WO2014198025A1 - Method and system for information retrieval - Google Patents

Method and system for information retrieval Download PDF

Info

Publication number
WO2014198025A1
WO2014198025A1 PCT/CN2013/077118 CN2013077118W WO2014198025A1 WO 2014198025 A1 WO2014198025 A1 WO 2014198025A1 CN 2013077118 W CN2013077118 W CN 2013077118W WO 2014198025 A1 WO2014198025 A1 WO 2014198025A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
retrieval
information
information retrieval
visual interface
Prior art date
Application number
PCT/CN2013/077118
Other languages
French (fr)
Chinese (zh)
Inventor
牛合庆
郝玺龙
陈金玉
丁海星
Original Assignee
天津海量信息技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 天津海量信息技术有限公司 filed Critical 天津海量信息技术有限公司
Priority to PCT/CN2013/077118 priority Critical patent/WO2014198025A1/en
Publication of WO2014198025A1 publication Critical patent/WO2014198025A1/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24Querying
    • G06F16/242Query formulation
    • G06F16/2428Query predicate definition using graphical user interfaces, including menus and forms

Abstract

Provided are a method and system for information retrieval. The method comprises configuring a corresponding relationship, on one side whereof are feature nodes within a visual interface and connections between the feature nodes, and on the other side whereof are retrieval units within an information retrieval formula and logic operations between the retrieval units. A visual interface is constructed according to an information retrieval objective, and comprises the feature nodes and their interconnections. The feature nodes within the visual interface and their interconnections are compiled into an information retrieval formula according to the corresponding relationship, and a search is conducted according to this formula. Applying the present technical solution, users can construct complex retrieval formulae, and in particular can retrieve retrieval units of specific types, thereby reducing the difficulty for users of constructing retrieval formulae, expanding information retrieval functionalities and improving information retrieval efficiency.

Description

An information retrieval method and system

FIELD

The present invention relates to the field of information retrieval, especially relates to a method and system for information retrieval. Background technique

With the advent of the information age, people enter the ocean of information. In the face of a flood of information, but people know what to do, it is difficult to find the information they need in a short time. However, the development of computer technology and network technology, information retrieval to provide some degree of help, people can build their desired search strategy, and the use of computer technology and network technology, to get the correct information.

The so-called search strategy is based on analysis of key questions on determining the retrieved data sources, search terms, and clearly retrieve the logical relationships between words and find scientific arrangement of steps. Retrieval formula (i.e., dubbed group retrieval words each operator expression) on the narrow search strategy is.

The retrieval process of the first link is to retrieve a clear demand, if the first step wrong, it would not last the accuracy of search results. Because users of their needs, especially potentially ambiguous requirements are not always very clear, and therefore needs to be analyzed, in order to achieve a complete and clear expression.

In the configuration retrieval process formula, to be involved in many aspects of knowledge and skills. Such as the extent to which the user explicitly retrieve the subject, the analysis of the retrieved subject; mastery of data sources and their system characteristics and functions; logical aspects of the preparation techniques to retrieve formulas and methods to retrieve adjustment formula will affect the overall effect of the user retrieval. Therefore, the development of retrieval formula is a very professional work, to construct a suitable and effective retrieval formula for non-professionals is very difficult. SUMMARY

Object of the present invention is to overcome the disadvantages and deficiencies of the prior art by providing a method and system for information retrieval by a visual way to help users configured to retrieve formula, thereby improving the efficiency of information retrieval.

An embodiment of the present invention provides a method of information retrieval comprising the steps of: setting a correspondence relationship, the correspondence relationship is a connection between one of the visual interface of the feature node and wherein the node, the another correspondence relationship It is one of the information retrieval operation between the logical formula and the search unit searching unit;

Visual interface constructed in accordance with the search target information, the connection between the feature node and wherein the node comprises a visual interface;

The connection between the feature node and wherein the node in the visual interface, according to the correspondence, compiled information retrieval formula;

The information retrieval formula for retrieving.

Preferably, the feature node includes logical nodes and node model identification, logical nodes for indicating logical operations, pattern recognition for the node representing the search unit, a logical node corresponds to less than a logical nodes and / or less than a recognition mode node.

Preferably, the identification mode node and / or logical nodes with lines pointing to its own logical node involvement.

Preferably, said node comprises a logic AND, OR, XOR, and non nodes, different logical nodes different icon.

Preferably, the node is a text recognition mode, images, audio and / or video data.

Preferably, the node is a data type recognition mode.

Preferably, the data type is text, image, audio and / or video.

Preferably, said contact is a data type, named entity, sentence, sentence, or emotional language.

