KR100834066B1 - Patent searching method and patent search system using query automatically containing hierarchically subordinate patent classifications with or operator - Google Patents

Abstract

A method and a system for providing a search service by automatically including a patent class code hierarchically subordinated and using an OR operator are provided to search patents by including a lower IPC(International Patent Classification) code hierarchically subordinated to an IPC code when the IPC code is included in a query form. A hierarchical IPC code database(130) stores information for parent/child relation between the IPC codes having a hierarchical structure. A patent information analyzing engine(140) processes the IPC codes by including a lower IPC code processing module(144) and a modified query generating module(146). The lower IPC code processing module obtains the information for a lower IPC code of the received IPC code when the IPC code is included in the received query. The modified query generating module modifies the received query to include the information for the lower IPC code information searched by the lower IPC codes processing module in a search result.

Description

[Patent Searching Method and Patent Search System Using Query Automatically Containing hierarchically Subordinate Patent Classifications With Or Operator}

1 is a diagram of an embodiment of a patent search service providing system of the present invention.

FIG. 2 is a diagram illustrating a flow of providing a patent search service for automatically including and searching a hierarchically dependent patent classification code according to the present invention.

3 is a diagram of another exemplary embodiment of a patent search service providing system according to the present invention.

4 is an exemplary diagram of a method for generating a hierarchical modified patent classification code database of the present invention.

5 is a diagram of one embodiment of a method for creating a tree structure from USPC patent classification codes.

FIG. 6 is a diagram illustrating an exemplary method of assigning a modified patent classification code corresponding to a tree structure with the USPC patent classification code of FIG. 5.

FIG. 7 is an exemplary diagram of a tree structure composed of modified patent classification symbols having the same structure as that of the USPC patent classification symbol of FIG. 5.

8 shows Index to U.S. Patent Classification (a.k.a., Classification Index File) file) is an exemplary view showing that a tree structure of a patent classification code as shown in FIG.

9 is an exemplary diagram of a system of the present invention for performing a search using a DBMS.

10 is yet another embodiment of the system of the present invention for performing a search using a DBMS for a modified patent classification code of the present invention.

Republic of Korea Patent Application No. 20000010784

Republic of Korea Patent Application No.20010024151

Republic of Korea Patent Application No. 20040081211

Republic of Korea Patent Application No.20020040459

Republic of Korea Patent Application No. 20030070373

The present invention relates to a patent search service providing system that automatically includes a hierarchically dependent patent classification code using a union operator, and a method of analyzing patent information using the patent search service providing system. In detail, it uses a patent search service providing system and a patent search service providing system that automatically search for and include a patent classification code that shares the lowest classification code notation in patent classification code notation but is hierarchically subordinated in content. It relates to a method for analyzing patent information.

All patent documents are assigned at least one IPC (International Patent Classification) symbol representing a technology included in the patent document as one symbol. The IPC is made in accordance with the Strasbourq Agreement on International Patent Classification and is assigned at least one IPC symbol to all patent documents. IPC notation has a classification notation system that is hierarchically represented by sections, classes, subclasses, groups, and subgroups.

In addition, the patent offices of some countries have their own unique patent classification system, such as USPC used by the US Patent Office, ECLA used by the European Patent Office, FI and FT used by the Japanese Patent Office.

USPC uses class and subclass notation, and FI in Japan has a structure that extends IPC vertically, and FT is the major theme of the theme and Fterm notation belonging to the theme's infrastructure. There is.

Patent classification code systems correspond to each classification code in a one-to-one correspondence section describing the contents of the corresponding patent classification code "title". For example, IPC H01S 3/00 corresponds to the title "Laser", and H01S 3/14 corresponds to the title "characteristic of the substance used as the active medium".

The hierarchical patent classification code and title system as described above share a common feature of dot structure. Using IPC as an example, the patent classification symbols corresponding to all subgroups have a "subclass" + "number / number" structure, and a title corresponding to this patent classification symbol is combined with dots. Mostly. (Except numbers / 00, most dots are combined, and usually numbers / 00 are symbols corresponding to groups.) The fewer dots combined in a title, the higher the title compared to its surroundings. It is a technical classification of concepts, and the larger the number of dots, the more the title is a technical classification of lower / detailed concepts when compared with its surroundings.

The following is an example of a hierarchical patent classification code system including a dot structure.

Section: H Electric

Class: H01 Basic electric element

Subclass: H01F Magnet

Main group: H01F 1/00 Magnet or magnetic material characterized by magnetic material

1-dot subgroup: 1/01 ● Inorganic materials

2-dot subgroups: 1/03 ●● Characterized by the coercivity

3-dot subgroup: 1/032 ●●● Of hard magnetic material

4-dot subgroup: 1/04 ●●●● Metal or alloy

5-dot subgroup: 1/047 ●●●●● Alloy characterized by composition

6-dot subgroup: 1/053 ●●●●●● Including rare earth metals

Even in a subgroup having the same patent classification code representation format as in the above example, it can be seen that there is a hierarchical structure of the upper and lower levels in terms of the contents of the technology.

However, since the classification code expression format is the same at the subgroup level, the following problem occurs when performing a patent search using the patent classification code.

First, when searching by a patent classification code, the search engine does not reflect the hierarchical structure, but searches for only documents that match the patent classification code inputted as a search word. In the example above, if you enter H01F 1/032, the search results that match the intent of the searcher are not only H01F 1/032, but also H01F 1 conceptually, such as H01F 1/04, H01F 1/047, H01F 1/053 It would be correct to see all documents that have been assigned a patent classification code corresponding to all subclasses included in / 032.

Second, even when a search result through a patent classification code is used for analysis, the hierarchical structure of the patent classification code is not reflected below a subgroup, and thus an analysis result including a logical error may occur. That is, when analyzing the application amount, the document belonging to H01F 1/032 should always be greater than or equal to the document belonging to H01F 1/04, regardless of the notation of the patent classification code appearing in the patent document. That is, a patent document to which H01F 1/04 is granted belongs to H01F 1/032, and since this document also belongs to H01F 1/04, which is one of the subordinates of H01F 1/032, the document contains H01F 1/04. H01F 1/04 is given instead of 032.

Therefore, if a patent classification symbol is included in a search expression inputted by a search word inputter including the dot structure information of a title corresponding to a patent classification symbol in a patent search or a patent analysis, the lower patent classification symbol corresponding to the patent classification symbol is automatically generated. It is necessary to provide a search result reflecting the lower patent classification code.

SUMMARY OF THE INVENTION The first technical problem to be achieved by the present invention is a method and system for providing a patent retrieval service for including a lower patent classification symbol that is hierarchically dependent on the patent classification symbol when a patent classification symbol is included in a search expression. To provide.

In order to achieve the technical problem to be achieved by the present invention, the step of receiving a search request that the patent classification code included in the predetermined patent classification code system included in the search formula; Obtaining a lower patent classification code of a patent classification code included in the search expression on the patent classification code system; Generating a modified search expression by reflecting the obtained lower patent classification code; And performing a search through a search engine in the modified search formula. Wherein the generating of the modified search expression comprises: enclosing all patent classification symbols below the patent classification symbol with an OR operator. It provides a way to provide a search service by including symbols automatically.

The receiving of the lower patent classification code may include: querying a hierarchical patent classification code database generated by reflecting a hierarchical structure included in title information that is a description corresponding to a patent classification code on the patent classification code system; The hierarchical structure data of the patent classification code system generated by reflecting the hierarchical structure included in title information which is a description corresponding to the patent classification code on the patent classification code system is generated, and then using a depth-first search technique. It is preferably any one selected from among methods of obtaining a list of patent classification symbols corresponding to the lower patent classification.

The patent classification code system is one of an international patent classification (IPC), a US patent classification (USPC), a Japanese file index (FI), a Japanese F-term (FT), and a European patent classification (ECLA). desirable.

The patent classification code included in the search formula is a dot level patent classification code, and the dot level patent classification code is title information which is a description corresponding to the patent classification code on the patent classification code system. It is preferable that the dot contains a dot.

When the patent classification code system is an international patent classification (IPC), the dot classification patent classification code has a dot in title information, which is a description corresponding to the patent classification code, among sub-group patent classification codes. In the case of the US Patent Classification (USPC), a dot is included in title information, which is a description corresponding to a patent classification symbol, among subclass classification or lower patent classification symbols, and a Japanese file index. In the case of (FI), a dot is included in title information, which is a description corresponding to a patent classification code, among the patent classification codes below a subgroup, and a theme in the case of a Japanese F-term (FT). The dot information is included in the title information, which is a description corresponding to the patent classification code, among the patent classification codes below the code. In the case of the European Patent Classification (ECLA), It is preferable that a dot is included in title information which is description corresponding to a patent classification code among the patent classification codes below a subgroup.

In order to achieve the technical problem to be achieved by the present invention, at least one or more patent classification code in a system for providing a patent search service that receives a search expression from a user, searches for an index constructed for patent information and provides a search result A hierarchical patent classification code database containing information on parent-child relationships between patent classification codes having a hierarchical structure in the system; A patent classification code processing engine that processes the patent classification code; The patent classification code processing engine includes a patent classification code when the patent classification code is included in the search expression, and when the patent classification code includes a dot in title information corresponding to the patent classification code, the inputted patent. A lower hierarchical patent classification code processing module for obtaining information on a lower patent classification code of a classification code; And a modified search expression generation module for modifying the input search expression to include information on the lower patent classification code found by the lower layer patent classification code processing module in the search result. It provides a patent search service providing system including a hierarchically dependent lower patent classification code using a union operator characterized in that it comprises a.

delete

When the lower layer patent classification code processing module includes a dot in title information which is a description corresponding to the input patent classification code, the method of receiving the lower patent classification code is a patent classification on the patent classification code system. A method of querying a hierarchical patent classification code database generated by reflecting a hierarchical structure included in title information, which is a description corresponding to a symbol, and a title corresponding to a patent classification code on the patent classification code system. The hierarchical data of the patent classification code system generated by reflecting the hierarchical structure included in the information is generated, and then a depth first search method is used to obtain a list of patent classification codes corresponding to the lower patent classification. It is preferable to be one.

