Secret page rank computing system and method thereof, secret computing device

Abstract

Transaction data held by a plurality of data sources is input, and high-precision page ranking is calculated in a state in which the transaction data of each data source is concealed. A data source device (1) calculates a transaction ratio for each combination of transaction subjects (S12). The data source device (1) encrypts the transaction ratio and transmits it to each secret calculation device (2) (S13). The secret calculation device (2) receives the ciphertext of the transaction ratio from a plurality of data source devices (1) (step S21). The secret calculation device (2) performs secret calculation on the ciphertext of the page ranking of the transaction subject which becomes a calculation target after decryption, using the ciphertext of the transaction ratio related to the transaction subject of the calculation target and the ciphertext of the page ranking of the transaction object (S22).

Description

Technical Field

[0001] This invention relates to secret computing techniques, and particularly to secret computing page sorting techniques for the purpose of reliability estimation. Background Technology

[0002] PageRank is a framework that uses the link relationships among web pages to determine the importance of a web page (see Non-Patent Document 1). Applying this framework, page ranking can also be calculated based on transaction data such as inter-company business flows, and used as an index of each company's reliability (see Non-Patent Document 2).

[0003] Existing technical documents

[0004] Non-patent literature

[0005] Non-patent literature 1: L. Page, S. Brin, R. Motwani and T. Winograd, "ThePageRank citation ranking: Bring order to the Web," Technical Report of theStanford Digital Library Technologies Project (1998).

[0006] Non-patent document 2: Onishi Ratsuo, Takayasu Hideki, Takayasu Misako, "Inter-enterprise cited information", Information Processing Society Research Report, Vol. 2010-MPS-81, No. 1, 2010 Summary of the Invention

[0007] The problem that the invention aims to solve

[0008] However, since many data circulation platforms now independently accumulate inter-enterprise transaction data, there is a possibility of information bias if page sorting is performed based on a single data source. Furthermore, this transaction data contains confidential information from each company, raising concerns about security issues such as data breaches if multiple data sources are simply merged.

[0009] In view of such technical challenges, the object of the present invention is to perform high-precision page sorting calculations by taking transaction data held by multiple data sources as input and concealing the transaction data of each data source.

[0010] Methods for solving problems

[0011] One aspect of the secret page ranking calculation system of the present invention is a secret page ranking calculation system comprising multiple secret calculation devices, each secret calculation device comprising: a page ranking storage unit for storing ciphertext of page rankings of multiple transaction subjects; a transaction ratio receiving unit for receiving ciphertext of transaction ratios from each of multiple data sources; and a page ranking calculation unit for performing secret calculations on the decrypted ciphertext of page rankings of transaction subjects that are the calculation objects, using the ciphertext of transaction ratios related to the transaction subjects of the calculation objects, and the ciphertext of page rankings of transaction subjects that have transacted with the transaction subjects of the calculation objects. The transaction ratio represents the proportion of transactions performed in each combination of transaction subjects within the overall transaction history maintained by the data source.

[0012] Invention Effects

[0013] According to the present invention, it is possible to perform high-precision page sorting calculations by taking transaction data held by multiple data sources as input and hiding the transaction data of each data source. Attached Figure Description

[0014] Figure 1 This is a diagram illustrating the functional structure of a secret page sorting calculation system.

[0015] Figure 2 This is a diagram illustrating the functional structure of a data source device.

[0016] Figure 3 This is a diagram illustrating the functional structure of a secret computing device.

[0017] Figure 4 This is a diagram illustrating the processing steps of the secret page sorting calculation method.

[0018] Figure 5 This is a diagram used to illustrate Example 1.

[0019] Figure 6 This is a diagram used to illustrate Example 2.

[0020] Figure 7 This is a diagram illustrating the functional structure of a computer. Detailed Implementation

[0021] The embodiments of the present invention will be described in detail below. Furthermore, in the accompanying drawings, structural parts with the same function are assigned the same reference numerals, and repeated descriptions are omitted.

[0022] [Implementation Method]

[0023] The present invention provides a secret page ranking calculation system and method that uses encrypted transaction data input from multiple data sources to generate page rankings while concealing the transaction data.

