WO2017138624A1

WO2017138624A1 - Information processing device and information processing method

Info

Publication number: WO2017138624A1
Application number: PCT/JP2017/004832
Authority: WO
Inventors: 山崎　俊彦; 康平大原
Original assignee: 国立大学法人東京大学
Priority date: 2016-02-12
Filing date: 2017-02-09
Publication date: 2017-08-17
Also published as: JP2017142723A; JP6661193B2

Abstract

(Problem) To provide an information processing device and an information processing method with which properties can be searched for in a suitable manner. (Solution) This information processing device is provided with: a means for managing graph information in which rooms and connections between rooms are represented as a graph structure including nodes and edges, for each set of property information; an input means for accepting input of a query graph indicating a graph structure to be searched for, created in accordance with a user operation; a calculating means for calculating a degree of similarity between the graph structures relating to each set of property information, and the graph structure relating to the query graph; and an output means for outputting the property information in accordance with the degree of similarity.

Description

Information processing apparatus and information processing method

Some aspects according to the present invention relate to an information processing apparatus and an information processing method for searching for real estate properties, for example.

2. Description of the Related Art In recent years, a property search service for searching for apartments for rent or sale, for example, and a dwelling unit such as a detached house is provided on a website on the Internet. In such a property search service, a search for a real estate property is usually performed by combining search conditions such as the nearest station, exclusive area, rent, and keyword. However, it is not possible to search on the condition of specific floor plans such as the size of individual rooms such as living dining and Japanese-style rooms and the connection method between rooms such as whether or not the living dining and Western rooms are directly connected.
Non-Patent Document 1 describes a method of confirming the influence of room layout on rent by creating a room layout as a graph and then performing graph mining.

However, the method described in Non-Patent Document 1 does not disclose any method for searching based on the room arrangement.

Some aspects of the present invention have been made in view of the above-described problems, and an object thereof is to provide an information processing apparatus and an information processing method capable of suitably searching for a property.

The information processing apparatus according to one aspect of the present invention includes a unit that manages graph information in which rooms and connections between rooms are expressed as a graph structure including nodes and edges for each piece of property information, and a user operation. An input unit that receives an input of a created query graph indicating a graph structure to be searched, a graph unit related to each property information, a calculation unit that calculates a similarity between the graph structure related to the query graph, and the similarity Depending on the degree, output means for outputting the property information is provided.

An information processing method according to one aspect of the present invention includes a step in which an information processing apparatus manages, for each piece of property information, graph information in which a room and a connection between rooms are represented as a graph structure including nodes and edges; A step of receiving an input of a query graph indicating a graph structure to be searched, created according to a user operation, a step of calculating a similarity between the graph structure according to each property information and the graph structure according to the query graph; And outputting the property information according to the degree of similarity.

In the present invention, “part”, “means”, “apparatus”, and “system” do not simply mean physical means, but “part”, “means”, “apparatus”, “system”. This includes the case where the functions possessed by "are realized by software. Further, even if the functions of one “unit”, “means”, “apparatus”, and “system” are realized by two or more physical means or devices, two or more “parts” or “means”, The functions of “device” and “system” may be realized by a single physical means or device.

It is a flowchart which shows the flow of a process of the information processing apparatus which concerns on embodiment. It is a figure which shows the specific example of the floor plan layout of search object. It is a figure which shows the specific example of a query graph. It is a figure for demonstrating the specific example of the identification method of the largest common subgraph. It is a figure for demonstrating the specific example of the identification method of the largest common subgraph. It is a block diagram which shows the function structure of the information processing apparatus which concerns on embodiment. It is a block diagram which shows the specific example of the hardware constitutions which can mount the information processing apparatus shown in FIG.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the embodiment described below is merely an example, and there is no intention to exclude various modifications and technical applications that are not explicitly described below. That is, the present invention can be implemented with various modifications without departing from the spirit of the present invention. In the following description of the drawings, the same or similar parts are denoted by the same or similar reference numerals. The drawings are schematic and do not necessarily match actual dimensions and ratios. In some cases, the dimensional relationships and ratios may be different between the drawings.