Preferably, the modified search expression generation module modifies the entered search expression to include all the patent classification symbols obtained by the lower layer patent classification code processing module and the input patent classification code by using an OR operator. .

The patent classification code system is one of an international patent classification (IPC), a US patent classification (USPC), a Japanese file index (FI), a Japanese F-term (FT), and a European patent classification (ECLA). desirable.

It will be described below in more detail with reference to the drawings.

1 is an exemplary diagram of a patent search service providing system 100 providing a search service including a hierarchically dependent lower patent classification code of the present invention. The patent search service providing system 100 of the present invention allows access of a computer of a user connected to a network through a wired / wireless network 400, receives a search expression from the user, and includes a patent classification code in the search expression. If so, the patent classification code is processed by the patent classification code processing engine 140 to generate a modified search expression, and the search engine 110 queries the search index 120 for the modified search expression. Produces Of course, there is a patent information database that can provide patent information corresponding to the patent search results in the system.

The search engine 110 of the present invention indexes the content information of the acquired patent document through a search indexer, and stores it in the form of a search index 120. At this time, the contents of the patent document may be indexed for each field. In addition, the search engine 110 extracts a searched patent document, and the patent search service providing system obtains all patent information related to a bibliography, a text, and / or drawings of the searched patent document from the patent information database 150. Can be transferred to the user computer (100).

Conventional patent documents include bibliographic information, textual information and drawing information including text of patent technology, and in some cases, administrative processing information on the patent document may be further included. The bibliographic information includes at least one patent classification code.

The patent classification code information and the title information of the above-described patent classification information correspond to 1: 1, and in the case of IPC, the upper and lower hierarchical information, which cannot be represented by the patent classification code, is represented by the number of dots in the subgroup or less. . In the case of the IPC, it is difficult to know the upper and lower relations only by the patent classification code information below the subgroup, so that the upper and lower hierarchical information is grasped by using the dots included in the title information corresponding to the patent classification code information. That is, below the subgroup, by identifying the dot structure included in the title, the upper and lower hierarchical structures between the patent classification symbols below the subgroup are identified. The upper and lower hierarchies between the patent classification codes have the form of a tree structure.

In the present invention, the tree structure between the patent classification symbols will be described by taking the IPC 7th edition standard main group H04B 7/00 (the title is a wireless transmission system) and the patent classification symbols corresponding to the lower part thereof. The method and description are not limited to this embodiment, but are applied equally or equally over the entire IPC area, as well as other patent categorization schemes including the upper and lower levels of the hierarchy as a dot structure in the title (eg, For example, USPC, FI, FT, or ECLA) may be applied in the same or equivalent manner.

As of May 5, 2006, the patent classification code corresponding to the main group H04B 7/00 and the lower part thereof and the title are as follows.

H04B 7/00 Radio transmission systems, ie using radiated electromagnetic fields

H04B 7/005 Control of transmissions; Equalization

H04B 7/01 ● Reduction of phase shift

H04B 7/015 Reduction of echo effects

H04B 7/02 Diversity System

H04B 7/04 Using a plurality of independent aerial lines arranged apart

H04B 7/06 ●●● Thing in sending country

H04B 7/08 ●●● Thing in recipient country

H04B 7/10 Using single airborne systems characterized by polarization or directional characteristics

H04B 7/12 Frequency Diversity System

H04B 7/14 Relay system

H04B 7 / 145Manual relay system

H04B 7/15 Active relay system

H04B 7/155 Ground-based stations

H04B 7/165 ●●●● Using angle modulation

H04B 7/17 ●●●● Using pulse modulation

H04B 7/185 Space-based or airborne stations

H04B 7/19 ●●●● Earthsynchrenons stations

H04B 7/195 ●●●● Asynchronous Station

H04B 7/204 multiple access

H04B 7/208 ●●●● Frequency division multiple access

H04B 7/212 ●●●● Time Division Multiple Access

H04B 7/216 ●●●● Code Division or Spread-Spectrum Multiple Access

H04B 7 / 22Scattered wave propagation system

H04B 7/24 For communication between two or more points

H04B 7/26 At least one point can be moved

If the tree hierarchical structure is made easier based on the dot, it can be expressed as shown in Table 1 below.

Main group 1-dot subgroup 2-dot subgroup 3-dot subgroup 4-dot subgroup title H04B 7/00 Wireless transmission system H04B 7/005 Control of transmission; Equalization H04B 7/01 Reduction of phase shift H04B 7/015 Reduction of echo effect H04B 7/02 Diversity System H04B 7/04 Using a plurality of independent aerial lines arranged apart H04B 7/06 ●●● Thing in sending country H04B 7/08 ●●● Thing in recipient country H04B 7/10 Using single airborne systems characterized by polarization or directional characteristics H04B 7/12 Frequency Diversity System H04B 7/14 Relay system H04B 7/145 Manual relay system H04B 7/15 Active relay system H04B 7/155 Groundbased stations H04B 7/165 ●●●● Using angle modulation H04B 7/17 ●●●● Using pulse modulation H04B 7/185 Space Agency or Aircraft Loading Station H04B 7/19 ●●●● Earth Drive Station H04B 7/195 ●●●● Asynchronous Station H04B 7/204 ●●● Multi-access H04B 7/208 ●●●● Frequency Division Multiple Access H04B 7/212 ●●●● Time Division Multiple Access H04B 7/216 Code Division or Spread-Spectrum Multiple Access H04B 7/22 Scattering wave propagation system H04B 7/24 For communication between two or more points H04B 7/26 At least one point can be moved

When a patent document indexes its contents, it is natural to index each patent office or patent document processing institution (especially EPO and WIPO).

The search engine 110 of the present invention receives a search expression from a subject using the search engine 110 to perform a search and generates a search result.

The patent classification code processing engine 140 of the present invention largely includes a detailed module for processing the following three tasks.

First, a hierarchical patent classification code database generation module 142, wherein the module targets patent classification information data including patent classification code information and title information containing description contents corresponding to the patent classification code information. A hierarchical patent classification code database 130 including a tree structure between patent classification codes is generated. The hierarchical patent classification code database 130 may be an example of the hierarchical patent classification code database 130 because it is generated by reflecting a hierarchical structure between patent classification codes.

The hierarchical patent classification code database generation module 142 has another patent except for the highest patent classification code (eg, section display patent classification code in the case of IPC and class display patent classification code in the case of USPC). Corresponding to the classification code, the upper patent classification code to which the patent classification code belongs (1: 1) or the lower patent classification code corresponding to the lower part of the classification code to the one patent classification code. (1: n) Perform the function you set. While both are computationally feasible, the electronic method of creating a parent 1: 1 relationship corresponding to one patent classification child is more reasonable. All of the patent classification codes establish a relationship with each other in the tree structure by the above correspondence.

The former is a method of mapping a parent to a child, and the latter is a method of mapping a parent to children belonging to the parent. Both can be used to find the children belonging to the parent. The first is to find the child pointing to the parent when the parent is given. You can use it. Therefore, the former saves the child-parent relationship 1: 1 based on the child, and the latter directly stores the immediate children belonging to the parent.

The method of generating the hierarchical patent classification code database 130 by the hierarchical patent classification code database 130 will now be described based on IPC classification.

1) From the section to the main group, the hierarchical patent classification code database 130 is created according to the classification code system, and its parent information is stored for itself, its own upper and lower patent classification codes, or its own lower level. Save child information.

2) For the main group patent classification code, for all patent classification codes before the next main group patent classification code is found, the patent classification code for which the title corresponding to the patent classification code has one dot is found. The main group patent classification code is usually easy to find because the slash ("/") is followed by 00. In the case of child-> parent (the way the child points to its only parent), each patent classification symbol in the 1-dot subgroup with one dot stores the patent classification symbol of the main group which is its only immediate parent. . In the case of parent-> child (a method in which all children under the child are stored), the patent classification symbol of the 1-dot subgroup having one dot belonging to the main group is associated with the patent classification symbol representing the main group. Save it.

At this time, the main group itself, its own subordinate patent classification code is stored its parent information, or its own child information.

3) For a patent classification code having one dot, a patent classification code having two dots in the title is searched for all the patent classification symbols belonging to the same main group and before the patent classification symbol having one dot. If there is no other patent classification symbol in the title that has one dot in the title (the main group has one subgroup, or the last subgroup in the subgroup), the title has one dot. A patent classification code having two dots in the title is found among all the patent classification symbols before the patent classification symbol and the patent classification symbol indicating the next main group. The way to save its parent and child is the same as in 2) above.

4) With respect to the patent classification code having two dots in the title, a patent classification code having three dots in the title is found by the same logic as in the above 3), and processed in the same manner as in the above 3).

Another hierarchical patent classification code database generation module 142 suggests a method of generating the hierarchical patent classification code database 130 as follows.