[0024] When using page ranking to measure inter-enterprise reliability, calculating the ranking using only a single data source may result in a company with low reliability being mistakenly classified as highly reliable. Therefore, using multiple data sources to calculate the ranking can potentially improve the accuracy of this reliability assessment. However, in generating page rankings using multiple data sources, it is essential to consider the technical aspects of preventing the information held by each data source from being leaked to other data sources. In particular, entities that collect and utilize data, such as data circulation platforms, find it difficult to readily provide this data to other entities because the accumulated data itself becomes valuable.

[0025] Therefore, in the secret page ranking calculation system of the implementation method, the above-mentioned problem is solved by secretly calculating the page ranking calculation formula that takes multiple data sources as input and by obtaining the page ranking in a state where the input data from each data source is hidden.

[0026] like Figure 1 As shown, the secret page sorting calculation system 100 of the embodiment includes N (≥2) data source devices 11,…, 1 N And K (≥1) secret computing devices 21, …, 2 K Data source devices 11,...,1 N And secret computing devices 21,..., 2 K Each device is connected to communication network 9. Communication network 9 is a line-switched or packet-switched communication network configured to enable communication between the connected devices, such as the Internet, Local Area Network (LAN), or Wide Area Network (WAN).

[0027] Data source devices 11, …, 1 N One of these devices acts as the calculation unit for the requested page ranking (hereinafter also referred to as the "request calculation device"). Additionally, data source devices 11, ..., 1 N Unlike any of the computing request devices, one of them plays the role of registering the initial values ​​for page sorting (hereinafter also referred to as the "initialization device").

[0028] When there are multiple secret computing devices 2 (i.e., K≥2), secret computing device 2 k (k∈{1,…, K}) For example, using secret distribution-based methods such as Shamir secret distribution or replicated secret distribution to coordinate other secret computing devices2 k(k'∈{1, …, K} and k≠k') and calculate the page order. In the case where the secret computing device 2 is a single unit (i.e., K=1), the secret computing device 2 calculates the page order, for example, using a secret computing method based on homomorphic encryption or other encryption methods.

[0029] like Figure 2 As shown, data source device 1 n (n=1, …, N) For example, it includes a transaction history storage unit 10, a transaction ratio calculation unit 12, and a transaction ratio sending unit 13. The data source device, which serves as the calculation request device, also includes... Figure 2 The page sorting decryption unit 14 and the convergence determination unit 15 are indicated by dashed lines. The data source device, serving as an initialization device, also includes... Figure 2 The page sorting initial value registration section 11 is represented by a dashed line. For example... Figure 3 As shown, secret computing device 2 k (k=1, …, K) For example, it includes a page sorting storage unit 20, a transaction ratio receiving unit 21, a page sorting calculation unit 22, a page sorting sending unit 23, and a page sorting updating unit 24. Through these data source devices 11, …, 1 N And secret computing devices 21, …, 2 K Mutual coordination and conduct Figure 4 The processing of each step shown implements the secret page sorting calculation method of the embodiment.

[0030] Data source device 1 n And secret computing device 2 k For example, it is a special device that reads special programs into a known or special-purpose computer equipped with a central processing unit (CPU) and main storage (random access memory (RAM)). Data source device 1n and secret computing device 2 k For example, each process is executed under the control of the central processing unit. This is achieved by inputting data source device 1. n And secret computing device 2 k Data, or data obtained through various processes, is stored, for example, in the main storage device. Data stored in the main storage device is read out as needed and used for other processing. Data source device 1 n And secret computing device 2 k At least a portion of each processing unit can also be constructed from hardware such as integrated circuits. Data source device 1 n And secret computing device 2 kThe various storage units can be, for example, main storage devices such as random access memory (RAM), auxiliary storage devices made of semiconductor storage elements such as hard disks, optical disks or flash memory, or middleware such as relational databases or key-value stores.

[0031] Reference Figure 4 The processing steps of the secret page ranking calculation method executed by the secret page ranking calculation system 100 in the embodiment will be described. Furthermore, in Figure 4 In the example, the secret page sorting calculation system 100 includes two data source devices 11 and 12, where data source device 11 is the calculation request device and data source device 12 is the initialization device.

[0032] In each data source device 1 n The transaction history storage unit 10 stores the data source device 1. n The transaction history that is processed and represents the content of transactions between two trading entities.

