JP2010277200A - Object retrieval device, method for processing the same, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an object retrieval device which can, even when a plurality of words as the reference of retrieval are included in information as retrieval conditions, and words showing abstract language expressions showing geographical characteristics are related with some of the words, perform detailed retrieval based on information included in the retrieval conditions. <P>SOLUTION: Unique names included in document information for retrieval and abstract language expressions showing geographical characteristics are extracted, and on the basis of the dependency of the abstract language expressions on the unique names, one or more combinations of the unique names and the abstract language expressions having the dependency are specified, and a retrieval range on a map with respect to the unique names included in the combinations for each of the combinations is determined based on the abstract language expressions corresponding to the unique names, and the final retrieval range, whose retrieval object is the final unique name appearing in the tail end of the document information for retrieval, is determined based on the retrieval range corresponding to the unique name having no dependency and the retrieval range for each of the combinations, and the retrieval object is retrieved from the final retrieval range. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an object search apparatus, an object search method, and a program thereof for searching a search object from a database based on search document information.

  When a telephone company provides telephone number services to customers, or when searching for addresses specified by customers during the service provision process, a worker can use a search device that can search for telephone numbers and addresses. And that search work is done. In a search device that performs such a search, an input of a word or a document indicating a facility or a place that the person in charge has grasped over the telephone is received from the person in charge of the search, performs a search process, and the process result is obtained. Output. Techniques relating to such a search technique are disclosed in Patent Document 1 and Patent Document 2.

JP 2007-65206 A JP 2005-181125 A

  Here, the technique of the above-mentioned patent document 1 searches a facility based on a combination of position designation words that designate a positional relationship between a place name and a facility name stored in map data. In addition, the position designation word for designating the positional relationship is, for example, the words “near”, “main part of”, “in”, “along”, and is clear in the technique of Patent Document 1. Even if you do not remember the address, you can search if you remember that the facility was near a certain area or that it was along a certain road.

  Moreover, the technique of the above-mentioned patent document 2 is a technique for enabling a peripheral facility search with directivity in the search range and performing a peripheral facility search according to the user's intention. In this technology, the number of search results is relatively large compared to the case where the user wants to have directivity such as the west side or the east side by listing and displaying the specified genre search in all directions from the search base point. It solves the problem of narrowing down by scrolling operation.

However, in the technique of Patent Document 1, in the case of a search process using the sentence “Ramen shop near XX station”, what kind of facility is the reference facility (“XX station” in the above example) However, the search range is limited based on the position designation word “near”, which may cause a re-search operation in which the search range is changed.
Further, in the technique of Patent Document 1 and the technique of Patent Document 2, a word (“Funabashi”, “375”, “Convenience Store”, such as “factory behind convenience store along No. 375 of Funabashi”), etc. ”,“ Factory ”, etc.), and a search cannot be performed in the case where an abstract language expression (such as“ along ”,“ back ”, etc.) indicating geographical features relates to a plurality of the words.

  Therefore, according to the present invention, the search condition information (speech, text, etc.) includes a plurality of words serving as search criteria, and a word indicating an abstract language expression indicating geographical features is included in a plurality of the words. Even in such a case, an object of the present invention is to provide an object search device, an object search method, and a program thereof that can perform a detailed search using information included in the search conditions.

  In order to achieve the above object, the present invention provides document information analysis means for extracting a unique name that is a word serving as a reference for search and an abstract language expression indicating a geographical feature included in search document information, Based on the dependency relationship of the abstract language expression with respect to the unique name, one or more combinations of the unique name and the abstract language expression having the dependency relationship are specified, and each combination is included in the combination. A search range on the map for the unique name is determined based on an abstract language expression corresponding to the unique name, and based on the search range corresponding to the unrestricted unique name and the search range for each combination , A search range determining means for determining a final search range on the map with the last unique name appearing at the end of the search document information as a search target, and the final search From 囲内, as an object search device, characterized in that it comprises a search means for searching for the search target.

  Further, the present invention provides the above-described object search device, wherein the position / address storage unit stores address information corresponding to the position information of the search range, and the address / unique name storage unit stores a unique name corresponding to the address. The search means specifies address information from the position / address storage means based on the position information corresponding to the final search range, and from the address / unique name storage means based on the address information. The unique name corresponding to the search target is specified.

  Further, the present invention is the object search apparatus described above, wherein the search range determination means is the map indicated by the unique name when the abstract language expression is an expression indicating that it is close to the unique name. A circle whose radius is a distance corresponding to the unique name with the upper position as the center is determined as the search range on the map.

  Further, the present invention is the object search device described above, wherein, when the abstract language expression is an expression indicating that the unique language name is along the unique name, the search range determining means A range expanded vertically from the end of the line or the surface by a distance corresponding to the unique name on the basis of a space graphic is determined as the search range on the map.

  Further, the present invention is the object search device described above, wherein the search range determination means, when the abstract language expression is an expression indicating that it is in the proper name, the map indicated by the proper name The upper range is determined as the search range on the map.

  In the object search device according to the present invention, the search range determination unit may indicate whether the abstract language expression indicates a direction, direction, or side of the second unique name when viewed from the first unique name. If it is an expression, the first unique name is identified as a first search criterion, the second unique name is identified as a second criterion, and a vertical line from the first search criterion to the second search criterion is centered A range up to a predetermined left-right angle is determined as the search range on the map.

  Further, the present invention is the object search device described above, wherein the search range determining means, when the abstract language expression is an expression indicating the opposite of the unique name, the position on the map indicated by the unique name In a circle having a radius corresponding to the unique name as a center, the spatial figure is identified as a line, and the side on which the unique name is present with the line as a reference, The search range on the map is determined.

  Further, the present invention is the object search device described above, wherein the search range determination means, when the abstract language expression is an expression indicating the back of the unique name, the position on the map indicated by the unique name In the circle whose radius is the distance according to the unique name, the spatial figure is identified as a line, and the side where the unique name is not present with respect to the line, The search range on the map is determined.

  Further, the present invention is the object search device described above, wherein, when the abstract language expression is an expression indicating the next to the unique name, the search range determining means is a position on the map indicated by the unique name. The search range is determined within a circle having a radius corresponding to the unique name as a center, and the search means specifies a search object that is close to the unique name as a search target. It is characterized by.

  Further, the present invention is the object search device described above, wherein, when the abstract language expression is an expression indicating a sequence of the unique names, the search range determining means is a position on the map indicated by the unique name. In the circle with the radius corresponding to the unique name as the radius, the space figure is closest to the unique name, which is a line, and vertically up and down from the line by the distance corresponding to the unique name A rectangular range widened is determined as the search range on the map.

  Further, the present invention is the object search device described above, wherein the search range determining means is an expression indicating that the abstract language expression is higher than the position of the unique name, the unique name is In a circle centered on the position on the map and having a radius corresponding to the unique name, a position higher than the position of the unique name is stored in the data storing each position on the map. Based on this, the specified item is specified as a search range.

