JP2020149120A - Information processing device, information processing method, and program - Google Patents

Abstract

To suitably analyze transaction data stored on a block.SOLUTION: An information processing device stores transaction data on a data block connected via a hash chain constructed in a network on the basis of a fact that the transaction has occurred. Then, the information processing device extracts data of a prescribed transaction among the stored transaction data, sets a specific transaction object among the transaction objects specified by the extracted transaction data to be a recommendation object person, performs a prescribed arithmetic operation on the basis of the transaction data for all the transaction objects related to the recommendation object person, and recommends an object matching the preference of the recommendation object person according to the result of the arithmetic operation.SELECTED DRAWING: Figure 14

Description

The present invention relates to information processing devices, information processing methods and programs.

In recent years, commercial transactions using virtual currencies such as Bitcoin (registered trademark) have been carried out. In commercial transactions using the virtual currency, a technology called blockchain is used to prevent fraud without the need for centralized management. In the blockchain, multiple transaction data, previous hash values and other information are defined as "blocks", and the trust of the information in the "blocks" by the process of consensus building within the network formed by all participants. The sex is guaranteed. However, when conducting commercial transactions of virtual currency by such blockchain technology, there is a problem that the number of cases that can be processed per second is small.

In order to deal with such a problem, for example, Non-Patent Document 1 discloses a technique called a hash graph.

"There is technology that goes beyond blockchain! What is the blockchain" Hashgraph "? , [Online], [Search on September 3, 2018], Internet <URL: https://coin-otaku.com/topic/7754>

The hash graph technology disclosed in Non-Patent Document 1 records transaction data with the approval of some randomly selected users, so that processing is faster than processing using blockchain technology. It will be feasible. However, all the techniques are common in that a plurality of blocks in which a plurality of transaction data are stored in a time series are connected, and the transaction data stored in the blocks is not sufficiently analyzed in a suitable manner.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an information processing device, an information processing method, and a program capable of suitably analyzing transaction data stored on a block. There is.

In order to achieve the above object, the information processing device according to the first aspect of the present invention is
A data storage means for storing the data of the transaction in a data block connected by a hash chain constructed in the network based on the occurrence of the transaction.
An extraction means for extracting the data of a predetermined transaction from the data of the transaction stored in the data storage means, and an extraction means.
An arithmetic means that performs a predetermined operation based on the transaction data extracted by the extraction means, and
It is provided with a recommendation means for recommending a target corresponding to the preference of the recommendation target person according to the calculation result by the calculation means.
The transaction data includes information that identifies the transaction target of the transaction.
The calculation means sets a specific transaction target among the transaction targets specified by the transaction data extracted by the extraction means as the recommendation target person, and all the transaction targets related to the recommendation target person. , Performs the predetermined calculation based on the transaction data.
It is characterized by that.

When the data of the transaction is newly stored in the data storage means, the extraction means newly extracts the data of the transaction.
The calculation means sets the transaction target specified by the transaction data newly extracted by the extraction means as the recommendation target person, and all the transaction targets related to the recommendation target person are converted into the transaction data. By performing the predetermined calculation based on the above, the calculation result already performed by the calculation means is updated.
You may do so.

The calculation means allocates an initial value to the recommendation target person as the predetermined calculation, and calculates all the transaction targets related to the recommendation target person from the transaction data based on the initial value. Calculate the value multiplied by the weight to be
You may do so.

The transaction data further includes attribute information indicating the transaction target attribute of the transaction.
The calculation means further allocates the initial value to the attribute indicated by the attribute information of the recommendation target person, and sets the initial value for all the transaction targets related to the transaction target having the same attributes as the recommendation target person. A value obtained by multiplying the reference by the weight calculated from the transaction data is further calculated.
The recommendation means recommends a target corresponding to the preference of the attribute of the recommendation target person in addition to the target corresponding to the preference of the recommendation target person according to the calculation result by the calculation means.
You may do so.

In order to achieve the above object, the information processing method according to the second aspect of the present invention is
It is an information processing method in an information processing device.
A data storage step for storing the data of the transaction in a data block connected by a hash chain constructed in the network based on the occurrence of the transaction.
An extraction step for extracting the data of a predetermined transaction from the data of the transaction stored in the data storage step, and
A calculation step that performs a predetermined calculation based on the transaction data extracted in the extraction step, and
It is provided with a recommendation step of recommending a target corresponding to the preference of the recommendation target person according to the calculation result of the calculation step.
The transaction data includes information that identifies the transaction target of the transaction.
In the calculation step, a specific transaction target among the transaction targets specified by the transaction data extracted in the extraction step is set as the recommendation target person, and all the transaction targets related to the recommendation target person are used. , Performs the predetermined calculation based on the transaction data.
It is characterized by that.

In order to achieve the above object, the program according to the third aspect of the present invention is
Computer,
A data storage means for storing the data of the transaction in a data block connected by a hash chain constructed in the network based on the occurrence of the transaction.
An extraction means for extracting a predetermined transaction data from the transaction data stored in the data storage means.
An arithmetic means that performs a predetermined operation based on the transaction data extracted by the extraction means,
It functions as a recommendation means for recommending a target corresponding to the preference of the person to be recommended according to the calculation result by the calculation means.
The transaction data includes information that identifies the transaction target of the transaction.
The calculation means sets a specific transaction target among the transaction targets specified by the transaction data extracted by the extraction means as the recommendation target person, and all the transaction targets related to the recommendation target person. , Performs the predetermined calculation based on the transaction data.
It is characterized by that.

According to the present invention, transaction data stored on a block can be suitably analyzed.

It is a block diagram which shows an example of the information processing system which concerns on embodiment of this invention. It is a block diagram which shows an example of the information processing apparatus which concerns on embodiment of this invention. It is a flowchart which shows an example of the initial registration process. It is a flowchart which shows an example of a data block registration process. It is a figure which shows the configuration example of a hash chain. It is a figure which shows an example of the transaction which occurred. It is a figure which shows the example of the registration contents of a hash chain. It is a flowchart which shows an example of a graph generation process. It is explanatory drawing for demonstrating the calculation example of a weight. It is a figure which shows an example of the directed bipartite graph generated by the graph generation process. It is a flowchart which shows an example of a recommendation process. It is a flowchart which shows an example of the recommendation value calculation process. It is explanatory drawing in the case of calculating the recommendation value. It is explanatory drawing in the case of calculating the recommendation value. It is explanatory drawing in the case of calculating the recommendation value. It is explanatory drawing in the case of calculating the recommendation value. It is explanatory drawing in the case of calculating the recommendation value. It is explanatory drawing for demonstrating the change of the recommendation value. It is explanatory drawing in the case of analyzing the tendency for each attribute. It is explanatory drawing in the case of estimating the attribute. It is explanatory drawing for demonstrating the effectiveness of the recommendation based on the attribute information. It is explanatory drawing when paying attention to the order of transaction data. It is explanatory drawing when paying attention to the order of transaction data.

An example in which the information processing device 100 in the present invention is applied to the information processing system 1 shown in FIG. 1 will be described. In the information processing system 1, as shown in FIG. 1, the information processing devices 100A to 100C (the information processing devices 100A to 100C are also simply referred to as the information processing device 100) and the terminals 100D of the tourists A to D each connect the network 210. It is connected so that it can communicate via. The terminals 100D of tourists A to D can each function as the information processing device 100 by themselves, but in order to facilitate understanding, in this embodiment, these are referred to as the information processing device 100. It will be explained separately.

The information processing device 100 is an information terminal (so-called computer) such as a mobile phone, a smartphone, a tablet, or a PC (Personal Computer), and constructs a distributed network 210 such as P2P (Peer to Peer). The information processing system 1 is not limited to the P2P type system, and may be, for example, a cloud computing type.

The information processing device 100 receives data (transaction data) indicating an event occurring between a tourist and a tourist destination (sightseeing spot) from the network 210, such as when tourist C visits tourist spot 2. Based on the above, it functions as a hash chain registration device that generates a new data block, stores the received data, and connects it to an existing hash chain. As described above, since the terminals 100D of the tourists A to D may function as the information processing device 100, the information processing device 100 has a function of generating transaction data and distributing it to the network 210. It may also function as a generator. The content registered by the hash chain registration device is shared by each information processing device 100 via the network 210. Further, usually, a certain number of transaction data is stored in one data block (a certain number of transaction data are collectively stored in one block), but in this embodiment, in order to facilitate understanding, It is assumed that one transaction data is stored in one data block.

In this embodiment, transaction data generated by an event occurring between a tourist and a tourist spot (generated by the terminal D functioning as a transaction generator) is stored in one unit called a data block. It is managed by each type of hash chain that concatenates the data blocks in chronological order. For example, the transaction data of an event such as tourist C visiting tourist spot 2 has the same contents as a new data block for each of the tourist spot C type hash chain and the tourist spot 2 type hash chain. Is registered and managed.