[0033] In step S11, the page sorting initial value registration unit 11 of the data source device 12 randomly generates the page sorting PR of each transaction subject i (i=1, …, I, where I is the total number of transaction subjects) in the transaction history stored in the transaction history storage unit 10, which is related to each transaction. i The initial value. Then, the page sorting initial value registration department 11 sorts the page PR of each transaction entity i. i The initial value is encrypted and sent to each secret computing device 2. k (k=1, …, K) Send the ciphertext [PR i Additionally, [·] represents the ciphertext of the value ·, and in the case of encryption via secret distribution, it represents the distributed value of the value ·. In the case of encryption via secret distribution, the page ranking PR of each transaction entity i is determined. i The dispersion of the initial values ​​[PR] i [Assigned to each secret computing device 21, …, 2] K To enable a secret computing device 2 k Maintain a share [PR] i ].

[0034] In step S20, each secret computing device 2 k The page sorting PR of each transaction entity i received from the data source device 12 is stored in the page sorting storage unit 20. i The ciphertext of the initial value [PR] i].

[0035] In step S12, each data source device 1 n The transaction ratio receiving unit 12, based on the transaction history stored in the transaction history storage unit 10, performs a combination (i, j) (j∈V) of each transaction subject. i V i (This refers to the set of trading partners of trading entity i) The transaction ratio k is calculated. n j→i Additionally, k n j→i Indicates the nth data source device 1 n The transaction ratio maintained is the proportion of transactions involving transaction entity i and transaction object j. The transaction ratio represents the proportion of transactions conducted through a combination of transaction entities within the overall transaction history. For example, if the data source is a bank and the transaction is a remittance, the value obtained by dividing the total amount of remittances processed by the bank from one company to other companies by the total amount of remittances recorded in the transaction history is the transaction ratio related to that company and other companies. This transaction ratio is calculated using all combinations of transaction entities recorded in the transaction history. The transaction ratio calculation unit 12 outputs the calculated transaction ratio k to the transaction ratio sending unit 13. n j→i .

[0036] In step S13, each data source device 1 n The transaction ratio sending unit 13 sends the input transaction ratio k. n j→i and the predefined weight α n product α n k n j→i Encryption is performed, and the encryption is sent to each secret computing device 21, …, 2 K Send the ciphertext [α] n k n j→i In the case of encryption through secret distribution, the distribution value of the transaction proportion [α] is... n k n j→i [Assigned to each secret computing device 21, …, 2] K To enable a secret computing device 2 k Maintain a share [α] n k n j→i ] k Assume that data is pre-installed in each data source device 11, …, 1 N The settings are configured to allow each data source device 11, …, 1 to... N The weights α1, …, α that are maintainedN Satisfying α1 + … + α N =1.

[0037] In step S21, each secret computing device 2 k The transaction ratio receiving unit 21 receives data from each data source device 11, …, 1 N The encrypted message [α1k] representing the proportion of received transactions 1 j→i ], …, [α N k N j→i ].

[0038] In step S22, each secret computing device 2 k The page sorting calculation unit 22 uses the ciphertext [α1k] related to the transaction proportion of the transaction subject i (∈{1, …, I}) of the calculation object. 1 j→i ], …, [α N k N j→i ], and the transaction object j (j∈V) that transacted with the transaction entity i. i PageRank (PR) j ciphertext [PR] j [To sort the PageRank of the pages that become the transaction subject i after decryption] i ciphertext [PR] i The page sorting calculation is performed secretly. Furthermore, the transaction subject being calculated does not need to be a single entity; page sorting can be calculated for all transaction subjects or for any selected multiple transaction subjects. The page sorting calculation unit 22 outputs the page sorting PR of transaction subject i to the page sorting sending unit 23. i ciphertext [PR] i ].

[0039] Specifically, the page ranking calculation unit 22 calculates the page ranking PR of transaction subject i by calculating the following formula. i ciphertext [PR] i ].

[0040] [Formula 1]

[0041]

[0042] In step S23, each secret computing device 2 k The page sorting sending unit 23 sends the page sorting PR of the input transaction subject i to the data source device 11, which is a calculation request device. i ciphertext [PR] i ].

[0043] In step S14, the page sorting and decryption unit 14 of the data source device 11 sorts the data from each secret computing device 21, ..., 2. K The page ranking PR of the received transaction subject i i ciphertext [PR] i Decrypt the data to obtain the page ranking PR of transaction entity i. i .

[0044] In step S15, the convergence determination unit 15 of the data source device 11 determines the page ranking PR of the transaction subject i. i Has convergence been achieved? Convergence is determined by comparing the difference between the page ranking before and after calculation with a specified threshold. If the page rankings of all transaction entities in the calculation are determined to have converged, the convergence determination unit 15 completes the process. If the page rankings of one of the transaction entities in the calculation are determined to have not converged, the calculated page ranking PR is adjusted. i Encryption is performed, and the encryption is sent to each secret computing device 21, …, 2 K Send the encrypted message [PR] i ].