  Further, the present invention is the object search device described above, wherein the search range determination means is an expression indicating that the abstract language expression is a position lower than the position of the unique name, the unique name is In the circle whose center is the position on the map shown and whose radius is the distance according to the unique name, the one that is lower than the position of the unique name is stored in the data storing the respective positions on the map. Based on this, the specified item is specified as a search range.

  Further, the present invention is the object search device described above, wherein the search range determining means is a representation of the space graphic of the proper name when the abstract language expression is an end of the proper name. The search on the map for a circle centered at an edge and having a radius corresponding to the unique name or a range in which the distance corresponding to the unique name is extended vertically from the edge of the space graphic of the unique name It is characterized by determining the range.

  Further, the present invention is the object search apparatus described above, wherein the search range determining means is an expression indicating that the abstract language expression is a corner of the unique name, A circle whose center is a corner and whose radius is a distance corresponding to the unique name is determined as the search range on the map.

  Further, the present invention is the object search device described above, wherein the search range determination means, when the abstract language expression is an expression indicating between a third unique name and a fourth unique name, 3 is specified as a third search criterion, the fourth unique name is specified as a fourth search criterion, and an elliptical range having a predetermined flatness ratio connecting the third search criterion and the fourth search criterion is defined as The search range on the map is determined.

  Further, the present invention is the above-described object search device, wherein the position / address storage means stores past address information corresponding to position information in a search range, and the address / unique name storage means stores the past A unique name on the map in the past corresponding to the address is stored.

  Further, the present invention is a processing method of an object search device, wherein the document information analysis means of the object search device includes a unique name included in the search document information and an abstract language expression indicating a geographical feature. The search range determining means of the object search device extracts one combination of the unique name and the abstract language expression having the dependency relationship based on the dependency relationship of the abstract language expression with respect to the unique name. One or more are identified, and for each combination, the search range on the map for the unique name included in the combination is determined based on the abstract language expression corresponding to the unique name, and the unsupported unique name is supported. Based on the search range for each combination and the search range for each combination, the final search on the map is searched for the last unique name appearing at the end of the document information for search. Range determines the search unit of the object retrieval apparatus, from within said final search range, a processing method characterized by searching for the search target.

  In addition, the present invention provides a computer of an object search apparatus, a document information analysis unit for extracting a unique name included in search document information and an abstract language expression indicating a geographical feature, and the abstract language expression for the unique name. Based on the dependency relationship, one or more combinations of the unique name and the abstract language expression having the dependency relationship are specified, and for each combination, a search on the map for the unique name included in the combination is performed. A range is determined based on an abstract language expression corresponding to the unique name, and based on a search range corresponding to the unrestricted unique name and a search range for each combination, the end of the search document information A search range determining means for determining a final search range on the map for which the last unique name appearing in the search object is to be searched; Is a program for causing to function as a retrieval means, to find the target.

  According to the present invention, morphological analysis is performed on search document information generated from speech and text search conditions, and a unique name attribute is indicated based on a predetermined unique name or a unique name and an abstract language expression of its dependency destination. A search range corresponding to the spatial figure on the map is specified, and when multiple unique names are included in the search document information, a search is performed from the final search ranges indicated by the AND condition of each search range. A search is made for a match with the last unique name appearing at the end indicated by the document information. Therefore, the search document information generated from the search conditions for speech and text includes a plurality of words serving as search criteria, and a word indicating an abstract language expression indicating geographical features is included in a plurality of the words. Even in this case, a detailed search can be performed based on information included in the search conditions.

It is a block diagram which shows the structure of a target object search apparatus. It is a figure which shows the example of data of an abstract language expression table. It is a figure which shows the example of data of a specific name table. It is a figure which shows the example of data of a search reference | standard space figure specific table. It is a figure which shows the example of data of an individual setting table. It is a figure which shows the example of data of a position and address corresponding | compatible table. It is a figure which shows the example of data of a search reference position storage table. It is a figure which shows a 1st morpheme table. It is a figure which shows a 2nd morpheme table. It is a figure which shows a 3rd morpheme table. It is a figure which shows the outline | summary of a search range determination process. It is a 1st figure which shows the process outline of a search range determination process and a search process. It is a 2nd figure which shows the process outline of a search range determination process and a search process. It is a 3rd figure which shows the process outline of a search range determination process and a search process. It is a 4th figure which shows the process outline of a search range determination process and a search process. It is a 5th figure which shows the process outline of a search range determination process and a search process. It is a 6th figure which shows the process outline of a search range determination process and a search process. It is a 7th figure which shows the process outline of a search range determination process and a search process. It is an 8th figure which shows the process outline of a search range determination process and a search process. It is a 9th figure which shows the process outline of a search range determination process and a search process. It is a 10th figure which shows the process outline of a search range determination process and a search process. It is an 11th figure which shows the process outline of a search range determination process and a search process. It is a 12th figure which shows the process outline of a search range determination process and a search process. It is a 13th figure which shows the process outline of a search range determination process and a search process. It is a 14th figure which shows the process outline of a search range determination process and a search process. It is a figure which shows the outline | summary of the process which determines whether it is contained in the last search range. It is a schematic diagram which shows two different processing orders of a target object search apparatus. It is a figure which shows the processing flow of a countermeasure search apparatus.

Hereinafter, an object search device according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of an object search apparatus according to the embodiment.
In this figure, reference numeral 1 denotes an object search device. The object search apparatus 1 includes a document information analysis unit 11, a search range determination unit 12, a search unit 13, an output unit 14, an abstract expression database 15, a unique name database 16, a search reference space graphic database 17, and an individual setting database 18. The processing unit and the storage unit of the list database 19 are provided. The document information analysis unit 11 includes a processing unit such as a geographic information extraction unit 111 and a dependency analysis unit 112, and the search range determination unit 12 includes a range narrowing unit 121 and a processing unit such as an abstract expression interpretation unit 122. Has been. Details of the processing of each processing unit and the contents of information recorded in each database will be described later.

In the present embodiment, the target object search apparatus 1 extracts a unique name included in search document information generated by character input, voice recognition, or the like, and an abstract language expression indicating a geographical feature. Then, the object search device 1 specifies one or a plurality of combinations of the unique name and the abstract language expression having the dependency relationship based on the dependency relationship of the abstract language expression with respect to the unique name, and for each combination. The search range on the map for the unique name included in the combination is determined based on the abstract language expression corresponding to the unique name, the search range corresponding to the unrestricted unique name, the search range for each combination, Based on the above, the final search range on the map is determined with the last unique name appearing at the end of the search document information as a search target. And the target object search apparatus 1 performs the process which searches a search target from the last search range.
Thereby, the object search device 1 according to the present embodiment includes a plurality of words serving as search criteria in the information (speech, text, etc.) serving as a search condition. Even when a word indicating an abstract language expression indicating a geographical feature is involved, a detailed search is performed using information included in the search conditions.

