JP4291874B1 - E-mail route search system, route search method, computer program - Google Patents

Abstract

A route search suitable for an image of a route desired by a user is facilitated.
An official route name of a transportation facility and a popular route name designated by a user are associated with each other and registered in a popular route DB. When a route search request mail in which a route search condition and a so-called route name are described in a natural sentence is received from the user terminal 30, the natural language processing tool 12 extracts a route search condition and a so-called route name from the request email. The route search tool 13 performs a route search based on the extracted route search conditions, and finds the official route name associated with the same known route name as the extracted known route name from the known route DB. The route candidates including the official route name searched from the plurality of types of route candidates obtained by executing the route search are narrowed down, and the narrowed route candidates are output. After that, a reply mail in which the narrowed route candidates and diagram information are described in a natural sentence is created, and this reply mail is returned to the user terminal 30 through an electronic mail.
[Selection] Figure 1

Description

  The present invention relates to a route search technique using electronic mail that makes it possible to realize a route search environment of a transportation facility that is comfortable for a user.

  About the search technology of the optimal operation route in transportation such as trains, the route search software `` Station Spago '' (registered trademark, non-patent literature) developed by the applicant of the present application in the Showa era and gained the support of many users. Since the advent of 1), many improvements have been attempted by the hands of multiple corporate engineers. In the route search of transportation facilities, the search process when a departure place and a destination are specified is generally performed as follows.

  Searching for the next station that minimizes the route operating cost (time, distance, fare, etc.) for all routes (transportation routes from one point to another point) and stations connected to the departure point are chained By repeating the above, the route search is performed in the form of ripples when a stone is thrown on the water surface. When a destination is specified in the search process, a route between the destination and the departure point is set as a route candidate desired by the user, and a plurality of route candidates derived from the route are searched and stored. Enable sorting according to specified search conditions. For example, if the search condition specified by the user is the route with the shortest time, the stored patterns are sorted by time, the route with the pattern with the shortest time is determined as the optimal route candidate, and further compared with this optimal route candidate Then, some valid alternative routes are identified, and these routes are displayed on the display device.

  Nowadays, mobile terminals such as mobile phones, communication networks, and e-mails have become widespread, and various mechanisms for searching for routes using these infrastructures have been proposed. For example, in the “departure information notification system and last train information notification system” described in Patent Document 1, the current position and the moving position are measured by a GPS (positioning system) mounted on a portable terminal, and the vehicle departs toward the nearest station. Information about the time, route, and nearest station to be notified is sent to the mobile terminal by e-mail. In this system, since the size of the display screen of the portable terminal is limited, it is necessary to divide and display information (see paragraph 0016 of FIG. 1, FIG. 8 and the like).

  In the “route information distribution system” described in Patent Document 2, when departure point information, destination information, and a transmission request are input through a predetermined input screen of the mobile terminal, the server The route search is performed based on the information, the search result is saved, and the route information is recorded for the e-mail address of the mobile terminal of “http://www.foobar.com/hogehge?ID=0123abcd” Such a URL is embedded in the mail body and returned, and the route information is transmitted to the portable terminal that requested the route transmission through the URL (see paragraphs 0022 and 23 of Patent Document 2). In this system, route information needs to be connected to the Internet to access a website, so that there is a problem that it takes time to obtain route information from condition input.

  Considering the point that the data input environment in the portable terminal is not sufficient as compared with a personal computer or the like, as in the “portable terminal device” described in Patent Document 3, the location and time of Templates with embedded granularity are presented and emails are generated based on the templates. However, in this portable terminal device, since a template must be prepared, advance preparation is necessary, and there remains a problem of lack of versatility.

The above technique uses electronic mail for route search condition input or transmission of route information. However, there is also proposed a technique in which condition input and result output are all performed by electronic mail. For example, in the “route guidance system” described in Patent Document 4, a route search is executed based on route search condition information included in an e-mail transmitted from a terminal, and route search result information is transmitted to the terminal by e-mail. At the same time, a guidance time for executing route guidance is set based on the transmitted route search result, and an e-mail including route guidance information is transmitted to the terminal at the set guidance time.
The content of the route search request or instruction is determined by the subject and body of the e-mail. If the subject is blank and there is a route search condition in the text, a new route search request is requested. If the subject is "Re ~" and the route guidance is quoted in the text, the route guidance is canceled, and the subject is "Re ~". If the route search result is quoted in the text, specify another route search under the same route search conditions. If the subject is "Re ~" and the text is blank, delete the searched route and provide route guidance. This is an instruction not to execute (paragraphs 0045 to 47).

