depth hash | |
| CN102339294A (en) | Searching method and system for preprocessing keywords | |
| JP2020512651A (en) | Search method, device, and non-transitory computer-readable storage medium | |
| CN102968419A (en) | Disambiguation method for interactive Internet entity name | |
| CN115563313A (en) | Knowledge graph-based document book semantic retrieval system | |
| CN101477555A (en) | Fast retrieval and generation display method for task tree based on SQL database | |
| CN103678513A (en) | Interactive search generation method and system | |
| CN115438274A (en) | False news identification method based on heterogeneous graph convolutional network | |
| CN114328800A (en) | Text processing method and device, electronic equipment and computer readable storage medium | |
| CN102314464A (en) | Lyrics searching method and lyrics searching engine | |
| CN103020206A (en) | Knowledge-network-based search result focusing system and focusing method | |
| CN104376000A (en) | Webpage attribute determination method and webpage attribute determination device | |
| CN107577690B (en) | Recommendation method and recommendation device for mass information 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130403 |
|
| RJ01 | Rejection of invention patent application after publication |