Travel plan creation support system and travel plan creation support method

Abstract

To enable proposing an appropriate itinerary that suits taste of a person and in which influence due to congestion is suppressed and travelling can be efficiently performed, to the person for whom a concrete travel destination is unconfirmed.SOLUTION: A travel plan creation support system 10 comprises a device 100 including: a storage device 101 that stores information associated with a travel destination candidate site; and an arithmetic device 104 that identifies a plurality of candidate sites which are adapted to a search condition on the basis of the search condition specified by a user and the information, selects a candidate site and moving means which satisfy a reference concerning a congestion status on the basis of information or the like associated with congestion of each of the plurality of candidate sites, creates a plurality of itinerary proposals defining a visit order and moving means thereof, evaluates each of itinerary proposals according to a congestion status of the candidate site and the moving means, and selects an itinerary proposal having relatively high evaluation to output.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a travel plan creation support system and a travel plan creation support method. [Background technology]

[0002] Many people who want to travel privately have a vague idea of ​​their purpose and destination, and plan their trip based on that idea through trial and error. In such cases, selecting specific travel destinations and creating a travel plan that fits their budget and time is a very time-consuming task. Furthermore, creating a travel plan that suits your preferences while taking into account the expected congestion at the destination, road congestion, and congestion on public transportation is even more time-consuming.

[0003] For this reason, various conventional technologies have been proposed to assist in creating travel plans, such as an information processing system (see Patent Document 1) that creates a travel plan for efficiently visiting designated spots even for a trip that spans multiple days.

[0004] This information processing system includes a travel plan creation condition acquisition means for acquiring travel plan creation conditions including a plurality of spots, a visiting order determination means for determining an order in which to visit the plurality of spots, a travel plan creation means for creating a travel plan by allocating the plurality of spots across a plurality of days based on the determined visiting order and daily activity time periods, and an information output means for outputting the travel plan. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2019-046256 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, the conventional technology is not structured to support the creation of appropriate travel plans for users who are unclear about the main elements of their trip, such as their destination and stopovers. In other words, the technology is intended only for users who have a clear destination and stopovers from the beginning, and it is difficult to say that it is immediately useful for other users.

[0007] Therefore, an object of the present invention is to provide a technology that can propose appropriate itineraries to people who have not yet decided on a specific travel destination, which match their preferences, reduce the impact of congestion, and enable them to travel efficiently. [Means for solving the problem]

[0008] The travel plan creation support system of the present invention, which solves the above problems, includes a storage device that stores information about potential travel destinations, and a user-specified Travel planning Based on the conditions and the information, Travel planning conditions a process of selecting an image set from a group of images for estimating a user's preferences among the information according to at least the user's image of the trip, a process of presenting the image set to the user and acquiring a selection result of images that are close to the image of the travel destination by the user, a process of calculating a similarity between each image included in the selection result and predetermined information regarding each spot among the information, and a list of spots whose similarity is equal to or exceeds a threshold as a list of candidate spots. A process of identifying Included in the list Information about congestion at each of the proposed locations; and The above Based on information on the means of transportation to each candidate site, at least the congestion situation and duration of stay at the candidate site Select candidate locations and transportation methods that meet the criteria for and The candidate site a process of generating a directed graph by setting links between nodes when movement between the nodes is possible based on the relationship between the positions of the departure point and the return point and the departure and arrival times, and the congestion situation and the stay time conditions are satisfied; and a process of applying a predetermined optimization algorithm to each route between the nodes shown in the directed graph to generate a directed graph by setting links between the nodes corresponding to the congestion situation and the stay time. A process of creating a plurality of itinerary plans that specifies the order of visits between candidate locations and the means of transportation, and a process of determining the congestion status of the candidate locations and the means of transportation included in each of the plurality of itinerary plans. and the duration of stay and a calculation device that executes a process of selecting an itinerary plan with a relatively high evaluation and outputting the selected itinerary plan to a specified device.

[0009] In addition, the travel plan creation support method of the present invention is a method in which an information processing system stores information on potential travel destinations in a storage device and Travel planning Based on the conditions and the information, Travel planning conditions a process of selecting an image set from a group of images for estimating a user's preferences among the information according to at least the user's image of the trip, a process of presenting the image set to the user and acquiring a selection result of images that are close to the image of the travel destination by the user, a process of calculating a similarity between each image included in the selection result and predetermined information regarding each spot among the information, and a list of spots whose similarity is equal to or exceeds a threshold as a list of candidate spots. A process of identifying Included in the list Information about congestion at each of the proposed locations; and The above Based on information on the means of transportation to each candidate site, at least the congestion situation and duration of stay at the candidate site Select candidate locations and transportation methods that meet the criteria for A process of The candidate site a process of generating a directed graph by setting links between nodes when movement between the nodes is possible based on the relationship between the positions of the departure point and the return point and the departure and arrival times, and the congestion situation and the stay time conditions are satisfied; and a process of applying a predetermined optimization algorithm to each route between the nodes shown in the directed graph to generate a directed graph by setting links between the nodes corresponding to the congestion situation and the stay time. A process of creating a plurality of itinerary plans that specifies the order of visits between candidate locations and the means of transportation, and a process of determining the congestion status of the candidate locations and the means of transportation included in each of the plurality of itinerary plans. and the said stay time and selecting an itinerary plan with a relatively high rating and outputting the selected itinerary plan to a specified device. Effect of the Invention

[0010] According to the present invention, it is possible to propose an appropriate itinerary to a person who has not yet decided on a specific travel destination, which matches the person's preferences, reduces the effects of congestion, and enables the person to travel efficiently. [Brief description of the drawings]