Next, data tables stored in each database will be described.
FIG. 2 is a diagram showing an example of data in the abstract language expression table stored in the abstract expression database.
As shown in this figure, the abstract language expression table is a table in which an abstract language expression indicating a geographical feature of a unique name and its ID are stored in association with each other. For example, in FIG. 2, the abstract language expressions “near” and “periphery” are ID1 of the pattern 1 section, and the abstract language expression “along” is ID2 of the pattern 1 section, “middle”, “ni”, “ In this example, the abstract language expression “is” is stored in association with ID3 of the pattern 1 section, and the abstract expressions “old” and “previous” are stored in association with ID4 of the pattern 1 section. Also, for example, the abstract language expression table shown in FIG. 2 stores the abstract language expression “back” in ID7 and the abstract language expression “inside, in” in association with ID4 of pattern 2 classification. An example is shown.

FIG. 3 is a diagram illustrating a data example of the unique name table stored in the unique name database.
As shown in this figure, in the unique name table, a unique name that is a reference for search and information on its attributes (location name, road name, facility name, etc.) are registered in association with each other. FIG. 3 shows an example in which unique names such as “Funabashi”, “No. 375”, “Convenience store”, “Factory” are registered. The attribute of “Funabashi” is “place name”, the attribute of “375” is “road”, and the attributes of “convenience store” and “factory” are “facility name”.

FIG. 4 is a diagram showing a data example of the search reference space graphic specifying table stored in the search reference space graphic specifying database.
As shown in FIG. 4, the search reference space specifying table stores a search reference space figure of an object represented by the unique name in association with each attribute of the unique name. For example, when a place name such as “Funabashi” is a search criterion, the spatial graphic on the map indicated by the search criterion is a surface, and therefore an identifier “A” indicating surface information is associated and registered. Yes. If a road name such as “375” is a search criterion, the spatial figure on the map indicated by the search criterion is a line, and therefore an identifier “B” indicating line information is registered in association with it. ing. In addition, when a facility name such as “convenience store” is a search criterion, since the spatial figure on the map indicated by the search criterion is a point, an identifier “C” indicating point information is associated and registered. Yes.
Based on the information of the search reference space graphic registered in the search reference space specifying table, the search range on the map when searching based on the search reference is determined.

FIG. 5 is a diagram showing an example of data in the individual setting table stored in the individual setting database.
As shown in FIG. 5, the individual setting table is a table that stores a search distance for each unique name. This search distance information is different for each unique name. For example, when the size of the object indicated by the unique name is large or wide, the search reference space figure specified by the unique name is used. The search distance from the position is set to be long, and the search range is specified based on the search distance information.

FIG. 6 is a diagram showing an example of data in the position / address correspondence table stored in the list database.
As shown in FIG. 6, the position / address correspondence table stores a position (latitude / longitude) corresponding to each address (prefecture name, city name, large letter / town chome, block code / lot number) in association with each other. is doing. The position / address correspondence table is used in the processing described later as information for specifying an address corresponding to the search range on the map.

FIG. 7 is a diagram illustrating an example of data in the search reference position storage table.
As shown in this figure, the search reference position storage table is data that stores the position of a unique name that serves as a search reference, in which an abstract language expression is in a document. For example, in the case of the abstract language expression “near”, since the unique name is located in front as in the document “near a convenience store ...”, the abstract language expression “near” is used as the dependency destination. “Previous” information is registered as the position of the unique name. Based on this search base point position storage table, a unique name having a dependency on an abstract language expression is specified, and this search name is used for specifying a search range on a map when the unique name is used as a search reference.

FIG. 28 is a diagram showing a processing flow of the countermeasure-target search device.
Next, details of the processing flow of the object search device 1 will be described with reference to FIG.
Hereinafter, an example in which the object search device 1 is used for a telephone number guidance service will be described.
First, it is assumed that the object search device 1 is used in a call center that provides a telephone number guidance service. In this case, it is assumed that a call center guide person receives a request for phone number guidance from a user who is provided with guidance, for example, “a factory behind a convenience store along No. 375 of Funabashi”. At this time, the object search device 1 accepts an input of search document information from “the factory behind the convenience store along the No. 375 of Funabashi” from the guidance person in charge (step S101). The input of this information may be accepted by character input using a keyboard, or may be accepted by voice. The document information analysis unit 11 of the target object search apparatus 1 performs morphological analysis (step S102), and divides the search document information into morphemes. Then, the document information analysis unit 11 generates a morpheme table as shown in FIG. 8 (a diagram showing the first morpheme table) (step S103) and stores it in a memory or the like.

  Next, the geographic information extraction unit 111 in the document information analysis unit 11 determines whether there is a morpheme registered as an abstract language expression in the abstract language expression table (FIG. 2) among the morphemes registered in the morpheme table. Determination is made (step S104). For example, an abstract language expression may be a single morpheme such as “along” or a combination of multiple morphemes such as “along” by connecting “ni” and “al”. Yes, in the abstract language expression table, if other morphemes “Ni” and “A” that may be combined with “Along” are registered in association with each other, they are used as one abstract language expression. It can also be determined. The geographic information extraction unit 11 associates the morphemes “along”, “ni”, and “al” with each other and registers them in the abstract language expression table. The language expression is read as an ID, and the ID is registered in association with the corresponding morpheme in the morpheme table (step S105). As a result, as shown in FIG. 8, ID (2) is registered in the morpheme table for each of the “along”, “ni”, and “al” morphemes.

  Similarly, the geographic information extraction unit 111 reads the ID of the abstract language expression specified by “back” and “no” from the abstract language expression table, and in the morpheme table, these “back” and “no”. The ID read in association with the morpheme is registered. As shown in FIG. 8, ID (7) is registered in the morpheme table for each “back” and “no” morpheme.

  Next, the geographic information extraction unit 111 determines whether there is a morpheme registered as a unique name that can be a search criterion in the unique name table among the morphemes registered in the morpheme table (step S106). If there is a unique name registered as a search criterion in the unique name table, the unique name attribute is read from the unique name table. The attribute is, for example, a place name, a road name, a facility name, or the like. Then, the geographic information extraction unit 111 registers the attribute of the morpheme registered as the unique name in the unique name table in the morpheme table (step S107). As shown in FIG. 8, the “place name” attribute is assigned to the morpheme “Funabashi”, the “road name” attribute is assigned to each of the “375” and “No.” morphemes, and “convenience store”, “factory”. For each morpheme, an attribute of “facility name” is registered in the morpheme table.

  Next, the dependency analysis unit 112 in the document information analysis unit 11 analyzes the relationship of each of a plurality of morphemes by a syntax analysis process (step S108). Then, by this analysis process, each morpheme is collected for each meaningful minimum unit (step S109), an ID is assigned to each minimum unit, and the morpheme table (showing the second morpheme table in FIG. 9). sign up. In addition, the dependency analysis unit 112 analyzes which minimum unit is the destination for each minimum unit, and an ID indicating the other minimum unit of the destination is displayed for each minimum unit in the morpheme table (see FIG. 9 morpheme table (2)). Note that if the other minimum unit has already been determined and the last remaining unit in the search document information that does not exist is the last minimum unit, there is no other destination. Register the information indicating.