You can do the following by looking up its parent, as the IPC moves up from the bottom of the table in order (record number): 1) The number of dots in the title information corresponding to the bottom patent classification code is checked, and the number of dots in the title of the patent classification code in the record number (just above IPC) of its record number -1 is checked. If the number of dots is one, it is its own parent. If the number of dots is the same, it is at the same level as itself. If a patent classification code of a subgroup or more is encountered, the patent classification code of the sub group or more is stored as its parent.

2) If the number of dots in the title of the patent classification code in the record number of the own record number -1 is equal to or greater than itself or the patent classification code is more than the subgroup, the parent relationship is ignored, and the self record number -1 2) is performed on the patent classification code of the record number.

In the above two embodiments, the method of generating the hierarchical patent classification code database 130 has been presented. However, this is only an embodiment. Generating own child information would be easy for those skilled in the art of developing a computing system.

The hierarchical patent classification code database generation module 142 generates a hierarchical patent classification code database 130 when there is a given patent classification code. An implementation thereof is as follows.

For the patent classification code below H04B7 / 00, an embodiment of the implemented results of both methods is presented.

First, Table 2 shows the results according to a method of matching the parent patent (child-> parent) which is the immediate classification code to which the classification code belongs.

Child Parent (his parent) H04B 7/005 7/00 H04B 7/01 7/00 H04B 7/015 7/00 H04B 7/02 7/00 H04B 7/04 H04B 7/02 H04B 7/06 H04B 7/04 H04B 7/08 H04B 7/04 H04B 7/10 H04B 7/02 H04B 7/12 H04B 7/02 H04B 7/14 7/00 H04B 7/145 H04B 7/14 H04B 7/15 H04B 7/14 H04B 7/155 H04B 7/15 H04B 7/165 H04B 7/155 H04B 7/17 H04B 7/155 H04B 7/185 H04B 7/15 H04B 7/19 H04B 7/185 H04B 7/195 H04B 7/185 H04B 7/204 H04B 7/15 H04B 7/208 H04B 7/204 H04B 7/212 H04B 7/204 H04B 7/216 H04B 7/204 H04B 7/22 7/00 H04B 7/24 7/00 H04B 7/26 H04B 7/24

Next, as shown in Table 3, one embodiment of the result of the method (parent-> child) that has all the child information about its own underneath is presented.

Parent Child (own child) H04B 7/00 H04B 7/005 H04B 7/01 H04B 7/015 H04B 7/02 H04B 7/14 H04B 7/22 H04B 7/24 H04B 7/005 null H04B 7/01 null H04B 7/015 null H04B 7/02 H04B 7/04 H04B 7/10 H04B 7/12 H04B 7/04 H04B 7/06 H04B 7/08 H04B 7/06 null H04B 7/08 null H04B 7/10 null H04B 7/12 null (no value) H04B 7/14 H04B 7/145 H04B 7/15 H04B 7/145 null H04B 7/15 H04B 7/155 H04B 7/185 H04B 7/204 H04B 7/155 H04B 7/165 H04B 7/17 H04B 7/165 null H04B 7/17 null H04B 7/185 H04B 7/19 H04B 7/195 H04B 7/19 null H04B 7/195 null H04B 7/204 H04B 7/208 H04B 7/212 H04B 7/216 H04B 7/208 null H04B 7/212 null H04B 7/216 null H04B 7/22 null H04B 7/24 H04B 7/26 H04B 7/26 null

Second, in the lower layer patent classification code processing module 144, a patent classification code obtained through a search expression input by a user is subordinate to the obtained patent classification code through the hierarchical patent classification code database 130. It finds the lower layer patent classification code.

In order to include the patent document corresponding to the lower layer patent classification code for the acquired patent classification code in the search results, one or more of the following three methods are required, and this is performed by the lower layer patent classification code processing module ( 144).

First, the hierarchical patent classification code database 130 finds all the lower hierarchy patent classification codes of the acquired patent classification code on the basis of the acquisition time point. The patent classification symbols on the hierarchical patent classification code database 130 are found in a tree structure. If the received search expression includes a specific patent classification code, it is possible to find all patent classification codes of the lower layer of the patent classification code along the tree structure, and transmit the result to the modified search expression generation module. , The modified search expression generation module may generate a new search expression reflecting all received patent classification codes and query the search engine 110.

Second, the lower layer patent classification symbols are found, matched (stored in a matching table) for all patent classification symbols in advance, and if there is a specific patent classification symbol in the search expression obtained, the lower level matching the specific patent classification symbol. Hierarchical patent classification symbols may be immediately sent to the modified search expression generation module.

There may be two broad ranges of the lower layer patent classification code extracted by the lower layer patent classification code extraction module. One is a method of extracting only the patent classification code immediately below the obtained patent classification code, and the other is a method of extracting all lower patent classification codes that depend on the obtained patent classification code. At this time, even when extracting only the subordinate patent classification symbols, the subordinate patent classification symbols are extracted again for each of the extracted subordinate patent classification symbols, and the method is subordinated to the patent classification symbols obtained by repeating such a method. All subpatent classification symbols can be extracted.

For example, when the patent classification code included in the search expression is H04B 7/15, the subordinate patent classification codes are H04B 7/155, H04B 7/185, and H04B 7/204, and all subordinate patent classifications. Symbols H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7/195, H04B 7/204, H04B 7/208, H04B 7/212, H04B 7 / 216.

It will be apparent to those skilled in the art that the above-described direct patent classification code and all lower patent classification codes can be easily extracted directly or recursively from the child-> parent correspondence or from the parent-> child correspondence. Of course it will.

The lower layer patent classification code extracting module combines and stores the extracted lower layer patent classification code. When the patent classification code included in the above search expression is H04B 7/15, when only the lower patent classification code is stored, " H04B 7/15, H04B 7/155, H04B 7/185, and H04B 7/204 " And all lower patent classification symbols are stored together with the "H04B 7/15, H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7 / 195, H04B 7/204, H04B 7/208, H04B 7/212, and H04B 7/216 "are stored together.

Third, the modified search expression generation module generates the modified search expression by integrating the associated patent classification code with the original existing search expression.

In the existing search formula including the patent classification code, at the position where the patent classification code was located, the patent classification code and the extracted lower layer patent classification code are grouped in an OR relationship, and the bundled patent classification code is used. It can be done in a manner that replaces the patent classification symbol place of the original existing search expression. In this case, if two or more patent classification codes are included in the original existing search expression, it is natural that the above patent classification codes may be integrated for each patent classification code.

For example, if the given search expressions are "{keyword = wireless and active} and {applicant = Samsung Electronics} and {IPC = H04B 7/15}, the modified search expression generation module may be configured as IPC H04B 7 / in the given search expression. In search of 15, obtain and coalesce the subordinate or all sub-patent classification symbols, "{Keyword = Wireless and Active} and {Applicant = Samsung Electronics} and {IPC = H04B 7/15 OR H04B 7/155 OR H04B 7 / 185 OR H04B 7/204} or "{Keyword = Wireless and Active} and {Applicant = Samsung Electronics} and {IPC = H04B 7/15 OR H04B 7/155 OR H04B 7/165 OR H04B 7/17 OR H04B 7 / 185 OR H04B 7/19 OR H04B 7/195 OR H04B 7/204 OR H04B 7/208 OR H04B 7/212 OR H04B 7/216} Generate a modified search expression that incorporates the classification symbols. The latter search is good.

Summarizing the role played by the lower layer patent classification code processing module 144, when the patent classification code is included in the obtained search expression, information related to the lower layer patent classification code of the patent classification code is found and a modified search expression is generated. Allow the module to handle it. Of course, if the patent classification code of the search expression obtained can be processed as a truncator, such as the case of more than the main group in the case of IPC, the lower layer patent classification code processing module 144 does not need to find the lower patent classification code. (For example, if only H04B7 / 00 or H04B is obtained, it can be processed as H04B7 / * or H04B *.) Therefore, the lower layer patent classification code processing module 144 is subgroup or less in the case of IPC. It is more useful when there is a dot in the title information, and it can also be useful when determining the position in the classification system by the number of dots in other classification system.)

2 sequentially shows the role played by the patent search service providing system 100 of the present invention as described above. The patent search service providing system 100 receives a search expression including a patent classification code from a user's computer (S110). At this time, the patent classification code included in the obtained search expression includes a dot in title information. In the case, the lower patent classification code is obtained on the patent classification code system through the patent classification code processing engine 140 (S120), the obtained patent classification code is bundled with OR (S130), and the lower patent classification bound with OR is obtained. A modified search expression including a symbol is generated (S140), and a search is performed by the modified search expression (S150). Each step is as described above.

If there is an insertion, deletion, or relocation in the patent taxonomy, the insertion can be treated similarly to the concept of adding a new word at a specific place in the dictionary, and the deletion can be handled by removing the word. In the case of relocation, it can be treated as a combination of insertion and deletion. Of course, whenever there is an insertion, deletion or relocation, the entire database can be updated or recreated for the entire changed patent taxonomy.

In the above, the description has been focused on generating a modified search expression by enclosing the lower patent classification codes in OR.

Patent classification symbols on the patent classification code system may need to be arranged in alphabetical order. In this case, the modified patent classification code arranged in alphabetical order must correspond to the original patent classification code system in a 1: 1 manner. In this case, the above-described method of generating a modified search expression by grouping the lower patent classification codes in OR may be performed on the modified patent classification code system. The patent classification system of the modified patent classification code is called a modified patent classification code system.

3 is another embodiment of the patent retrieval service providing system 100 according to the present invention. The symbol on the patent classification code system may be used for at least one patent classification code system selected from among IPC, USPC, FI, FT, and ECLA. It is characterized in that the modified patent classification code is utilized without using as is.