[0011] [Figure 1] 1 is a diagram illustrating an example of the configuration of a travel plan creation support system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram showing an example of the configuration of a reference image database in the present embodiment. [Diagram 3] FIG. 2 is a diagram showing an example of the configuration of a travel article database in the present embodiment. [Figure 4] FIG. 2 is a diagram illustrating an example of a configuration of an area information database in the present embodiment. [Diagram 5] FIG. 2 is a diagram showing an example of the configuration of a spot information database in the present embodiment. [Figure 6] FIG. 2 is a diagram showing an example sequence of a travel plan creation support method according to the present embodiment. [Figure 7] FIG. 4 is a diagram showing an example of a screen in the present embodiment. [Figure 8] FIG. 4 is a diagram showing an example of a screen in the present embodiment. [Figure 9] FIG. 11 is a diagram showing an example of a flow of a candidate area extraction process in the present embodiment. [Figure 10] FIG. 11 is a diagram illustrating an example of a flow of a keyword estimation process in the present embodiment. [Figure 11] FIG. 4 is a diagram showing an example of a flow of a process for creating itinerary candidates in the present embodiment. [Figure 12] FIG. 2 is a diagram showing an example of itinerary candidate information according to the embodiment; [Figure 13] FIG. 4 is a diagram showing an example of itinerary information in the present embodiment. [Figure 14] FIG. 2 is a diagram showing an example of a passenger behavior network creation process in this embodiment. [Figure 15] FIG. 2 is a diagram showing an example of a passenger behavior network in this embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] <System configuration> An embodiment of the present invention will be described in detail below with reference to the drawings. Fig. 1 is a diagram showing an example of the configuration of a travel plan creation support system 10 according to this embodiment. The travel plan creation support system 10 shown in Fig. 1 is a computer system that can propose an appropriate itinerary to a person who has not yet decided on a specific travel destination, which matches the person's preferences, reduces the impact of congestion, and enables the person to travel efficiently.

[0013] 1, in a travel plan creation support system 10 of this embodiment, a server 100 and a terminal device 200 are communicably connected via a network 1. These are collectively referred to as a travel plan creation support system.

[0014] The terminal device 200 is a terminal of a user who wishes to create and present a travel plan. Specifically, the terminal device 200 may be a smartphone, a tablet terminal, a personal computer, or the like, but is not limited to these.

[0015] <Hardware configuration> The server 100 (travel plan creation support device) that mainly constitutes the travel plan creation support system 10 of this embodiment includes a storage device 101, a main storage device 103, an arithmetic device 104, The system includes a communication device 105 and a database group 106 .

[0016] Of these, the storage device 101 is configured with an appropriate non-volatile storage element such as an SSD (Solid State Drive) or a hard disk drive.

[0017] The main memory device 103 is composed of a volatile memory element such as a RAM (Random Access Memory).

[0018] The arithmetic device 104 is a CPU (Central Processing Unit) that reads out a program stored in the storage device 101 into the main storage device 103 and executes it to generally control the device itself and also performs various types of judgment, calculation, and control processing.

[0019] The communication device 105 is connected to the network 1 and is responsible for communication processing with the terminal device 200, and may be, for example, a network interface card.

[0020] Furthermore, the database group 106 is a collection of appropriate databases, such as an RDB (Relational Database), and in this embodiment includes a reference image database 125, a travel article database 126, an area information database 127, a spot information database 128, and a route information database 129.

[0021] Of these, the reference image database 125 and the travel article database 126 are databases that store information used in processing by the candidate area extraction program 111 (described later).

[0022] The area information database 127 is a database that stores area information (described later), and the spot information database 128 is a database that stores spot information (described later).

[0023] The route information database 129 is a database that accumulates information used by, for example, the itinerary candidate creation program 112 (described later) in route searches related to departure points, destinations, and travel between spots. The information accumulated in the route information database 129 includes congestion information by day and time period. The congestion information is a value statistically predicted from past performance, and is appropriately acquired and updated from an external traffic control system or the like at a predetermined time interval, such as once a day. Of course, the information may be updated in real time based on operation information of public transportation such as trains and airplanes.

[0024] The server 100 and the terminal device 200 may be integrated without using the network 1. Furthermore, the travel plan creation support system 10 can be configured as a stand-alone machine by using pre-stored information for information updated via the network 1, such as the database group 106, or updating via an external medium such as a CD (compact disc).

[0025] Also stored in storage device 101 are a travel plan creation program 110, a candidate area extraction program 111, and an itinerary candidate creation program 112 as programs for executing functions required by the travel plan creation support device of this embodiment.

[0026] Among these, the travel plan creation program 110 is a program for controlling the overall flow of the travel plan creation support method of the present invention. While the travel plan creation program 110 is executed by the arithmetic device 104, the candidate area extraction program 111 and the itinerary candidate creation program 112 are executed by the arithmetic device 104 to perform necessary processing. become.

[0027] In addition, the candidate area extraction program 111 is a program that extracts candidate areas to be visited based on the travel plan creation conditions specified by the user and the selected image information selected by the user according to his / her travel image, etc.

[0028] In addition, the itinerary candidate creation program 112 is a program that appropriately selects candidate visiting spots based on the candidate areas extracted by the candidate area extraction program, creates a passenger behavior network taking into account the congestion situation (details will be described later, see Figures 14 and 15), evaluates each route, i.e., the itinerary, in the passenger behavior network taking into account the congestion situation and stay time, and selects an itinerary candidate that is suitable for the travel plan creation conditions specified by the user.

[0029] Additionally, as information used when executing these programs 110 to 112 or obtained by executing these programs, travel plan creation conditions 120, a preference estimation image set 121, selected image information 122, and itinerary candidate list 123 are also held in the storage device 101. Details of this information will be described later.

[0030] On the other hand, the terminal device 200 includes a control unit 201 , a communication unit 202 , an operation input unit 203 , and a display unit 204 .

[0031] Of these, the control unit 201 is a CPU (Central Processing Unit) that executes programs stored in an appropriate storage device to perform overall control of the device itself, as well as various types of judgments, calculations, and control processes. Processing Unit).

[0032] The control unit 201 has the functions of a condition acquisition unit 210 and an information output unit 211. Of these, the condition acquisition unit 210 cooperates with the server 100 to accept and acquire user input of conditions for creating a travel plan. The information output unit 211 outputs information required for processing by the condition acquisition unit 210 to the display unit 204, and also outputs processing results (e.g., itinerary candidates, etc.) by the server 100 to the display unit 204.