According to the system disclosed in Patent Document 4, a user can request a route search using widely-distributed email without installing dedicated software or preparing a dedicated device, and the result can be sent to the email. There is an advantage that can be received immediately. This is convenient for business operators that provide route guidance services, as it can reduce capital investment (paragraph 0063).
However, in the system of Patent Document 4, when a route search is canceled or another route re-search is instructed, the route guidance mail or the route search result is quoted in the text (paragraphs 0045 to 47). There will be at least one round trip, which not only wastefully uses the communication line, but also requires the server to check whether there is a change in the content of the text.
Also, when there is a lot of time information in the route search result, the user is bothered to send the route information about the same route many times to the terminal unless the user bother to send and instruct the user. is there. Furthermore, in this system, when searching for a new route, the user adds “departure place” “from” “destination” “to” “date and time (year / month / day / minute)” “condition (“ departure ”) in the body of the e-mail. , “Arrival”, “final”, etc.) ”must be correctly described (paragraph 0045), and it takes effort to finish the input.

"Electrical Spa" Fuunroku Amagasaki and others Nikkei Business Publications, Inc. March 20, 2006 JP 2002-131075 A JP 2002-296067 A JP 2008-046840 A JP 2007-232441 A

In the systems described in the above-mentioned patent documents, when the route search result is confirmed on the mobile terminal, the display screen of the mobile terminal is small, so that many routes cannot be displayed on one screen. Therefore, even if there is an image of a route that the user wants to use, it is difficult for the user to find a route including the route on the display screen.
Although it is conceivable to specify a route name in advance as a search condition, it is difficult for the user to accurately input a formal route name (displayed route name). In addition, the route name is generally called as a common name, but the way of common name varies depending on the person and is highly ambiguous. Therefore, the route name is used in an interactive environment such as a Web service. In the conventional route search system, a route name could not be specified except for selecting from the list.

  In the present invention, even when a non-interactive medium such as e-mail is used, the route name of the operation route can be specified by an ambiguous character string, and optimized according to the user's image. Another object of the present invention is to provide a route search technique capable of obtaining a route search result.

In order to solve the above problems, the present invention provides a route search system, a route search method, and a computer program.
The route search system according to the present invention includes an e-mail means for sending and receiving e-mails to and from a user terminal, a so-called route DB (in which a formal route name of a transportation facility and a so-called route name designated by a user are recorded in association with each other. DB is an abbreviation of database, the same applies hereinafter), and when the e-mail means receives a route search request mail in which the route search conditions including the departure point , destination and time data and the so-called route name are described in natural text The natural language processing tool for extracting the route search condition and the so-called route name by performing natural sentence analysis on the request mail, and performing a route search based on the extracted route search condition. The official route name associated with the same known route name as the extracted known route name is retrieved from the known route DB and obtained by executing the route search. A route search tool that narrows down the route candidates including the official route name searched out from several types of route candidates, outputs a narrowed route candidate, and creates a reply e-mail describing the narrowed route candidates and diagram information in natural text And a control means for returning the reply mail to the user terminal through the electronic mail means.

  The user can receive a route candidate associated with the so-called route name as a reply mail only by sending a request email describing the route search condition and so-called route name in natural text from the user terminal. In addition to remarkably reducing the user's operation work when requesting a route, it is possible to obtain route search results narrowed down to route candidates that match the image of the route to be used. Thereby, the use environment of the route search which becomes comfortable for a user is realizable.

  The natural language processing tool determines whether the hiragana corresponds to a constituent character of the route search condition expressed by characters other than hiragana when the request mail includes hiragana. The natural sentence analysis is performed after hiragana is converted into the constituent characters. Thereby, since the user can issue a request mail by hiragana, the conversion work into kanji or the like becomes unnecessary, and the operation burden on the user is reduced.

  In one embodiment, a landmark name representing a specific region, place, or facility input by a user, a nearest station of the landmark name, and a required time to reach the landmark name are registered in association with the identification information of the user. The natural language processing tool performs a natural sentence analysis using the registered landmark name as one of the constituent characters of the route search condition. This makes it possible to search for a route at a location away from the station.

In one embodiment, the route search tool accesses the operation information DB that stores operation information for route search including route information and timetable information of transportation when triggered by the input of the route search condition. by the execution, to acquire the diagram information of the path candidate, one or more of the time later than the output Hatsuji time at the departure point in the same line as the path candidates Hatsuji time out of the departure diagram information of the alternative pathway, or the path candidate and the operation information of the diagram information of the alternative pathway to the time earlier than the destination arrival Chakuji time and arrival Chakuji time of the destination in the route candidate of the same line Obtained from the DB , the control means merges the obtained route candidate and the diagram information of the alternative route, and generates the reply mail with the merged diagram information.
As a result, not only a specific route candidate but also diagram information of alternative routes when the train is delayed or moved with a margin can be collectively displayed as a transfer list on one screen.

The route search method of the present invention relates to a step of associating an official route name of a transportation facility with a common route name designated by a user and recording it in a predetermined memory, a route search condition including a departure place , a destination, and time data , Receiving a route search request mail in which the common route name is described in a natural sentence; and performing a natural sentence analysis on the received request mail with a natural language processing tool, thereby obtaining the route search condition and the common route In the route search tool for performing route search for transportation facilities, a route search based on the extracted route search condition is executed and associated with the same known route name as the extracted popular route name. The official route names that have been searched are retrieved from the memory, and the route candidates that include the searched official route names are narrowed down from a plurality of types of route candidates obtained by executing the route search. A route search method using electronic mail, comprising: outputting a narrowed route candidate; and creating a reply mail describing the narrowed route candidate in a natural sentence and returning the reply mail to the user terminal. It is.