The modified patent classification code processing engine 160 of the system of the present embodiment generates the modified patent classification code system, processes the lower layer patent classification code by using the modified patent classification code system, and modifies the search expression obtained. Creates a modified search expression. The modified patent classification code processing engine 160 includes a hierarchical modified patent classification code database generation module 162, a modified patent classification code utilization lower layer patent classification code processing module 164, and a modified search expression generation module.

The method of generating a modified patent classification code may be as follows. The IPCs such as H04B 7/005, H04B 7/01, and H04B 7/185 may be modified as H04B70000050, H04B70000100, H04B70000185 and the like, respectively. Modification rules may vary, but based on the present embodiment, a 4-digit number is allocated to a group and a 4-digit number is allocated to a sub group. (Of course, it is preferable to accept the classification system related to the section and class which is higher classification system than the group as it is.) Of course, when there is an extension symbol under the sub group like FI, the corresponding digit is allocated. Modified input patent classification code can be generated.

Therefore, in the case of having a direct input patent classification code system that can be transformed without major modification such as IPC, the matching table can generate a modified patent classification code system even with a very simple rule.

On the other hand, even if it does not have a direct input patent classification code system like the USPC, it is possible to generate a matching table. A method of generating a matching table for the USPC will be described later. The hierarchical modified patent classification code database generation module 162 is responsible for generating the modified patent classification code system as described above. The modified patent classification code system generated by the hierarchical modified patent classification code database generation module 162 is stored in the hierarchical modified patent classification code database 135.

The modified patent classification code processing engine 160 correspondingly assigns the classification symbols on the modified patent classification code system 1: 1 by one to the classification symbols on the given classification system, and assigns the modified patent classification code system to which the indexer is assigned. To build a search index 120. When the user inputs a search expression into the given patent classification code system, the modified patent classification code utilization lower layer patent classification code processing module 164 finds a modified patent classification code corresponding to the input patent classification code, Modified patent classification code utilization modified search expression generation module modifies the search expression through the modified patent classification code, and performs a search through the modified search expression to derive a search result. It is not necessary that the modified patent classification code appear in the search results.

Next, the USPC will be described in more detail. USPC, USPTO's patented technology classification system, has difficulty in grasping up and down relationships and correlations between subclasses using only subclass numbers without a dot structure. In other words, unlike the IPC, the USPC is arranged irrespective of the alphabetical order and requires more systematic modification. USPC or USPC Patent Classification System (http: //www.uspto.gov/web/patents/classification / ... or Classification Index to US Patent Classification (aka) published by the US www.uspto.gov as of June 2006. , The Classification Index File) file can be viewed as a variant input. In this case, the visually understood USPC code is not in alphabetical order as IPC, so modifications are necessary.

The following two methods can be modified.

First, a tree structure is created based on title information including a dot structure for the entire USPC. The tree structure should be arranged such that the USPC forms nodes of the tree structure, reflecting parent-child relationships between USPCs. A depth-first search of the USPC tree structure is followed by a modified patent classification code including an alphabetical ordering for each node. Of course, it would be natural to match the USPC of each node with the modified patent classification code corresponding to the USPC and store it in the modified patent classification database.

Second, the Index to U.S. Based on the Patent Classification file (this file contains information about the USPC, description, and number of dots for each title (depth information based on root) and its parent.) After creating a tree structure of the entire USPC, the modified patent classification symbols are assigned to each node in alphabetical order using a search technique such as a depth-first search technique, and the modified patent classification symbols for each title are matched and stored. Release.

The above two examples are just an embodiment, and another method of using the title information including the dot structure of the USPC to correspond to the modified patent classification code in each individual USPC, but arrange the system of the modified patent classification code in alphabetical order. There may be many depending on the variant algorithm. One such example is presented in the above table of IPCs reflecting title information, and in the form of modified patent classification symbols arranged in alphabetical order through the table. Therefore, not only it is impossible to describe all such modification methods, but such modifications will be apparent to those skilled in the art such as computerized experts. It will be obvious that the modified patent classification code system modified by the above-described algorithm is included in the present invention.

Hereinafter, it demonstrates in detail through the following example.

The following is for the first part of subclasses of US class 002 apparel. This is a copy from http://www.uspto.gov/web/patents/classification/uspc002/sched002.htm. Class numbers 002 are omitted before the numbers below, but all numbers should be considered to have 002 before them.

1 MISCELLANEOUS

455 GUARD OR PROTECTOR

456Body cover

457Hazardous material body cover

458Thermal body cover

2.11Astronaut`s body cover

2.12 Having relatively rotatable coaxial coupling component

2.13 ●●● Having convoluted component

2.14Aviator`s body cover

2.15 Underwater diver`s body cover

2.16Having an insulation layer

2.17 ●●● Having a garment closure

459Shoulder protector

It can be seen from the dot structure that there are 457, 458, 2.11, 2.14, 2.15 directly below the 456 body cover, and 2.12 and 2.13 are directly below the 2.11 body cover. It is obvious that the above table can be made by reflecting the dot structure, and since the above description is provided in the description of the IPC, the description of this USPC will be omitted.

5 shows a tree structure of the USPC reflecting the dot structure. FIG. 6 shows that the modified patent classification code, which may be arranged in alphabetical order, may correspond to each node of the USPC tree in FIG. 5. FIG. 6 shows that the modified patent classification code corresponding to the "=>" mark can be assigned to the USPC on each tree structure. In FIG. 6, the information corresponding to "002" corresponding to the class display in the modified patent classification code is omitted. The display corresponding to "002" corresponding to the class display may be in various forms such as "002 /", "0020", "USPC + 002", and "002". FIG. 7 shows that the modified patent classification code may correspond to each of the USPCs in 1: 1. FIG. 7 shows that "0020" is employed among "002 /", "0020", "USPC + 002", and "002" as a display corresponding to "002" which is the class display. That is, when the processing shown in FIGS. 6 and 7 is performed, the modified patent classification symbols arranged in alphabetical order to each USPC can be corresponded to each other in a 1: 1 manner. It has a tree structure. FIG. 8 shows that the tree structure shown in FIG. 5 can be completed with the modified patent classification code. That is, it can be seen that the tree structure of FIG. 5 and the tree structure of FIG. 8 have a completely corresponding structure. As described above, each of the USPCs can be stored in correspondence with the modified patent classification code. An example of such storage may be as shown in the following table. Table 4 below may be an embodiment of the modified patent classification database of the present invention.

Subclass before transformation Dots Modified patent classification code (including class number) title One 002001000 MISCELLANEOUS 455 0 002002000 GUARD OR PROTECTOR 456 One 002002100 Body cover 457 2 002002110 Hazardous material body cover 458 2 002002120 Thermal body cover 2.11 2 002002130 Astronaut`s body cover 2.12 3 002002131 Having relatively rotatable coaxial coupling component 2.13 3 002002132 Having convoluted component 2.14 2 002002140 Aviator`s body cover 2.15 2 002002150 Underwater diver`s body cover 2.16 3 002002151 Having an insulation layer 2.17 3 002002152 Having a garment closure 459 One 002002200 Shoulder protector

And the Index to U.S. The Patent Classification file is expressed as follows.

002000000 0 1 APPAREL

002001000 1 2 000 000 MISCELLANEOUS

002455000 1 3 000 000 GUARD OR PROTECTOR

002456000 2 4 45 5000 Body cover

002457000 3 5 45 6000 Hazardous material body cover

002458000 3 6 456000 Thermal body cover

002002110 3 7 456000 Astronaut`s body cover

002002120 4 8 002110 Having relatively rotatable coaxial coupling component

002002130 4 9 002110 Having convoluted component

002002140 3 10 456000 Aviator`s body cover

002002150 3 11 456000 Underwater diver`s body cover

002002160 4 12 002 150 Having an insulation layer

002002170 4 13 002 150 Having a garment closure

002459000 2 14 455000 Shoulder protector

Index to U.S. Patent Classification is described. The first three digits of the first nine digits are the class number, and the last six digits represent the subclass number. On the Internet, there is a decimal point or a point like 2.11, but it can be seen that it is represented by 002110. The next digit represents the relative depth from the root. Apparel of class 002 is 0 depth, MISCELLANEOUS without dot represents 1depth directly below apparel, and thermal body cover with 2 dots represents 3 depth. The column after depth is the serial number, and the next column is the number representing its parent. That is, 455000 immediately preceding the body cover, such as "455000 body cover", indicates that its parent has 455000 and that the GUARD OR PROTECTOR is its parent. This can be seen as a representation corresponding to the dot structure. For example, it can be seen that, in the "456000 Aviator's body cover", the parent of the Aviator's body cover is a body cover using 456000.

Since the USPC classification code is not arranged in alphabetical order, reprocessing is required to arrange the alphabetical order.

Index to U.S. It would be easy for a person skilled in the art to make a tree structure as shown in FIG. 5 by using the information about the parent and the relative depth information in each USPC in the Patent Classification file, and the tree structure of the modified patent classification code as shown in FIG. Creating a tree structure would be easy. That is, Index to U.S. It is easy to create a tree structure as shown in FIG. 5 with information provided by Patent Classification, and assign a modified patent classification code to the tree structure as shown in FIG. 6, and uses the modified patent classification code as shown in FIGS. Of course, it is possible to create a tree structure of the modified patent classification code of the form shown in FIG. As described above, each of the USPCs can be stored in correspondence with the modified patent classification code. An example of the storage may be as Table 5 below. The table below may be an embodiment of the modified patent classification database of the present invention.