[0033] The communication unit 202 is a network interface card or the like that is connected to the network 1 and handles communication processing with the server 100 .

[0034] The operation input unit 203 receives key input or voice input from the user, and is configured with input devices such as a keyboard and a mouse. Note that the operation input unit 203 does not necessarily need to be configured with a keyboard or the like as hardware, and may be configured as a combination of hardware and software, such as a combination of a touch panel and a software keyboard.

[0035] The display unit 204 is configured as an output device such as a display that displays data processed by the control unit 201 .

[0036] <Data structure example> Next, a description will be given of various types of information used by the travel plan creation support system 100 of this embodiment. Fig. 2 shows an example of the configuration of the reference image database 125 in this embodiment.

[0037] The reference image database 125 of this embodiment is a database that stores a group of images that are delivered to the terminal device 200 and displayed to present as a selection of images that fit the user's travel image. Note that, as an example of this group of images, the image set for preference estimation in FIG. The values ​​are shown as .

[0038] This reference image database 125 is a collection of records that link data such as the feature vector obtained for an image and the image file of the image, using an image ID that uniquely identifies each image as a key.

[0039] The above-mentioned feature vector is obtained by inputting the original image (image file) to a machine learning model (hereinafter also referred to as a "feature vector extractor") (not shown) (obviously, this is not limited to this). The above-mentioned model is, for example, an image classification model made up of a multi-layer neural network that is machine-learned (supervised learning using association between image data and tags) based on a predetermined large-scale data set (a group of images tagged with objects that appear in the image), and then classified into multiple classes using a softmax function or the like in the final layer. The parts being classified are removed. The output value (embedding vector) of such a model (feature vector extractor) becomes the feature vector. The concept of feature vector explained here The methods for obtaining the training data and the training results will be the same in the following explanations (a common trained model will be used).

[0040] 3 shows an example of the configuration of the travel article database 126 in this embodiment. The travel article database 126 in this embodiment is a database that stores information on travel articles published in a predetermined medium (including the Internet). This travel article information is, for example, periodically acquired and updated by the server 100 by accessing an appropriate travel information site via the network 1.

[0041] This travel article database 126 is a collection of records that link together, for example, an article ID that uniquely identifies each article as a key, image data obtained for the article (e.g., an image of a travel destination), text data (e.g., an explanatory text describing the conditions at the travel destination) that is placed around the image, and a feature vector of the image data. This feature vector is obtained using a feature vector extractor similar to that used in the reference image database 125 described above.

[0042] 4 shows an example of the configuration of the area information database 127 in this embodiment. The area information database 127 in this embodiment is a database that stores information on articles introducing local areas that are published in media or the like different from the data resources of the above-mentioned travel article database 126. The articles introducing local areas are, for example, periodically acquired and updated by the server 100 by accessing the network 1 from appropriate local information sites.

[0043] Such area information database 127 is a collection of records that link together, using the area ID as a key, data such as a spot ID list for identifying spots such as parks and art museums included in the area identified by the area ID, a local introduction article ID for identifying a local introduction article that introduces the area, text data included in the local introduction article (e.g., explanatory text describing the appearance of popular spots, etc. of the area), a keyword list extracted from the text data (e.g., a list of vocabulary expressing various phenomena such as the atmosphere, quality, grandeur, richness, convenience, etc. of nature, history, culture, etc.), and the expected congestion level by time period and the acceptable congestion level for the area.

[0044] Among these, the expected congestion level by time period is, for example, the value of the congestion level in the area predicted by day and time period. This value is obtained by using past statistical values ​​observed regarding the congestion level in the area, or is periodically obtained and updated from an external demand forecasting system.

[0045] The acceptable congestion level indicates the level of congestion that the local government or tourism business operator that manages the area can tolerate for the area. The acceptable congestion level is, for example, a value such as the ratio of the number of visitors to the maximum expected number of visitors in the area, and is a value that the local government or tourism business operator has determined in advance with the intention of avoiding overtourism.

[0046] The value of the acceptable congestion level is used to determine whether or not to accept travelers in the area. The acceptable congestion level may be managed at a granularity by time period.

[0047] 5 shows an example of the configuration of the spot information database 128 in this embodiment. The spot information database 128 in this embodiment is a database that stores information on each spot that exists in each area and that also appears in each record of the area information database 127 described above.

[0048] Such a spot information database 128 is a collection of records that link, for example, a spot ID that uniquely identifies the spot as a key, data such as the spot's arrival / departure point pair list, visit time period, acceptable congestion level, expected congestion level by time period, standard stay time by congestion level, keyword list, standard stay cost, and spot feature vector.

[0049] The arrival / departure point pair list is information for dealing with spots where the arrival point and departure point are different, such as river boat trips or sightseeing boat trips, and is a list of frequently used pairs of arrival points and departure points for multiple arrival points to the spot and multiple departure points from the spot.

[0050] A passenger behavior network (described later) is created on the assumption that to travel from another spot to the spot in question, passengers will arrive at one of the arrival points registered in the arrival / departure point pair list, and to travel from the spot to another spot, passengers will depart from one of the departure points registered in the arrival / departure point pair list.

[0051] The visitable time period is information that indicates the time period during which the spot can be used. For example, if the spot of interest is an art museum, this information corresponds to the opening and closing times of the art museum. The acceptable congestion level indicates the level of congestion that the local government or business entity that manages the spot can tolerate for the spot. The acceptable congestion level is, for example, a value such as the ratio of the number of visitors to the maximum capacity of the spot, and is a value that the local government or business entity has determined in advance with the intention of avoiding overtourism, for example. The value of the allowable congestion level is used to determine whether or not to accept travelers at the spot. The allowable congestion level may be managed at a granularity by time period. The expected congestion level by time period is, for example, a value of the congestion level at the spot predicted by day and time period. This value is obtained by using past statistical values ​​observed regarding the congestion level at the spot, or is periodically obtained and updated from an external demand forecasting system. The keyword list is, for example, a list of keywords extracted from text data included in an article introducing the spot, or a list of keywords registered in advance by the spot operator.