In addition, the information processing device 100 in this embodiment presents recommended tourist spots to the target tourists by targeting the registered data specified by the user from the data blocks included in each type of hash chain. It also functions as a recommendation device.

The function of the information processing device 100 as a recommendation device in this embodiment will be described in detail later, but transaction data specified by the user is extracted from a plurality of registered data blocks to perform a predetermined operation and is directed. By generating a two-part graph and calculating the recommendation value, recommended tourist spots are presented to the target tourists.

Next, the configuration of the information processing apparatus 100 in this embodiment will be described with reference to FIG. In the illustrated example, the information processing device 100A is taken as an example, but the same applies to the case where the information processing devices 100B and 100C and the terminal 100D function as the information processing device 100, and thus the description thereof will be omitted. Although not shown, a functional unit having a function as a transaction generator may be provided.

As shown in FIG. 2, the information processing device 100A (hereinafter, information processing device 100) includes a storage unit 110, a control unit 120, an input / output unit 130, a communication unit 140, and a system bus that connects them to each other. (Not shown) and.

The storage unit 110 includes a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The ROM stores a program executed by the CPU (Central Processing Unit) of the control unit 120 and data necessary for executing the program in advance (not shown). Specifically, in this embodiment, a program for making the information processing device 100 function as a hash chain registration device and a recommendation device is pre-installed. RAM stores data that is created or modified during program execution. The storage unit 110 stores various types of hash chains 111 as main information used by the program executed by the control unit 120.

The hash chain 111 of each type is a hash chain corresponding to each type of tourists A to D and tourist spots 1 to 4, and each hash chain is a concatenation of data blocks in which transaction data is stored. Is. The hash chain 111 of each type is stored in the storage unit 110 in the data block registration process described later, and is referred to by the graph generation process and the recommendation process described later.

The control unit 120 is composed of a CPU, an ASIC (Application Specific Integrated Circuit), and the like. The control unit 120 operates according to the program stored in the storage unit 110, and executes the process according to the program. The control unit 120 includes an initial registration unit 121, a synchronization processing unit 122, a data block registration unit 123, a data acquisition unit 124, and a graph generation unit as main functional units provided by a program stored in the storage unit 110. A unit 125 and a recommendation processing unit 126 are provided.

The initial registration unit 121 realizes a function of registering initial information necessary for generating a new type of hash chain based on an operation on the user's input / output unit 130 and a request received via the communication unit 140. It is a functional part to do. Although details will be described later, the initial registration unit 121 has a function of registering, for example, the address, index value, etc. of tourist C as initial information necessary for generating a hash chain of the type of tourist C (that is, described later). It has a function to register the base information of the hash chain to be used.

The synchronization processing unit 122 is a functional unit that realizes a function of synchronizing various different hash chains stored in the information processing device 100 with various different hash chains stored in the other information processing device 100. Although details will be described later, the synchronization processing unit 122 has a function of synchronizing the transaction data with the other information processing device 100 before and after registering the transaction data in the data block of the information processing device 100 in the data block registration process described later. are doing. Further, the synchronization processing unit 122 has a function of synchronizing the base information of the hash chain with another information processing device 100 in the initial registration process described later.

The data block registration unit 123 is a functional unit that realizes a function of adding a data block including transaction data of a generated transaction to a hash chain of a corresponding type among various hash chains 111 of the information processing apparatus 100. is there. As will be described in detail later, for example, when tourist C visits tourist spot 2, the data block registration unit 123 adds the content (transaction data) to the hash chains of each type of tourist C and tourist spot 2. , Has a function to register as a new data block.

The data acquisition unit 124 is a functional unit that realizes a function of acquiring transaction data included in a data block included in each registered hash chain 111 based on a user's designation.

The graph generation unit 125 is a functional unit that realizes a function of generating a directed bipartite graph based on transaction data acquired by the data acquisition unit 124.

The recommendation processing unit 126 is a functional unit that organizes a directed bipartite graph generated by the graph generation unit 125 and realizes a function of presenting recommended tourist spots to the recommendation target person.

Each of these functional units cooperates to make the information processing device 100 function as a hash chain registration device and a recommendation device.

The input / output unit 130 is composed of a keyboard, a mouse, a camera, a microphone, a liquid crystal display, an organic EL (Electro-Luminence) display, and the like, and is a device for inputting / outputting data.

The communication unit 140 is a device for communicating with another information processing device 100 or terminals C to G via a network 210.

The above is the configuration of the information processing device 100. Subsequently, the operation of the information processing apparatus 100 and the like will be described with reference to FIGS. 3 to 17. First, the operation of the information processing device 100 when functioning as a hash chain registration device will be described with reference to FIGS. 3 to 7.

First, the initial registration process shown in FIG. 3 is performed based on the operation (input operation) on the input / output unit 130 of the user and the request (registration request) received from the terminal 100D. The initial registration process is a process of registering the initial information necessary for generating a new type of hash chain. In this example, in order to facilitate understanding, a case where the initial registration of tourist C is performed will be described. When the initial registration process is started, the information processing apparatus 100 generates type data by the function of the initial registration unit 121 (step S101). Specifically, in step S101, the information of tourist C and the address of tourist C, which are the type information included in the input operation and the registration request, are used as the information for generating the hash chain of the type of tourist C. Register in the storage unit 110. The information may be registered as a hash chain of type C among the hash chains 111 of each type. Further, the type information may not be registered and only the address may be registered. In this embodiment, the addresses of tourists A to D are 01 to 04, and the addresses of tourist spots 1 to 4 are X1 to X4, respectively.

Next, the information processing apparatus 100 registers the initial value “0” in the index value of the registered tourist C type hash chain by the function of the initial registration unit 121 (step S102). When the index value is "0", it indicates that the hash chain is in the initial state (the state in which no data block is registered). That is, in the processes of steps S101 and S102, it can be said that a hash chain (hash chain base information) in which the data blocks are not registered is generated although the data blocks are not connected by a chain. In this state, since the data blocks are not connected by a chain, there is a problem that the data can be deleted and it cannot be verified that the data has been deleted. Therefore, in order to improve the tamper resistance, a special transaction may be generated for a hash chain selected at random, and a process of concatenating data blocks having an index value of "0" in the chain may be performed.

After executing the process of step S102, the information processing device 100 registers the content in the other information processing device 100 by the function of the synchronization processing unit 122 (step S103), and ends the initial registration process. Specifically, in step S103, the contents of the hash chain (base information of the hash chain) of the tourist C type in the information processing device 100 are converted into various different hash chains 111 in the storage unit 110 of the other information processing device 100. Copy to. As a result, the hash chain (base information of the hash chain) of the tourist C type is shared by all the information processing devices 100.

Next, the data block registration process for registering the generated transaction in the data block of the corresponding type of hash chain will be described. The data block registration process is executed by receiving transaction data corresponding to the generated transaction. Although the transaction data is received by the terminal 100D, it may be input to the input / output unit 130. In this embodiment, as described above, one data block is created and registered for one transaction. Therefore, the data block registration process is repeatedly executed when there are a plurality of transactions to be registered. Will be done.

FIG. 4 is a flowchart showing an example of the data block registration process. In this example, in order to facilitate understanding, the data block registration process when a transaction that tourist C visits tourist spot 2 occurs (transition from the state of FIG. 5 (A) to the state of (B)). Case) will be described.

When a transaction that tourist C visits tourist spot 2 occurs, transaction data corresponding to the transaction is transmitted from the terminal 100D of tourist C. As a result, the information processing apparatus 100 starts the data block registration process by the function of the data block registration unit 123. Note that, for example, the data block registration process may be started by receiving transaction data having the same contents from both the terminal 100D of the tourist C and the terminal of the tourist spot 2. In this case, if transaction data with different contents is received, an error may be output.

When the data block registration process is started, the information processing apparatus 100 first uses the function of the synchronization processing unit 122 to cause the target hash chain in the information processing apparatus 100 and the target hash chain in the other information processing apparatus 100. It is determined whether or not they are synchronized (step S201). Specifically, in the process of step S201 in this example, the index value and hash value of the latest data block included in the hash chain of the tourist C type among the hash chains 111 of each type, and the type of the tourist spot 2 Whether or not the index value and the hash value of the latest data block included in the hash chain of the above are synchronized by determining whether or not they are the same as those stored in all the other information processing devices 100. Is determined.

Explaining using the example shown in FIG. 5 (A), among the hash chains 111 of each type of the information processing device 100 (referring to the information processing device 100 that has received transaction data, the same applies hereinafter), tourist C The data block with the index value "N-1" is registered as the latest data block in the hash chain of the tourist spot 2, and the data block with the index value "Z-1" is the latest data in the hash chain of the tourist spot 2. It is registered as a block. As shown in the figure, the hash chain data block (mth data block) in this embodiment includes the registered data (transaction data) and the data block at the end (one before) of the current hash chain (one before). The point that the hash value (256 bits) of the m-1th data block) is included is the same as the conventional blockchain mechanism. Further, the data block includes, for example, additional data for analysis such as a time stamp indicating the date and time when the registered data was registered in the data block (not shown), which is the same as the conventional blockchain mechanism. is there.