Class + subclass before transformation depth Subclass of pre-modification parent Variant Patent Classification Symbol Subclass of variant patent taxonomy parent Title 002000000 0 002000000 APPAREL 002001000 One 000000 002001000 000000 MISCELLANEOUS 002455000 One 000000 002002000 000000 GUARD OR PROTECTOR 002456000 2 455000 002002100 002002000 Body cover 002457000 3 456000 002002110 002002100 Hazardous material body cover 002458000 3 456000 002002120 002002100 Thermal body cover 002002110 3 456000 002002130 002002100 Astronaut`s body cover 002002120 4 002110 002002131 002002130 Having relatively rotatable coaxial coupling component 002002130 4 002110 002002132 002002130 Having convoluted component 002002140 3 456000 002002140 002002100 Aviator`s body cover 002002150 3 456000 002002150 002002100 Underwater diver`s body cover 002002160 4 002150 002002151 002002150 Having an insulation layer 002002170 4 002150 002002152 002002150 Having a garment closure 002004590 2 455000 002002200 002002000 Shoulder protector

Of course, it is possible to correspond to the modified patent classification code having the alphabetical order in the USPC system including the given title information through an algorithm different from the above, which is stored as a matching table of the modified patent classification code corresponding to the USPC as described above. Of course it can be.

4 illustrates a flow for an embodiment in which the hierarchical modified patent classification code database generation module 162 generates the hierarchical modified patent classification code database 135.

The hierarchical modified patent classification code database generation module 162 obtains all patent classification codes on the patent classification code system (S310), arranges all the acquired patent classification codes in a tree structure with reference to dot information of a title ( In operation S320, the patent classification code system arranged in the tree structure is searched using a depth-first search method, and the modified patent classification code is corresponded to each patent classification code node (S330), and the corresponding patent classification code and the corresponding modified patent classification are classified. The relationship of symbols is stored in the hierarchical modified patent classification code database 135 (S340).

Of course, the search index 120 should include information on the modified patent classification code corresponding to each USPC. The modified patent classification code may be formed of only digits having arbitrary digits, but may be formed including numbers and letters. In this case, when there is something added in the USPC, it is possible to generate a modified patent classification code so as to have a proper arrangement position when arranging in alphabetical order without reflecting the dot structure. The method for generating such an alphabetical ordering position is the same known technique as having a legally ordered position in alphabetical order when a new word is added to the dictionary, and will be readily implemented by those skilled in the art. In addition, if more subdivisions are created in alphabetical order, you can continue to expand the backseat. For example, when a plurality of patent classification symbols occur in multiple stages at the bottom of 002 / 2.12, additional backseats such as 0020021310a and 0020021310b, which are lower than alphabetical order of modified patent classification symbols 002002131 corresponding to 002 / 2.12, are higher than 002002132, etc. As you expand, you will continue to be able to generate variant patent classification symbols.

The search index 120 for the search engine 110 preferably includes information in which the original patent classification code is processed and information in which the modified patent classification code is processed. If the user does not want a search result for the lower layer patent classification code, it is not necessary to use an index on which the modified patent classification code is processed.

Subsequently, the modified search expression generation module of the present invention generates the modified search expression by integrating with the original modified search expression with the obtained modified patent classification symbols.

The core of the patent classification code processing engine 140 of the present invention as described above, when there is a given patent classification code, automatically processes the lower patent classification in every child or immediate hierarchy of the patent classification code. It's a feature that allows it to be reflected in search results. This is also obvious for the modified patent classification code.

Hereinafter, the configuration of the patent search service providing system 100 of the present invention using the patent classification code processing engine 140 as described above and the patent information is searched using the system, and the patent information processing based on the retrieved patent information Give a way.

First, a patent information search method through the patent search service providing system 100 of the present invention is presented. The core of the patent information retrieval method of the present invention is to provide patent information including a patent classification code under the patent classification code when the patent classification code is included in the search expression of the patent information.

As a first embodiment of patent information retrieval, an information processing method when a search classification includes a patent classification code of less than the main group is presented.

When there is a patent classification code less than the main group (For a patent classification code less than the main group, it is usual that at least one or more dots are included in the title information. Therefore, the inventive idea of this embodiment includes a dot structure as well. Of course, the patent classification code may be obtained by the patent classification code processing engine 140, and the hierarchical patent classification code database 130 may be inquired to obtain the patent classification code. Find out information about all patent classification symbols below the patent classification symbol given in the search expression, and based on this, integrate the obtained patent classification symbol with the obtained patent classification symbol to form a modified search expression, and the patent with the modified search expression Perform a search. Performing a patent search by a search formula including a modified patent classification code as described above is referred to herein as a "search containing a modified patent classification code".

Of course, if the patent classification symbol included in the search expression is the main group, the / 00 of the main group is cut out of / 00 and the searcher is used to search for all patent classifications below the main group. For example, given IPC = H04B 7/00, conducting a search for all patent classification symbols beginning with H04B 7 by converting => IPC = H04B 7 /? The present invention is importantly utilized at the patent classification symbol level (normally, the patent classification symbol level having a dot structure) whose upper, lower, and equivalent levels are not known in the patent classification symbol itself. .

When the search result is output by the modified search expression, the display of the search result may be provided by clustering according to the searcher's selection. Clustering is a method of grouping search results and expressing them in group units, and when applied to the present invention, are as follows. When there is a clustering window that represents the clustered structure, or when a clustering level can be selected, the search results are displayed only in the search results of the dot subgroup and subgroups corresponding to the preset or preselected clustering level. When a lower or upper clustering level is selected (the dot level corresponds to the clustering level), a search result corresponding to the selected clustering level is expressed.

For example, if clustering at the 1-dot subgroup level is selected through the search result related to H04B 7/00, the clustering window corresponds to 7/005, H04B 7/01, H04B corresponding to the 1-dot subgroup. 7/015, H04B 7/02, H04B 7/14, H04B 7/22, H04B 7/24 appear in the form of a subdirectory folder of H04B 7/00, and the search results show all results corresponding to H04B 7/00 . At this time, when the searcher selects H04B 7/14, the subordinate H04B 7/145 and H04B 7/15 appear in the form of a sub-directory folder. In this case, the search result corresponds to H04B 7/14 and the lower patent classification code. Only search results are displayed. If the searcher selects H04B 7/15 again, the subordinates H04B 7/155, H04B 7/185, and H04B 7/204 are displayed, and the search result corresponds to H04B 7/15 and its lower patent classification code. Only the results are displayed.

At this time, for the convenience of the searcher, the number of patent classification symbols below and the number of all the sub-patent classification symbols, together with each of the clustering items (the folder name of the directory) in the clustering window, are displayed together or which Only one can be displayed. The number of patent documents corresponding to the patent classification code and all lower patent classification codes of the patent classification code may be displayed. At this time, both the number of applied documents and the number of registered documents may be displayed, or only the number of applied documents or one registered document may be displayed. For example, in the case where the patent classification code is H04B 7/15, the subordinate patent classification code is H04B 7/155, H04B 7/185, and H04B 7/204, and thus, "3" is displayed or all lower patent classification codes. H04B 7/155, H04B 7/165, H04B 7/17, H04B 7/185, H04B 7/19, H04B 7/195, H04B 7/204, H04B 7/208, H04B 7/212, H04B 7/216 Therefore, "10" can be displayed. It is advisable to give both indications because there may be less classifications directly underneath, but there may be an enormous size classification underneath them. This is especially true in emerging emerging technologies, or in high-tech areas where technology divergence rates or subdivisions are increasing rapidly. If only the number of sub-categories is displayed, it may be because there is a user who ignores the categorization because it cannot be grasped even though the technical field is important when the numerical value is small.

The patent classification code information may be arranged in a directory manner, and the user may systematically access the lower patent classification code of the specific patent classification code by browsing the user. Currently, www.delphion.com provides this service. However, the site provides information on the number of subordinate patent classification symbols, but 1) does not provide the total number of subordinate patent classification symbols of the user, and 2) the number of patent documents for all patent classification symbols. And 3) it does not provide a patent document that contains all direct sub-patent classification symbols that correspond to that number when clicked. The present invention can solve this problem through the hierarchical patent classification code database 130 and the modified patent classification code search method.

When the patent classification code information is arranged in a directory manner, a symbol such as a color or an icon may be displayed in each directory to improve accessibility of the user. In other words, if there are many sub-categories, all sub-categories, deep, deep patents, many patent documents, etc., different colors or preset icons are assigned to the user. The number of all subcategories, the depth of categorization, the number of patent documents, etc.). In particular, when the number of all sub-classifications of a specific patent classification code directory is large or the depth of classification is large, it is more important than the directory that is not because it has many patent applications to the area and proves that various technologies are applied. This is because it is often larger.

In the above, the method and system for automatically performing search by including the lower patent classification code for the search engine 110 have been described. However, the idea of the present invention may be applied not only to the case of using the search engine 110 but also to the case of searching using a DBMS. While search engines use search queries for search engines, DBMSs use DB queries for DBMSs against databases. Hereinafter, the case of using a DBMS will be described in detail. However, overlapping with the case of using the search engine is omitted.

Changing the DB query for a DBMS (which can be referred to as a search expression for a DBMS) makes no difference in terms of generating a modified search expression and applying the spirit of the present invention. That is, in the case of using a DBMS, the DBMS 170 may be used instead of the search engine 110 and the search index 120 in FIGS. 1 and 3. The configuration is shown in FIGS. 11 to 12.