[0052] In addition, the standard stay time by congestion level is a value used when adjusting the stay time at a spot according to the congestion level when it is expected that a queue will occur depending on the congestion level at the spot. For example, when the congestion level at a spot is 50%, no queue will occur and the standard stay time is 1 hour, but when the congestion level is 180%, it is expected that a queue will occur, so the standard stay time is set to 90 minutes, taking into account the 30 minute wait time while waiting in line. The pattern is as follows:

[0053] The standard cost of living is the standard cost of living according to the conditions for creating a travel plan. Specifically, the average cost of living shown in the past travel records is registered by clustering the past travel records based on the number of people traveling together, etc.

[0054] The spot feature vector is information used to select candidate spots for inclusion based on the degree of match with the travel plan creation conditions (traveler information, evaluation viewpoint information). A weight is registered for each element of the spot feature vector to reflect the situation related to the traveler or evaluation viewpoint in the evaluation value of the itinerary candidate. For example, if an element "family-friendly" is set as one element of the spot feature vector and a weight of "1" (family-friendly) is registered, the evaluation for that element is expected to be "1" (favorable) if the number of adults and children registered in the traveler information is 1 or more, and "0" (no particular positive evaluation) if there are only adults. The evaluation value of the spot is calculated as the sum of the evaluation values ​​obtained for each element of the spot feature vector according to the situation related to the traveler or evaluation viewpoint. <How to support travel plan creation>

[0055] The actual procedure of the method for supporting travel plan creation in this embodiment will be described below with reference to the drawings. Various operations corresponding to the method for supporting travel plan creation described below are mainly realized by the arithmetic unit 104 of the server 100 reading out a program stored in the storage device 101 into a memory or the like and executing it. However, in the following description, in order to clarify whether an operation is performed on the server 100 side or the terminal device 200 side, the operation will be described as being performed by the server 100 or the terminal device 200, for example, "the server 100 creates an image set for estimating preferences."

[0056] 6 is a diagram showing an example of a sequence of the travel plan creation support method in this embodiment. In this case, the terminal device 200 displays a predetermined input screen (see FIG. 7) on the display unit 204, and acquires travel plan creation conditions specified by the traveler through this input screen (s100).

[0057] The input screen shown in FIG. 7 is a screen for specifying travel plan creation conditions, such as information on travel conditions (travel condition information) such as departure point, return point, departure date, departure time, return date, return time, estimated budget, and conditions for specifying the image desired for the travel plan (for example, "automatic" which does not specify anything in particular, and "key phrase" which specifies the image of the trip that the traveler has), information on the attributes of the traveler (traveler information) such as the number of adults and children by gender, and information on evaluation viewpoints (spot congestion, travel congestion, road congestion, and other preferences) (evaluation viewpoint information). The information on each evaluation viewpoint specified here is used in each process described later. For example, if a traveler using this system is a person with travel restrictions, by selecting "avoid" for spot congestion and "avoid" for travel congestion, the process will be such that spots and means of transportation that are more crowded than a specified standard will be excluded from the options. If "avoid if possible" or "don't care" is selected, the options will not be excluded, but the process will be such that the traveler's preferences are reflected in the evaluation of the itinerary candidates by reflecting them in the weight value when calculating the evaluation value. The details of these processes will be described later.

[0058] The terminal device 200 also transmits the travel plan creation conditions obtained in s100 to the server 100 (s101).

[0059] When the server 100 receives the travel plan creation conditions from the terminal device 200, it creates an image set for preference estimation (s102). Note that, when creating this image set, if the traveler selects "automatic" as the travel plan creation condition, the server 100 uses the reference image database 12 5, and use the images as a set of images for preference estimation. On the other hand, if the traveler specifies a "key phrase" as a condition for creating a travel plan, the server 100 searches the travel article database 126 for the specified key phrase, extracts images used in the corresponding articles, and uses the extracted images as a set of images for preference estimation.

[0060] Next, the server 100 transmits the created image set for preference estimation to the terminal device 200 (s103).

[0061] When the terminal device 200 receives the image set for preference estimation from the server 100, it displays the received image set for preference estimation on the display unit 204 (s104). An example of the display screen at this time is shown in Fig. 8. In this example, a screen example is shown that allows the traveler to select an image that is closest to the image of the destination from among eight images.

[0062] Next, the terminal device 200 acquires the result of image selection by the user from the image set for preference estimation (s105). At this time, the image selected may be one or more images.

[0063] After acquiring the result of the image selection by the user, the terminal device 200 transmits the selection result acquired in s105 to the server 100 (s106).

[0064] When the server 100 receives the selection result transmitted from the terminal device 200 (s107), it executes a candidate area extraction process (s108). The candidate area extraction process of s108 will be described in detail later (FIG. 9).

[0065] Next, the server 100 executes an itinerary candidate creation process (s109). The details of this itinerary candidate creation process will be described later (FIG. 11).

[0066] Next, the server 100 transmits information on a recommended travel plan from among the itinerary candidates identified in s109, that is, recommended travel plan information, to the terminal device 200 (s110).

[0067] Such a recommended travel plan is selected from among the itinerary candidates obtained in s109, either from the one with the highest evaluation value, or a predetermined number of itineraries in descending order of evaluation value. Alternatively, the itinerary candidate with the highest evaluation value for each candidate area may be selected. The concept of the evaluation value will be described later.

[0068] When the terminal device 200 receives the recommended travel plan information from the server 100, it outputs it to the display unit 204 (s111) and ends the process. If the terminal device 200 can detect that a certain plan has been adopted by inputting a hotel reservation or the like after outputting the recommended travel plan, the adoption result may be used to adjust the parameters of the evaluation function (for example, the relative magnitude of the coefficients of each term, etc., when an evaluation function expressed by a linear combination is used) to improve the adoption rate the next time a plan is proposed.