[0045] In step S24, each secret computing device 2 K If the page sorting update unit 25 receives the calculated page sorting PR from the data source device 11, which is the calculation request device... i ciphertext [PR] i If the page sorting is specified, the page sorting PR stored in the page sorting storage unit 10 will be updated using the encrypted page sorting information. i ciphertext [PR] i Then, the process returns to step S22. Thereafter, the page sorting calculation and convergence determination are performed again.

[0046] [Variation Example 1]

[0047] In the secret page sorting calculation system 100 of the implementation method, multiple data source devices 1 are configured. n Any one of them can act as the computation requesting device. However, the computation requesting device can also be configured to work with any data source device 1. n Each is a separate device. In this case, the secret page sorting calculation system 100 also includes... Figure 1 The calculation request device 3 is indicated by a dashed line. The calculation request device 3 includes a page sorting decryption unit 14 and a convergence determination unit 15, which are present in the data source device 11 that serves as the calculation request device in the secret page sorting calculation system 100 of this embodiment. In Modification 1, the secret page sorting calculation system only includes this unit in each of the secret calculation devices 21, ..., 2. KThe page sorting sending unit 23 sends the encrypted page sorting result to the calculation request device 3 instead of the data source device 11 of the calculation request device, which is different from the secret page sorting calculation system 100 in the embodiment.

[0048] [Variation Example 2]

[0049] In the secret page sorting calculation system 100 of the implementation method, the weight is set as α. n In data source device 1 n The transaction ratio is pre-set and sent to the 13th transaction ratio k. n j→i With weight α n The structure is calculated using the product of [variables]. However, the weights α1, …, α N Alternatively, it can be omitted from data source device 1. n Instead of being set in the central location, it is located in secret computing devices 21, …, 2. K Pre-configured. In this case, data source device 1 n The transaction ratio transmission unit 13 sends the data to each secret computing device 21, ..., 2. K Send transaction ratio k n j→i The ciphertext [k] n j→i Each secret computing device 21, …, 2 K The page sorting calculation unit 22 performs secret calculations from the data source device 1 n Received transaction ratio k n j→i The ciphertext [k] n j→i Multiply by weight α n Furthermore, based on this, a secret calculation is performed on the page sorting formula.

[0050] [Variation Example 3]

[0051] In the secret page ranking calculation system 100 of the embodiment, the structure is configured such that, if the page ranking of a transaction subject determined to be one of the calculation objects has not converged, the convergence determination unit 15, which is the data source device 11 of the calculation request device, sends a request to each secret calculation device 21, ..., 2. K Send calculated page ranking PR i ciphertext [PR] i ], each secret computing device 2 k The page sorting update department 25 updates the page sorting PR received. i ciphertext [PR] i The page sorting PR stored in the page sorting storage unit 10 is used to sort the pages. i ciphertext [PR]i Updates are performed. However, the structure can also be configured such that, if the page ranking of the transaction subject determined to be one of the computational objects has not converged, the convergence determination unit 15 of the data source device 11 sends updates to each secret computing device 21, ..., 2. K Each secret computing device 2 only sends a signal instructing the calculation to continue. k When the page sorting update unit 25 receives a signal instructing the calculation to continue, it updates the page sorting PR calculated by the page sorting calculation unit 22. i ciphertext [PR] i The page sorting PR stored in the page sorting storage unit 10 is used to sort the pages. i ciphertext [PR] i Updated.

[0052] [Example 1]

[0053] Example 1 is an embodiment of applying the secret page ranking calculation system of the implementation method to the credit assessment of a bank. For example... Figure 5 As shown, the secret page ranking calculation system of Embodiment 1 consists of Bank A, Bank B, and a secret calculation system. The secret calculation system contains multiple secret calculation devices. Bank A and Bank B each maintain their own transaction history between their client companies. Figure 5 In the example, firms α, β, γ, δ, and ε are the client firms of each bank. Figure 5 The arrows connecting the various companies represent bank transfers (transactions). The amount recorded near the arrow is the transfer amount. For example, in Bank A's transaction history, there is a record of a transfer of 100 yen from company α to company γ. That is, Bank A and Bank B are equivalent to data source devices.