In addition, the dependency analysis unit 112 identifies a minimum unit having another minimum unit as a destination among the minimum units in the morpheme table, and becomes a search reference before the abstract language expression in the minimum unit. A case where a unique name exists and a case where a unique name serving as a search criterion exists after an abstract language expression are distinguished (step S110). In the smallest unit, since the unique name exists before the abstract language expression, the unique name that is the destination of the abstract language expression is behind the abstract language expression, and the “attribute” of the unique name in the morpheme table. "" Is registered (that is, a unique name that can be a search criterion), the unique name is specified as a search criterion (step S111; FIG. 9). Also, since the unique name exists after the abstract language expression in the smallest unit, the unique name that is the destination of the abstract expression is before the abstract language expression, and the “attribute” of the unique name in the morpheme table is In the case where it is registered (that is, when it is a unique name that can be a search criterion), the unique name is similarly specified as a search criterion.
If there is a unique name serving as a search criterion both before and after the abstract language expression, the unique name sandwiched between the two preceding and following abstract language expressions is specified as the search criterion.

  Next, based on the attribute of each unique name in the morpheme table, the range narrowing unit 121 of the search range determining unit 12 obtains a space graphic when the facility represented by the specific name is used as a search reference, as a search reference space graphic table. It is read from (FIG. 4) and registered in the morpheme table as a space graphic for the unique name (step S112). For example, in the example of the morpheme table shown in FIG. 10 (a diagram showing the third morpheme table), the space figure on the map indicated by the search criterion of the unique name “Funabashi” is the plane, and the uniques “375” and “No.” The space figure on the map indicated by the name search criteria is a line, the space figure on the map indicated by the search name of the unique name “Convenience Store”, and the space figure on the map indicated by the search name of the unique name “Factory” Since they are points, they are registered in association with the corresponding unique names in the morpheme table.

FIG. 11 is a diagram showing an outline of the search range determination process.
Next, as shown in FIG. 11, the search range determination unit 12 sequentially calculates the logical product (AND condition) of the figures in the order of the IDs of the other smallest units specified for each smallest unit of the morpheme table. The final search range is determined by obtaining (step S113). At this time, the number of unique names may be any number, but if the attribute of the search base point is given to the unique name, the abstract expression is interpreted and the range on the map indicated by the morpheme in the smallest unit The specific identification of is performed. In this embodiment, the case where the search document information is composed only of positive expressions is described. However, when the search document information includes a negative expression such as “other than...” For example, an exclusive logical product may be taken. Then, the search unit 13 performs information search processing from the specified search range (step S114). Then, the output unit 14 outputs the search processing result (step S115).

Here, details of the search range determination process and the search process will be described.
<Example of the case where an abstract language expression is near / into / near>
FIG. 12 is a first diagram showing an overview of the search range determination process and the search process.
When the abstract expression interpretation unit 122 determines that the abstract language expression is “near”, “near”, or “near”, the range narrowing unit 121 uses the abstract language expression as a dependency destination. The search range is determined according to the reference unique name. In the determination of the search range, the range narrowing unit 121 reads the search distance registered in association with the unique name serving as the search reference in the individual setting table (FIG. 5), and displays it on the map of the unique name serving as the search reference. A circle whose center is the position of and whose radius is the search distance is specified as the search range.
Note that if the search document information contains only a unique name that serves as a search criterion with this abstract language expression (“near”, “near”, “near”) as a dependency destination. The search unit 13 specifies information corresponding to the information indicated by the unique name of the last dependency in the document information for search from the information corresponding to the circle within the specified search range, and outputs it as a search result To do. If a search result is not obtained within the search range, the search unit 13 may again determine a search range in which the search distance is gradually increased by a predetermined distance, and perform search processing in the search range ( (A) of FIG. As shown in FIG. 12C, when the search reference space figure corresponding to the attribute of the unique name that is the search reference is a point, the point is centered on the circle, and when it is a line, the line segment is The center may be the center of the circle, and if it is a surface, the center (or center of gravity) of the surface may be the center of the circle. Note that the search distance varies depending on the unique name serving as a search reference. For example, in the case of telephone number guidance at a call center, the target object search device 1 specifies an address on a map included in a circle whose radius is a search range centered on the location of a facility with a unique name serving as a search reference. Then, among the facilities in the address, the telephone number of the one that matches the facility indicated by the unique name of the last person in the search document information is specified.

<Example of using alongside +, as an abstract language expression for Nii / Nori>
FIG. 13 is a second diagram showing an overview of the search range determination process and the search process.
When the abstract language expression determined by the abstract expression interpreter 122 is “along”, “along”, “along”, etc., the uniqueness of the search criteria that uses the abstract language expression as a destination There is a high possibility that the attribute of the name is a line or a surface of the search reference space graphic. For example, among roads, tracks, and geographical shapes (along rivers, mountains, and seas), roads, railroads, and rivers are search criteria space graphics, and mountains and oceans are search criteria space graphics. It is. In this case, the range narrowing unit 121 determines a search range corresponding to a unique name serving as a search criterion with the abstract language expression as a dependency destination. In determining the search range, the range narrowing unit 121 reads the search distance registered in association with the unique name serving as the search reference in the individual setting table (FIG. 5). If the space graphic of the search reference is a line, the search range up to the search distance perpendicular to the line is specified as the search range (FIGS. 13C and 13D). Further, when the search reference space graphic is a surface, a range further expanded by the search distance from the edge of the space graphic is specified as the search range.
If there is only a unique name serving as a search criterion in which the abstract language expression is a dependency in the search document information, the search unit 13 selects information corresponding to the specified search range. Among them, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. If a search result is not obtained within the search range, the search unit 13 may gradually re-determine a search range in which the search distance is increased by a predetermined distance, and perform search processing in the search range. . Note that the search distance varies depending on the unique name serving as a search reference. Here, if the specific processing in this case is, for example, telephone number guidance at a call center, the object search device 1 is included in a search range that is enlarged by the search distance from the space graphic indicated by the unique name serving as the search reference. The address on the map is identified, and the telephone number of the facility in the address that matches the facility indicated by the unique name of the last person in the search document information is identified.

<Example of using an intermediate language as an abstract language expression>
FIG. 14 is a third diagram showing an outline of the search range determination process and the search process.
When the abstract language expression determined by the abstract expression interpretation unit 122 is “middle”, “middle”, “middle”, etc., the search criteria specific to the abstract language expression as a destination This is a case where the name attribute is a face of the search reference space graphic. For example, in the case of a sentence such as in an amusement park, 1 in 1 chome. The range narrowing unit 121 determines the range of the surface indicated by the space graphic with the unique name as a search criterion that uses the abstract language expression as a destination as the search range. If you do not know the surface on the map, such as “in a group of high-rise buildings”, for example, a building of 25 floors (predetermined number of fixed floors) or more for identifying high-rise buildings. A range on the map surrounding the position is set as a search range ((c) in FIG. 14). In the case of a document “in a residential area”, a building may be specified based on the information of the number of floors equal to or lower than the second floor, and a range on the map surrounding the position of the building may be set as a search range.
Note that if the search document information includes only a unique name that serves as a search criterion with the abstract language expression as a destination, the search unit 13 includes the information corresponding to the specified search range. From the search document information, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. For example, if the specific processing in this case is telephone number guidance at a call center, the target object search device 1 may include an address on a map included in the search range within the space figure indicated by the unique name serving as a search reference. And the telephone number of the facility in the address that matches the facility indicated by the unique name of the last person in the search document information is identified.