Preferably, the node is a pattern for identifying the data type, identified by adding a label or using a preset icons.

Another embodiment of the present invention to provide an information retrieval system, comprises means, an input means, a database, compiling means and retrieval means, wherein the display,

Display means for displaying a visual interface, said visual interface constructed in accordance with the search target information, the connection between the feature node and includes a feature node;

Input means for inputting a connection between the feature node and wherein the node in the visual interface; a database for storing a correspondence relationship, one of the correspondence relationship between the feature node is connected to the node and wherein the visual interface, said corresponding side of the relationship is a logical operation between the information retrieval unit retrieves formulas and retrieval unit;

Compiling means for connecting between the feature node and wherein the node in the visual interface, according to the correspondence, compiled information retrieval formula;

Retrieval means for retrieving information according to a formula, for retrieval. Aspect 'of the present invention

Without complicated retrieval formula prepared, can be configured to retrieve a complex formula, in particular, the search unit can retrieve some type, thereby reducing the difficulty of retrieving the user to build formulas, the extended information retrieval function, improve information retrieval effectiveness. BRIEF DESCRIPTION

Figure 1 is a flow chart of information retrieval aspect of the invention;

FIG 2 a schematic diagram of a visual interface of the embodiment of the present invention;

3 visual interface schematic diagram of a second embodiment of the present invention;

FIG 4 is a schematic configuration information retrieval system according to an embodiment of the present invention. detailed description

DETAILED DESCRIPTION Hereinafter, embodiments of the present invention are described in detail in conjunction with the accompanying drawings. However, embodiments of the present invention is not limited thereto.

Aspect of the present invention is primarily provided by the correspondence between the visual interface and information retrieval formula, so that the user can build the lines and only a visual interface between the feature node, can according to the aforementioned correspondence relationship information compiled retrieval formula, complete information retrieval.

Figure 1 is a flow chart of information retrieval aspect of the invention. As shown in FIG. 1, the information retrieval process comprising the steps of:

Step 101, the correspondence relationship set in advance. One of the correspondence relationship between the feature node is connected to the node and wherein the visual interface, and the other of the correspondence relationship between the information retrieval operation is a logical formula and the search unit searching unit.

The visual interface, comprising several independent feature node, wherein the nodes are connected by these lines.

Wherein the feature node includes the role of logical nodes and node identification pattern two types of nodes, the node identification pattern is information retrieval is a retrieval unit of the formula, i.e. in retrieval word (keyword) in the traditional information retrieval formulas, it is a logical node information retrieval formula represents a logical operation between the respective retrieval unit, comprising a logical node, or, XOR, and other non nodes represent different logical operations, different logical nodes using different icons represent, for example, "and" logical node may the use of "X",

"Or" logical node a "+", "NOT" logical node using "a", and the like.

This visual interface, the overall structure is a tree, a root node or to spread (and logical node) that represents the root node is the ultimate goal of information retrieval. A logical node corresponds to one or more logical nodes and / or identifying a pattern or a plurality of nodes, these nodes identify patterns and / or logical nodes with lines pointing to its own logical node involvement.

Step 102, visual interface constructed in accordance with the search target information, the connection between the feature node and wherein the node comprises a visual interface.

When the user starts information retrieval, confirm their first search target information, and then based on the search target information, select the desired mode of logical nodes and node identification, and then using these feature node lines are connected together to complete the information Construction retrieve visual interface.

In order to expand and improve the functionality and efficiency of information retrieval, the technical solutions of the present invention, the content and form of logical nodes and node identification pattern have been expanded to include the following a few examples of the visual interface power ports to be described.

Example a user wants to retrieve "Siemens in 1995, the production of refrigerators or ice reject the model," which is the target user's information retrieval, information retrieval how to build this target corresponding visual interface yet.

Firstly, the target information retrieval, pattern recognition can identify nodes need to retrieve, including "Siemens", "1995 ,,," refrigerator "," Ice reject "and" model ", in order to improve the recall rate, taking into account" Siemens "is a German company, which is formerly known as" SIEMENS ", the" SIEMENS "can be used as identification mode node.

Second, a logical operation is to find the relationship between these nodes recognition mode, between "Siemens" and "the SIEMENS" yes "or" relationship, "refrigerator" and "ice repellent" between "or" relation "" Siemens "or" SIEMENS "", "1995 years", "" refrigerator "or" ice reject "", "model" among these four is "aND". It is necessary "or" and "and" logical node.

Then, by lines (such as lines with arrows) and these identifying patterns logical nodes connected nodes.

As shown, the visual interface of the tree structure is 2, so the build up process is the opposite of the above analysis procedure. First, from the logical node provided in advance, select "and" nodes, the root node 2.