The computer program of the present invention relates to a computer having a memory by associating an electronic mail means for transmitting and receiving an electronic mail with a user terminal, an official route name of a transportation facility, and a common route name designated by a user. When a route search request mail that describes the route search conditions including the route name registration means, departure point , destination, and time data and the so-called route name recorded in natural text is received in response to the request mail A natural language processing tool that extracts the route search condition and the so-called route name by performing a natural sentence analysis, and performs a route search based on the extracted route search condition, and the extracted so-called route name A plurality of types of route candidates obtained by executing the route search by searching the memory for the official route name associated with the same common route name as Narrow down the route candidates that include the searched route name from above, create a route search tool that outputs the narrowed route candidates, create a reply email describing the narrowed route candidates and diagram information in natural text, A computer program that functions as a control unit that sends a reply to the user terminal through the electronic mail unit.

  In the present invention, one or more common route names are associated with each formal route name and stored in the common route DB, so the formal route name can be designated with various names. Therefore, the user can simply narrow down the route search result with the official route name associated with the popular route name by simply describing the common name that he / she always uses in the request mail in natural text along with the route search condition. Since the result can be received by a reply mail, it becomes easy to find a route candidate suitable for the image on the display screen. This makes it possible to realize route search in a usage environment that is comfortable for the user.

An embodiment will be described in which the present invention is applied to a network type route search system that enables a route search service of a transportation system for portable terminals.
FIG. 1 is an overall configuration diagram of the route search system of this embodiment, and only characteristic portions are shown. In the present embodiment, an example of a route search system 1 that is connected to a mobile phone network N1 and the Internet N2 and that realizes a route search service for trains by using both e-mail and Web communication is shown. A user accesses the route search system 1 with a mobile terminal 30 having a mail function and a Web browser function, for example, a mobile phone.

  This route search system 1 has a server main body provided with a storage device, and the server main body executes the computer program of the present invention, whereby the server main body is divided into the main control unit 10, the mail transmitting / receiving unit 11, the natural language processing. The tool 12, the route search tool 13, the natural sentence editing tool 14, the registration tool 15, and the communication control unit 16 are made to function as well as a rule DB (DB is an abbreviation of database, the same applies hereinafter) 21, a user information DB 22, The operation information DB 23, the station name candidate DB 24, the landmark DB 25, and the so-called route DB 26 are constructed.

  The rule DB 21 includes a grammar file referred to by the natural language processing tool 12, various dictionary files, a knowledge base for semantic analysis, an extraction rule for extracting a plurality of condition candidate data as candidates for a route search condition, and a plurality of extraction rules. The candidate selection rule for selecting the optimum condition candidate data from the condition candidate data is stored. Both the condition candidate data and the optimum condition candidate data can be one of the route search conditions input to the route search tool 13.

The route search condition is a combination of a plurality of constituent characters, that is, a combination of a departure place, a destination, a waypoint, time data, a landmark name and a route name described later (including a common route name described later). is there. However, it is not a route search condition that includes all of these, but usually some combinations. What is essential is a combination of the starting point and the destination.
For the sake of convenience, in this embodiment, the condition candidate data is assumed to be station name candidate data, landmark names, time data, and route names that can be departure points, destinations, and waypoints.

Condition candidate data extraction rules include, for example, rules for correspondence of formal expression names corresponding to Hiragana / Katakana / Romaji / mixed characters, candidate expressions corresponding to other stations of the same name (for example, Nakano (Tokyo), Nakano (Gunma) Prefecture), Nakano (Nagano Prefecture), Nakano Sakaue, Nakano Sakae ...), the search condition rule when transmitting to the route search tool 14 together with the route search execution request, that is, when only two station name candidate data are input Is interpreted as searching for an average route between stations. In addition, the following rules for distinguishing the departure point from the destination are also included in the extraction rule.
・ Proverb between proper nouns “Shibuya to Shinjuku” → Departure: Shibuya, Destination: Shinjuku • Proper noun positional relationship “Shibuya Shinjuku” → Departure: Shibuya, Destination: Shinjuku / Japanese dependency Depart from Shibuya at 11:00, Shinjuku is the arrival station ”→ Departure: Shibuya, Destination: Shinjuku / Noun symbols“ Departure: Shibuya, Arrival: Shinjuku ”→ Departure: Shibuya, Destination: Shinjuku“ Shinjuku ← Shibuya ”→ Departure: Shibuya, Destination: Shinjuku, other stations based on the confirmed contents of“ To Shinjuku, Shibuya ”→ Departure: Shibuya, Destination: Shinjuku / Bunbun for forced resolution“ Shibuya is Arrival, Shinjuku arrives ”→ Departure: Shibuya, Destination: Shinjuku“ Shibuya departs from Shinjuku ”→ Departure: Shibuya, Destination: Shinjuku“ Ikebukuro to Nakano ”→ Departure:“ Ikebukuro ”, Destination : "Nakano (Tokyo)"