[0054] Bank A requests the calculation of its clients' page rankings from the secret computing system for credit assessment. In other words, Bank A acts as the calculation requesting device. Prior to this, Bank B registers the initial page ranking values ​​for each client with the secret computing system (step 0). That is, Bank B acts as the initialization device. Bank B randomly generates the initial page ranking values, encrypts these random values ​​through secret distribution, and registers the distribution value with the secret computing system. Bank B's registration of the initial page ranking values, even without making the calculation request, is to prevent transaction-related information held by Bank B from being leaked to Bank A, which is making the calculation request.

[0055] Bank A and Bank B, based on their respective transaction histories, total the remittance amounts for each enterprise's portfolio and calculate the proportion of that amount within the overall transaction history (hereinafter referred to as the "remittance amount proportion"). Bank A and Bank B encrypt the calculated remittance amount proportion using a secret distribution method and register this distribution value with the secret computing system (step 1). Bank B can either perform steps 0 and 1 simultaneously, or register both the distribution value of the remittance amount proportion and the distribution value of the initial page sorting value with the secret computing system at the same time.

[0056] Based on the calculation request from Bank A, the secret calculation system uses the dispersion value of the registered remittance amount ratio and the dispersion value of the page ranking to secretly calculate and rank the pages of each enterprise (step 2). The secret calculation system sends the calculated dispersion values ​​of the page rankings for each enterprise to Bank A, which made the calculation request.

[0057] Bank A reconstructs the page ranking dispersion values ​​received from the secret computing system for each enterprise, obtaining the plaintext page ranking. Then, it determines whether the obtained page ranking has converged. If it determines that the page rankings for each enterprise have not converged, Bank A sends the page ranking dispersion values ​​to the secret computing system. The secret computing system updates its maintained page ranking dispersion values ​​for each enterprise using the received dispersion values ​​(step 3) and recalculates the page ranking (step 2). If it determines that the page rankings for each enterprise have converged, Bank A uses the converged page rankings to determine the creditworthiness of the enterprises.

[0058] [Example 2]

[0059] Example 2 is an embodiment of applying the secret page ranking calculation system of Modification 1 to providing credit results to external organizations. When conducting new transactions with enterprises included in the transaction history of various banks, the external organization requests page ranking calculations for credit assessment. Figure 6 As shown, the secret page calculation system in Embodiment 2, based on Bank A, Bank B, and the secret calculation system, also includes an external organization X. The secret calculation system sends the calculated dispersion value of the page ranking to the external organization X, and the external organization X determines the convergence of the page ranking (step 3). If the page ranking is determined to be converged, the external organization X uses the converged page ranking to determine the creditworthiness of the enterprise. Other processing is the same as in Embodiment 1.

[0060] The embodiments of the present invention have been described above, but the specific structure is not limited to these embodiments. Even if appropriate design changes exist without departing from the spirit of the present invention, they are obviously included in the present invention. The various processes described in the embodiments can be executed not only in the order described in time, but also in parallel or individually depending on the processing capability of the device performing the process or as needed.

[0061] [Program, Recording Medium]

[0062] When implementing the various processing functions of the devices described in the above embodiments using a computer, the processing content of the functions that each device should have is described by a program. Furthermore, by making... Figure 7 The computer's storage unit 1020 reads the program and causes the arithmetic processing unit 1010, input unit 1030, output unit 1040, etc. to operate, thereby realizing the various processing functions of the above-mentioned devices on the computer.

[0063] This describes a program that can be recorded on a computer-readable recording medium. Computer-readable recording media include, for example, non-transitory recording media such as magnetic recording devices and optical discs.

[0064] Furthermore, the program can be circulated, for example, by selling, transferring, or lending portable recording media such as DVDs and CD-ROMs containing the program. Alternatively, the program can be circulated by storing it in the storage device of a server computer and forwarding it from the server computer to other computers via a network.

[0065] A computer executing such a program may first temporarily store the program recorded on a removable recording medium or the program forwarded from a server computer in its own auxiliary storage unit 1005, which is a non-temporary storage device. Then, during processing, the computer reads the program stored in its own auxiliary storage unit 1005 and executes the processing according to the read program. Alternatively, as another way of executing the program, the computer may directly read the program from the removable recording medium and execute the processing according to that program. Furthermore, it may execute the processing according to the received program sequentially each time a program is forwarded from the server computer to the computer. Furthermore, the structure may be configured such that the forwarding of the program from the server computer to the computer is not performed, but the processing function is achieved solely through the execution instructions and result retrieval, using a so-called ASP (Application Service Provider) type service. Additionally, the program in this embodiment may include information provided for computer-based processing and program-dependent information (data, etc., that, while not direct instructions to the computer, have the nature of specifying the computer's processing).