In the root node ( "and" node) below the floor, there are four results of participation "and" operation, namely, "" Siemens "or" SIEMENS "", "1995", "" refrigerator "or" ice reject " "," model ,, wherein "" Siemens "or" the SIEMENS "" is between "Siemens" and "the SIEMENS" "or" operation, it is necessary to select the first "or" node 21, at first, "or" node below, is connected to the node identification pattern "Siemens" node 211 and a recognition mode "SIEMENS" 212, as the second layer; "" refrigerator "or" ice repellent "" is between "refrigerator" and "ice repellent" " or "operation, it is necessary to select the second" or "node 22, a second" or "beneath node identification pattern is connected to the node" refrigerator "node 221 and a recognition mode" ice repellent "222, also the second layer. Further recognition mode, two nodes "1995" 23 and "model" 24 belonging to the first layer, directly connected to the root node.

Then again the lines with arrows, respective logic operations involved in recognizing patterns or logical node corresponding to the node pointing to logical nodes, thus forming a visual interface for information retrieval.

In order to improve the precision of information retrieval precision, or can also increase the recognition mode node, such as text, image, audio or video. If the user in the search "Siemens refrigerator manufactured in 1995 or refuse ice models" have been some need to find a refrigerator or ice-repellent images, you can use these images as recognition mode node will identify these images first point mode node "or" logical node, then the newly added second logical node point "or" logical node 22, so you can find the "Siemens in 1995 with the production of these same types of imaging devices." , even if the target file does not appear in the "refrigerator" or "ice reject" vocabulary.

Meanwhile, in the logic operation relationship, you can also add other logical operation, such as a user when retrieving "Siemens in 1995, the production of refrigerators or ice reject the model" in the hope to retrieve only Siemens's own production of refrigerators or ice-repellent, to exclude Siemens OEM or ODM other companies refused to file a refrigerator or ice, you can add "not" logical nodes, by adding "OEM" and "ODM" recognition mode two nodes and two nodes pointing recognition mode "or" logic node, "or" logical node point "not" logical node, and finally the root node, so that you can be Siemens refrigerator to those from other companies OEM or ODM or ice refused to exclude files in the target file, improve the precision of search results information, reducing the workload of the user to read.

The second embodiment, the user wants to retrieve "produced by Siemens refrigerator or ice refused models, and these models refrigerator or ice-repellent, sorted by year" for the target user's information retrieval, information retrieval how to build this target corresponding visual interface yet.

Firstly, the target information retrieval, pattern recognition can identify nodes need to retrieve, including

"Siemens", "SIEMENS", "refrigerator", "Ice reject" and "model" to identify these patterns nodes, but how the year reflect it, because if a document retrieved, there is no information of the year for which the information retrieval target is useless, and what now is not limited in the year, while not bringing all year by "or" operation into the visual interface.

Then you need to use the data type instead of a specific data pattern to identify the node. For the above-described information retrieval target no specific year, year data types may be increased as a node model identification, data substituted for the specific year.

Further, for the above-described information retrieval target, some of the file if there is no "model" word, while data models, it may not be detected, the model may be increased as a data type node model identification, in place of the specific data model .

On this basis, the step of constructing a visual interface with Example A similar embodiment, the final result shown in FIG. In the ( "and" node) below the root layer 3, there are four results participation "and" operation, namely, "" Siemens "or" the SIEMENS "", year data type, "" refrigerator "or" ice repellent " "" "model" or model data types. "

Wherein "" Siemens "or" the SIEMENS "" is between "Siemens" and "the SIEMENS" "or" operation, it is necessary to select the third "or" node 31, under the third "or" node, is connected to the identification mode node "Siemens" node 311 and a recognition mode "SIEMENS" 312, as the second layer; "" refrigerator "or" ice repellent "" is between "refrigerator" and "ice-repellent", "oR" operation is necessary to select the first four "or" node 32, under the fourth "or" node, is connected to the node identification pattern "refrigerator" node 321 and a recognition mode "ice repellent" 322, also the second layer.

Year data type identification pattern 33 as a node, connected directly to the root node, as a first layer. To the "year" word recognition mode node formed distinguished using ## define together, showing that there is a year as long as the type of data, such as 1995, 2000 and so on, even if the subject.

"" Type "data type, or model" is "" in between model and data type "model or" logical operation, it is necessary to select the fifth "or" node 34, the fifth "or below" node is connected recognition mode node "model" 341 and a node identification pattern model data types 342, as a second layer. Similarly node model identification pattern ## is defined datatypes up, as long as it represents a model type of data is present, even if the check, not necessarily a "model ,, the word. Such data identifying the type of pattern nodes, according to user We need to be set from different angles.