The candidate selection rules are as follows.
・ The station combination that minimizes the operating cost is divided into `` Nakano Ikebukuro '' → `` Nakano '' and `` Ikebukuro ''. From `` Nakano '', `` Nakano (Tokyo) '', `` Nakano (Gunma) '', "Nakano (Nagano)", "Nakano Sakaue" and "Nakano Sakae" are derived, but the operation cost is the smallest at "Nakano (Tokyo)" and at "Ikebukuro" is there.
・ If there is a history referred to in the past, use it from `` Nagano Nakano to Ikebukuro '' → `` Nakano '' is `` Nakano (Nagano Prefecture) ''
・ Priority is given to railways and airports ・ Combination of stations closest to prefectures is given highest priority

  The user information DB 22 stores user specific information, registered e-mail addresses, usage logs, and the like in association with each other by a user ID or the like.

  The operation information DB 23 stores all the operation information necessary for searching for an operation route, including routes of transportation facilities (railway / bus / aircraft), schedule information (time information), fare information, and a station list. This operation information is updated at any time.

  The station name candidate DB 24 stores all station name candidate data specified by natural sentence analysis by the natural language processing tool 12. These data are stored in association with a search ID that is assigned when an e-mail is received from the mobile terminal 30. When reading, this search ID is used as a read key.

  In the landmark DB 25, a place, facility, etc. arbitrarily selected by the user is represented by a landmark name, and the nearest station from there and the required travel time are stored in association with the user's unique information. For example, the landmark name “Val” is registered in association with the unique information of the Val Institute, Inc., and the information about this “Val” is that the nearest station is Koenji Station and the travel time to it is a 6-minute walk. sign up. Landmark names are handled in the same way as station name candidate data when searching for routes, and are used to select the total required time when the distance from the nearest station is long, and to select the optimum route when there are multiple nearest stations. .

  In the common name line DB 26, an official route name (for example, “JR Chuo / Ome Line Ome Special”) and an arbitrary common name that the user can easily remember (for example, “JR Line” or “Ome Line”) are stored in the user specific information. Stored in association with. Since the mobile terminal 30 cannot display many routes on the display screen, it is difficult to find the route even if there is an image of the route to be used. Even when a route name is desired to be specified as a route search condition (restriction condition), it is difficult for the user to input an official route name (displayed route name) accurately. Therefore, a single official route name is registered by associating a plurality of common route names with simplified route names, so that one official route name can be designated by various names. The common route name may not be unique in the system. By handling a plurality of official route names on an equal basis from common route names designated by the user, it is possible to easily narrow down route search results.

  The main control unit 10 controls operations including activation of the following functional blocks 11 to 16 in order to realize various processes related to route search. The control procedure will be described later.

The mail transmission / reception unit 11 exchanges an electronic mail using the registered mail address with the mobile terminal 30 via the mobile phone network N1. Hereinafter, an e-mail including a route search request text is referred to as a request mail, and an e-mail including the response text is referred to as a reply mail.
At the time of reception, the registered mail address of the user is extracted from the header part of the request mail and temporarily recorded in an address table (not shown). When replying, a reply mail is sent to the registered mail address recorded in this address table. The communication control unit 16 performs control for providing a Web service to the mobile terminal 30 accessed via the Internet N2 based on authority information described later.

  The natural language processing tool 12 is based on a grammar file, a dictionary file, a knowledge base, and various rules stored in the rule DB 21 for a request sentence described in, for example, a body part of a request mail received by the mail transmission / reception unit 11. Then, a natural sentence analysis process is performed to extract a plurality of station name candidate data, time data, and the like.

FIG. 2 is a procedure explanatory diagram of natural sentence analysis processing by the natural language processing tool 12. Here, it is assumed that a natural sentence (partly from Tokyo at 3 o'clock) (partial half-width characters, registered landmarks) is input.
The natural language processing tool 12 first performs notation control in preprocessing. Notation control is a process of changing a half-width character to a full-width character, deleting a binary or line feed, and cutting the number of characters by a specified number of characters. As a result, the grammar for parsing can be simplified.

  Then, hiragana query determination is performed. This process is a process of interpreting (extracting) a hiragana query as a constituent character representing a station name (or landmark) using a request sentence as a query (a request command that can be recognized by a computer). Further, when the extracted station name candidate data is one, it is interpreted that the use is unknown, and when there are two station name candidate data, there is a departure place and a destination. When there are three station name candidate data, it is interpreted that there is a departure point, a waypoint, and a destination. After hiragana query determination, landmark name conversion is performed. Landmark conversion is a process for finding a landmark name from a query and converting the landmark name into a character string for morphological analysis.