  Although the present invention has been described in detail about querying a search engine with a modified search formula that automatically includes lower patent classification codes, the present invention provides a modified query (corrected search for DBMS) that automatically includes lower patent classification codes. The query is not essentially different from querying the DBMS 170. In other words, the present invention can be applied not only to search engines but also to DBMSs.

Assuming the search engine 110, the lower hierarchical patent classification code processing module 144, the hierarchical patent classification code database, the modified search expression generation module 146, and the modified patent classification code utilization subordinate of the present invention. The method of generating a modified search expression using the hierarchical patent classification code processing module 164 and the hierarchical modified patent classification code database 135 may be used to generate a modified query for a DBMS. That is, the patent classification code processing engine 140 and the modified patent classification code processing engine 160 of the present invention are not dependent on the search engine 110 or the DBMS 170, but the search engine 160 or the like. It can be seen that the pre-processing operation for processing the information of the DBMS (170). That is, the patent classification code processing engine 140 and the modified patent classification code processing engine 160 of the present invention perform a preprocessing operation independently of the search engine 110 or the DBMS 170, and the preprocessed result is Reflected in the hierarchical patent classification code database 130 and the hierarchical modified patent classification code database 135. The preprocessing result reflected in the hierarchical patent classification code database 130 and the hierarchical modified patent classification code database 135 is reflected in the search index 120 or the patent information database 150. Therefore, when the search engine 110 performs a search on the search index 120, the preprocessing result is displayed in the search index 120 and / or the patent information database 150. As reflected in, only the search expression queried by the user needs to be modified. Similarly, if only the search expression is queried when the user queries the patent information database 150 through the DBMS 170, the DBMS 170 outputs a search result corresponding to the modified search expression. Will be done.

When the DBMS 170 method is used as shown in FIGS. 11 to 12, the contents of the hierarchical patent classification code database 130 and the hierarchical modified patent classification code database 135 are included in the patent information database 150. It is desirable to have this physical or logical integration. Logically integrated means that even if the tables are separated, they can be combined logically through the commands supported by the DBMS. At this time, the supporting command may include a command to JOIN a table that can be logically related, this method is easy for those skilled in the art of handling a DBMS.

The modified search expression generation module 146 processes a patent classification code included in a search expression (DB query) obtained from the user to generate a search expression (DB query) to include a lower patent classification code of the patent classification code. Create That is, when querying a DBMS using the DBMS query, the patent classification code is included in the DBMS search expression included in the query, and the dot information is included in the title information of the patent classification code. It is possible to automatically obtain the lower patent classification code of the patent classification code of the present invention to generate a modified query for the DBMS. That is, the above-described method of finding the lower patent classification code and grouping the patent classification code and the lower patent classification code by OR may be applied as it is, or 2) the above-described method of the modified patent classification code may be applied as it is. Could be You can query the DBMS with the modified query just as you query the search engine with the modified search expression.

Therefore, the processes of FIGS. 2 and 4 are equally applied even when the DBMS 170 is used.

The patent information database 150 includes information on bibliographic matters, full texts, and drawings obtained from patent documents for each country. In particular, various types of bibliographic matters to be analyzed are arranged for each field. . In addition, the counted information may be arranged in the patent information database, and the counted information includes the number of claims, the number of drawings, the number of families, the number of inventors, the number of applicants, and the patent classification code. Examples are numbers. In addition, the counted information may further include the number of foreign applications, citing numbers, cited numbers, etc. related to the patent document.

If one or more criteria selected from a predetermined country, period, applicant, technical field (including patent classification code), a specific technical keyword, and an inventor are added to the DB query, more detailed search results can be obtained. Of course I will.

At this time, whether to use the method of using the first search engine or the method of using the second DBMS may be a problem in search efficiency. In general, when a DBMS is used, the DBMS supports full-text search but is slower than a search engine. Therefore, if a keyword includes a keyword search word, a patent document that satisfies the search expression through the search engine is used. It would be relatively efficient to know the number of.

Next, a patent analysis method based on a patent classification code using a modified patent classification including search method of the present invention will be described. The patent analysis method is performed by the patent information analysis engine 200 of the present invention. Analysis of all patent information performed by the patent information analysis engine of the present invention is intended for the following two cases.

The first is an analysis method based on the number of patent documents satisfying the obtained search expression when the search engine 110 queries the search engine 110 for a specific search expression including a patent classification code.

The second is an analysis method based on the number of patent documents obtained by querying the DBMS of the present invention.

The first and second analysis results may be derived in time series (including a specific time / period unit). In this case, the concept of speed may be generated by dividing the analysis result by time / period unit. Analysis involving this concept of speed is called speed analysis. Next, a speed analysis method will be described. An engine that performs the speed analysis is referred to as a speed analysis engine 210.

Speed analysis based on patent taxonomy is to obtain 1) numerical information and 2) subject information based on patent taxonomy on patent taxonomy supported by each country's patent information for a specific country or several countries. I say something.

The numerical information may include 1) quantity, 2) increase rate of quantity, 3) increase rate, and 4) increase acceleration in a predetermined period. Knowing only the quantity information in the preset period, the rate of increase of quantity is calculated as {quantity of the period-quantity of the previous period} / {quantity of the period before}} 100, and the speed is {quantity of the period-quantity of the period before } / {Calculated length of period}, the acceleration may be calculated as the rate of change of the speed. Therefore, for the sake of convenience of explanation, the specification will describe the speed analysis based on the quantity, but it is assumed that the above descriptions 2), 3) and 4) are naturally included. The quantity may be any quantity that can be derived from the patent information, for example, 1) the amount related to the simple counting number, such as application amount, registration amount, 2) ratio amount such as registration rate, share, concentration rate, 3) the amount of the index, such as the patent activity index, 4) the amount of subjects, such as the number of applicants, inventors, agents, etc., 5) the number of countries involved in a particular patent document, or a number of patent documents in one country. The number of administrative agencies, such as the number of possible agencies (primary judging bodies such as the Patent Office and the Patent Tribunal, second higher judging bodies such as the Patent Court, and the Supreme Court, etc.), 6) the number of claims, the number of pages of the specification, Life cycle of a patent application such as number of citation documents, number of citation documents, number of citation documents, number of citation documents, etc. For example, the number of events occurring during the period and the amount of the result of the function value, such as a combination of each of them or an importance evaluation function using each of them as an input variable.

Subject information means: 1) applicant or owner, 2) inventor, 3) a patent document based on 1) quantity, 2) rate of increase of quantity, 3) rate of increase, and 4) increase acceleration subject to the analysis. Information about the subject such as an agent.

As an example of a simple speed analysis, the number of applications or registrations can be calculated for each patent classification per year for the entire patent classification or for a specific hierarchy of patent classification (for example, H04B 7/00 group). Here, the IPC is described as an example, but other patent classification systems having a dot structure can achieve the desired purpose in a similar way. In addition, for the convenience of explanation, the target analysis target is described using the number of applications as an example, but the target of the other target analysis target can be achieved in a similar way.

First, the patent classification of a specific layer will be described. The entire patent classification will correspond to the entire technical area as the highest set of patent classifications of a particular hierarchy.

First, since the dot structure is obtained from a subgroup after receiving the specified information about the specific layer, when the obtained patent classification is subclass or higher, the patent classification of the subclass or higher is decomposed to the main group unit. For each of the decomposed main groups, one dot subgroup in the subgroup is found (the main group reference immediate hierarchy patent classification symbols), and from the one dot subgroup, the search containing the modified patent classification symbol of the present invention is performed. Should be performed. (Of course, since the search using the modified search formula of the present invention is possible even from the main group or above, the search method including the modified patent classification code of the present invention is naturally applied. Below the subgroups it is possible only by the method of the present invention.)

For example, if the obtained patent classification is H04B at the subclass level, the main group H04B 1/00, H04B 3/00, H04B 5/00, H04B 7/00, H04B 10/00, which are subclasses corresponding to H04B. Are divided into H04B 11/00, H04B 13/00, H04B 14/00, H04B 15/00, and H04B 17/00. Then, for each main group, find a single dot subgroup belonging to the main group again. From the 1-dot subgroup, a search including the lower layer patent classification code of the 1-dot subgroup can be performed only when the search method including the modified patent classification code unique to the present invention is applied.

If the main class is H04B 7/00, then for each of the seven 1-dot subgroups, which are the lower-level hierarchies of the main class, out of all 25 sub-layer patent classification symbols belonging to the main class, all 2 belonging to the sub-class A search containing the modified patent classification code of the present invention is performed on the dot subgroup, the 3-dot subgroup, and / or the 4-dot subgroup.

As described above, the patent classification code group to be analyzed may be several, and the group may be at a level (for example, every subclass unit), and two or more different levels (some are subclasses, some other classes). May be a main group, some of which are class units).

When the patent classification code group is expanded, the patent information analysis engine 200 may process the entire patent classification code system. That is, a search including the modified patent classification code of the present invention may be performed for every single patent classification code below the subgroup.

One example is to obtain IPCs of all subgroup units for H04B 7/00 and calculate the number of applications per year for each of them.

In this case, the search expressions obtained may not only include patent classification codes, but may further include keywords, applicant information, and the like. Therefore, the search expression can derive the search result through the " AND " operation of the keyword or the applicant information and the patent classification code. The search engine 110 may transmit a combined search expression to a query for all fields placed in the index by the search engine 110, and if the combined search expression has a patent classification code, the patent classification By applying the modified patent classification symbol search method, the modified search expression is transmitted to the query. The search engine 110 derives a search result based on the modified search formula. In general, the search engine 110 quickly outputs the number of search results corresponding to a search expression (comparing with providing a list of search results incurs a relatively large load on a server or a network). When counting for analysis, you can achieve what you want without spending a lot of time.