Data identifying the type of pattern nodes can be text type, an image type, and other types of audio or video type. For example, in Example II, users need to check files embodiment, the product must contain an image, otherwise unnecessary, the image data may be increased as a type of pattern recognition node and the root node points, so that the detection of the target file must be there are images, those files without images will not be detected.

Identifying the type of node data pattern may be contact type, named entity type, type of sentence, sentence type, a type of emotion or language type. Second embodiment In the example embodiment, the user need files are checked to evaluate the product, the type of emotion recognition mode may be increased node, the detection of such a target file, must emotional words or meaning.

These data identifying the type of node to its literal mode distinguish requires using a preset label or icon representation, of course, indicates the method can be set according to their needs.

Step 103, the visual connection between the feature node and the node interface feature, step 101 in accordance with a preset correspondence relationship, compiled information retrieval formula.

Example embodiment of a visual interface embodiment, the correspondence relation, the following information can be compiled into the retrieval formula = ((Siemens or SIEMENS) and in 1995 and (refrigerator or ice-repellent) and type).

In this information retrieval formula, identifying patterns nodes visual interface of Siemens, SIEMENS, 1995, refrigerator, ice refused, models have become the retrieval unit or search terms, logical node "and", "or" have become a logical operation character, and the line becomes directed brackets, which indicates which logical operation involved in retrieval unit.

The visual interface according to a second embodiment, according to the correspondence relationship, the following information can be compiled into the retrieval formula = ((Siemens or SIEMENS) and # and # year (freezer or ice-repellent) and (Model Model # or #).

Step 104, the information retrieval formula is compiled directly operation, information retrieval formula thus obtained, to retrieve.

It should be noted that the pattern recognition logic nodes and nodes may also include other ways.

To achieve the above process, embodiments of the present invention also provide an information retrieval system, as shown in FIG. 4, the information retrieval system includes a display device 401, input device 402, a database 403, and retrieval apparatus 404 compiling means 405. Wherein the display means displays a visual interface, visual interface constructed in accordance with the search target information, the connection between the feature node and includes a feature node.

Input means connected between the feature node and wherein the node in the visual interface.

Database storing correspondence relationship, one of the correspondence relationship between the feature node is connected to the node and wherein the visual interface, and the other of the correspondence relationship between the information retrieval operation is a logical formula and the search unit searching unit.

Compile means connected between the visual feature node and wherein the node interface, based on the correspondence relationship, compiled information retrieval formula.

Retrieval means for retrieving information according to a formula, for retrieval.

Since the compilation step compiling the finished device according to the preset corresponding relation to convert the user to build visual interface to information retrieval formula, this step for the user, there is no need of concern, the user can visually according to their information retrieval target build complex formulas of retrieval, the user only reduces the difficulty of constructing search of formula, but also improves the efficiency of information retrieval, and data type as the introduction of a retrieval unit, greatly expanding the information retrieval function. The above-described preferred embodiment of the present invention embodiment, but the embodiment of the present invention is not limited to the above embodiments, changes made to any other without departing from the spirit and principle of the present invention, modifications, substitutions, combinations, tube of, replacement pattern of equivalent effect, are included within the scope of the present invention.

Claims

Claims
1. An information retrieval method, characterized by comprising the steps of:
Setting a correspondence relationship, one of the correspondence relationship between the feature node is connected to the node and wherein the visual interface, the other of the correspondence between the information retrieval operation is a logical formula and the search unit searching unit;
Visual interface constructed in accordance with the search target information, the connection between the feature node and wherein the node comprises a visual interface;
The connection between the feature node and wherein the node in the visual interface, according to the correspondence, compiled information retrieval formula;
The information retrieval formula for retrieving.
2. An information retrieval method according to claim 1, wherein said feature node includes logical nodes and node model identification, logical nodes for indicating logical operations, pattern recognition is used to represent a node searching unit, a least a logical node corresponds to logical nodes and / or less than a recognition mode node.
3, a method of retrieving information according to claim 2, characterized in that the pattern recognition node and / or logical nodes with lines pointing to its own logical node involvement.
4. The method of claim 2 or the information retrieved according to claim 3, wherein said node comprises a logic AND, OR, XOR, and non nodes, different logical nodes different icon.
5. The method of claim 2 or the information retrieved according to claim 3, wherein said node is a text recognition mode, images, audio and / or video data.
6. The method of claim 2 or the information retrieved according to claim 3, wherein said identification pattern is a data type node.
7. The method of claim 6, said information retrieval claims, characterized in that the data type is text, image, audio and / or video.
8. The method of retrieving information according to claim 6, wherein said data type is a contact, named entity, sentence, sentence, or emotional language.
9. The method of retrieving information according to claim 6, wherein the node is a pattern for identifying the data type, identified by adding a label or using a preset icons.
10. An information retrieval system, characterized in that the device, an input device, a database, compiling means and retrieval means, including a display,
Display means for displaying a visual interface, said visual interface constructed in accordance with the search target information, the connection between the feature node and includes a feature node;
Input means for inputting a connection between the feature node and wherein the node in the visual interface; a database for storing a correspondence relationship, one of the correspondence relationship between the feature node is connected to the node and wherein the visual interface, said corresponding side of the relationship is a logical operation between the information retrieval unit retrieves formulas and retrieval unit;
Compiling means for connecting between the feature node and wherein the node in the visual interface, according to the correspondence, compiled information retrieval formula;
Retrieval means for retrieving information according to a formula, for retrieval.
PCT/CN2013/077118 2013-06-10 2013-06-10 Method and system for information retrieval WO2014198025A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2013/077118 WO2014198025A1 (en) 2013-06-10 2013-06-10 Method and system for information retrieval