<Examples of abstract language expression in the past, before and after +>
FIG. 15 is a fourth diagram showing an overview of the search range determination process and the search process.
When the abstract language expression determined by the abstract expression interpretation unit 122 is “old”, “old”, “old”, “previous”, “previous”, “previously”, etc. May not currently have a unique name for a search criterion that depends on the abstract language expression. For example, there may be a sentence such as “a post office near a candy store that used to be”. In this case, the range narrowing unit 121 instructs the search unit 13 to perform search processing from past information. For the search range, an example of specifying another search range (a search range specifying process using abstract language expressions such as “near”, “along”, and “inside”) is used.
In other words, the search reference object does not exist due to rebuilding, etc., but when the search reference object exists at the same position at a certain point in time in the past, the current position (point) corresponding to the search reference object A search is performed using the surrounding area as a search range. The range of years that “old” points back may be limited to a range in which past map information exists, and it is only necessary to enable search at any specified number of years. This example is useful even when the organization itself, such as a company or a public office, does not change its position, and even if it has existed for a long time, the name of the organization has been changed and repeated. By using the location / address table and name / address table according to past map information, even if you only know the name of the old organization, search for the current organization name corresponding to it. It can also be used as a reference point. As a result, since the time axis is specified and the past map information and list are also searched, the abstract language expression “old” can be used.

<Example when abstract language expression is in direction / direction / side +
FIG. 16 is a fifth diagram showing an overview of the search range determination process and the search process.
The abstract language expressions determined by the abstract expression interpreter 122 are A to B “direction”, A to B “direction”, A to B “direction”, and A to B “direction”. , A to B “in the direction”, A to B “in the direction”, A from the B “side”, A from the B “side”, A from the B “side” , And so on, the range narrowing unit 121 specifies the unique name A of the search criterion that uses the abstract language expression as the destination as the first search criterion and B as the second search criterion. Then, the range narrowing unit 121 reads the search distance registered in association with the unique name as the first search criterion in the individual setting table (FIG. 5), and from the first search criterion A to the read search distance. And a range from the position of the first search criterion A to the second search criterion B up to a predetermined left and right angle is specified as the search range. The predetermined left-right angle is, for example, 90 degrees.
If there is only a unique name serving as a search criterion in which the abstract language expression is a dependency in the search document information, the search unit 13 selects information corresponding to the specified search range. Among them, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. If a search result is not obtained within the search range, the search unit 13 may determine again the search range in which the search distance is gradually increased by a predetermined distance, and perform search processing in the search range. Note that the search distance varies depending on the unique name serving as a search reference. Here, if the specific processing in this case is, for example, telephone number guidance in a call center, the target object search device 1 uses the unique name of the first search criterion and the unique name of the second search criterion as the search criteria. Identifies the address on the map included in the specified search range, and matches the facility indicated by the unique name of the final recipient in the search document information among the facilities in the address. Identify the phone number.

<An example of abstract language expression that is in the opposite side, behind / inside>
FIG. 17 is a sixth diagram illustrating an outline of the search range determination process and the search process.
The abstract language expression determined by the abstract expression interpreter 122 is “opposite” of A, “opposite” of A, “opposite” of A, “back of A”, “back of” of A, If A is “behind” or the like, the range narrowing unit 121 identifies the unique name A of the search criterion that uses the abstract language expression as a dependency as the search criterion. Then, the range narrowing unit 121 reads the search distance associated with the unique name A serving as the search reference and registered in the individual setting table (FIG. 5), and sets the read search distance as a radius around the search reference. Specify a circle. Next, the range narrowing unit 121 determines whether other unique names whose space graphic information is a line or a surface are registered in the search target group in all directions in the specified circle. For example, if the search target group is map information, the space figure is registered as a line from the unique names of all objects (facility, road, river, mountain, sea, etc.) existing within the circle in the map Identify the unique name. For example, roads, tracks, etc. Then, the range narrowing unit 121 determines, based on the line in the circle, the region that does not have the search criterion, and the region that has the search criterion is the back (FIG. 17).
If there is only a unique name serving as a search criterion in which the abstract language expression is a dependency in the search document information, the search unit 13 selects information corresponding to the specified search range. Among them, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. If a search result is not obtained within the search range, the search unit 13 may determine again the search range in which the search distance is gradually increased by a predetermined distance, and perform search processing in the search range. Note that the search distance varies depending on the unique name serving as a search reference. Here, if the specific processing in this case is, for example, telephone number guidance in a call center, the target object search apparatus 1 uses a unique name of the search criterion, a circle whose center is the search criterion, and whose radius is the search distance. Of the addresses in the circle, the spatial figure present in the circle is based on the object of the line, the address of the area on the search reference side is behind, the address of the area on the side without the search reference is facing, Based on which abstract language expression is “back”, the telephone number of the thing corresponding to the facility indicated by the unique name of the last receiver in the search document information is specified from the corresponding area.

<Example of the case where an adjacent language + is an abstract language expression>
FIG. 18 is a seventh diagram illustrating an overview of the search range determination process and the search process.
When the abstract language expression determined by the abstract expression interpretation unit 122 is “next to” A, “next to” A, “next to” A, or the like, the range narrowing unit 121 The unique name A of the search criterion that uses the abstract language expression as the receiver is specified as the search criterion. Then, the range narrowing unit 121 reads the search distance associated with the unique name A serving as the search reference and registered in the individual setting table (FIG. 5), and sets the read search distance as a radius around the search reference. Is specified as a search range. Next, the search unit 13 specifies a unique name that is registered in association with the latitude and longitude included in the range within the specified circle, and among the unique names, the distance from the unique name A that is the search reference is the longest. A specific number of unique names in the order from the unique name close to or the one closest to the unique name A serving as a search reference is specified and output as a search result.
If a search result is not obtained within the search range, the range narrowing unit 121 again determines a search range in which the search distance is gradually increased by a predetermined distance, and the search unit 13 performs a search process in the search range. You may do it. Note that the search distance varies depending on the unique name serving as a search reference. Here, if the specific processing in this case is, for example, telephone number guidance in a call center, the target object search apparatus 1 uses a unique name of the search criterion, a circle whose center is the search criterion, and whose radius is the search distance. Among the buildings registered in association with the latitude and longitude in the specified number of facilities in order, starting from the facility closest to the search criteria facility or closest to the search criteria facility Among these facilities, the telephone number of the facility that matches the facility indicated by the unique name of the last person in the search document information is specified.