[0069] <Candidate area extraction process> The candidate area extraction process (s108) will now be described in detail with reference to Fig. 9, which is a diagram showing the flow of the candidate area extraction process in this embodiment.

[0070] In this case, the server 100 uses the candidate area extraction program 111 to extract a set of corresponding keywords based on the information of the selected image obtained in s105 (s200).

[0071] The extraction of this keyword set is performed, for example, by executing the process shown in the flow of Figure 10. The process of extracting a set of keywords will be described below with reference to FIG.

[0072] Specifically, the server 100 first calculates a feature vector for each selected image (s300). As already described, the feature vector is calculated using a feature vector extractor.

[0073] Next, the server 100 calculates the similarity between the feature vector of the image included in each article in the travel article database 126 and the feature vector obtained in s300 (s301). The feature vector of the image included in each article may be obtained by referring to the travel article database 126 and previously calculating the feature vector in the same manner as in s300. As a measure of the similarity, for example, cosine similarity is used.

[0074] The server 100 registers the similarity obtained in s301, for example, in a similarity list provided in the main storage device 103 (s302). The similarity list is intermediate data that is temporarily created, and stores the similarity obtained in s301 in association with a pair of identification information of the travel article and the selected image.

[0075] Next, the server 100 refers to the above-mentioned similarity list and determines that pairs whose similarity is equal to or greater than a predetermined value, or pairs whose similarity is included in a predetermined number of top rankings, are highly similar pairs (s303).

[0076] As a result of the above-mentioned determination, the server 100 identifies pairs with high similarity (s303: Yes), and by referring to the travel article data indicated by the identified pair, extracts text adjacent to the image data of the travel article, and registers it in a context information list provided in the main memory device 103 (s304).

[0077] The context information list is intermediate data that is temporarily created to store text extracted from travel articles in the pairs identified in s303, i.e., a collection of text from travel articles that are likely to describe images that are highly similar to the selected image.

[0078] Next, the server 100 refers to the above-mentioned context information list and estimates a group of keywords from the text registered therein. For example, based on the text registered in the context information list, i.e., the text obtained from travel articles for each of the multiple selected images, the server 100 extracts keywords (e.g., "mountain walking," "mountain stream," "stylish," "great view," "luxurious meal," etc.) common to multiple pieces of the text, for example, by using a text mining technique such as morphological analysis.

[0079] Returning to the explanation of the flow in FIG. 9, the server 100 extracts the value of the keyword list column for each area in the area information database 127, and obtains the similarity between this keyword and each keyword in the keyword group obtained for the selected image in s200 (s201). For example, cosine similarity is used as a measure of similarity in comparing the keyword groups. Any known technology such as Word2Vec may be used. In this way, by performing a process of determining the similarity by comparing the keyword groups estimated using the entire text of the travel article related to each of the multiple images, it is possible to obtain a selection result that more reflects the similarity in terms of what travelers are interested in, such as atmosphere and moving content, compared to a process of simply determining the similarity of images.

[0080] Next, the server 100 judges whether the similarity calculated in s201 is equal to or greater than a predetermined threshold (s202). If the result of this judgment is that the similarity is equal to or greater than the predetermined threshold (s202: Yes), the server 100 determines the expected congestion level for the target area based on the travel plan creation conditions by referring to the information on the congestion level by time period in the area information database 127; It is determined whether the congestion level of the area is below the allowable congestion level (s203).

[0081] If the result of this determination is that the congestion level of the area is equal to or lower than the allowable congestion level (s203: Yes), the server 100 adds the target area as a candidate for a destination to a list of candidate areas provided in the main storage device 103 (s204). The server 100 generates a list of candidate areas by repeating the above-mentioned series of processes.

[0082] <Itinerary candidate creation process> Next, the itinerary candidate creation process (s109) will be described in detail with reference to Fig. 11, which is a diagram showing the flow of the itinerary candidate creation process in this embodiment.

[0083] First, the server 100 empties the itinerary candidate list provided in the main storage device 103 (s400). Next, the server 100 executes the following processes of s401 to s407 for each candidate area determined in the flow of FIG.

[0084] The server 100 first selects candidates for the main spot (s401). For example, by referring to the spot information database 127, the server 100 selects, from among the spots linked to the candidate areas, spots that have a relatively high similarity between the keywords indicated in the travel plan creation conditions (those specified by the traveler or those in the selected image keyword list) and the values ​​indicated in the keyword list field registered in the spot information.

[0085] Next, the server 100 selects candidates for the sub-spots (s402). For example, from past cases with similar travel plan creation conditions, one or more spots (other than the main spot) that are estimated to have a high probability of being visited within the candidate area of ​​interest are selected.

[0086] Next, the server 100 executes a process for creating a passenger behavior network (s403). The process for creating the passenger behavior network will be described in detail below with reference to FIG.

[0087] 14 is a diagram showing an example of the passenger behavior network creation process in this embodiment. The passenger behavior network creation process is a process for creating a passenger behavior network, which is a directed graph that expresses the possibility of state transitions related to a traveler's stay or movement and the degree of evaluation of the stay or movement.

[0088] First, the server 100 creates nodes corresponding to the departure point and the destination point (s500). At least the geographical position of the departure point and information about the departure time are linked and stored in the node corresponding to the departure point. Similarly, at least the geographical position of the destination point and information about the curfew time are linked and stored in the node corresponding to the destination point.

[0089] Next, the server 100 creates arrival nodes and departure nodes corresponding to each discretized time for the main spot candidates and the sub-spot candidates (s501).

[0090] Each node, such as an arrival node and a departure node, corresponds to each spot in the area of ​​interest. The arrival node includes at least the identification information (spot ID) of the spot, the geographical position of the arrival point, and information on the arrival time. Similarly, the departure node includes at least the identification information (spot ID) of the spot, the geographical position of the departure point, and information on the departure time.