In the process of step S201 of FIG. 4, the index value "N-1" and the hash value of the N-1 and the index value "Z-1" and the hash value of the Z-1 are all other information processing devices. By determining whether or not it matches what is stored in 100, it is determined whether or not it is synchronized. For example, when there is an information processing device 100 in which the index value "N" and the hash value of the N are stored, the hash chain information of the tourist C type stored in the information processing device 100 is stored. It is determined to be old (in the case of the index value "Z" and the hash value of the Z, the hash chain of the tourist spot 2 type). On the other hand, when there is an information processing device 100 in which the index value "N-2" and the hash value of the N-2 are stored, the hash of the type of tourist C stored in the information processing device 100. It is determined that the chain information is old (in the case of the index value "Z-2" and the hash value of the Z-2, the hash chain of the tourist spot 2 type).

When not synchronized (step S201; No), the information processing device 100 uses the function of the synchronization processing unit 122 to transfer the contents of the latest hash chain to the information processing device 100 in which the hash chain determined to be old is registered. The synchronization process for copying is executed (step S202). In the process of step S202, the entire hash chain may be copied, or only the missing data blocks (including the index value and the hash value) may be copied.

When it is determined in step S201 that the data is synchronized (step S201; Yes), or when the process of step S202 is executed, the information processing apparatus 100 stores a data block in which the same registered data is stored in the target hash chain. Each is added (step S203). As described above, when adding a new data block (mth data block) to the hash chain, the registered data (transaction data) and the data block at the end (one before) of the current hash chain (first data block). The point that the hash value (256 bits) of the m-1 data block) is stored in the data block and connected to the data block (the previous m-1 data block) at the end (one before) of the current hash chain is It is the same as the conventional blockchain mechanism.

Specifically, in the process of step S203, as shown in FIG. 5B, the data block having the index value “N” is the data block having the index value “N-1” in the hash chain of the tourist C type. In addition to being linked to, the data block having the index value "Z" is linked to the data block having the index value "Z-1" in the hash chain of the tourist spot 2 type. Each concatenated data block contains the same transaction data. Specifically, the content indicating that the tourist C visited the tourist spot 2 (From: 03) (To: X2) is registered. As shown, "From" and "To" indicate which tourists visited which tourist spot. In addition, for example, when indicating that the customer has purchased the product at the store, the magnitude of the value may be indicated by further registering data indicating, for example, the amount of movement of coins. The hash value of N-1 is stored in the data block of the index value "N" in the hash chain of the tourist C type, and the data block of the index value "Z" in the hash chain of the tourist spot 2 type is stored. Stores the hash value of Z-1. As described above, the information processing apparatus 100 in this embodiment adds the same transaction data to each corresponding hash chain.

Returning to FIG. 4, after executing the process of step S203, the information processing device 100 registers the hash chain of the type to which the data block is added in step S203 with the other information processing device 100 by the function of the synchronization processing unit 122. By doing so, the information processing apparatus 100 and the other information processing apparatus 100 are synchronized (step S204), and then the data block registration process is terminated. In the process of step S204, the contents of each of the types of hash chains of tourist C and tourist spot 2 in the information processing device 100 are copied to various different hash chains 111 in the storage unit 110 of the other information processing device 100. To do. In the process of step S205, the entire hash chain may be copied, or only the missing data block (including the index value and the hash value) may be copied, as in the process of step S202.

In this way, each time a transaction occurs, the data block registration process shown in FIG. 4 is performed, and the transaction data is registered in the target hash chain. For example, when 9 transactions occur as shown in FIG. 6 (A) (10 in the case shown in FIG. 6 (B)), the data block registration process shown in FIG. 4 is performed 9 times (FIG. 6). In the case shown in (B), it is performed 10 times), and each transaction data is registered in the hash chain of the target type. FIG. 7 shows the registration contents of each hash chain when the data block registration process shown in FIG. 4 is performed 9 times for the 9 transactions shown in FIG. 6 (A). In the example shown in FIG. 7, the hash value and the additional data for analysis stored in each data block are omitted for ease of understanding.

Specifically, the transaction data that the tourist A corresponding to the first transaction in FIG. 6 (A) visited the tourist spot 1 is shown in FIG. 7 by performing the data block registration process shown in FIG. It is registered in the data block (1) in the hash chain of the tourist A type and the data block (1) in the hash chain of the tourist spot 1 type. The transaction data that the tourist A corresponding to the second transaction of FIG. 6 (A) visited the tourist spot 4 is the tourist shown in FIG. 7 by performing the data block registration process shown in FIG. 4 again. It is registered in the data block (2) in the hash chain of the type A and the data block (1) in the hash chain of the tourist spot 4. Then, the transaction data that the tourist B corresponding to the third transaction of FIG. 6 (A) visited the tourist spot 2 is shown in FIG. 7 by performing the data block registration process shown in FIG. 4 again. It is registered in the data block (1) in the hash chain of the tourist B type and the data block (1) in the hash chain of the tourist spot 2 type. By performing the data block registration process for the number of times corresponding to the number of transactions in this way, each transaction data is linked to the corresponding type of hash chain (for the fourth and subsequent transactions in FIG. 6 (A)). The same applies).

The above is the operation of the information processing device 100 when it functions as a hash chain registration device. Subsequently, the operation of the information processing device 100 when functioning as the recommendation device will be described with reference to FIGS. 8 to 17. In order to facilitate understanding, in the following example, nine transactions shown in FIG. 6A are generated, and nine data block registration processes are performed corresponding to each transaction, which is shown in FIG. 7. It is assumed that the data block of each hash chain is registered as described above. Further, after that, the tenth transaction shown in FIG. 6B occurs, and the content of the tenth transaction (tourist D visits tourist spot 2) is the hash chain of the type of tourist D and the hash chain. It will be described below assuming that each of the tourist spots 2 is registered in the hash chain of the type.

FIG. 8 is a flowchart showing an example of the graph generation process. The graph generation process is started based on an operation (input operation) on the input / output unit 130 of the user. The graph generation process is a process for acquiring transaction data for a period desired by the user from each type of hash chain 111 and generating a directed bipartite graph. Details of the directed bipartite graph will be described later. In this embodiment, the data block registration process shown in FIG. 4 generates a hash chain of the registered contents shown in FIG. 7 (that is, nine transactions shown in FIG. 6A are generated). A case where a directed bipartite graph for all the blocks shown in FIG. 7 is generated will be described as an example in a state where the data block registration process is performed 9 times corresponding to each transaction).

When the graph generation process is started, the information processing apparatus 100 first determines whether or not the graph generation target is specified by the function of the data acquisition unit 124 (step S301). Specifically, in step S301, it is determined whether or not the target period (start time and end time) is specified by the user's input operation.

When it is determined in step S301 that the target period is specified (step S301; Yes), the information processing apparatus 100 uses the function of the data acquisition unit 124 to perform transaction data (transaction data of the data block corresponding to the specified period). Target data) is acquired (step S302). In step S302, the data block corresponding to the period specified by the user's input operation is specified based on the time stamp included in the non-analyzable data, and the transaction data (target data) registered in the data block is acquired. To do.

On the other hand, when it is determined in step S301 that the target period is not specified (step S301; No), the information processing apparatus 100 uses the function of the data acquisition unit 124 to perform transaction data (target data) of all data blocks. (Step S303). As described above, in this embodiment, since the directed bipartite graph for all the blocks shown in FIG. 7 is generated, in step S303, the transaction data of all the blocks shown in FIG. 7 is acquired. To.

After executing the process of step S302 or step S303, the information processing apparatus 100 calculates the weight for each type based on the contents of the transaction data by the function of the graph generation unit 125 (step S304). Specifically, in the process of step S304, as shown in FIG. 9A, tourist C has visited tourist spot 2 four times and tourist spot 3 once, for a total of five times. In step S304, as shown in FIG. 9B, the number of visits to each tourist spot is divided by the total value to calculate the weight. Since tourist C has visited tourist spot 2 four times and visited tourist spot 3 once, the weights are 0.80 and 0, respectively, as shown in FIG. 9 (B). It becomes 20. In step S304, this process is performed for all tourists, and the same process is performed for tourist spots. For example, with respect to the tourist spot 1, since only the tourist A visits twice, the weight from the tourist spot 1 to the tourist A is 1.00 (see FIG. 10). As for tourist spot 2, since tourist B visits once and tourist C visits four times, the weight of tourist spot 2 to tourist B is 0.20 and the weight of tourist C is 0.80. (See FIG. 10).