<Example of using an abstract language expression for a line + ///>
FIG. 19 is an eighth diagram showing an overview of the search range determination process and the search process.
When the abstract language expression determined by the abstract expression interpretation unit 122 is “line-up” of A, “line-up” of A, “line-up” of A, or the like, the range narrowing-down unit 121 uses the abstract language. The unique name A of the search criterion that uses the expression as a destination is specified as the search criterion. Then, the range narrowing unit 121 reads the search distance associated with the unique name A serving as the search reference and registered in the individual setting table (FIG. 5), and sets the read search distance as a radius around the search reference. Specify a circle. Next, the range narrowing unit 121 searches for the unique name registered in association with the latitude and longitude included in the range within the specified circle and whose attribute space pattern is a line. And, if the space figure has a unique name for a line in the circle, it is within the circle and within a plane that is vertically extended from the line in the circle by a predetermined distance (that is, a rectangle centered on the line) Is specified as the search range. In addition, when there are a plurality of data having a unique name of a space figure in the circle, the line having the shortest distance from the search standard to the line is used.
Note that if the search document information includes only a unique name serving as a search criterion with the abstract language expression as a destination, the search unit 13 then determines the latitude included in the specified search range. The unique name registered in association with the longitude is identified, and among the unique names, the one that matches the information indicated by the unique name of the last dependency destination in the search document information is identified as the search result. Output. If a search result is not obtained within the search range, the search unit 13 may determine again the search range in which the search distance is gradually increased by a predetermined distance, and perform search processing in the search range. Note that the search distance varies depending on the unique name serving as a search reference. Here, the specific processing in this case is, for example, telephone number guidance at a call center, the object search device 1 has a space figure as a line in a circle centered on the search reference and having a search distance as a radius. A facility that is registered in association with the address of the latitude and longitude of the search range is searched for a range that is vertically extended from the object (river, road, or track) as a search range. Among these facilities, the telephone number of the facility that matches the facility indicated by the unique name of the last person in the search document information is specified.

<Example when abstract language expression is higher than XX and lower than XX>
FIG. 20 is a ninth diagram showing an overview of the search range determination process and the search process.
When the abstract language expression determined by the abstract expression interpretation unit 122 is XX “higher”, XX “lower”, etc., the range narrowing unit 121 uses the abstract language expression as a dependency destination. The unique name is specified as a search criterion. Then, the range narrowing unit 121 reads the search distance associated with the unique name serving as the search reference and registered in the individual setting table (FIG. 5), and sets the read search distance as the radius around the search reference. Identify the circle to be. Next, when the abstract language expression is “higher”, the range narrowing unit 121 converts the altitude data (data in which the unique name and the altitude value are associated) in all directions within the specified circle. Based on this, an altitude region higher than the altitude of the search reference is determined as the search range. Further, when the abstract language expression is “lower”, the range narrowing unit 121 is based on altitude data (data in which a unique name and an altitude value are associated) in all directions in the specified circle. Thus, an altitude area lower than the altitude of the search reference is determined as the search range.
If there is only a unique name serving as a search criterion in which the abstract language expression is a dependency in the search document information, the search unit 13 selects information corresponding to the specified search range. Among them, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. If a search result is not obtained within the search range, the search unit 13 may determine again the search range in which the search distance is gradually increased by a predetermined distance, and perform search processing in the search range. Note that the search distance varies depending on the unique name serving as a search reference. Here, if the specific processing in this case is, for example, telephone number guidance in a call center, the target object search device 1 is on a map included in a circle with a unique name serving as a search reference as a center and a search distance as a radius. Of the facility in the address that is registered as an altitude higher than the altitude of the search criteria, and matches the facility indicated by the unique name of the last person in the search document information Identify the phone number.

<Example of abstract language expression at the edge, corner + //>
FIG. 21 is a tenth diagram showing an overview of the search range determination process and the search process.
The abstract language expressions determined by the abstract expression interpreter 122 are A's "at the end", A's "at the end", A's "at the end", A's "corner", A's "corner", In the case of “in the corner” of A, etc., the range narrowing unit 121 identifies the unique name having the abstract language expression as a destination as the search criterion. Here, there is a high possibility that the space graphic of the attribute indicated by this unique name is a line or a surface. Then, when the space graphic of the attribute indicated by the unique name is a line, the range narrowing unit 121 is individually set in association with the unique name serving as a search criterion, with the two end points at both ends of the line as the centers. Two circles whose radius is the search distance registered in the table (FIG. 5) are determined as search ranges. Alternatively, when the space graphic having the attribute indicated by the unique name is a surface, the range narrowing unit 121 determines a surface range enlarged by the search distance specified by the unique name from the edge of the space graphic as the search range. .
If there is only a unique name serving as a search criterion in which the abstract language expression is a dependency in the search document information, the search unit 13 selects information corresponding to the specified search range. Among them, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. If a search result is not obtained within the search range, the search unit 13 again determines two circles whose search distances are gradually increased as a search range, and performs search processing within the search range. May be. Note that the search distance varies depending on the unique name serving as a search reference. Here, specific processing in this case is, for example, telephone number guidance at a call center, and if the unique name that is the search criterion is XX bridge, center the both ends of the XX bridge and set the search distance to a radius. The address on the map included in the two circles is identified, and the telephone number of the facility in the address that matches the facility indicated by the unique name of the final recipient in the search document information is identified. To do.

  On the other hand, if the unique name is a surface and the abstract language expression is “end” of A and “end” of A, the range narrowing unit 121 starts from the edge (boundary, outer periphery) of the surface. Then, the range in which the search distance that is associated with the unique name serving as the search reference and registered in the individual setting table (FIG. 5) is widened is determined as the search range. Alternatively, the range narrowing unit 121 has a unique name of a rectangular surface, and when the abstract language expression is “corner” of A and “corner” of A, the four corners of the rectangular surface 4 circles, each of which is associated with a unique name serving as a search criterion and having a search distance as a radius registered in the individual setting table (FIG. 5), is determined as a search range.

<Example of abstract language expression in between + ///>
FIG. 22 is an eleventh diagram showing an outline of the search range determination process and the search process.
If the abstract language expression determined by the abstract expression interpretation unit 122 is “between” A and B, “between” A and B, “between” A and B, etc., the range The narrowing-down unit 121 specifies the unique name A of the search criterion that uses the abstract language expression as a dependency as the first search criterion and B as the second search criterion. Then, the range narrowing unit 121 reads the search distance registered in association with the unique name as the first search criterion in the individual setting table (FIG. 5), and from the first search criterion A to the read search distance. And passing through the first search criterion A and the second search criterion B around the direction of the vertical line from the position of the first search criterion A to the second search criterion B, and a predetermined flat rate f The range of the ellipse is specified as the search range.
If there is only a unique name serving as a search criterion in which the abstract language expression is a dependency in the search document information, the search unit 13 selects information corresponding to the specified search range. Among them, the search document information that matches the information indicated by the unique name of the last dependency destination is identified and output as a search result. If a search result is not obtained within the search range, the search unit 13 may gradually decrease the flatness ratio, determine the expanded search range again, and perform search processing in the search range. Here, for example, in the case of telephone number guidance at a call center, the specific processing in this case is based on the unique name of the first search criterion, the unique name of the second search criterion, and the flat rate f as search criteria. Identify the address on the map that is included in the specified elliptical search range, and call the facility that matches the facility indicated by the unique name of the last person in the search document information among the facilities in the address Identify the number.