[0066] Furthermore, in this embodiment, the device is configured by executing a prescribed program on a computer, but it may also be configured to implement at least a portion of these processing contents in hardware.

Claims

1. A secret page sorting calculation system, comprising multiple secret computing devices, each secret computing device comprising: The page sorting storage unit stores the encrypted page sorting of multiple transaction entities; The transaction ratio receiving unit receives encrypted transaction ratios from various data sources across multiple data sources; and The page ranking calculation unit secretly calculates the page ranking of the transaction subject, which becomes the calculation object, using the ciphertext of the transaction ratio related to the transaction subject of the calculation object, and the ciphertext of the page ranking of the transaction subject that has transacted with the transaction subject of the calculation object. The transaction ratio represents the proportion of transactions conducted through each combination of the transaction entities within the overall transaction history maintained by the data source. [·] represents the ciphertext of the value ·, i is the serial number of the transaction subject of the computation object, j is the serial number of the transaction object, and V i It is the collection of the transaction objects, PR j α1, ..., α2 are the page sortings of transaction object j, N is the number of data sources, and α1, ..., α2 are the data sources. N It satisfies α1+α 2+ …+ α N The weight k = 1 n j→i (n=1, 2, …, N) represents the proportion of transactions related to transaction subject i and transaction object j received from the nth data source. The page ranking calculation unit calculates the ciphertext of the page ranking for transaction subject i using the following formula. [Formula 2] 。 2. The secret page sorting calculation system according to claim 1, wherein, The secret page ranking calculation system includes a calculation request device for any one of the plurality of data sources or different from the plurality of data sources. The computation request device includes: The page sorting decryption unit decrypts the ciphertext of the page sorting of the transaction subject of the computing object received from each secret computing device; and The convergence determination unit determines whether the page ranking of the transaction entities of the computational object has converged. Each secret computing device also includes: The page sorting sending unit sends the encrypted page sorting information of the transaction subject of the computing object to the computing request device; and The page sorting update unit updates the encrypted page sorting of the transaction subject of the computational object stored in the page sorting storage unit when the page sorting of the transaction subject of the computational object has not converged.

3. The secret page sorting calculation system according to claim 2, wherein, The secret page sorting calculation system includes an initialization device that is not part of the calculation request device in the data source. The initialization device includes a page sorting initial value registration unit, which encrypts the initial value of the randomly generated page sorting of the transaction object and sends the ciphertext to each secret computing device.

4. A secret page sorting calculation method, which is a secret page sorting calculation method executed by a secret page sorting calculation system comprising multiple secret computing devices. In the page sorting storage unit of each secret computing device, the encrypted text of the page sorting of multiple transaction entities is stored. Each secret computing device's transaction ratio receiving unit receives ciphertext of transaction ratios from various data sources across multiple data sources. Each secret computing device's page ranking calculation unit uses ciphertext related to the transaction proportion of the transaction subject being calculated, and ciphertext of the page ranking of the transaction subject that has transacted with the transaction subject being calculated, to secretly calculate the ciphertext of the page ranking of the transaction subject that becomes the transaction subject being calculated after decryption. The transaction ratio represents the proportion of transactions conducted through each combination of the transaction entities within the overall transaction history maintained by the data source. [·] represents the ciphertext of the value ·, i is the serial number of the transaction subject of the computation object, j is the serial number of the transaction object, and V i It is the collection of the transaction objects, PR j α1, ..., α2 are the page sortings of transaction object j, N is the number of data sources, and α1, ..., α2 are the data sources. N It satisfies α1+α 2+ …+ α N The weight k = 1 n j→i (n=1, 2, …, N) represents the proportion of transactions related to transaction subject i and transaction object j received from the nth data source. The page ranking calculation unit calculates the ciphertext of the page ranking for transaction subject i using the following formula. [Formula 2] 。 5. A secret computing device, which is used in the secret page sorting computing system according to any one of claims 1 to 3.

6. A computer program product for causing a computer to perform the steps of the secret page sorting calculation method of claim 4.

Images