  After the above series of processing is completed, Japanese analysis is performed. Japanese analysis is a process of converting a query sentence into a sequence of Japanese morphemes and phrases. In this embodiment, the morpheme is converted into a phrase ID and stored in a memory area (not shown). As a result, the result of Japanese analysis is a sequence of phrase IDs. The phrase ID is obtained by assigning numerical values in the order of appearance to words that appear in the grammar file stored in the rule DB 21. By handling the ID, it is possible to perform the extraction process of the station name candidate data that takes a very large load at a high speed.

  When the Japanese language analysis is completed, the natural language processing tool 12 acquires the Japanese language analysis result and extracts case candidates using the dictionary file of the rule DB 21. “Case” is intended to give meaning to morphemes and clauses (“He is” is a case, “He” is a case), and “Candidate” is a group of clauses that can form a “case”. Refers to the structure of the extracted data for all possibilities. Since they are treated as candidates, those that are not finally selected as “cases” are also extracted. For example, since “Tokyo” can be interpreted as “Tokyo Station” in addition to the place name, two case candidates are extracted. At 3 o'clock, the expression is unified as “h = 3”.

  After the case candidate is extracted, the signature case candidate is deleted. The signature is a fixed sentence written at the end of the body part of the e-mail. If there is a person name or address notation that can be taken as a station name, such as “Taro Shinagawa”, there is a possibility of misrecognizing it as station name candidate data, so the case candidate is deleted from the line that seems to be a signature. Then, case candidate optimization is performed. In other words, the combination of the extracted case candidates is evaluated as a brute force to determine the best interpretation for the entire sentence. Semantic analysis is performed on the case candidates determined in this way. Here, in particular, time and date are interpreted, and specific dates and times are expressed numerically. As a result, the natural language processing tool 12 provides an output on the upper right side of the figure.

  In this way, the words, phrases, and sentences included in the request mail are extracted, and a plurality of station name candidate data are extracted based on the extracted words, phrases, or sentences and the condition candidate extraction rules stored in the rule DB 21. To do. If route names and time data are included, these are also extracted. In addition, based on the candidate selection rules stored in the rule DB 21, among the combinations of a plurality of station name candidate data (combinations including these when a route and time data are included), the above condition selection rules are followed. Any one set that is most likely is selected as a route search condition that is a combination of optimum station name candidate data that is estimated to reflect the user's intention.

  The route search tool 13 is triggered by the input of the route search conditions, and based on the information accumulated in the operation information DB 23, the landmark DB 25, and the so-called route DB 26, the next station (or the lowest operation cost by transportation) (or By repeating the search process of (landmark name) like a ripple when a stone is thrown on the water surface, a route candidate according to the route search condition and its diagram information are specified. The operation cost is usually for time, but distance and fare may be used, or a combination of these may be used. The route search tool 13 is also equipped with an inference engine, and if it becomes clear that the route becomes unrealistic during the search, the route search tool 13 stops the search after the next station on the route. In the route search tool 13, for example, the route search function portion of the route search software “Station Spat” (registered trademark) provided by the applicant of the present application can be used.

  In the present embodiment, the natural language processing tool 12 selects an optimum combination among the plurality of station name candidate data (including the route name and time data if present) extracted from the request sentence as described above. Is input to the route search tool 13, but when there are a large number of route candidates, the route candidates are narrowed down by a common route name.

This will be described in detail with reference to FIGS.
As described above, a plurality of common route names are registered in advance in the common name DB 26 from official route names used for display. When performing a route search, as shown in FIG. 3, the user designates the name of the route he / she wants to use in addition to the departure point and destination (step S <b> 11). The route search tool 13 performs a route search (step S12), and in the process of specifying a plurality of route candidates (step S13), accesses the so-called route DB 26 to perform a so-called route name search (step S14), and is designated by the user. All the official route names associated with the known route name that matches the known route name are acquired (step S15). Then, the plurality of route candidates specified in step S13 are narrowed down to only route candidates including the acquired official route name (step S16), and the narrowed route candidates are output as search results (step S17).

In the example of FIG. 4, if the user designates the so-called route name A, the official route name # 1 and the official route name # 2 associated with the so-called route name A are acquired from the so-called route DB 26. “Route # 1”, “Route # 2”, and “Route # 3” are specified as the route search results, and “Route # 1” includes the official route name # 3 and the official route name # 4. “2” includes the official route name # 1 and the official route name # 5, and “route # 3” includes the official route name # 6. As a result of the search, only “route # 2” is searched. Is output as
The above is an example in the case of using a route candidate designated by the user. However, if the designated route candidate is not used, the route candidate including the official route name may not be output.

  When the route candidate and the diagram information thereof are specified, the route search tool 13 determines whether the next optimum diagram (an efficient diagram that departs after the departure time + 1 minute) or the previous one on the same route as the candidate route. Assemble the schedule information of the alternative train that will be the optimal schedule (an efficient schedule that can arrive by arrival time minus 1 minute). There may be a plurality of route candidates. In this case, “route 1”, “route 2”, and “route 3” are specified in ascending order of operation cost, and the schedule information of the alternative route is specified in each route.