1 to 5 are diagrams for explaining the embodiment. Hereinafter, embodiments will be described along the following flow with reference to these drawings. First, an outline of the property search apparatus according to the embodiment will be described with “1”. Subsequently, the processing of the property search apparatus will be described in “2” with a specific example. “3” describes the functional configuration of the property search apparatus, and “4” describes a specific example of a hardware configuration capable of realizing the property search device. Finally, after “5”, the effects according to the embodiment and other embodiments will be described.

(1. Overview)
2. Description of the Related Art In recent years, a property search service for searching for apartments for rent or sale, for example, and a dwelling unit such as a detached house is provided on a website on the Internet. In such a property search service, it is common to perform a refined search using search conditions combined with AND / OR such as the nearest station, exclusive area, route, rent, and keyword.

Here, there is a demand to search for a property by paying attention to the layout of the property according to the specific lifestyle of each user. For example, there is a structure in which a child always goes through the living room when leaving the child room so that the child does not go out of the house, that is, when the user is looking for a property where each room is a living-in. However, the conventional property search service cannot cope with the search based on the layout of each room.

Therefore, the property search device according to the present embodiment enables a search focusing on the layout of the floor plan. More specifically, first, the user inputs the layout of the floor plan that the user wants to search as a graph structure (hereinafter also referred to as a query graph). The property search device calculates a similarity between the input query graph and a graph structure indicating a layout of each property included in the database. In addition, the property search apparatus creates a property list according to the calculated similarity and presents the list to the user. Thereby, the said property search apparatus provides the search function according to the layout of a floor plan to a user. Details will be described below.

(2. Process flow)
First, referring to FIG. 1, the processing of the property search apparatus according to the present embodiment will be described with a specific example. FIG. 1 is a flowchart showing a processing flow of the property search apparatus.

(2.1. Input of query graph)
First, the property search apparatus causes a user to input a layout of a floor plan to be searched as a graph structure (S101). In this regard, a specific example of a query graph created by the property search device according to the present embodiment in response to a user operation will be described with reference to FIGS. 2A and 2B. FIG. 2A is a specific example of a layout of a floor plan that the user wants to search, and FIG. 2B is a specific example of a graph structure showing the layout.

In the floor plan layout that the user wants to search shown in FIG. 2A, a 6-cm Western-style room and a Japanese-style room, and a 4-cm washroom are connected to an 8-cm dining kitchen. In addition, a 4cm bath and a 3cm toilet are connected to the washroom. In the example of FIG. 2A, the entrance is not described, but this is because the user does not particularly emphasize the entrance position in the floor plan search and does not include it in the search target.

FIG. 2B shows a specific example of a graph structure showing the layout of the floor plan shown in FIG. 2A, that is, a query graph input by the user. In the graph structure shown in FIG. 2B, a dining kitchen (DK), a Japanese-style room (JR), a Western-style room (WR), a washroom (PR), a bath (Ba) A room such as a toilet (WC: Water Closet) is represented as a node having a label indicating the type and size of each room. Note that “DK”, “JR”, “WR”, “PR” and the like indicating the room type of each node correspond to the label of the node.

Also, in the graph structure of FIG. 2B, the connection between rooms that can be moved between rooms through a door or the like is expressed as an edge connecting the nodes. In the example of FIG. 2B, for example, between the node “DK8” indicating the dining kitchen of 8 cm and the node “JR6” indicating the 6-inch Japanese-style room, or the node “DK8” and the node “DK8” indicating the 6-inch Western-style room “ An edge is provided between “WR6” and the like. Although not included in the graph of FIG. 2B, when the user wants to search a layout of the floor plan including the arrangement of the corridor and stairs, the corridor and stairs are expressed as nodes, What is necessary is just to connect with the node which shows this room by an edge.

In the property search apparatus according to the present embodiment, the user can set a room and a connection between the rooms that are desired to be emphasized when searching. For example, if the user wants to emphasize the presence of a dining kitchen or direct connection between a Western-style room and a dining kitchen, the user is “Important” for the node or edge of the corresponding portion, as shown in FIG. 2B. Can be set. Nodes and edges that have been shown to be important in this way are given a large weight when calculating the similarity described later.