[0091] Here, the geographical location of the arrival node and the departure node are not necessarily the same. For example, in the case of a spot where you can experience river boating, the arrival node may correspond to a spot that is relatively far from the river. The arrival node may be a location upstream of the destination, and the departure node may be a location relatively downstream in the river. Even if the geographical location of the arrival point is the same, multiple arrival nodes may be created due to differences in arrival time. The same applies to the departure node. The method of discretizing the time may be simply "every hour."

[0092] If there is a special time, such as the departure time of a pleasure boat, the time should be adjusted to that time. In that case, if a queue is expected to occur, the time corresponding to the arrival node is set earlier by the amount of the waiting time.

[0093] Next, the server 100 creates a link (sometimes called a "directed edge" or "arc") if a state transition is possible for the stay at each spot (s502). Specifically, the server 100 refers to the time of the start point (arrival node) and the end point (departure node) and creates a link if both of the following conditions are met:

[0094] Condition 1) The congestion is tolerable in light of the travel plan creation conditions. For example, if the travel plan creation conditions specify that the spot congestion should be "avoided," if the congestion is greater than a predetermined threshold, it is determined to be "unacceptable." In this determination, for example, the maximum congestion between the time corresponding to the arrival node and the time corresponding to the departure node is used to determine whether it is acceptable. Condition 2) The stay time is equal to or greater than the minimum stay time. Here, the minimum stay time is a value that may change depending on the degree of congestion. For example, if the spot of interest is a theme park, the stay time required to have an equivalent experience is estimated in advance according to the degree of congestion in response to the case where it is expected that after arriving at the site, a queue will be formed before using each attraction, and the estimated time is registered in the spot information database 128.

[0095] The links corresponding to stays at each spot include information about the stay (including rating values). Here, the evaluation value is, for example, a spot feature vector corresponding to the spot. The travel score is calculated by comparing the information on the travel destination and the traveler information and evaluation criteria information indicated in the travel plan creation conditions with the expected congestion level by time of day and the standard stay time by congestion level at the spot, determining the evaluation value for each evaluation criteria, and adding up these results.

[0096] Next, the server 100 creates a link if a state transition is possible for travel between the departure point, each spot, and the return point (s503). Specifically, the start point and end point times are referenced, and a link is created only when both of the following conditions are met: Condition 1) The congestion is acceptable in light of the travel plan creation conditions, and Condition 2) The predicted arrival time is earlier than the arrival time corresponding to the end node. Here, the predicted arrival time is a value that may change depending on congestion (traffic congestion impact, etc.). In addition, multiple links may exist for the same pair of nodes. For example, each means of travel, such as a train and a car, or another train, corresponds to a link. Each link holds information (including an evaluation value) about the travel route. Here, the evaluation value is calculated using the travel plan creation conditions. In addition, the reflection of the impact of traffic congestion, etc. may be performed using a known route search (route search) technology that includes a traffic congestion prediction function. Also, in order to reduce calculation time, not all possible links are created, but rather a few travel routes are selected from the top when a route search is performed, giving priority to a specific evaluation criteria, such as travel time.

[0097] After completing the process of s503, the server 100 ends the passenger behavior network creation process.

[0098] An example of a passenger behavior network generated in this way is shown in Figure 15. In the passenger behavior network shown in Figure 15, multiple nodes are created for the same spot at discrete times based on a predetermined rule (e.g., arrival and departure at multiple different times for a park).

[0099] Additionally, dashed lines are links corresponding to stays, and solid lines are links corresponding to travel. Taking into consideration the degree of congestion, links are created only in areas where transitions are possible. Additionally, links corresponding to stays are managed by linking them to spot information, and links corresponding to travel are managed by linking them to travel route information. This is because such information is referenced when calculating the evaluation value.

[0100] In addition, if there are multiple routes, such as arriving at the park at 16:00 from the departure point, Corresponding links are generated. Even if the arrival and departure points are different, such as in the case of river boating, if that is the purpose of the visit itself, it is treated as a spot stay. Each path that can be followed on this passenger behavior network from the departure point via the main spot to the destination corresponds to a rough itinerary plan.

[0101] Returning to the explanation of the flow in Fig. 11, the server 100 selects a general itinerary plan (s404). This general itinerary plan is selected by, for example, k-optimization for each route indicated by the passenger behavior network. By solving the shortest route problem, the k routes with the highest evaluation values ​​can be obtained as rough itinerary plans.

[0102] Next, the server 100 confirms the time for each outline itinerary plan and makes it an itinerary plan (s405). In this case, the server 100 recalculates the predicted time from departure to arrival to create the itinerary plan. The travel route connecting each spot is searched for under conditions such as arriving at the earliest possible time (taking congestion into consideration) and departing from the visited spot at the latest time that allows for making it in time for the next trip (taking congestion into consideration). In this way, after creating an outline itinerary plan, the detailed times are recalculated to find actual itinerary candidates, which results in a coarse-to-fine approach overall, making it possible to find itinerary candidates efficiently. The server The itinerary plan 100 stores the itinerary plan thus created in the itinerary candidate information 129 (FIG. 12).

[0103] Next, the server 100 calculates a comprehensive evaluation value of the itinerary plan (s406). Specifically, the server 100 recalculates the evaluation values ​​of the travel route and the visiting spots using the traveler information and the evaluation viewpoint information based on the confirmed time, and calculates a comprehensive evaluation value of the itinerary plan.

[0104] Next, the server 100 creates itinerary candidate information and registers it in the itinerary candidate list provided in the main storage device 103 (s407). The itinerary identification information, the overall evaluation value, and specific itinerary information are combined into a set to form itinerary candidate information (see FIG. 12), which is then added to and registered in the itinerary candidate list.

[0105] As shown in Fig. 12, the itinerary candidate information is composed of itinerary identification information, an overall rating value, and itinerary information. The itinerary identification information is a serial number among all the itinerary candidates created in the travel plan creation process. Alternatively, it may be a combination of an area ID and a serial number within the area. The overall rating value is the rating value of the entire itinerary calculated in the process of s406 described above.