The natural sentence editing tool 14 merges (integrates) the route candidate and alternative route diagram information obtained by the route search tool 13. When there are a plurality of route candidates, each route candidate is merged with the diagram information of the alternative route. Then, the route candidate and the merged diamond information are used as an initial search result, and the initial search result, the search ID, and authority information for the sender of the request mail to access the route search tool 13 later by Web communication are obtained. Then, it is edited as an answer sentence to be displayed on one screen of the portable terminal 30, and a reply mail including the answer sentence is created. At this time, since the display of the portable terminal 30 is small and the resolution is low, it is possible to remove unnecessary information and display it with as few packets as possible and on one screen as much as possible, but edit so that necessary information is included.
This reply mail is returned from the mail transmitting / receiving unit 11 to the sender of the request mail.

  The registration tool 15 accepts registration of user specific information, e-mail address, landmark name, popular route name, various rules, service condition setting, and the like. The registered user specific information is stored in the user information DB 21. The registered email address becomes the registered email address. The landmark name is stored in the landmark DB 25. The so-called route name is stored in the so-called route DB 26.

[Operational form]
Next, an operation mode example of the route search system 1 configured as described above will be described.

  The user starts the operation of the route search system 1 by sending a request mail to the dedicated address of the route search system 1 using the registered mail address. In this example, as shown in FIG. 5, it is assumed that the user has transmitted a request mail describing a request sentence such as “Val Tsukuba Express Asakusa”. “Val” is the landmark name of the Val Institute, which takes 5 minutes to reach JR Koenji Station, and Tsukuba Express is a route. This landmark name is also treated as one of the station name candidate data.

  The main control unit 10 starts the control operation in the procedure of FIG. That is, when this request mail is received (step S101: Yes), the registered mail address of the sender is recorded in an address memory (not shown) (step S102). Further, a search ID is assigned to the request mail (step S103), and this search ID is stored in the station name candidate DB 24 (step S104).

  When the search ID is stored, the main control unit 10 activates the natural language processing tool 12. The natural language processing tool 12 extracts a plurality of candidate station name data by performing a natural sentence analysis on the request sentence described in the request mail. The extracted plurality of station name candidate data is stored in association with the search ID stored in the station name candidate DB 24 (step 105). When the storage is completed, the natural language processing tool 12 selects a set of station name candidate data to be optimum condition candidate data from among a plurality of station name candidate data combinations.

  In the case of this example, the optimum condition candidate data is selected with “Val (landmark name)” as the departure point, “Tsukuba Express” as the route, and “Asakusa Station” as the destination. When the optimum condition candidate data is selected, among the station name candidate data stored in the station name candidate DB 24, the selected tag is assigned to the corresponding station name candidate data, and the route search is performed using the optimum condition candidate data as the route search condition. Input to the tool 13 (step S106). The tag is, for example, to prevent “Val” from being selected again as a starting point, but is not essential.

  When the route search result is received from the route search tool 13 (step S107: Yes), the main control unit 10 activates the natural sentence editing tool 14. The natural sentence editing tool 14 merges route candidate and alternative route diagram information included in the route search result, and authority information and search ID for accessing the route search tool 13 from the user terminal 30 through Web communication. It is edited as a reply message for mail (step S108). As the authority information, the registered user's unique information and the above search ID are parameterized and used as a set with the URL (Uniform Resource Locator) of the route search tool 14.

  When the response sentence is edited by the natural sentence editing tool 13, the main control unit 10 controls the mail transmitting / receiving unit 11 to read the registered mail address of the sender (step S109), and the answer sentence is registered address of the sender. A reply is made to the address (step S110).

  The reply sentences included in the reply mail are illustrated in FIGS. In the example of FIG. 8, the display screen 50 of the mobile terminal 30 shows route candidates for [Route 1] and diagram information of the alternative routes. FIG. 9 shows a route candidate for [Route 2] and its schedule information, and FIG. 10 shows a route candidate for [Route 3] and its schedule information. Whether or not the route search condition does not conform to the user's intention is immediately known from the search result for [Route 1]. Therefore, authority information 52 is described at the bottom of the first display screen 50.

  When the display screens shown in FIGS. 8 to 10 are intended, the user completes a series of processes for route search without taking any action. In other words, the user's purpose is achieved by one-time transmission of an e-mail by a natural sentence by the user and a reply e-mail due to an automatic response. For this reason, a route guidance e-mail is added to the text as in the system of Patent Document 4, for example. There is no need to quote and reply, not only the operation is simplified, but also waste of communication resources is avoided.

  On the other hand, if the search result is not intended, the cursor is placed on the authority information 52 described at the bottom of the display screen 50 in FIG. 8 and clicked, via the network N2 and the communication control unit 16. Thus, it is possible to seamlessly access the route search tool 13 on the Web. FIG. 6 is a processing procedure diagram of the main control unit 10 during Web access.