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2013/077118 WO2014198025A1 (en) 2013-06-10 2013-06-10 Method and system for information retrieval

Publications (1)

Publication Number Publication Date
WO2014198025A1 true WO2014198025A1 (en) 2014-12-18

Family

ID=52021555

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/077118 WO2014198025A1 (en) 2013-06-10 2013-06-10 Method and system for information retrieval

Country Status (1)

Country Link
WO (1) WO2014198025A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1808428A (en) * 2005-01-22 2006-07-26 鸿富锦精密工业(深圳)有限公司 Information searching criteria presentation and editing system and method
CN1851696A (en) * 2005-10-26 2006-10-25 华为技术有限公司 Correlation inquiry system and its method
CN1904884A (en) * 2005-07-29 2007-01-31 株式会社理光 Graph inquiring structuring apparatus for isomerization media and method thereof
CN101458697A (en) * 2007-12-11 2009-06-17 国际商业机器公司 Supporting creation of search expressions employing a plurality of words

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1808428A (en) * 2005-01-22 2006-07-26 鸿富锦精密工业(深圳)有限公司 Information searching criteria presentation and editing system and method
CN1904884A (en) * 2005-07-29 2007-01-31 株式会社理光 Graph inquiring structuring apparatus for isomerization media and method thereof
CN1851696A (en) * 2005-10-26 2006-10-25 华为技术有限公司 Correlation inquiry system and its method
CN101458697A (en) * 2007-12-11 2009-06-17 国际商业机器公司 Supporting creation of search expressions employing a plurality of words

Similar Documents

Publication Publication Date Title
US8510327B2 (en) Method and process for semantic or faceted search over unstructured and annotated data
CN104854583B (en) Search results ranking and presentation
EP1672537B1 (en) Data semanticizer
US20020178184A1 (en) Software system for biological storytelling
US9396262B2 (en) System and method for enhancing search relevancy using semantic keys
Jänicke et al. On Close and Distant Reading in Digital Humanities: A Survey and Future Challenges.
CN101479728A (en) Visual and multi-dimensional search
US20160162476A1 (en) Methods and systems for modeling complex taxonomies with natural language understanding
Soibelman et al. Management and analysis of unstructured construction data types
Giunchiglia et al. A large dataset for the evaluation of ontology matching
CN103927354A (en) Interactive searching and recommending method and device
US9613317B2 (en) Justifying passage machine learning for question and answer systems
CN103309797B (en) The user interface method and apparatus for automatic testing
CN103425714A (en) Query method and system
Hyvönen et al. CultureSampo–Finnish culture on the semantic web 2.0. Thematic perspectives for the end-user
Deng et al. Managing UI pattern collections
US20160110446A1 (en) Method for disambiguated features in unstructured text
US20100174704A1 (en) Searching method and system
JP6286549B2 (en) Recommended result exhibition method and apparatus
US8185563B2 (en) Data-visualization system and method
CN103605706B (en) A resource Retrieval Based on Knowledge Map
Ferrandez et al. The QALL-ME framework: A specifiable-domain multilingual question answering architecture
Hao et al. An end-to-end model for question answering over knowledge base with cross-attention combining global knowledge
McKenzie et al. A weighted multi-attribute method for matching user-generated points of interest
CN102541975B (en) Analysis of the structure of an object, such as benefits and provider contracts and the like

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13886851

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13886851

Country of ref document: EP

Kind code of ref document: A1