FIG. 23 is a twelfth diagram showing an overview of the search range determination process and the search process.
FIG. 24 is a thirteenth diagram showing an overview of the search range determination process and the search process.
FIG. 25 is a fourteenth diagram showing an overview of the search range determination process and the search process.
Having described a plurality of different search range determination processes and search processes in that case, returning to an example where the document information for search is “the factory behind the convenience store along the 375 of Funabashi”, Based on a unit composed of two morphemes “Funabashi”, an area “Funabashi” is identified from the map (FIG. 23A). Next, in the area of “Funabashi”, based on the unit consisting of the following five morphemes “along 375”, the line was expanded vertically from the 375 road line by the distance specified by the road. An area along No. 375 is specified (FIG. 23B). Next, a convenience store along 375 is identified as a search criterion based on a unit composed of two morphemes “convenience store” based on the region along 375 ((a) of FIG. 24). In the example of FIG. 24a, two convenience stores, convenience store A and convenience store B, are specified as search criteria. Then, based on a unit composed of two morphemes “back”, the areas behind the convenience store A and convenience store B that are search criteria are specified as the final search range. Then, the search unit 13 specifies a facility that matches the “factory” that is the unique name of the final dependency in the search document information from the final search range. In this specification, the search unit derives the address corresponding to the final search range from the position (latitude and longitude) / address correspondence data, and the address of the facility name at the address corresponding to the final search range from the address / facility name correspondence data. A facility name whose attribute is a factory is specified, and a final list of factory telephone numbers is output from the facility name / phone number correspondence data. As a result, in the telephone number guidance service, the object search device can output the search target telephone number based on the search document information “factory behind convenience store along No. 375 of Funabashi”. .

FIG. 26 is a diagram showing an outline of processing for determining whether or not the image is included in the final search range.
In addition, the process of extracting the search target candidate included in the final search range is, as shown in FIG. It is determined whether the number of intersections between the straight line and the final search range is odd or even. If the number of intersections between the straight line to the right of the facility and the final search range is an odd number, the facility is included in the final search range, and if it is an even number, the facility is included in the final search range. Can be determined. Note that this processing technique can be used not only for polygons as shown in FIG. 26 but also for various closed regions.

FIG. 27 is a schematic diagram showing two different processing orders of the object search apparatus.
As processing examples of the object search device 1, there are two types of processing forms. As described above, first, the spatial figure is captured based on the attribute of the unique name in the search document, the search range is specified according to the abstract representation of the dependency of the unique name, and the latitude and longitude of the specified search range Address information corresponding to the address is searched from the position / address correspondence table, and based on the name, the name matching the searched address information matches the unique name of the final dependency in the search document information. Is identified. This example is the processing of the object search device described above.
On the other hand, as another example, the latitude and longitude on the map of the address of each unique name is calculated based on the correspondence table of the address and the unique name, and mapped as mapping information on the map. Then, it captures the space figure based on the attribute of the unique name in the search document, specifies the search range according to the abstract representation of the dependency of the unique name, and the address and unique name of the mapping information included in the search range From the list of correspondence relationships, the one that matches the unique name of the last dependency destination in the search document information is specified.

  As described above, the embodiment of the present invention has been described. However, according to the processing of the above-described target object search device, the search document information generated from the search conditions for speech and text is morphologically analyzed, and a predetermined unique name or unique name is obtained. Based on the abstract language expression of the dependency destination, the search range corresponding to the space figure on the map indicated by the unique name attribute is specified, and when a plurality of unique names are included in the search document information, From the final search ranges indicated by the AND conditions of the respective search ranges, a search is made for a match with the unique name of the final dependency destination indicated by the search document information. Therefore, the search document information generated from the search conditions for speech and text includes a plurality of words serving as search criteria, and a word indicating an abstract language expression indicating geographical features is included in a plurality of the words. Even in this case, a detailed search can be performed based on information included in the search conditions.

Further, according to the present invention, it is possible to narrow down the search range on the map by using the abstract expression with directivity, and thereby specify the search range with fine directivity according to the user's intention. Therefore, even when the number of times of specifying the search document information is small, it is possible to reduce the time and effort of searching for the search again and again by detailing the explanation in the search document information.
Further, according to the present invention, when narrowing down the search range in the direction with the second search criterion (including the direction) as seen from the first search criterion, the direction can be narrowed down when the direction is known. When it is not possible to specify only the area searched by the direction search, the search range can be expanded and narrowed down step by step. At this time, even if the direction is not immediately known, if the place name or facility on the extension line is remembered, the search range in a specific direction can be narrowed down using this as the second search criterion.
In addition, according to the present invention, when narrowing down from the search criteria to the opposite side (or the near side) with respect to some target, the requester is said to be near / behind the search criteria. The target is likely to exist. And the target is specified even if it does not specify the target. Then, with the target as the boundary, the side where the search criterion exists is “behind”, and the side where the search criterion does not exist is “opposite”, so that narrowing down according to the intention of the user can be performed.
In addition, according to the present invention, it is possible to narrow down in the direction adjacent to the search criterion, thereby determining whether the adjacent address is difficult to determine in the conventional address / unique name correspondence table list due to the demarcation of roads and the like. It can be performed.
In addition, according to the present invention, it is possible to narrow the search range along the linear target located near the search reference, so that the linear target is not clearly indicated in the search document information. You can narrow down according to your will.
In addition, according to the present invention, by enabling narrowing down to a direction corresponding to a range that is higher or lower than the search criterion, it is possible to determine whether the altitude is higher or lower than the search criterion even if the address or place name is unknown. If you remember, you can narrow down the search range in that direction.
In addition, according to the present invention, by enabling narrowing down to the direction of the corner or end of the search criterion, the corner of the point (surface) of the search criterion, such as a place name or range, where the address or place name is unknown. If you remember that it is located at the edge, you can narrow down in that direction.
In addition, according to the present invention, it is possible to narrow down an area between a plurality of search criteria, so that even if an address or a place name is not known, a certain point (or building) and another point ( If you remember that it was located between the buildings), you can narrow down the area.
Also, according to the present invention, the current name is currently unknown due to rebuilding or organization name change, but when the search criteria that existed in the past by going back the time axis are known, the search criteria is used as a clue. Can do.
Further, according to the present invention, even if the abstract language expression is the same, the search range can be expanded and contracted according to what the search reference object is, so that the number of search repetitions can be reduced and convenience can be improved. Search criteria can be set based on the results of a user questionnaire, for example, according to the case of point, line, surface (small), and surface (large). / "Near", "Old") and National Highway No. 375 (Line / "Along") can describe the search range according to the specific equipment and abstract expression.
In addition, when the object search device 1 is applied to a call center, a sentence that has been converted into a text by voice recognition can be processed for voice uttered by a customer using a mobile phone or a fixed telephone. . In this case, using a GPS mobile phone as a search base, not only can the current position be obtained from a GPS device like a car navigation system, but if a fixed phone is used, a rough one is estimated from the local phone number of the outgoing phone number. It can also be used as a search base point.