(2.2. Identification of MCS)
When a user inputs a graph structure (query graph) of a layout to be searched and requests a search (Yes in S103 in FIG. 1), the property search apparatus displays a graph corresponding to the floor layout of each property from the database (DB). The structure is read (S105). In addition, the property search device specifies a maximum common subgraph (MCS) between the graph structure of the read property information and the query graph that is the search condition (S107).

Here, there are various methods for specifying the MCS, and the calculation time required for each method varies depending on the number of nodes, the number of edges, the density of the graph, and the like included in the graph. For example, the Durand-Pasari algorithm, the Balas Yu algorithm, the McGregor algorithm, and the like can be cited as methods for specifying the MCS. Any method may be used to specify the MCS, but here, the McGregor algorithm is modified and used. The outline | summary of the method in which the property search apparatus in this embodiment calculates | requires MCS is demonstrated, referring FIG. 3A and FIG. 3B.

FIG. 3A shows an example of a graph structure to be compared. Here, MCS of graph 1 and graph 2 shown in FIG. 3A is obtained. In the graph 1, three

nodes id

0, 1, and 2, respectively labeled “A”, “B”, and “C”, are connected in series. In the graph 2, four nodes id 0 ′, 1 ′, 2 ′, and 3 ′, which are respectively labeled “A”, “B”, “A”, and “C”, are connected in series.

First, the nodes are read from graph 1 in ascending order of id. In the example of FIG. 3B, the node with the id “0” of the label “A” is read from the graph 1. Then, the nodes of id0 ′ and 2 ′ having the same label “A” as the node read from the graph 1 are read from the graph 2 and added to the search tree as MCS candidates (

nodes

31A and 31B).

Next, the node of id1 in graph 1 is read. Since the label of the node with id1 is “B”, the node with id1 ′ having the label “B” is also read out in the graph 2. In addition, if there is an edge between the id0 and id0 ′ nodes whose correspondence has been specified in the

graph nodes

31A and 31B of the MCS candidates, and the id1 and id1 ′ nodes read this time Both the read node and the edge are added as search tree candidates. In the example of

graphs

1 and 2 in FIG. 3A, since an edge exists between the node with id0 and the node with id0 ′ and the node with id1 and the node with id1 ′ read this time, the node with id1 and the node with id1 ′ Are added to the search tree as MCS candidates along with edges (nodes 33A and 33B).

Subsequently, the id2 node is read from the graph 1. Since the label of the node with id2 is “C”, the node with id3 ′ having the label “C” is also read out in the graph 2. In addition, there is an edge between the id0 and id0 ′ nodes identified in the graph nodes 33A and 33B of the MCS candidates and the id1 and id1 ′ nodes and the id2 and id3 ′ nodes read this time. Check if it exists. In the example of

graphs

1 and 2 of FIG. 3A, no such edge exists. Therefore, the node of id2 and the node of id3 'read this time are added to the search tree as MCS candidates (

nodes

35A and 35B).

In the normal McGregor algorithm, if there is no edge between the graph node of the MCS candidate and the read node, the read node is not included in the MCS candidate. That is, when the MCS is obtained by a normal McGregor algorithm, the nodes id2 and id3 'are not added to the search tree of the MCS candidate in the

nodes

35A and 35B. However, as described above, the property search device according to the present embodiment adds a node to an MCS candidate even if there is no edge. This is because the specification of MCS in this embodiment is used to calculate the similarity of the floor plan. When calculating the similarity of the floor plan, even if the rooms are not connected, the same type of room itself is used. This is because it is considered that the existence fact should be reflected in the similarity.

(2.3. Calculation of similarity)
When the MCS is specified between the query graph corresponding to the search condition and the graph structure related to the read property information, the property search device determines the similarity between the query graph and the graph structure related to the read property information. Calculate (S109 in FIG. 1). There are various methods for calculating the degree of similarity. For example, it can be calculated by the following mathematical formula.