  In FIG. 6, when the main control unit 10 detects an access (web access) from the mobile terminal 30 (step S201: Yes), the main control unit 10 displays an input screen for route search conditions on the screen of the mobile terminal 30 (step S202). At this time, the station name candidate data stored in the station name candidate DB 24 and specified by the search ID is read from the station name candidate DB 24, and these can be selected by the cursor for each departure place, waypoint, destination, and route. (Step S203).

  FIG. 11 is an example of an input screen for route search conditions. Four input areas are formed on the display screen 50. From each input area, already extracted station name candidate data (including landmark names) and routes can be selected. As default, the top data, that is, the optimum station name candidate data is automatically entered. “Not specified” is included because the transit point and route are not mandatory input conditions. When “Not specified” is selected, it is treated as no condition.

When the user selects one target station name candidate data for each input area by a cursor operation and clicks the execution button image 55 on the display screen 50 to instruct (Step S204: Yes), the main control unit 10 changes. The route search condition thus entered is input to the route search tool 14 (step S205). Thereafter, the main control unit 10 ends the web communication (web access) (step S206). The subsequent processing procedure is the same as that after step S107 in FIG. That is, the user can receive the route search result according to his / her intention by a reply mail including the answer sentences as shown in FIGS.
Note that the route search result after the condition change may be displayed on the display screen 50 in a state where Web communication is maintained instead of the reply mail again.

  As described above, in the route search system 1 according to the present embodiment, all station name candidate data is transferred to the station name candidate DB 24 at the stage where the station name candidate data is obtained even if it is a Japanese natural sentence that is ambiguous for the user. Since the station name candidate data originally intended by the user can be selected using Web communication when the interpretation of the natural language processing tool 12 is incorrect, the convenience of route search by e-mail can be selected. Can be supplemented by Web communication. Although the currently established natural language processing technology does not necessarily provide a complete result, according to the present embodiment, after receiving a reply mail, the interpretation can be further changed, which is an effect that cannot be obtained from the conventional technology. Be expected.

  In the present embodiment, the natural language processing tool 12 makes a hiragana query determination, so that even when the request mail includes hiragana, the natural language processing tool 12 can convert the hiragana into natural characters such as kanji. Since sentence analysis is performed, the work of character conversion is reduced when creating a request sentence.

  In this embodiment, since the route information and the alternative route diagram information that can be used before and after the route candidate are merged and displayed on one display screen 50, the route guidance information can be displayed even when the interactive interface is not provided. Can be provided efficiently.

  In the present embodiment, a landmark name representing a specific region, place, or facility input by the user, a nearest station of the landmark name, and a time required to reach the landmark name are associated with the identification information of the user and the landmark. Since the DB 25 is provided and the natural language processing tool 12 includes the registered landmark name as one of the starting point, the waypoint or the destination, the natural language analysis is performed. It is possible to quickly specify conditions including the above. In the case of the portable terminal 30, the act of inputting characters itself is troublesome, and it is desirable to input a route search condition with as short a word as possible. However, according to the present embodiment, this is possible. Since the landmark name may be a private facility such as a home, a company, or a friend's house, a route search specialized for the user can be performed.

  In the present embodiment, for each official route name of the transportation facility, one or a plurality of common route names arbitrarily designated by the user are registered in the common route DB 26 in association with the identification information of the user, Since the search tool 13 uses any registered common route name as a narrowing-down condition when there are a plurality of route candidates, it is possible to “narrow down” the search result according to the user's preference. In particular, in the case of the mobile terminal 30, the operating environment is not sufficient, and many routes cannot be displayed on the display screen 50, so even if there is an image of the route to be used, it is difficult to find the route. According to the embodiment, it becomes easy.

  In the present embodiment, the description has been made on the assumption that the interpretation of a natural sentence having ambiguity is not limited to one. However, due to a shift in the extraction area of words from a request mail, Even when characters other than the station name are erroneously recognized as the station name candidate data, the present invention can be applied as a procedure for changing the interpretation to the correct station name candidate data.

  In this embodiment, an example in which station name candidate data is used as one of the condition candidate data is shown. However, when a bus is incorporated as a transportation facility, the station name candidate data becomes a bus stop candidate data, and an aircraft is also incorporated. In this case, the station name candidate data is airport name candidate data.

  In the present embodiment, an example in which the route search service is provided using the electronic mail of the mobile terminal 30 has been described. However, electronic mail by a personal computer can also be used.