In the case where Samsung is inputted as an applicant, one example is to obtain IPCs of all subgroup units for H04B 7/00, and calculate the number of applications for each one by year.

It will be apparent that the patent information analysis engine 200 of the present invention is not capable of calculating only the number of applications per year. The patent information analysis engine 200 can easily find numerical information primarily. Such numerical information includes, in addition to basic numerical information such as the number of applications, registrations, and registration rates by year or by predetermined period, various numerical information that can be found by calculating directly or indirectly in the patent literature of that country. The various numerical information is 1) the search engine 110 can respond through direct or simple processing without special prior information processing by a combination of search expressions, 2) special prior information processing, or 3) other There may be things that can be answered via the patent information analysis engine 200.

The example of 1) corresponds to the information in most bibliographic details, and the number of applications or registrations by country is representative, and the registration rate can be calculated by calculating the number of registrations compared to the number of applications. However, the calculation of the registration rate is of the type 3) when it is determined whether an individual filed patent is registered, rather than the number of registrations / applications by year.

Representative example of the 3) may be the ranking information for each applicant subject to a specific search condition. In order to output the ranking information for each applicant, a list of search results is obtained, the list of applicants included in the result list is identified, the frequency of occurrence of a specific applicant is calculated, and the higher the frequency, the higher the ranking. Can be. In this case, the ranking calculation engine processes the ranking.

Representative examples of 2) may include the number of important patents. An important patent may have several criteria, but in one exemplary method, (1) a patent with many foreign applications, (2) a patent with many claims, (3) a patent with a large number of pages in a specification, and (4) a highly cited patent Assuming that patents and the like have higher importance, weights may be assigned to each of them, and the importance may be evaluated by calculating whether the patent document corresponds to the above item and calculating the sum of the weights. When these scores are calculated in advance for each reference for each patent document, the number of patent documents corresponding to the reference among the search results may also be grasped.

In this manner, the search engine 110 performs a search on the received search expression and returns a search result including the number of search results.

Hereinafter, a method of finding meaningful information from patent information based on the number of search results through a method called speed analysis is disclosed. Speed analysis refers to analyzing a trend in which a search object changes in a predetermined time unit for each patent classification code constituting the group in a group including a patent classification code. The purpose of the search is an index that is the purpose of the search, such as the number of applications, the number of registrations, and the like, and the changing trend may be, for example, the number, the change (speed) of the number, the speed (acceleration) of the number change, and the like.

If the purpose of the search is the trend of application by technology field (patent classification code), the interpretation of the trend may be changed according to the following example.

1) Although a specific technical field has a small number of applications, the application growth rate is high and the application acceleration is likely to be a pioneering new technology area, and the applicants and technologies applied to the area are monitored. It needs to be great. In particular, if there is the same pattern abroad or the same pattern in some specific countries, it is likely that the pioneering competition is in the world.

2) There is a small number of applications in a particular technical field, and the rate of increase of the application is relatively large. If the acceleration is negative, it is likely that the niche area is occupied or occupied by a few occupants.

3) If a certain technical field is low in application volume and the rate of application growth is negative, it may be far from market interest, or it may be a small market and a place where defensive application is already made to defend the technical area after important technology development is completed. This is big.

4) If a particular technology sector has a large volume of applications, a high rate of application growth, and a high rate of acceleration of application growth, it is likely that this is an area in which large companies are entering the market due to the large market size and high future profitability.

5) If a particular technical field has a large application volume and a relatively high application speed, and if the acceleration is negative, it may be a pattern when the market workers concentrate the application toward the applied technology rather than the right of the core technology.

6) If a specific technical field has a large application volume, but the application growth rate is low or negative, and the application increase acceleration is negative, the area for which the protection technology is applied for the right of the process technology or the improvement technology is applied. Likely to be.

Therefore, various useful information can be known by measuring the number, the change (speed), and the speed (acceleration) of the number change for all patent classification codes corresponding to all technical areas. Through this, a distribution curve or a distribution map according to the application speed for a specific technical field or an entire technical field can be made.

Similarly, a pattern-wise analysis such as 1) to 6) above is possible for the number of applicants or the number of higher applicants through secondary processing instead of the application amount.

If the number of applicants is the same as 1) above, that is, if the number of applicants is small, but the applicant growth rate is high, and the applicant growth acceleration is positive, several leading technology companies are leading investments. Due to the prospects, the level of interest of participants is gradually increasing, which may be an area of technology development that is likely to be bright and profitable.

If the number of applicants is the same as above 4), the area is likely to be a region with strong red ocean character, which is becoming increasingly competitive.

Similarly, if the above speed analysis is performed on the entire technical domain or a specific technical domain (specific patent classification code) based on the specific applicant, the following results can be expected.

If a specific applicant has a small number of applications in a specific technical area, but the application growth rate is high and the increase acceleration is positive (+), the applicant considers the area as a future new business and develops intensive technology. high portential.

If a specific applicant has a large standard deviation of the application amount in a specific technology area, it may mean that the technology development in that area is not effective. When technology development cruises toward a predetermined target value, a patent application also shows a steady increase pattern, and the standard deviation of the application amount tends to be small.

If a specific applicant has a large application volume in a specific technical area, a high application growth rate, and a small application acceleration, the technical area may be an area that is likely to have passed the high development period.

As such, when individual indicators such as application amount, registration amount, number of applicants, or two or more indicators are mixed, the individual indicators or mixed indicators can be reflected in the search expression (for example, (Applicant = Samsung E) and (IPC = H04B 7/15), meaningful results can be obtained when the search results are arranged on the time axis or output in a table or the like.

Through the above-mentioned speed analysis, 1) application trend information for each technical field, 2) application trend information for each specific technology field, 3) application trend information for each technical field of a specific applicant, and 4) application trend information for a specific technology field of a specific applicant In addition, it can be visualized in a graph or table, and furthermore, if the above 3) and 4) for two or more applicants are performed and compared to a table or a graph, a comparison table or a comparison graph, and further a comparison map (two-dimensional or 3D, one or more of the applicant's axis of comparison, the technical field axis, and the year axis, and the height axis indicates the quantity).

In addition, if there is an item that is equal to or higher than a predetermined criterion for the result of the speed analysis corresponding to the whole or a specific technical area (a technical area or a technical area of a specific applicant unit), it can be reported to the outside. One example of the predetermined criteria is 1) filing of 50 or more applications annually, 30% or more annual application growth, 10% or more acceleration of application growth, or 2) specific applicants and / or specific technical areas (specific patent classification symbols). It may be a reference value obtained for at least one of the application amount, application growth rate, application growth acceleration. Reporting takes the form of reporting, which includes: 1) criteria satisfaction data for specific applicants and / or specific technical areas that satisfy the criteria; Any one or more of the list of search result patents corresponding to the criteria satisfaction may be included.

Such speed analysis can also be performed through a patent database of two or more countries. If more than one patent database is involved, there are largely the following series.

The first series analyzes the speed of a specific technology area by country. This can be used to find out what technology areas are concentrated in each country. In particular, when associated with the top applicants with a high share in each country for each particular technology area, which company in a country is leading in that technology area, there may be applicants with a high share in two or more countries. In that case, it is likely that the applicant is leading that particular technical area globally.

To this end, the speed analysis engine 210 forms a country search expression by including the same patent classification code in a search expression for a country-specific patent database, and then queries the search engine 110 with the search results for each country. Perform a search with If there is a keyword or the like, it is natural to prepare a country search expression including the translation of the keyword into the corresponding language of each country.

The second series is a series that analyzes how many applications are filed in a particular second country to a particular second country. For this analysis, a search formula for a specific applicant's entire technical area or a specific technical area for a first country is created (in the case of the entire technical area, a patent classification symbol is not specified, and a specific technical area is searched for). The patent classification code corresponding to the technical area may be limited to the expression.) After performing a search, a first list of all or specific technical areas is generated, and the same search expression for the second country is applied. It would be natural to change a person's name to the language of a second country), perform a search and then generate a full or specific list of secondary technical areas, and either the secondary list and the secondary list in whole or Analysis of the second series can be accomplished by comparing by specific technical area.

Another method of achieving the analysis purpose of the second series is a method of utilizing priority information. After generating the primary list, it is checked whether the application number existing in the primary list is in the patent information of the second country. (This is usually possible when filing a foreign application in the first country and filing the first country's application number with the second country while claiming priority within the period of priority claim (one year from the best date in the business). As a result of the review, a list of patents with priority information in a second country can be compared in whole or in specific technical areas. This method is particularly advantageous when the applicant's representative name is not sufficient, or when the applicant files in the name of another subject in the second country.

The third series is a cross-search series with different patent classification codes. To do this, a concordance table is essential. Www.uspto.gov, run by the US Patent and Trademark Office, operates a cross-conversion service between the USPC and IPC, and publishes a concordance table. It also sells FT themes and concordance tables for IPC in relation to Japanese patent information. If there is a concordance table, perform a rate analysis on the USPC in the United States, and then, for a particular applicant and / or a specific technical area that meets a predetermined criterion, the USPC is referred to the concordance table and the corresponding IPC or other After the patent classification code is changed, the modified patent classification code is included in a search expression for a second country other than the United States to perform a search for speed analysis and analysis of the search result. (Of course, when there is an applicant or keyword, it includes translation of the applicant or keyword into a second language.)