In this equation, q and g indicate a graph structure G, and can include elements of a node set V, an edge set E, a map L that returns a node label, and a map W that returns a width.

In the above formula, mcs is a maximum common subgraph (MCS) between q and g.

Γ is a weighting coefficient set according to the importance. For example, 100 can be set if the edge or node is set as important by the user, and 1 can be set if the edge or node is not set. The weighting value set by γ is not limited to this, and can be adjusted as appropriate. Further, in this embodiment, the importance (weight) that can be set for each node or edge is two stages, but the present invention is not limited to this, and the user may be able to set the importance in multiple stages. .

λ is a coefficient for determining how much the difference in area between rooms is reflected in the similarity of the calculation target. For example, when λ = 0.1 is set, the influence on the similarity given by the difference of 10 squares in the total size of the room becomes equal to the influence given by the difference in the number of nodes or the number of edges on the similarity. When it is desired to extract a room as close as possible to the size of the room set by the user in the query graph (to increase the similarity), λ may be set large. v _q, v _g is the corresponding node of each graph q and g.

That is, in the above formula for calculating the similarity sim (q, g), in the first term, the number of nodes that match between the query graph q and the graph structure g related to the read property information, and the importance of those nodes A value is calculated according to whether or not the degree is specified. In the second term, similarly, a value corresponding to the number of matching edges between the query graph q and the graph structure g related to the read property information, and the presence / absence of importance designation of those edges is calculated. In the third term, a value corresponding to the difference in the size of each room between the query graph q and the graph structure g is calculated. By summing these values, the similarity between the query graph q and the graph structure g is calculated.

Thus, after calculating the similarity between the graph structure related to the property information read from the DB and the query graph (S109 in FIG. 1), the property information for which the similarity is not yet calculated in the DB. If it remains (No in S111), the property search apparatus newly reads unprocessed property information from the DB and performs the same processing (S105 to S109). When the similarity calculation for all the property information in the DB is completed (Yes in S111), the property search apparatus sorts the property information according to the similarity and outputs a list of property information (S113). ).

(3. Functional configuration of the property search device)
Hereinafter, the functional configuration of the property search apparatus 100 according to the present embodiment will be described with reference to FIG. FIG. 4 is a functional block diagram illustrating a specific example of a functional configuration of the property search apparatus. In the present embodiment, the property search apparatus 100 includes a query graph input unit 111, a query graph display unit 113, a maximum common partial graph specification unit 115, a graph similarity calculation unit 117, a search result output unit 119, and a database (DB) 120. Including.

Here, as described above, the property search apparatus 100 is connected to an information processing apparatus (user terminal) such as a PC (Personal Computer) or a smartphone operated by a user via a network, for example, For example, a property search service can be provided as a website. In this case, the user checks and inputs various information on a browser operating on the user terminal.

However, the method for providing the property search service is not limited to this. For example, by installing a dedicated application on a user terminal such as a PC or a smartphone operated by each user and the application and the property search apparatus 100 cooperate, the property search service can be provided to the user. In this case, for example, when the user inputs a query graph to the user terminal and selects a predetermined search button, the query graph input unit 111 described later receives the query graph and searches for a property. As for the search result, the search result output unit 119 described later may transmit the search result to the application of the user terminal, and the application may display the search result on the display device.

Alternatively, for example, it may be possible to provide a user with a property search function as an application that operates on an information processing device such as a PC or a smartphone operated by each user who operates alone. In this case, for example, the property information 121 stored in the DB 120 may be sequentially updated by an external server or the like.

In the example of FIG. 4, the property search device 100 is described as one device. However, the present invention is not limited to this, and the function of the property search device 100 is realized by a plurality of cooperating computers. Is also possible.

The query graph input unit 111 receives a graph structure corresponding to a layout to be searched for a user to search for a property, that is, a query graph created according to a user operation. In addition, the query graph display unit 113 sequentially displays the query graph created by the user on the display device. Thereby, the user can create a query graph while checking the graph structure on the display device.