Returning to FIG. 8, after calculating the weights for all the types in the process of step S305, the information processing apparatus 100 uses the function of the graph generation unit 125 to set the start point (From) and the end point (To) of each type. A directed bipartite graph is generated (step S305) in which each of the transaction targets is set as a vertex and each vertex is connected by an edge, and the graph generation process is completed. Specifically, in the process of step S305, the vertices starting from each tourist and ending at each tourist spot are connected by edges according to the presence or absence of a visit, and a directed bipartite graph as shown in FIG. 10 is generated. To do. In the illustrated example, in order to facilitate understanding, the side is represented only by the arrow from the tourist side toward the tourist spot, and the weight calculated in step S304 of FIG. 8 is used for sightseeing with the tourist side. Although it is described for each spot side, in reality, the weight in one side (from the tourist to the tourist spot) is represented by the thickness of the side (arrow) (the thicker the side, the heavier the weight). In this case, the weight from the tourist spot side (from the tourist spot to the tourist direction) may be expressed by the thickness of the side (arrow) from the tourist spot side to the tourist side. That is, in step S305, the vertices may be connected from each direction by the sides (arrows) having the thickness corresponding to the weight calculated in step S304.

Since the tourist A in this embodiment is visiting the tourist spot 1 and the tourist spot 4 (see FIG. 6 (A) and FIG. 7), as shown in FIG. 10, the top of the tourist A is sightseeing. It is connected to Spot 1 and Tourist Spot 4. Since tourist B is visiting tourist spot 2 (see FIG. 6 (A) and FIG. 7), the apex of tourist B is connected to tourist spot 2. Since tourist C visits tourist spot 2 and tourist spot 3 (see FIG. 6 (A) and FIG. 7), as shown in FIG. 10, the top of tourist C is tourist spot 2 and tourist spot. It is tied with 3. As shown in FIG. 10, the directed bipartite graph includes the weight calculated in step S304 of FIG. 8 according to the number of visits (the thickness of the side (arrow) from the tourist side and the thickness of the side (arrow)). The weight from each direction is represented by the thickness of the side (arrow) from the tourist spot side).

By executing the graph generation process in this way, the directed bipartite graph shown in FIG. 10 is generated. In such a state, when the tenth transaction shown in FIG. 6B occurs, the content of the tenth transaction (tourist D visits tourist spot 2) by the data block registration process of FIG. Is registered in the hash chain of the tourist D type and the hash chain of the tourist spot 2 type (not shown). In such a situation, the recommendation processing when presenting a recommended tourist spot to tourist D (when recommending a recommended tourist spot with tourist D as a recommendation target (also referred to as a recommendation target)) is as follows. explain. Normally, since various transactions occur, the directed bipartite graph shown in FIG. 10 is also updated in real time in accordance with this, but in this embodiment, in order to facilitate understanding, sightseeing It is assumed that only the transaction in which the customer D has visited the tourist spot 2 occurs, and the other transactions do not occur. That is, the directed bipartite graph shown in FIG. 10 will be described below assuming that it has not been updated in real time.

FIG. 11 is a flowchart showing an example of the recommendation process. The recommendation process is executed by receiving an execution instruction from the terminal 100D, but may be started based on an operation (input operation) on the input / output unit 130 of the user.

When the recommendation processing shown in FIG. 11 is started, the information processing apparatus 100 acquires the transaction data of the designated data block (target block) by the function of the recommendation processing unit 126 (step S401). In this example, since the tenth transaction (tourist D visited tourist spot 2) shown in FIG. 6B is the target, the hash chain of the tourist D type and the tourist spot 2 type Acquire transaction data of the relevant contents from the hash chain. Regarding the acquisition of the target transaction data, for example, the contents and period of From and To may be specified.

After executing the process of step S401, the information processing apparatus 100 organizes the directed bipartite graph generated by the graph generation process of FIG. 8 by the function of the recommendation processing unit 126 (step S402). In this example, since the transaction that tourist D visited the tourist spot 2 was added, the directed bipartite graph corresponding to the new transaction data is added to the directed bipartite graph shown in FIG. To. Specifically, in step S402 in this example, as shown in FIG. 13, a process of connecting the apex of the tourist D with the tourist spot 2 is performed. In this example, since the transaction was that a new tourist D visited the tourist spot 2, an example of connecting with a new side (arrow) was shown. For example, a tourist spot that an existing tourist has already visited. In the case of visiting again, that is, when the number of visits increases, the thickness of the side (arrow) may be organized into a thicker mode according to the weight.

Further, in this example, since the case where the transaction that the tourist D visited the tourist spot 2 is added, the directed bipartite graph corresponding to the new transaction data is shown in FIG. Although the example of being added to the part graph is shown, the directed bipartite graph corresponding to the past transaction data may be subtracted. For example, when subtracting the directed two-part graph corresponding to the first transaction shown in FIG. 6 (A) (when deleting the transaction that tourist A has visited tourist spot 1), tourist A However, since the number of visits to tourist spot 1 is reduced by 1 (from 2 times to 1 time), the thickness of the side (arrow) from tourist A to tourist spot 1 in FIG. 10 is reduced according to the weight. (Note that the side (arrow) from tourist spot 1 to tourist A is organized because the weight is not changed because only tourist A is visiting the tourist spot 1. It doesn't have to be). Further, when subtracting the directed bipartite graph corresponding to the second transaction shown in FIG. 6 (A) (when deleting the transaction that tourist A has visited tourist spot 4), tourist A Since the number of visits to tourist spot 4 is 0, the side (arrow) from tourist A to tourist spot 4 in FIG. 10 is deleted, and the side from tourist spot 4 to tourist A ( Organize to delete by matching the arrow).

After executing the process of step S402 of FIG. 11, the information processing apparatus 100 executes the recommendation value calculation process of calculating the recommendation value by the function of the recommendation processing unit 126 (step S403). Specifically, in step S403, the score (value 1) of the tourist spot visited by the recommended tourist is calculated, and the score (value 2) of the tourist who visited the tourist spot is calculated based on the value 1. Then, the score of the tourist spot visited by the tourist is used as the recommended value, and is calculated based on the value 2. In other words, the score of the tourist who is performing the behavior similar to the recommended person is calculated, and the behavior (sightseeing spot) for the recommended tourist is calculated from the score and the behavior of the tourist (visited tourist spot). Calculate the recommendation value. Hereinafter, the recommendation value calculation process in step S403 of FIG. 11 when the tourist D visits the tourist spot 2 will be described with reference to FIGS. 12 to 14.

FIG. 12 is a flowchart showing an example of the recommendation value calculation process executed in the process of step S403 of FIG. In the recommendation value calculation process shown in FIG. 12, the information processing device 100 first gives a value of 1.00, which is an initial value, to the tourists (recommendation target persons) to be recommended, and the apex connected to the recommendation target person. (That is, the transaction target related to the recommended person) is multiplied by the weight in the forward direction (from the tourist to the tourist spot direction) to calculate the value 1 (step S501). Specifically, in step S501, as shown in FIG. 13, a value of 1.00 is given to the tourist D who is the recommended target. Since only tourist spot 2 is connected to tourist D, the value of tourist spot 2 seen from tourist D is calculated as 1.00 by multiplying by the weight of 1.00 (Fig.). Value 1) of tourist spot 2 shown in 13. After executing the process of step S501 in FIG. 12, the information processing apparatus 100 reverses the direction (from the tourist spot) with respect to the apex connected to the apex having the value 1 (that is, the transaction target related to the transaction target having the value 1). The value 2 is calculated by multiplying the weight in the tourist direction (step S502). Specifically, in step S502, the weight from the tourist spot 2 side is 0.20 for the tourist B and 0.80 for the tourist C. Therefore, the sightseeing seen from the tourist spot 2 The value of tourist B is calculated as 0.20 (value 2 of tourist B shown in FIG. 13) by multiplying 1.00 of value 1 by 0.20 of the weight, and the value of tourist C is set to value 1. Multiply 1.00 by 0.80 of the weight to calculate 0.80 (value 2 of tourist C shown in FIG. 13).

After executing the process of step S502 of FIG. 12, the information processing apparatus 100 forwards (from the tourist) about the vertex connected to the vertex having the value 2 (that is, the transaction target related to the transaction target having the value 2). The recommendation value is calculated by multiplying the weight of the tourist spot direction (step S503), and the recommendation value calculation process is completed. Specifically, in step S502, as shown in FIG. 14, each value of tourist B and tourist C (value 2 of tourist B and tourist C shown in FIG. 14) is set to a tourist spot from the tourist side. Multiply the weight in the lateral direction to calculate the recommendation value for each tourist spot. More specifically, the information processing apparatus 100 calculates 0.20 by multiplying 0.20 of the value 2 of tourist B by 1.00 the weight to the tourist spot 2 by the function of the recommendation processing unit 126. Further, since there is a side from tourist C to tourist spot 2, 0.64 is calculated by multiplying 0.80 of the value 2 of tourist C by 0.80 of the weight to tourist spot 2. Then, these values are totaled (0.20 + 0.64) to calculate the recommendation value of the tourist spot 2 of 0.84 (the recommended value of the tourist spot 2 shown in FIG. 14). Further, the information processing apparatus 100 calculates 0.16 by multiplying 0.80 of the value 2 of tourist C by 0.20 of the weight to the tourist spot 3 by the function of the recommendation processing unit 126. Since only tourist C has visited the tourist spot 3 and tourist B has not visited the tourist spot 3, the value of 0.16 is used as it is as the recommended value of the tourist spot 3 (the recommended value of the tourist spot 3 shown in FIG. 14). ).