Although the above examples have been described focusing on the application volume, a similar method can be applied to other analysis indicators. Other analytical indicators may include: 1) concentration, 2) activity index (AI), and 3) market share, and the number of registrations, registration rates, etc. corresponding to the number of applications may be an indicator. The inventor or agent may be an indicator. In the above, the degree of concentration is an indicator of how much a particular applicant is concentrating in which area, and "the number of all applications filed for a specific patent classification code / all applications * of the applicant" based on a specific year or a predetermined period of time for a specific country. 100 ". A patent activity index is defined as "{(the number of applications filed by a applicant for a particular patent classification code / all applications filed by that applicant) / (all applications for a specific patent classification code) on a specific year or predetermined period basis for a specific country. Number / all applications) * 100}.

On the other hand, the following matters should be considered when counting in the number of speed analysis and the like. What should be done when two or more 1) patent classification symbols, 2) applicants, 3) inventors, and 4) agents correspond to one patent application? This is the case where the policy problems are all individually counted, when they are equalized by the corresponding number, and in the third case. Individual counting is a method of weighting each applicant by 1 if the applicant is 2, and equality is a method of counting the weight by 0.5 for the application. In the case of equality, the number information may have a value with a decimal point.

On the other hand, when two or more patent classification symbols appear in view of the modified patent classification symbol including the core of the present invention, the following methods in addition to the above two methods may be considered. Two or more patent classification symbols in the patent document can be grouped based on the patent classification symbols included in the search expression. For example, when five patent classification codes correspond to a specific patent document, the five patent classification codes are assigned to the first group belonging to the patent classification code itself or lower layer patent classification code included in the search expression and the search. It may be divided into a second group to which a patent classification code belonging to a separate or independent node / layer belongs to a higher layer patent classification code of a patent classification code included in an expression. In this case, there may be a method of varying the weight of the first group and the second group according to the number or distribution ratio of the constituent patent classification symbols belonging to the group. For example, if a patent document corresponds to H04B 7/005, H04B 7/01, H04B 7/145, H04B 7/204, H04B 5/00, the current search expression (the search expression at the time of analysis, the search H04B 7/14 in the formula contains a patent classification code), H04B 7/14 is H04B 7/145 and H04B 7/204 in the lower hierarchical information. Patent documents containing 7/145, H04B 7/204 will be included in the results of the search, and the weight for the patent document can be given 2/5, not 0, for the search formula H04B 7/14 criteria. have. In such a case, if the conventional method other than the search method including the modified patent classification code includes H04B 7/14 in the search expression, the patent document is not included in the search result.

In the case of applying the search method including the modified patent classification code, it may be usefully used not only for the analysis but also for monitoring and watching services. The monitoring service is to register a specific search expression, and if there is new patent information corresponding to the search expression, transmit the new patent information to a person who receives the monitoring service. In the present invention, the monitoring is performed by the patent information monitoring engine 300. The patent information monitoring engine 300 registers a search expression to be monitored, and sets the search expression on the search engine 110 or DBMS 170 at a predetermined time unit (daily, weekly, monthly, quarterly, and / or annually). Inquiries into the monitoring service user of the facts, the documents, or the facts or the information on the documents. do. In this case, the notification method may be any one or more of a method of transmitting by e-mail or other message transmission method to a person who will receive the monitoring result, or notifying when the monitoring service user accesses the monitoring service.

In order to perform a monitoring service, 1) registering a search expression for monitoring; Since the purpose is to query the search engine 110 or the DBMS 170 for a predetermined time unit and to find a new document when querying, the time range (for example, a date including a search query based on a publication date). It is common to limit the past 6 days to the time range) to the search formula.) If there is a new patent document, it may be sent to the surveillance service recipient. In this case, when the patent classification code is included in the search expression, the monitoring service may be performed by adopting a search method including a modified patent classification code. In the monitoring service using the modified patent classification code containing retrieval method of the present invention, the patent classification code included in the search expression in the patent document although it exists logically and hierarchically under the patent classification code system included in the search expression. In the absence of a fault, faults excluded from monitoring can be prevented.

The effects of the present invention are as follows.

First, at a patent classification code level having at least one dot in title information describing a patent classification code, a hierarchical relationship may exist between patent classification codes having a patent classification code notation of the same type. In this case, the conventional search engine finds only patent documents that match the input patent classification code. However, when the present invention is applied, not only the patent documents matching the input patent classification code but the patent classification code system are used. When there is a lower patent classification code that is dependent on the patent classification code, a patent document matching the lower patent classification code may be found as a search result. Through this, it is possible to minimize the occurrence of patent documents included in the patent classification code logically input, but excluded because the patent classification code is different.

Second, the present invention can be applied to the case of a patent monitoring that performs a search using a predetermined search expression for a new patent document or an updated patent document and a patent analysis that statistically analyzes the searched results. Through this, although the patent documents included in the logically input patent classification code, even the patent documents excluded because the patent classification code is different will be included in the monitoring and analysis target. In addition, the accuracy of monitoring and the set of analysis targets are increased, and thus the precise monitoring and analysis target set can be accurately analyzed.

Third, by applying the spirit of the present invention, it is possible to accurately count the number of patent documents related to a specific patent classification code. Therefore, it is possible to accurately measure various physical quantities (speed, acceleration, increase rate, occupancy rate, concentration rate, AI, etc.) and statistics (average, standard deviation, etc.) related to the number of documents.

Claims (10)

Receiving, by the patent search service providing system, a search request in which a patent classification code included in a preset patent classification code system is included in a search expression; Receiving, by the patent search service providing system, a lower patent classification code of a patent classification code included in the search expression on the patent classification code system; Generating a modified search expression by reflecting the obtained lower patent classification code by the patent search service providing system; And Performing, by the patent search service providing system, a search through a search engine in the modified search formula; Including; Generating the modified search expression includes enclosing all patent classification symbols below the patent classification symbol with an OR operator, The obtaining of the lower patent classification code may include: querying a hierarchical patent classification code database generated by reflecting a hierarchical structure included in title information that is a description corresponding to a patent classification code on the patent classification code system; After generating the hierarchical data of the patent classification code system generated by reflecting the hierarchical structure included in the title information, which is a description corresponding to the patent classification code on the patent classification code system, the hierarchical structure of the depth classification search method is described. A search service by automatically including a hierarchically dependent lower patent classification symbol using a union operator, which is selected from a second method of obtaining a list of patent classification symbols corresponding to the lower patent classification using How to give. delete The method of claim 1, The patent classification code system may be one of an international patent classification (IPC), a US patent classification (USPC), a Japanese file index (FI), a Japanese F-term (FT), and a European patent classification (ECLA). A method of providing a search service by automatically including hierarchically dependent lower patent classification symbols using a characteristic union operator. The method of claim 3, wherein The patent classification code included in the search expression is a dot classification patent classification code, The dot classification patent classification code is hierarchically using a union operator, characterized in that a dot is included in title information, which is a description corresponding to a patent classification code on the patent classification code system. A method of providing a search service by automatically including dependent subpatent classification symbols. The method of claim 4, wherein The patent classification code of the dot level is the patent classification code system In the case of International Patent Classification (IPC), a dot is included in title information, which is a description corresponding to a patent classification code, among sub-group patent classification codes. In the US Patent Classification (USPC), a dot is included in title information, which is a description corresponding to a patent classification symbol, among subclass classification or lower patent classification symbols. In the case of the Japanese file index FI, a dot is included in title information, which is a description corresponding to a patent classification code, among sub-group patent classification codes. In the case of a Japanese F-term (FT), a dot is included in title information, which is a description corresponding to a patent classification code, among patent classification codes below a theme code. In the case of the European Patent Classification (ECLA), a dot operator is included in title information, which is a description corresponding to a patent classification code, among sub-group patent classification codes. A method of automatically providing a search service by automatically including hierarchically subordinate patent classification codes. In the patent search service providing system that receives a search expression from the user, searches the index built on the patent information and provides a search result, A hierarchical patent classification code database containing information on parent-child relationships between patent classification codes having a hierarchical structure on at least one patent classification code system; A patent classification code processing engine that processes the patent classification code; Including; When the patent classification code is included in the search expression, the patent classification code processing engine is a lower level of the input patent classification code when the patent classification code includes a dot in title information corresponding to the patent classification code. A lower layer patent classification code processing module for obtaining information on the patent classification code; And a modified search expression generation module for modifying the input search expression to include information on the lower patent classification code found by the lower layer patent classification code processing module in the search result. Including When the lower layer patent classification code processing module includes a dot in title information that is a description corresponding to the input patent classification code, a method of obtaining the lower patent classification code,  A hierarchy of the patent classification code system generated by reflecting a hierarchical structure included in title information which is a description corresponding to a patent classification code on the patent classification code system and a first method of querying the hierarchical patent classification code database. Hierarchically dependent using a union operator, characterized in that the structural data is generated and then selected from a second method of obtaining a list of patent classification symbols corresponding to a lower patent classification using a depth-first search technique. A patent search service providing system that automatically searches for lower patent classification codes. delete delete The method of claim 6, The modified search expression generation module may modify the input search expression to include all the patent classification symbols obtained by the lower layer patent classification code processing module and the input patent classification code by binding to an OR operator. A patent search service providing system for automatically including a hierarchically dependent lower patent classification code using a union operator. The method of claim 6, The patent classification code system may be one of an international patent classification (IPC), a US patent classification (USPC), a Japanese file index (FI), a Japanese F-term (FT), and a European patent classification (ECLA). A patent retrieval service providing system for automatically including hierarchically subordinate subordinate patent classification symbols using a union operator characterized by a feature.

Images