The maximum common partial graph specifying unit 115 specifies the maximum common partial graph (MCS) between the query graph created by the query graph input unit 111 and the graph structure indicated by the graph data 123 included in the property information 121. . To specify the MCS, a node label (corresponding to the type of room) and an edge included in both the query graph and the graph structure related to the property information 121 are used. That is, it is not always necessary to specify the size of the room and the importance indicating whether or not the node or edge is important for the search by the user in order to specify the MCS. Since the specific method for specifying the MCS has been described in “2.2. Specifying the MCS”, the description thereof is omitted here.

The graph similarity calculation unit 117 calculates the similarity between the query graph created by the query graph input unit 111 and the graph structure indicated by the graph data 123 included in the property information 121. The degree of similarity includes the number of nodes and edges included in the common subgraph (MCS) between the two, whether or not the nodes and edges are designated as important by the user, the difference in the size of each room, etc. Can be calculated based on The specific method for calculating the similarity is described in “2.3. Calculation of similarity” above, and thus the description thereof is omitted here.

The search result output unit 119 outputs a search result obtained by searching the property information 121 using the created query graph in order to present it to the user according to the similarity calculated by the graph similarity calculation unit 117. For example, on the display screen that displays the search result by the output of the search result output unit 119, the list of the property information 121 can be shown after sorting the property information 121 in descending order of similarity to the query graph.

The DB 120 manages property information 121 that is information related to each property to be searched. The property information 121 includes at least graph data 123 indicating a layout of each room as a graph structure. Other properties that are necessary for users to grasp the detailed contents of the property, such as exclusive area, nearest station, route and seller, contractor, address, floor plan, and photos showing the room and surrounding environment, etc. Information 121 can be included.

(4. Specific example of hardware configuration)
Hereinafter, a specific example of the hardware configuration of the property search apparatus 100 will be described with reference to FIG. As shown in FIG. 5, the property search apparatus 100 includes a control unit 501, a communication interface (I / F) unit 505, a storage unit 507, a display unit 511, and an input unit 513, each of which is a bus line. Connected via 515.

The controller 501 includes a CPU (Central Processing Unit, not shown), a ROM (Read Only Memory, not shown), a RAM (Random Access Memory) 503, and the like. The control unit 501 is configured to execute the control program 509 stored in the storage unit 507 so that the above-described property search process can be realized in addition to the function as a general computer. For example, the query graph input unit 111, query graph display unit 113, maximum common subgraph specifying unit 115, graph similarity calculation unit 117, and search result output unit 119 described with reference to FIG. 4 are temporarily stored in the RAM 503. Above, it is realizable as the control program 509 which operate | moves on CPU.

Further, the RAM 503 temporarily holds part or all of the property information 121 in addition to the code included in the control program 509. The RAM 503 is also used as a work area when the CPU executes various processes.

The communication I / F unit 505 is a device for performing data communication with another information processing apparatus such as a user terminal operated by the user by wire or wireless. For example, the communication I / F unit 505 can perform various inputs from the user terminal for creating the query graph and output of the created query graph and search results to the user terminal.

The storage unit 507 is a nonvolatile storage medium such as an HDD (Hard Disk Drive) or a flash memory. The storage unit 507 stores an operating system (OS), applications, and data (not shown) for realizing functions as a general computer. The storage unit 507 stores a control program 509. As described above, the query graph input unit 111, the query graph display unit 113, the maximum common subgraph specifying unit 115, the graph similarity calculation unit 117, and the search result output unit 119 illustrated in FIG. 4 are realized by the control program 509. Can do.

The display unit 511 is a display device for presenting information to the administrator. Specific examples of the display unit 511 include a liquid crystal display and an organic EL (Electro-Luminescence) display. The input unit 513 is a device for receiving input from the administrator. Specific examples of the input unit 513 include a keyboard, a mouse, and a touch panel.

Note that the property search apparatus 100 does not necessarily include the display unit 511 and the input unit 513. The display unit 511 and the input unit 513 may be connected to the property search apparatus 100 from the outside via various interfaces such as a USB (Universal Serial Bus) and a display port.