As described above, in the recommendation value calculation process shown in FIG. 12, a value of 1.00 is given to the recommendation target person, and the score (value 2) of the person who has taken an action similar to that of the recommendation target person is used for the recommendation target person. Calculate the recommendation value of.

Returning to FIG. 11, after calculating the recommendation value in the process of step S403, the information processing apparatus 100 uses the function of the recommendation processing unit 126 to select a recommended tourist spot based on the recommendation value calculated in the process of step S403. By displaying the information on the input / output unit 130, the tourist D is recommended (step S404), and the recommendation process is completed. In the example shown in FIG. 14, the recommendation value of the tourist spot 2 is 0.84, and the recommended value of the tourist spot 3 is 0.16. In step S404, the tourist spot with the highest recommended value is usually displayed, but since tourist D has already visited tourist spot 2 most recently, in this example, tourist spot 3 which is the runner-up is displayed. .. If the tourist spot with the highest recommended value has not been visited, the tourist spot may be displayed. Furthermore, even if the tourist spot with the highest recommended value has not been visited most recently, if the tourist spot has already been visited, the next tourist spot may be displayed. Further, for example, it may be decided whether or not to display according to the number of visits, such as not displaying when visited three times or more. Further, the number of tourist spots to be displayed does not have to be one, and the top three may be displayed. In this case, if the tourist spots that have already been visited are included, the tourist spots with the next recommended value may be displayed (such as the top four). In step S404, when the recommendation process is executed by receiving an execution instruction from the terminal 100D, it is displayed on the input / output unit 130 of the terminal 100D, and an operation (input operation) on the user's input / output unit 130 is performed. ), It may be displayed on the input / output unit 130.

As described above, the information processing apparatus 100 in this embodiment extracts data for an arbitrary period from the transaction data stored in the data blocks connected by the hash chain, and generates a directed bipartite graph. Then, the transaction data generated after the directed bipartite graph is generated is extracted from the data block to generate a new directed bipartite graph, and the directed bipartite graph is organized by adding them to the latest. Can be transformed into. Further, by subtracting the directed bipartite graph of the portion corresponding to the past transaction data, the influence of the past portion over time can be reduced from the directed bipartite graph. That is, by performing addition or subtraction, it is possible to arbitrarily adjust the time axis such as which period of the transaction is to be recommended. Therefore, the transaction data can be preferably analyzed, and further, it is possible to recommend to the target person based on the analysis result.

Next, the transition of the recommendation value due to the change in the behavior of tourists will be described with reference to FIGS. 15 to 18. The contents of the tenth transaction (Tourist D visited Tourist Spot 2) shown in FIG. 6 (B) were registered in the hash chain of the tourist D type and the hash chain of the tourist spot 2 type, respectively. After that, by executing the recommendation process, the directed two-part graph shown in FIG. 14 is generated as described above, and the tourist spot 3 is recommended to the tourist D. A case where the tourist D further visits the tourist spot 1 in such a state will be described below as an example.

When a new transaction occurs in which tourist D further visits tourist spot 1, the data block registration process of FIG. 4 changes the content of the transaction to the hash chain of tourist D type and the tourist spot 1 type. Each is registered in the hash chain (not shown). Then, when the recommendation process shown in FIG. 11 is executed, in the process of step S401, the information processing device 100 obtains transaction data having the contents from the hash chain of the tourist D type and the hash chain of the tourist spot 1 type. get.

Subsequently, the information processing apparatus 100 organizes the directed bipartite graph shown in FIG. 14 by the process of step S402 (see FIG. 15). Specifically, the side (arrow) connecting the apex of tourist D and tourist spot 1 is added to the directed bipartite graph shown in FIG. 14 (see FIG. 15). In this case, the number of sides of tourist D is two, and the number of visits is one in each case. Therefore, the weight from tourist D to tourist spot 1 is 0.5, and to tourist spot 2. The weight of is 0.5. Therefore, in the process of step S402, the thickness of the side from the tourist D to the tourist spot 2 in FIG. 14 is organized into a thickness corresponding to the weight of 0.5, and the thickness of the side to the tourist spot 1 is formed. Is added with a thickness corresponding to a weight of 0.5 (see FIG. 15).

Subsequently, in the process of step S403 of FIG. 11, that is, the recommendation value calculation process of FIG. 12, the information processing device 100 gives a value of 1.00 to the tourist D who is the recommendation target, as shown in FIG. Since the sides connected to tourist D are tourist spot 1 and tourist spot 2, the value of tourist spot 1 seen from tourist D is 0.50 by multiplying each weight by 0.50. (Value 1 of tourist spot 1 shown in FIG. 15) and value of tourist spot 2 are calculated as 0.50 (value 1 of tourist spot 2 shown in FIG. 15) (step S501 in FIG. 12).

Further, since the weight from the tourist spot 1 side is 1.00 to the tourist A, the value of the tourist A seen from the tourist spot 1 is set to a value 1 in the process of step S502 of FIG. Is 0.50 multiplied by 1.00 of the weight to obtain 0.50 (value 2 of tourist A shown in FIG. 15). Further, since the weight from the tourist spot 2 side is 0.20 for the tourist B and 0.80 for the tourist C, in the process of step S502 of FIG. 12, from the tourist spot 2 The value of tourist B seen is calculated as 0.10 (value 2 of tourist B shown in FIG. 15) by multiplying 0.50 of the value 1 by 0.20 of the weight, and the value of tourist C is calculated. It is calculated as 0.40 by multiplying 0.50 of the value 1 by 0.80 of the weight (value 2 of tourist C shown in FIG. 15).

Then, in the process of step S503 of FIG. 12, the information processing apparatus 100 has the respective values of tourist A, tourist B and tourist C (tourist A and tourists shown in FIG. 16) as shown in FIG. The recommendation value of each tourist spot is calculated by multiplying the value 2) of B and tourist C by the weight from the tourist side to the tourist spot side.

Specifically, in the process of step S503 of FIG. 12, the information processing apparatus 100 uses the function of the recommendation processing unit 126 to multiply the value 2 of the tourist A by 0.50 by the weight of the tourist spot 1 by 0.67. 0.34 is calculated and used as the recommended value of tourist spot 1 (recommended value of tourist spot 1 shown in FIG. 16). Further, the information processing device 100 calculates 0.17 by multiplying 0.50 of the value 2 of the tourist A by 0.33 of the weight to the tourist spot 4 by the function of the recommendation processing unit 126, and calculates 0.17 of the tourist spot 4. Let it be a recommendation value (recommendation value of tourist spot 4 shown in FIG. 16). Further, the information processing device 100 calculates 0.10 by multiplying 0.10 of the value 2 of tourist B by 1.00 the weight to the tourist spot 2. Further, since there is also a side from tourist C to tourist spot 2, 0.32 is calculated by multiplying 0.40 of the value 2 of tourist C by 0.80 of the weight to tourist spot 2. Then, these values are totaled (0.10 + 0.32) to obtain a recommendation value of tourist spot 2 of 0.42 (recommendation value of tourist spot 2 shown in FIG. 16). Further, the information processing device 100 calculates 0.08 by multiplying 0.40 of the value 2 of tourist C by 0.20 of the weight to the tourist spot 3 by the function of the recommendation processing unit 126, and calculates 0.08 of the tourist spot 3. Let it be a recommendation value (recommendation value of tourist spot 3 shown in FIG. 16).

After that, in the process of step S404 of FIG. 11, the information processing apparatus 100 displays recommended tourist spots on the input / output unit 130 based on the recommendation value calculated in step S403 by the function of the recommendation processing unit 126. Make recommendations to tourist D. As shown in FIG. 16, as for the recommendation value in this example, the recommendation value of tourist spot 2 (0.42) is the largest, followed by the recommendation value of tourist spot 1 (0.33), and then the recommendation value of tourist spot 4. The value is (0.17), and the recommendation value (0.08) of the tourist spot 3 is the smallest. In this case, since the tourist spot 1 is the most recently visited by the tourist D, the information processing device 100 displays the tourist spot 2 having the largest recommended value as the recommended tourist spot. In addition, all of the tourist spots 1 to 4 may be displayed in order from the one with the highest recommendation value. In that case, it may be displayed in the order of recommended values together with additional elements such as a numerical value and the number of stars so that the recommendation level can be visually recognized.