  Note that the above-described object search apparatus has a computer system inside. Each process described above is stored in a computer-readable recording medium in the form of a program, and the above process is performed by the computer reading and executing the program. Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

  The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

DESCRIPTION OF SYMBOLS 1 ... Object search device 11 ... Document information analysis part 12 ... Search range determination part 13 ... Search part 14 ... Output part 15 ... Abstract expression database 16 ... Unique name database 17 ... Numerical map information database 18 ... Individual setting database 19 ... List database 111 ... Geographic information extraction unit 112 ... Dependency analysis unit 121 ... Range narrowing unit 122 ... Abstract expression Interpreter

Claims (18)

Document information analysis means for extracting a unique name, which is a word used as a search reference, included in the search document information, and an abstract language expression indicating a geographical feature,
Based on the dependency relationship of the abstract language expression with respect to the unique name, one or more combinations of the unique name and the abstract language expression having the dependency relationship are specified, and each combination is included in the combination. A search range on the map for the unique name is determined based on an abstract language expression corresponding to the unique name, and based on the search range corresponding to the unrestricted unique name and the search range for each combination A search range determining means for determining a final search range on the map for searching for the last unique name appearing at the end of the search document information;
Search means for searching for the search object from within the final search range;
An object retrieval apparatus comprising: Position / address storage means for storing address information corresponding to the position information of the search range;
An address / unique name storage means for storing a unique name corresponding to the address,
The search means specifies address information from the position / address storage means based on position information corresponding to the final search range, and based on the address information, from the address / unique name storage means, the search object The object search device according to claim 1, wherein a unique name corresponding to is specified. When the abstract language expression is an expression indicating that the abstract language expression is close to the unique name, the search range determining means has a distance according to the unique name with the position on the map indicated by the unique name as a center. The object search device according to claim 1, wherein a circle having a radius of is determined as the search range on the map. When the abstract language expression is an expression indicating that the unique name is in line with the unique name, the search range determining means starts from the end of the line or the surface with reference to a space figure of the line or the surface indicated by the unique name. 3. The object search device according to claim 1, wherein a range expanded vertically by a distance corresponding to the unique name is determined as the search range on the map. When the abstract language expression is an expression indicating that the unique language name is included in the unique name, the search range determining means determines that the range on the map indicated by the unique name is the search range on the map. The object search device according to claim 1, wherein the object search device is determined. When the abstract language expression is an expression indicating either the direction, direction, or side of the second unique name when viewed from the first unique name, the search range determining means determines the first unique name. The first search criterion and the second unique name are specified as the second criterion, and a range from the first search criterion to the second search criterion up to a predetermined left-right angle is defined on the map. The object search device according to claim 1 or 2, wherein the search range is determined. When the abstract language expression is an expression indicating the opposite of the unique name, the search range determining means is centered on a position on the map indicated by the unique name, and a distance corresponding to the unique name is a radius. In the circle, the spatial figure is identified as a line, and the side where the unique name exists is determined as the search range on the map with the line as a reference. The target object search apparatus of Claim 1 or Claim 2. When the abstract language expression is an expression indicating the back of the unique name, the search range determining means is centered on a position on the map indicated by the unique name, and a distance corresponding to the unique name is a radius. In the circle, the spatial figure is identified as a line, and the side where the unique name does not exist is determined as the search range on the map with the line as a reference. The target object search apparatus of Claim 1 or Claim 2. When the abstract language expression is an expression indicating the next to the unique name, the search range determining means is centered on a position on the map indicated by the unique name, and a distance corresponding to the unique name is a radius. The search range is determined as the search range,
The object search device according to claim 1, wherein the search unit specifies a search object that is close to the unique name in the search range. When the abstract language expression is an expression indicating the sequence of the unique names, the search range determining means is centered on the position on the map indicated by the unique name, and the distance corresponding to the unique name is a radius. In the circle to be identified, the space figure is closest to the unique name that is a line, and a rectangular range that is vertically expanded from the line by a distance corresponding to the unique name is defined as the search range on the map. The object search device according to claim 1, wherein the object search device is determined. When the abstract language expression is an expression indicating that the position of the abstract language is higher than the position of the unique name, the search range determining means uses the position on the map indicated by the unique name as a center, In a circle having a radius corresponding to the specified distance, a position higher than the position of the unique name is specified based on data stored on the map, and the specified position is specified as a search range. The object search device according to claim 1, wherein the object search device is an object search device. When the abstract language expression is an expression indicating that the position of the abstract language is lower than the position of the unique name, the search range determining means uses the position on the map indicated by the unique name as a center, In a circle having a radius corresponding to the distance, a position lower than the position of the unique name is identified based on data stored on the map, and the identified position is identified as a search range. The object search device according to claim 1, wherein the object search device is an object search device. In the case where the abstract language expression is an expression indicating that the end of the unique name, the search range determining means sets the distance corresponding to the specific name as a radius around the end of the space graphic of the specific name. Or a range in which a distance corresponding to the unique name is extended vertically from an edge of the space graphic of the unique name is determined as the search range on the map. Item 3. The object search device according to any one of Items 2 to 3. When the abstract language expression is an expression indicating that the unique name is a corner of the unique name, the search range determining means sets the distance corresponding to the unique name as a radius around the corner of the space graphic of the unique name. The object search device according to claim 1, wherein the circle is determined as the search range on the map. When the abstract language expression is an expression indicating between the third unique name and the fourth unique name, the search range determining means uses the third unique name as the third search criterion, and the fourth unique name as the fourth unique name. The unique name is specified as the fourth search criterion, and an ellipse range having a predetermined flatness ratio connecting the third search criterion and the fourth search criterion is determined as the search range on the map. The target object search apparatus of Claim 1 or Claim 2. The position / address storage means stores past address information corresponding to position information in a search range,
The target address search apparatus according to any one of claims 2 to 15, wherein the address / unique name storage unit stores a unique name on the map in the past corresponding to the past address. A processing method for an object search device, comprising:
The document information analysis means of the object search device extracts a unique name included in the search document information and an abstract language expression indicating a geographical feature,
The search range determining means of the object search device may include one or a plurality of combinations of the unique name and the abstract language expression having the dependency relationship based on the dependency relationship of the abstract language expression with respect to the unique name. For each combination, the search range on the map for the unique name included in the combination is determined based on the abstract language expression corresponding to the unique name, and the search range corresponding to the unrestricted unique name And, based on the search range for each combination, determine the final search range on the map to be searched for the last unique name that appears at the end of the search document information,
The processing method, wherein the search means of the object search device searches for the search object from within the final search range. The computer of the object search device
Document information analysis means for extracting a unique name included in the search document information and an abstract language expression indicating a geographical feature,
Based on the dependency relationship of the abstract language expression with respect to the unique name, one or more combinations of the unique name and the abstract language expression having the dependency relationship are specified, and each combination is included in the combination. A search range on the map for the unique name is determined based on an abstract language expression corresponding to the unique name, and based on the search range corresponding to the unrestricted unique name and the search range for each combination A search range determining means for determining a final search range on the map with the last unique name appearing at the end of the search document information as a search target;
Search means for searching for the search object from within the final search range,
A program characterized by functioning as

Images