(5. Effects according to the present embodiment)
For example, when searching for a property such as a condominium or a detached house, the property search device 100 according to the present embodiment can search for a property based on the layout of the floor plan. For this purpose, when a query graph expressing the layout desired by the user is created, the property search apparatus 100 creates a property list including a layout having a high similarity to the layout indicated by the query graph and presents the list to the user. This makes it easy for the user to search for a property having a specific floor plan structure, such as when there are two or more rooms that are living-in.

Moreover, in the property search apparatus 100 according to the present embodiment, when creating a query graph, it is possible to specify a room (node) and a connection (edge) between the rooms that are emphasized by the user. By searching for a property by the property search device 100 according to the importance, it is possible to easily find a property that is close to the floor plan layout that is emphasized by the user.
Furthermore, the property search apparatus 100 according to the present embodiment performs property search in consideration of the size of the room. This makes it easier for the user to extract, for example, a property with a wider living room.

(6. Other embodiments)
In the above embodiment, the user himself / herself creates a graph structure (query graph) to be searched, and the property search apparatus 100 is described as receiving the created query graph. However, the present invention is limited to this. is not.

For example, after preparing in advance the query graph related to a plurality of types of floor plans with different characteristics (which may be virtual ones) in the property search apparatus 100, the user inputs whether they like or dislike each floor plan. It can be considered. For example, when a user evaluates N floor plans as being liable and M dislikes floor plans (N and M are integers of 0 or more), a query graph relating to those N and M floor plans. MCS identification and similarity calculation are performed between the actual floor plan and the search target. Furthermore, the user's implicit preference and the actual search target using the similarity between the query graph determined to be liable and the query graph determined to be disliked and the actual floor plan to be searched. The similarity SIM with the floor plan is calculated. In this case, for example, the following formula can be used to calculate the similarity SIM.

In this equation, pq is a floor plan query graph determined by the user to like, and nq is a floor plan query graph determined by the user to dislike. About sim (q, g), it is the same as the above-mentioned formula.

After the similarity is calculated in this way, the method for presenting the user with the property in which the property having a high similarity SIM is arranged at the top and the property having a property with a low similarity SIM arranged at the bottom is shown as a search result. It is the same.
By doing in this way, even if the user himself / herself does not input a graph, it is possible to search for an actual property with a floor plan according to the user's preference by simply inputting whether or not the user likes or dislikes.

(7. Appendix)
Note that the configurations of the above-described embodiments may be combined or some components may be replaced. The configuration of the present invention is not limited to the above-described embodiment, and various modifications may be made without departing from the scope of the present invention.

100: Property search device 111: Query graph input unit 113: Query graph display unit 115: Maximum common partial graph specification unit 117: Graph similarity calculation unit 119: Search result output unit 120: Database 121: Property information 123: Graph data 501 : Control unit 503: RAM
505: Communication interface unit 507: Storage unit 509: Control program 511: Display unit 513: Input unit 515: Bus line

Claims

Means for managing, for each property information, graph information in which rooms and connections between rooms are represented as a graph structure including nodes and edges;
An input means for receiving an input of a query graph created in response to a user operation and indicating a graph structure to be searched;
A calculation means for calculating a similarity between the graph structure related to each property information and the graph structure related to the query graph;
An information processing apparatus comprising: output means for outputting the property information according to the similarity.
The query graph includes information for weighting at least one of rooms or connections between rooms,
The calculating means calculates the similarity in consideration of the weighting;
The information processing apparatus according to claim 1.
The query graph includes room size information,
The calculating means calculates the similarity in consideration of the size of the room;
The information processing apparatus according to claim 1 or 2.
The calculation means calculates the similarity based on a maximum common partial graph between a graph structure related to each property information and a graph structure related to the query graph,
The information processing apparatus according to any one of claims 1 to 3.
The output means outputs a list of the property information according to the similarity.
The information processing apparatus according to any one of claims 1 to 4.
Information processing device
Managing the graph information expressed as a graph structure including nodes and edges, and the connection between rooms, for each property information;
Receiving a query graph created in response to a user operation and indicating a graph structure to be searched;
Calculating the similarity between the graph structure related to each property information and the graph structure related to the query graph;
An information processing method that performs the step of outputting the property information according to the similarity.