In this way, only one new transaction in which tourist D visits tourist spot 1 occurs, the influence of the tourist who visits tourist spot 2 is reduced, and the recommendation value changes. As mentioned above, various transactions usually occur, and the directed bipartite graph is updated in real time accordingly, so it is more affected by the transaction data that has occurred, and the recommendation value is one by one. It will be changed. Therefore, it is possible to make a more accurate recommendation to the person to be recommended.

Next, as an example showing the change in the recommendation value due to the change in the behavior of the tourist, from the state shown in FIG. 16, the case where the tourist D visits the tourist spot 1 once more and then visits the tourist spot 4 is taken as an example. explain. In other words, from the state shown in FIG. 6B, after the transaction that tourist D visited tourist spot 1 occurred, the transaction that tourist D visited tourist spot 1 once more, and sightseeing. The case where the transaction of visiting the spot 4 occurs will be described. In this state, tourist D visits tourist spot 1 twice and tourist spots 2 and 4 once each, and the behavior is similar to that of tourist A (tourist A is tourist spot). I visited 1 twice and tourist spot 4 once).

In this case, the process of step S402 in FIG. 11 organizes a directed bipartite graph as shown in FIG. Specifically, the side (arrow) connecting the apex of tourist D and the tourist spot 4 is added to the directed bipartite graph shown in FIG. 16 (see FIG. 17). In addition, since there are three sides of tourist D, and the number of visits is twice for tourist spot 1 and once for each of tourist spot 2 and tourist spot 4, the number of visits is from tourist D to tourist spot 1 direction. The weight is 0.50, and the weights in the tourist spot 2 and tourist spot 4 directions are 0.25, respectively. Therefore, in the process of step S402, the thickness of the side from the tourist D to the tourist spot 2 in FIG. 16 is organized into a thickness corresponding to the weight of 0.25, and the thickness of the side to the tourist spot 4 is formed. Is added with a thickness corresponding to a weight of 0.25 (see FIG. 17).

Subsequently, in the process of step S403 of FIG. 11, that is, the recommendation value calculation process of FIG. 12, the information processing apparatus 100 gives a value of 1.00 to the tourist D who is the recommendation target, as shown in FIG. The value 1 and the value 2 at each vertex are calculated, and the recommended value of the tourist spot 1 to the tourist spot 4 is calculated (details of the calculation are omitted). In this way, when a transaction that tourist D visits tourist spot 1 again and a transaction that visits tourist spot 4 occur, the recommended values of tourist spots 1 to 4 are shown in FIG. The value changes from the value shown in FIG. As a result, in the state shown in FIG. 16, the recommended value of tourist spot 2 was the highest, followed by the recommended value of tourist spot 1, but in the state shown in FIG. 17, the recommended value of tourist spot 1 was the highest. Next, the recommendation value of the tourist spot 4 becomes higher, and the recommendation ranking for the tourist D becomes closer to the action of the tourist A.

In this way, when tourist D takes some action, the target of the recommended tourist spot is added or the ranking is changed, and the tourist D who is the recommended tourist is more accurate sightseeing. Spots are being recommended. It should be noted that the recommendation value of each tourist spot with respect to the tourist D changes not only by the behavior of the tourist D who is the target of the recommendation but also by the behavior of other tourists. For example, in the state shown in FIG. 17, when tourist C visits tourist spot 1, the recommendation value of each tourist spot to tourist D changes as shown in FIG.

In the state shown in FIG. 17, when a transaction that tourist C visits tourist spot 1 occurs, a directed bipartite graph is organized as shown in FIG. 18 by the process of step S402 in FIG. Specifically, the side (arrow) connecting the apex of tourist C and the tourist spot 1 is added to the directed bipartite graph shown in FIG. 17 (see FIG. 18). In this case, there are three sides of tourist C, and the number of visits is one for tourist spot 1, four for tourist spot 2, and one for tourist spot 3, so that tourist spot C is visited once. The weight in one direction is 0.17, the weight in two directions of tourist spots is 0.67, and the weight in three directions of tourist spots is 0.17. Therefore, in the process of step S402, the thickness of the side from the tourist C to the tourist spot 1 in FIG. 18 is added by the thickness corresponding to the weight of 0.17, and the thickness of the side to the tourist spot 2 is added. The thickness corresponding to the weight of 0.67 is knitted, and the thickness of the side to the tourist spot 3 is knitted to the thickness corresponding to the weight of 0.17 (see FIG. 18).

Subsequently, in the process of step S403 of FIG. 11, that is, the recommendation value calculation process of FIG. 12, the information processing apparatus 100 gives a value of 1.00 to the tourist D who is the recommendation target, as shown in FIG. The value 1 and the value 2 at each vertex are calculated, and the recommended value of the tourist spot 1 to the tourist spot 4 is calculated (details of the calculation are omitted). Here, since tourist C newly visited tourist spot 1, the weight from tourist spot 1 to tourist A is 0.67 (tourist A visits tourist spot 1 twice), and tourist spot 1 The weight in the direction of tourist C is 0.33 (tourist C visits tourist spot 1 twice). Therefore, tourist D is also influenced by the behavior of tourist C. Therefore, in the state shown in FIG. 17, the recommended value of the tourist spot 4 was larger than the recommended value of the tourist spot 2, but in the state shown in FIG. 18, the recommended value of the tourist spot 2 is the recommended value of the tourist spot 4. As the ranking of the recommendation value, the tourist spot 2 will be replaced with the tourist spot 4.

As described above, the recommendation value of each tourist spot with respect to the tourist D changes not only by the behavior of the tourist D who is the recommendation target but also by the behavior of other tourists. Therefore, by incorporating transaction data indicating the behavior of the person to be recommended and other persons, suitable data analysis becomes possible, and more qualified recommendations can be made. In addition, by adding (adding) newly generated transaction data, the directed bipartite graph used for recommendation can be updated, and the processing load can be reduced. Further, for example, by deleting (subtracting) "today's transaction data one year ago" from the directed bipartite graph, the past data can be easily excluded. Furthermore, since the recommendation result is derived from the relationship between the transaction data indicating each action history and the weighting, it is possible to logically explain the recommendation result.

(Modification example)
The present invention is not limited to the above embodiment, and various modifications and applications are possible. For example, the information processing apparatus 100 does not have to have all the technical features shown in the above-described embodiment, and has been described in the above-described embodiment so as to solve at least one problem in the prior art. It may have a structure of parts. In addition, at least a part of each of the following modifications may be combined.

In the above embodiment, an example is shown in which an event occurring between a tourist and a tourist destination (sightseeing spot) is stored as transaction data in a data block connected by a hash chain, and the transaction data is used for recommendation. However, this is just an example. As long as it can be expressed as a transaction, it is not limited to an event that occurs between a tourist and a tourist destination (sightseeing spot). For example, events that occur between a customer and a store (number of visits, purchase amount, etc.) are stored as transaction data in a data block linked by a hash chain, and the transaction data is used to recommend products and stores to the customer. May be recommended.

Further, in the above embodiment, in the process of step S301 of FIG. 8, it is determined whether or not the target period (start time and end time) is specified by the input operation of the user, and within the period specified by the user. An example of recommending a tourist spot based on transaction data is shown, which is an example. In step S301, conditions such as "bad weather days", "only days when the average temperature exceeds a certain level", and "daytime hours" are specified, not whether or not the target period is specified. It may be determined whether or not it is. According to this, it is possible to preferably make recommendations not only for a designated period but also under certain conditions. Also, for example, after specifying conditions such as "daytime hours on days when the weather is bad and the average temperature falls below a certain level" and obtaining recommendation results (recommended tourist spots output according to those conditions), all The tendency under the conditions may be analyzed by comparing with the recommendation result based on the transaction data of.

Further, in the above embodiment, the recommendation process when the transaction that the tourist D visits the tourist spot 2 is added has been described. However, for example, the state shown in FIG. 10 (no new transaction has occurred). In the case), the recommendation process shown in FIG. 11 may be executed. In this case, instead of executing the processes of steps S401 and S402, in the process of step S403, a value of 1.00 is given to the tourists of the recommendation target, and the recommendation value calculation process shown in FIG. 12 is executed. Good.

Further, in the above embodiment, an example in which a value of 1.00 is given to a tourist who is a recommendation target person in step S501 of the recommendation value calculation process shown in FIG. 12 is shown, but in addition to this, for example, As shown in FIG. 19, by giving a value for an attribute such as "gender", the tendency can be further analyzed for each attribute. For example, when examining the tendency of "where is the tourist spot that men like", 1.00 is given to the attribute of "male" as shown in FIG. Regarding the attribute (attribute information), it is assumed that the attribute of "male" is registered in the data block whose index value is "0" in the hash chain of the tourist A and tourist D types. Further, in the data block where the index value in the hash chain of the tourist spot 2 and tourist spot 3 types is "0", the tourist spot with the attribute of "gourmet" is the index value in the hash chain of the tourist spot 1 and tourist spot 4 types. It is assumed that tourist spots with the attribute of "cultural facility" are registered as attribute information in the data block of "0". Further, in the graph generation process shown in FIG. 8, it is assumed that a directed bipartite graph including the attribute is generated as shown in FIG. The attribute information is not limited to the data block whose index value is "0", and may be registered in the data block of all index values.