1 is an overall configuration diagram of a route search system to which the present invention is applied. Explanatory drawing of the procedure of the natural sentence analysis process by a natural language processing tool. Explanatory drawing when nickname route name is specified and route candidates are narrowed down. Explanatory drawing which showed the state by which the route candidate was narrowed down. The figure which showed the control procedure of the main control part until a user requests route search by request mail and receives the result by reply mail. The figure which showed the control procedure of the main control part when a user corrects the conditions of route search by Web access and reacquires a route search result. The figure which showed the example of the request sentence contained in request mail. The figure which showed the example of the reply sentence containing the route candidate about [route 1] displayed on the display screen of a portable terminal, and the diagram information of the alternative route. The figure which showed the example of the reply sentence containing the route candidate about [route 2] displayed on the display screen of a portable terminal, and the diagram information of the alternative route. The figure which showed the example of the reply sentence containing the route information about [route 3] displayed on the display screen of a portable terminal, and the diagram information of the alternative route. An input screen for route search conditions displayed on the display screen of the mobile terminal by Web access.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Route search system, 10 ... Main control part, 11 ... Mail transmission / reception part, 12 ... Natural language processing tool, 13 ... Path search tool, 14 ... Natural sentence editing tool, DESCRIPTION OF SYMBOLS 15 ... Registration tool, 16 ... Communication control part, 21 ... Rule DB, 22 ... User information DB, 23 ... Operation information DB, 24 ... Station name candidate DB, 25 ... Landmark DB, 26 ... Commonly known route DB, 30 ... Mobile terminal, 50 ... Display screen, 52 ... Authority information, 55 ... Execution button image, N1 ... Mobile phone network, N2 ···the Internet.

Claims (6)

An email means for sending and receiving emails to and from the user terminal;
A well-known route DB in which the official route name of the transportation facility and the well-known route name designated by the user are recorded in association with each other
When the e-mail means receives a route search request mail in which a route search condition including a starting point , a destination, and time data and the so-called route name are described in a natural sentence, a natural sentence analysis is performed on the request mail. A natural language processing tool for extracting the route search condition and the so-called route name from the request mail;
The route search is executed based on the extracted route search condition, and the official route name associated with the same popular route name as the extracted popular route name is retrieved from the popular route DB, and the route is searched. A route search tool that narrows down the route candidates including the searched official route name from a plurality of types of route candidates obtained by performing the search, and outputs the narrowed route candidates;
A control means for creating a reply mail describing the narrowed route candidates and diagram information in a natural sentence, and for replying the reply mail to the user terminal through the electronic mail means;
An operation route search system. The natural language processing tool determines whether the hiragana corresponds to a constituent character of the route search condition expressed in characters other than hiragana when the hiragana is included in the request mail. The natural sentence analysis is performed after the hiragana is converted into the constituent characters.
The route search system according to claim 1. A landmark DB that records a landmark name representing a specific area, place, or facility input by the user, the nearest station of the landmark name, and the time required to reach the landmark name in association with the identification information of the user. And
The natural language processing tool performs the natural sentence analysis by including a landmark name recorded in the landmark DB in one of the constituent characters of the route search condition.
The route search system according to claim 1 or 2. The route search tool, when triggered by the input of the route search conditions , accesses the operation information DB storing the operation information for route search including route and diagram information of transportation, and executes the route search. It acquires the diagram information of the path candidate, diagram information of one or more alternative paths for the slower time than out Hatsuji time at the departure point in the same line as the path candidates and outgoing Hatsuji time the departure point, or acquires diagram information of the alternative pathway to the time earlier than the destination arrival Chakuji time and its destination arrival Chakuji time in the route candidate of the same line as the path candidate from the operation information DB,
The control means merges the acquired route candidate and the diamond information of the alternative route, and generates the reply mail with the merged diamond information,
The route search system according to claim 1, 2 or 3. Correlating the official route name of the transportation with a common route name specified by the user in a predetermined memory;
Receiving a route search request mail describing a route search condition including a departure point , a destination, and time data and the so-called route name in a natural sentence;
Extracting the route search condition and the so-called route name from the request email by performing natural sentence analysis on the received request email with a natural language processing tool;
In the route search tool for performing route search for transportation, the route search based on the extracted route search condition is executed, and the official route name associated with the same popular route name as the extracted popular route name is Searching from the memory, narrowing down the route candidates including the searched official route name from a plurality of types of route candidates obtained by executing the route search, and outputting the narrowed route candidates;
A route search method by electronic mail, comprising: creating a reply mail describing the narrowed route candidates in a natural sentence and returning the reply mail to the user terminal. A computer with memory,
E-mail means for sending and receiving e-mail to and from user terminals;
A so-called route name registration means for associating and recording the official route name of the transportation facility and the so-called route name designated by the user in the memory;
By performing a natural sentence analysis on the request mail when receiving a route search request mail describing the route search condition including the departure point , destination and time data and the so-called route name in a natural sentence, A natural language processing tool for extracting a route search condition and the so-called route name;
The route search is executed based on the extracted route search condition, and the official route name associated with the same popular route name as the extracted common route name is retrieved from the memory, and the route search is executed. A route search tool that narrows down the route candidates including the official route name searched out from a plurality of types of route candidates obtained by the above, and outputs the narrowed route candidates;
Control means for creating a reply mail describing the narrowed route candidates and diagram information in a natural sentence, and for replying the reply mail to the user terminal through the electronic mail means;
As a computer program.

Images