[0106] The itinerary information is a list of a series of information on transportation means (what transportation means, what time of departure and what time of arrival) from the departure point to the return destination, and information on visiting spots (what time of arrival and what time of departure). Note that the itinerary information may also include expected congestion status, etc.

[0107] 13 shows an example of itinerary information 1291. The itinerary information 1291 shown here is an itinerary plan associated with values ​​such as a list of means of transportation, departure point / arrival point, departure time / expected arrival time, expected congestion level, traffic congestion waiting time, number of transfers, and fare.

[0108] The transportation list is information on trains, buses, cars, etc., used from departure to arrival. This is a list of transportation methods. In the case of trains and buses, the flight to be used and whether or not a reserved seat is required are also specified.

[0109] The departure time is the scheduled time if the first transportation means is a transportation means that operates according to a timetable, such as a train or bus. In other cases, it is the latest departure time that is expected to allow arrival at the desired arrival time.

[0110] In addition, the expected arrival time is an arrival time that reflects the effects of traffic congestion and operational disruptions, predicted based on information previously obtained from the operation management systems of transportation operators and route guidance service providers (it may be a predicted value based not only on past information, but also on information for the current day).

[0111] For information on the expected congestion level, traffic jam waiting time, number of transfers, and fare, the values ​​used to select candidate routes for inclusion based on the degree of match with the travel plan creation conditions (traveler information, evaluation viewpoint information) are registered. These values ​​are also used to calculate the route evaluation value. In order to be able to present how conditions such as fewer transfers, lower costs if possible, and congestion avoidance tendencies are reflected, values ​​are stored for each evaluation item (such as spot congestion, congestion during travel, and road congestion, as shown in Figure 7).

[0112] In addition, when selecting itinerary candidates, the server 100 will eliminate duplicate registrations of candidates that are extremely similar. In addition, if there is ambiguity in the return time to the destination, in order to avoid the phenomenon in which the longer the itinerary and the more spots are packed into it, the higher the evaluation, the more the itinerary candidates will be compared after converting them into "evaluation value per unit time." In this case, the curfew times can be discretized, the top k itinerary candidates for each can be created, and the evaluation values ​​can be normalized before comparison.

[0113] Although the best mode for carrying out the present invention has been specifically described above, the present invention is not limited to this, and various modifications are possible without departing from the spirit and scope of the present invention.

[0114] According to this embodiment, it is possible to create multiple specific itinerary plans for a traveler who has not yet decided on a specific travel destination, by selecting multiple areas (multiple spots) as candidate destinations based on the traveler's preferences and appropriately determining the order in which the spots will be visited and the means of transportation between the spots based on the congestion status, etc. Itinerary plans created in this way are of high quality and are configured to include destinations that match the preferences of the traveler, for a traveler who has not yet decided on a specific destination. In addition, it is likely that it will have the effect of mitigating the phenomenon of travelers concentrating at famous tourist spots, since itinerary plans that guide travelers to areas and spots that are relatively less crowded are more likely to be selected.

[0115] In other words, for those who have not yet decided on a specific travel destination, it is possible to propose an appropriate itinerary that matches their preferences, reduces the impact of congestion, and enables efficient travel.

[0116] Furthermore, the description of this specification makes at least the following clear: That is, in the travel plan creation support system of this embodiment, when identifying the plurality of candidate locations, the arithmetic device may perform a process of searching a predetermined database for existing travel information including an image similar to an image selected by a user as an image of a travel destination from among a predetermined plurality of images, based on the information of the image, and extracting a group of keywords included in the travel information, and identifying information introducing predetermined areas that includes a group of keywords similar to the group of keywords, and identifying areas indicated by the identified area introduction information as the plurality of candidate locations.

[0117] According to this, for a user who wants to be presented with a proposed itinerary but is not sure of the destination, etc., It is possible to estimate travel destinations that match the vague image of a travel destination that a user has in their mind, and create an appropriate itinerary based on this. In addition, for those who have not yet decided on a specific travel destination, it is possible to propose an appropriate itinerary that matches their preferences, reduces the impact of congestion, and allows them to travel efficiently.

[0118] In addition, in the travel plan creation support system of this embodiment, when creating the multiple itinerary plans, the calculation device may select candidate locations and means of transportation that meet criteria in terms of congestion and length of stay at candidate locations based on information regarding congestion at each of the multiple candidate locations and information regarding means of transportation to reach each of the multiple candidate locations, and create multiple itinerary plans that specify the order of visiting and means of transportation between the candidate locations.

[0119] This makes it possible to create and present appropriate itinerary proposals that ensure sufficient time spent at candidate locations, and ultimately to propose efficient and appropriate itineraries that match the preferences of those who have yet to decide on their travel plans.

[0120] In the travel plan creation support system of this embodiment, when performing the evaluation, the calculation device may evaluate each of the multiple itinerary plans based on the congestion situation and the duration of stay of the candidate locations and the means of transportation included in that itinerary plan.

[0121] This makes it possible to select and present to users an itinerary that takes into consideration the time required to fully enjoy the travel destination and travel efficiency. In addition, it makes it possible to propose an appropriate itinerary that matches the preferences of those who have not yet decided on their travel plans and is also highly efficient.

[0122] Furthermore, in the travel plan creation support system of this embodiment, when performing the evaluation, the calculation device may divide, for each of the multiple itinerary plans, an evaluation value determined using predetermined rules for the congestion situation and the duration of stay at the candidate locations and means of transportation included in that itinerary plan by the total number of hours in that itinerary plan to calculate an evaluation value per unit time, and make the selection based on the size of the evaluation value per unit time.

[0123] This makes it possible to avoid situations where itineraries with unnecessarily long total travel times are highly rated, and to create and propose suitable itineraries that are not dependent on total travel time. In turn, it makes it possible to propose efficient and appropriate itineraries that match the preferences of those who have not yet decided on their travel plans.