Then, in the same manner as in the above embodiment (similar to the recommendation value calculation process shown in FIG. 12), each value is multiplied by a weight to calculate the value a and the value b shown in FIG. Specifically, since there are two tourists, tourist A and tourist D, who are tied to the male attribute, 0.50 each is distributed. The tourist spots connected to the tourist A are the tourist spot 1 and the tourist spot 4, and the weight on the tourist spot 1 side is 0.67. Therefore, 0.50 distributed to tourist A is multiplied by a weight of 0.67 to obtain 0.335. On the other hand, since the weight on the tourist spot 4 side is 0.33, the weight 0.5 is multiplied by the weight 0.33 to obtain 0.165. Next, the tourist spots connected to the tourist D are the tourist spot 1, the tourist spot 2, and the tourist spot 4, and the weight on the tourist spot 1 side is 0.5. Therefore, 0.50 distributed to tourist D is multiplied by a weight of 0.5 to obtain 0.25. Further, since the weights of the tourist spot 2 and the tourist spot 4 are each 0.25, the weight 0.5 is multiplied by the weight 0.25 to obtain 0.125, respectively. Since tourist A has not visited tourist spot 2, the value a of tourist spot 2 is set to 0.13.

Then, the calculated value 0.335 from the tourist A to the tourist spot 1 and the calculated value 0.25 from the tourist D to the tourist spot 1 are added to obtain 0.59 as the value a of the tourist spot 1. Further, the calculated value 0.165 from the tourist A to the tourist spot 4 and the calculated value 0.125 from the tourist D to the tourist spot 4 are added to obtain 0.29 as the value a of the tourist spot 4. Subsequently, since the tourist spot 1 and the tourist spot 4 are each an attribute of the "cultural facility", the respective values a are added to obtain 0.88 as the value b of the attribute of the "cultural facility". Further, since the tourist spot 2 is an attribute of "gourmet", the value a is used as it is as the value b (since the value a is not calculated for the tourist spot 3 this time), 0.13 is obtained. The value b is the same as the recommended value in the above embodiment. Therefore, since the attribute value b of "cultural facility" is higher than the attribute value b of "gourmet", it can be seen that "male" tends to prefer "cultural facility" to "gourmet". .. According to this, for example, when tourist D is a "male", all tourist spots belonging to the attribute of "cultural facility" are recommended, and the tourist spots are not limited to the tourist spots based on the recommendation value in the above embodiment. It is also possible to perform attribute-based recommendations (recommendations based on attribute information) together (that is, both the tourist spots based on the recommendation value in the above embodiment and the tourist spots recommended based on the attribute information are used by the user. Display identifiable for).

In this way, it is possible to analyze the tendency at the attribute level by having attribute information for each type of hash chain and generating a directed bipartite graph including the information. In this example, only the "gender" attribute is used for ease of understanding, but a plurality of attributes such as "gender" and "nationality (Japanese and foreigners)" may be used. According to this, it is possible to analyze trends for each attribute, such as "where are the tourist spots preferred by Japanese men" and "where are the tourist spots preferred by foreign women". Further, in this example, the value is calculated from the attribute of "gender" in the direction of the attribute of the tourist spot (also referred to as upstream to downstream), but by performing the same calculation in the opposite direction, for example. , "Tourists who like cultural facilities" can be extracted. Then, it is also possible to check the tourist spots visited by the extracted tourists (check the behavior history) and send a guidance email when they become prospective tourists (for example, tourists who have not visited yet). It will be possible.

In addition, if there is a sufficient hash chain with attribute information, it is possible to estimate the attributes of the target tourist using the attributes of the tourist who behaves much like the target tourist. it can. For example, as shown in FIG. 20, when tourist A is a male and tourist B and tourist C are female (each attribute information is in the data block of the index value "0" of each type of hash chain. (When is registered), the attribute of tourist D can be estimated by using the value 2. The value 2 shown in FIG. 20 is the same as the value 2 shown in FIG.

Specifically, the value 2 calculated when the value 1.00 is given to the tourist D is the largest when the value 2 of the tourist A is 0.75 (that is, the behavior most similar to the tourist D). The tourist who is doing this is Tourist A). Therefore, the value X of the male attribute of the tourist A is 0.75. On the other hand, the value 2 of tourist B is 0.05 and the value 2 of tourist C is 0.20. Since these belong to the female attribute, the value X of the female attribute is 0. It becomes 25. From this, it can be estimated that the probability that tourist D is male (the probability that tourist D is female is 25%).

In this way, if there is a sufficient hash chain with attribute information, the attributes of the target tourist can be estimated using the attributes of the tourist who behaves much like the target tourist. be able to. However, if there are not enough hash chains with attribute information, the accuracy of this estimation will drop. When the accuracy of the estimation is lowered in this way, the accuracy of the recommendation based on the attribute information is also lowered. Therefore, in such a case, it is difficult to effectively utilize the attribute information even if it is managed by a program. Therefore, when the effectiveness of the attribute information is evaluated and the evaluation score is close to "0" and falls below a preset threshold value, it is considered that the recommendation based on the attribute information is not effective, and the recommendation in the above embodiment is performed. It is possible to display only the recommendations of tourist spots based on the value, and not to display the tourist spots recommended based on the attribute information. The calculation of the evaluation score will be described below.

In the following example, the gender ratio of tourists registered in the hash chain is male: female = 1: 1 (2 males, 2 females). For example, if the attribute of the tourist you want to check is "male" and the genders of other tourists close to that tourist are "male probability = 0.75" and "female probability = 0.25". Since the expected value of being a man is "1 / (1 + 1) = 0.50", 0.75-0.50 = 0.25 is set as the evaluation score. On the other hand, if the attribute of the tourist to be examined is "female", the evaluation score is 0.25-0.50 = -0.25. In the state shown in FIG. 20, when the tourist D has a male attribute, the evaluation score of the tourist D is 0.25. Such an evaluation score is calculated not only for tourist D but also for tourists A to C.

In the state shown in FIG. 20, since the tourist D is the closest to the tourist A, when the value 1.00 is given to the tourist A, the value 2 of the tourist D becomes 1.00. Since both tourist A and tourist D are male attributes, the evaluation score of tourist A is 1.00-0.50 = 0.50. In addition, since tourist C is the closest to tourist B (the weight from tourist spot 2 to tourist C is 0.8, and the weight to tourist D is 0.2), the value for tourist B is high. When 1.00 is given, the value 2 of tourist C becomes 0.8. Since both tourist B and tourist C are female attributes, the evaluation score of tourist B is 0.80-0.50 = 0.30. Also, since tourist B and tourist D are the closest to tourist C (tourist C connects tourist spot 2 and tourist spot 3, but tourist spot 3 is connected to tourist C). Exclude because it is only. And exclude tourist C from the destination where tourist spot 2 is connected. After performing this process, the weight from tourist spot 2 to tourist B is 0.5, and to tourist D. The weight of is 0.5), and when the value 1.00 is given to the tourist C, the value 2 of the tourist B and the value 2 of the tourist D are 0.5, respectively. Since both tourist C and tourist B are female attributes, the evaluation score of tourist C based on the tourist B side is 0.50-0.50 = 0.00. Further, since tourist C has a female attribute and tourist D has a male attribute, the evaluation score of tourist C based on the tourist D side is 0.50-0.50 = 0. It becomes 00.

The total evaluation score of each of tourist A to tourist D is 0.50 + 0.30 + 0.00 + 0.25 = 1.05, and the average value is 0.26, which is used as the attribute evaluation score. (Note that the maximum evaluation score is 0.50). Here, when the predetermined threshold value is 0.25, 0.26> the threshold value, so it may be determined that the recommendation based on the “gender” attribute is effective.

On the other hand, in the state shown in FIG. 21, the ratio of Japanese to foreigners is 3: 1 (3 Japanese and 1 foreigner). Therefore, the expected value of being Japanese is 0.75, and the expected value of being a foreigner is 0.25. When a value of 1.00 is given to tourist D, the value 2 of tourist A is 0.75 and the value 2 of tourist C is 0.20. Since both are Japanese, the evaluation score of tourist D Is 0.75 + 0.20-0.75 = 0.20. Similarly, when the value 1.00 is given to the tourist A, the evaluation score of the tourist A is 1.00-0.75 = 0.25. Further, since tourist C is the closest to tourist B, when the value 1.00 is given to tourist B, the value 2 of tourist C becomes 0.80 and the value 2 of tourist D is 0. It becomes .20. Since tourist B has the attribute of "foreigner" and tourist C and tourist D have the attribute of "Japanese", the evaluation score of tourist B is 0.20 + 0.80-0.25 = It becomes 0.75. Further, since tourist B and tourist D are closest to tourist C, when the value 1.00 is given to tourist C, the value 2 of tourist B and the value 2 of tourist D are 0, respectively. It becomes .5. Since tourist B has the attribute of "foreigner" and tourist C and tourist D have the attribute of "Japanese", the evaluation score of tourist C is 0.50-0.75 = -0. It becomes .25.