[0124] In the travel plan creation support method of this embodiment, when the information processing system identifies the multiple candidate locations, it may search a specified database for existing travel information containing images similar to an image selected by a user as an image of the travel destination from a specified multiple images, based on the information of the image, and extract a group of keywords contained in the travel information; and identify information from specified area introduction information that contains a group of keywords similar to the group of keywords, and identify the area indicated by the identified area introduction information as the multiple candidate locations.

[0125] Furthermore, in the travel plan creation support method of this embodiment, when the information processing system creates the multiple itinerary proposals, it may select candidate locations and means of transportation that meet criteria in terms of congestion and length of stay at candidate locations based on information about congestion at each of the multiple candidate locations and information about means of transportation to reach each of the multiple candidate locations, and create multiple itinerary proposals that specify the order of visiting and means of transportation between the candidate locations.

[0126] In the travel plan creation support method according to the present embodiment, the information processing system When performing the evaluation, each of the multiple itinerary plans may be evaluated according to the congestion status and the duration of stay of the candidate locations and the means of transportation included in that itinerary plan.

[0127] Furthermore, in the travel plan creation support method of this embodiment, when the information processing system performs the evaluation, it may divide, for each of the multiple itinerary plans, an evaluation value determined using a predetermined rule for the congestion situation and the duration of stay at the candidate locations and means of transportation included in that itinerary plan by the total time in that itinerary plan to calculate an evaluation value per unit time, and make the selection based on the size of the evaluation value per unit time. [Explanation of symbols]

[0128] 1 Network 10 Travel planning support system 100 Server (Travel plan creation support device) 101 Storage device 103 Main storage 104 Arithmetic equipment 105 Communication equipment 106 Databases 110 Travel Planning Program 111 Candidate Area Extraction Program 112 Itinerary candidate creation program 120 Travel plan creation conditions 121 Image set for preference estimation 122 Selected Image Information 123 Itinerary candidate list 125 Reference Image Database 126 Travel Article Database 127 Area Information Database 128 Spot Information Database 129 Itinerary candidate information 1291 Itinerary information 200 Terminal Equipment 201 Control section 202 Communications Department 203 Operation input section 204 Display section 210 Condition Acquisition Section 211 Information output section

Claims

1. A storage device for storing information related to potential travel destinations; A process of selecting an image set from a group of images for estimating a user's preferences among the information, according to at least the user's image of a trip indicated by the travel plan creation conditions based on a user-specified travel plan creation condition and the information, a process of presenting the image set to the user and obtaining a selection result of images that are close to the user's image of the travel destination, a process of calculating a similarity between each image subject to the selection result and predetermined information regarding each spot among the information, a process of specifying a list of spots having the similarity equal to or higher than a standard as a list of candidate locations, and a process of specifying a candidate location that satisfies at least a standard with respect to congestion and a stay time at the candidate location based on information regarding congestion at each of the candidate locations included in the list and information regarding a means of transportation to reach each of the candidate locations. a process of selecting candidate locations and means of transportation for each of the candidate locations as departure and destination nodes, a process of generating a directed graph by setting links between the nodes when travel between the nodes is possible based on the relationship between the positions of the departure location and the destination location and the departure and arrival times, and the congestion and stay times are satisfied; a process of applying a predetermined optimization algorithm to each route between the nodes shown in the directed graph, and creating a plurality of itinerary plans that specify the order of visits and means of transportation between the candidate locations corresponding to the nodes; a process of evaluating each of the plurality of itinerary plans according to the congestion and stay times of the candidate locations and the means of transportation included in the itinerary plan; and a calculation device that executes the process of selecting an itinerary plan with a relatively high evaluation and outputting it to a predetermined device. A travel plan creation support system including a device having the above-mentioned.

2. The computing device includes: When performing the evaluation, for each of the multiple itinerary plans, an evaluation value is calculated by determining the congestion status and the duration of stay of the candidate locations and the means of transportation included in the itinerary plan according to a predetermined rule, and the evaluation value per unit time is calculated by dividing the evaluation value by the total time in the itinerary plan, and the selection is made based on the size of the evaluation value per unit time.

2. The travel plan creation support system according to claim 1.

3. The information processing system Store information about potential travel destinations in a storage device, A process of selecting an image set from a group of images for estimating a user's preferences among the information, according to at least the user's image of a trip indicated by the travel plan creation conditions based on a user-specified travel plan creation condition and the information; a process of presenting the image set to the user and obtaining a selection result of images that are close to the user's image of the travel destination; a process of calculating a similarity between each image included in the selection result and predetermined information regarding each spot among the information; a process of specifying a list of spots having a similarity equal to or higher than a standard as a list of candidate locations; and a process of calculating a basis for at least the congestion situation and the duration of stay at the candidate locations based on information regarding congestion at each of the candidate locations included in the list and information regarding a means of transportation to each of the candidate locations. a process of selecting candidate locations and means of transportation that satisfy the criteria, a process of treating each of the candidate locations as a departure location and a return location, and generating a directed graph by setting links between the nodes when movement between the nodes is possible based on the relationship between the positions of the departure location and the return location and the departure and arrival times, and the conditions of the congestion level and the stay time are satisfied, a process of applying a predetermined optimization algorithm to each route between the nodes shown in the directed graph, and generating a plurality of itinerary plans that specify the order of visits and means of transportation between the candidate locations corresponding to the nodes, a process of evaluating each of the plurality of itinerary plans according to the congestion level and the stay time of the candidate locations and the means of transportation included in the itinerary plan, and a process of selecting an itinerary plan with a relatively high evaluation and outputting it to a predetermined device. A travel plan creation support method comprising:

4. The information processing system comprises: When performing the evaluation, for each of the multiple itinerary plans, an evaluation value is calculated by determining the congestion status and the duration of stay of the candidate locations and the means of transportation included in the itinerary plan according to a predetermined rule, and the evaluation value per unit time is calculated by dividing the evaluation value by the total time in the itinerary plan, and the selection is made based on the magnitude of the evaluation value per unit time.

4. The method for supporting travel planning according to claim 3.

Images