The total of the evaluation scores of Tourist A to Tourist D is 0.20 + 0.25-0.75-0.25 = -0.55, and the average value (attribute evaluation score) is -0. It becomes 14. Here, when the predetermined threshold value is 0.20, -0.14 <threshold value, so it may be determined that the recommendation based on the attribute of "nationality" is invalid. In such a case, the tourist spots recommended based on the attribute information may not be displayed. The threshold value may be arbitrarily set for each attribute. For example, it suffices if an arbitrary value can be set according to the age, gender, nationality, and the type of transaction of the user (specifically, transactions that occur between tourists and tourist spots, and customers and stores. Set a different threshold than the transaction that occurred in (for example).

Further, in addition to the recommendation processing in the above embodiment, for example, by paying attention to the order of transaction data, further analysis becomes possible. Specifically, tourists who visit tourist spot 2 are more likely to visit tourist spot 1 next, and few tourists visit tourist spot 1 to tourist spot 2 in that order. After visiting Spot 1, it is possible to analyze that there are many tourist spots d.

For example, as shown in FIG. 22, paying attention to the tourist spot 2, transaction data (tourist B to tourist D) of tourists connected to the tourist spot 2 are extracted. In this way, "sightseeing spot 2" is included in the series of transaction data linked to the extracted tourist hash chain, and what is the transaction data before and after that, and how long the time interval is, is analyzed. Is possible. In the example shown in FIG. 21, in order to simplify the explanation, the transaction data registered in the hash chain of the tourist C type is different from the order shown in FIG. 6 (different from FIG. 7). It is assumed that Tourist C visited the tourist spot 2, the tourist spot 3, the tourist spot 2, and the tourist spot 1 in this order.

Then, with respect to the extracted transaction data, the transaction data before and after the "sightseeing spot 2" of interest is arranged as shown in FIG. 23. This makes it possible to specify what the transactions before and after the transaction data of interest are, that is, the type of transaction data with high frequency. In the illustrated example, since the number of tourist spots 1 arranged on the box of tourist spot 2 is large, the visitor of "sightseeing spot 2" who paid attention may visit "sightseeing spot 1" next. It turns out to be expensive.

Note that, for example, when a certain period of time (which may be an arbitrarily set period) elapses after visiting the tourist spot 2, the transactions before and after may not be arranged. For example, in the example shown in FIG. 23, when the transaction data indicating that the tourist spot 1 connected to the hash chain of the tourist C type has been visited has passed one year or more since the visit to the tourist spot 2. The transaction data (the box of the tourist spot 1 in the illustrated example) indicating that the tourist spot 1 has been visited may not be arranged. In this case, it is sufficient to determine whether or not a certain period of time has passed, and to arrange the array only when it has not passed.

In addition, by applying this, the average staying time of tourists who visited "Sightseeing Spot 2" is calculated, and for tourists who visited Tourist Spot 2, when the average staying time has elapsed, "Visit to Tourist Spot 1" "Why don't you do it?" It will be possible to make recommendations such as push notifications. At that time, referring to the time, for example, if it is a "time zone when there is no visit data" such as outside the business hours of the tourist spot 1, the tourist spot 4 may be recommended instead of the tourist spot 1. Good. For the average staying time, for example, for each tourist's hash chain that visited "Sightseeing Spot 2", a data block in which "Sightseeing Spot 2" was registered and a tourist spot visited next were registered. It may be calculated by the average value of the difference between each time stamp and the data block.

Further, by applying this, for example, the order of transaction data such as "view information of tourist spot x", "view information of travel product y of tourist spot x", and "application (purchase) of travel product y" can be obtained. If the application is not made within 10 minutes from the occurrence of the transaction such as "viewing the information of the travel product y of the tourist spot x" by the analysis and the analysis based on the time interval of the transaction data, the application is made. It is possible to analyze that the probability is very low and use it such as changing the priority of recommendations (when the transaction of "view information of travel product y of tourist spot x" occurs, "travel product" Prioritize specific recommendations such as "application (purchase) for y").

In addition, when the above-mentioned function is realized by sharing the OS (Operating System) and the application, or by cooperating with the OS and the application, only the part other than the OS may be stored in the medium.

It is also possible to superimpose a program on a carrier wave and distribute it via a communication network. For example, the program may be posted on a bulletin board system (BBS, Bulletin Board System) on a communication network, and the program may be distributed via the network. Then, by starting these programs and executing them in the same manner as other application programs under the control of the operating system, the above-mentioned processing may be executed.

1 Information processing system, 100, 100A to C information processing device, 100D tourist A to D terminals, 110 storage unit, 111 hash chains of each type, 120 control unit, 121 initial registration unit, 122 synchronization processing unit, 123 data Block registration unit, 124 data acquisition unit, 125 graph generation unit, 126 recommendation processing unit, 130 input / output unit, 140 communication unit, 210 network

Claims (6)

A data storage means for storing the data of the transaction in a data block connected by a hash chain constructed in the network based on the occurrence of the transaction.
An extraction means for extracting the data of a predetermined transaction from the data of the transaction stored in the data storage means, and an extraction means.
An arithmetic means that performs a predetermined operation based on the transaction data extracted by the extraction means, and
It is provided with a recommendation means for recommending a target corresponding to the preference of the recommendation target person according to the calculation result by the calculation means.
The transaction data includes information that identifies the transaction target of the transaction.
The calculation means sets a specific transaction target among the transaction targets specified by the transaction data extracted by the extraction means as the recommendation target person, and all the transaction targets related to the recommendation target person. , Performs the predetermined calculation based on the transaction data.
An information processing device characterized by this. When the data of the transaction is newly stored in the data storage means, the extraction means newly extracts the data of the transaction.
The calculation means sets the transaction target specified by the transaction data newly extracted by the extraction means as the recommendation target person, and all the transaction targets related to the recommendation target person are converted into the transaction data. By performing the predetermined calculation based on the above, the calculation result already performed by the calculation means is updated.
The information processing apparatus according to claim 1. The calculation means allocates an initial value to the recommendation target person as the predetermined calculation, and calculates all the transaction targets related to the recommendation target person from the transaction data based on the initial value. Calculate the value multiplied by the weight to be
The information processing apparatus according to claim 1 or 2. The transaction data further includes attribute information indicating the transaction target attribute of the transaction.
The calculation means further allocates the initial value to the attribute indicated by the attribute information of the recommendation target person, and sets the initial value for all the transaction targets related to the transaction target having the same attributes as the recommendation target person. A value obtained by multiplying the reference by the weight calculated from the transaction data is further calculated.
The recommendation means recommends a target corresponding to the preference of the attribute of the recommendation target person in addition to the target corresponding to the preference of the recommendation target person according to the calculation result by the calculation means.
The information processing apparatus according to claim 3. It is an information processing method in an information processing device.
A data storage step for storing the data of the transaction in a data block connected by a hash chain constructed in the network based on the occurrence of the transaction.
An extraction step for extracting the data of a predetermined transaction from the data of the transaction stored in the data storage step, and
A calculation step that performs a predetermined calculation based on the transaction data extracted in the extraction step, and
It is provided with a recommendation step of recommending a target corresponding to the preference of the recommendation target person according to the calculation result of the calculation step.
The transaction data includes information that identifies the transaction target of the transaction.
In the calculation step, a specific transaction target among the transaction targets specified by the transaction data extracted in the extraction step is set as the recommendation target person, and all the transaction targets related to the recommendation target person are used. , Performs the predetermined calculation based on the transaction data.
An information processing method characterized by this. Computer,
A data storage means for storing the data of the transaction in a data block connected by a hash chain constructed in the network based on the occurrence of the transaction.
An extraction means for extracting a predetermined transaction data from the transaction data stored in the data storage means.
An arithmetic means that performs a predetermined operation based on the transaction data extracted by the extraction means,
It functions as a recommendation means for recommending a target corresponding to the preference of the person to be recommended according to the calculation result by the calculation means.
The transaction data includes information that identifies the transaction target of the transaction.
The calculation means sets a specific transaction target among the transaction targets specified by the transaction data extracted by the extraction means as the recommendation target person, and all the transaction targets related to the recommendation target person. , Performs the predetermined calculation based on the transaction data.
A program characterized by that.

Images