Multi-party data processing methods, multi-party joint recommendation methods, equipment and readable media

Abstract

This disclosure relates to a multi-party data processing method, a multi-party joint recommendation method, an apparatus, and a readable medium, and relates to the fields of multi-party secure computation and privacy computation. The method determines the intersection potential of the requester's first sample set and the data party's second sample set, filtered by an initial threshold for feature values, through anonymous privacy intersection calculation, without disclosing plaintext information from any party, thus avoiding privacy leaks. Furthermore, it calculates business indicators using the intersection potential and the number of first samples from the first user sample, and evaluates the filtering effect of the initial threshold by the selected intersection quantity to determine a target threshold. This enables the determination of a feature value distribution that fully meets the requester's business needs without disclosing plaintext information. Finally, the data party divides different fourth sample sets based on the target threshold and feature values, corresponding to different business indicator levels of the requester, thereby allowing the data party to support the requester's differentiated business needs in subsequent multi-party collaborations.

Description

Technical Field

[0001] This disclosure relates to the fields of multi-party secure computation and privacy computation technology, and more specifically, to a multi-party data processing method, a multi-party joint recommendation method, an apparatus, and a readable medium. Background Art

[0002] In scenarios involving multiple parties, data silos exist among them.

[0003] Currently, in order to break down data silos, it is usually necessary for all parties to exchange and analyze at least some of their data in plaintext, which poses a certain risk of privacy leakage in multi-party collaborative scenarios. Summary of the Invention

[0004] The purpose of this disclosure is to provide a multi-party data processing method, a multi-party joint recommendation method, an electronic device, and a computer-readable medium that can prevent the leakage of privacy data in multi-party collaboration.

[0005] According to a first aspect of this disclosure, a multi-party data processing method is provided. This method can be applied to a requesting party, which forms a first sample set with a first number of user samples. The method may include: initiating an anonymous privacy intersection with the data party based on each initial threshold; calculating the intersection potential of the first sample set and a second sample set corresponding to the initial threshold; the second sample set being obtained by the data party after filtering a local third sample set based on the initial threshold and feature values; the third sample set including second user samples, each second user sample having a corresponding feature value; determining a business indicator corresponding to the initial threshold based on the intersection potential and the number of first samples in the first sample set; determining at least one target threshold based on the business indicator within the initial threshold; and providing the target threshold to the data party so that the data party divides the third sample set into at least two fourth sample sets based on the target threshold and feature values.

[0006] Optionally, based on each initial threshold, an anonymous privacy intersection is initiated with the data provider to calculate the intersection potential of the first sample set and the second sample set corresponding to the initial threshold. This includes: initiating an anonymous privacy intersection with the data provider based on each initial threshold; performing random rearrangement, predetermined elliptic curve mapping, and encryption based on the first private key on the first sample set to obtain a first encryption result; sending the first encryption result to the data provider; obtaining a second encryption result provided by the data provider, which is obtained by the data provider performing random rearrangement, predetermined elliptic curve mapping, and encryption based on the second private key on the second sample set corresponding to the initial threshold; performing random rearrangement and encryption based on the first private key on the second encryption result to obtain a third encryption result; obtaining a fourth encryption result provided by the data provider, which is obtained by the data provider performing random rearrangement and encryption based on the second private key on the first encryption result; and comparing the third encryption result and the fourth encryption result to obtain the intersection potential of the first sample set and the second sample set.

[0007] Optionally, the first sample set includes a positive sample subset and a negative sample subset. Anonymous privacy intersection is initiated with the data provider based on each initial threshold. The intersection potential of the first sample set and the second sample set corresponding to the initial threshold is calculated. This includes: initiating two anonymity privacy intersections with the data provider based on each initial threshold, respectively calculating the positive sample intersection potential of the positive sample subset and the second sample set corresponding to the initial threshold, and calculating the negative sample intersection potential of the negative sample subset and the second sample set corresponding to the initial threshold. The first sample quantity includes the number of positive samples and the number of negative samples. Based on the intersection potential and the number of first samples in the first sample set, the business indicator corresponding to the initial threshold is determined. This includes: determining the business indicator corresponding to the initial threshold based on the positive sample intersection potential, the negative sample intersection potential, the number of positive samples in the first sample set, and the number of negative samples in the first sample set.

[0008] Optionally, if the number of second samples in the second user samples in the third sample set is greater than the number of first samples, before initiating anonymity and privacy intersection with the data provider based on each initial threshold, and calculating the positive and negative sample intersection potentials of the first sample set and the second sample set corresponding to the initial threshold, the method further includes: obtaining the number of second samples provided by the data provider; and aligning the number of samples in the first sample set and the second sample set based on the number of second samples.

[0009] Optionally, before initiating anonymity and privacy intersection with the data provider based on each initial threshold, and calculating the positive and negative sample intersection potentials of the first sample set and the second sample set corresponding to the initial threshold, the method further includes: obtaining the feature value range provided by the data provider; determining at least one initial threshold based on the feature value range; and providing the initial threshold to the data provider.

[0010] Optionally, business metrics may include at least one of hit rate, accuracy, and improvement.

[0011] Optionally, the data parties include at least two; wherein each data party corresponds to the same feature type; or, at least one data party corresponds to a feature type different from the feature types corresponding to the other data parties.

[0012] According to the second aspect of this disclosure, a multi-party joint recommendation method is provided. This method can be applied to the requesting party and may include: initiating an anonymous query to the data party for the user to be recommended, and determining the priority corresponding to the user to be recommended; each priority corresponds to a fourth sample set divided by the data party based on the target threshold and feature values ​​after providing a target threshold to the data party in the multi-party data processing method of the first aspect; and delivering recommendations to the user to be recommended based on the priority.

[0013] According to a third aspect of this disclosure, an electronic device is provided that can act as a requester, the requester forming a first sample set with a first number of first user samples. The electronic device may include: a privacy intersection module, configured to initiate anonymity privacy intersection with a data party based on each initial threshold, and calculate the intersection potential of the first sample set and a second sample set corresponding to the initial threshold; the second sample set is obtained by the data party after filtering a local third sample set according to the initial threshold and feature values; the third sample set includes second user samples, each second user sample having a corresponding feature value; an indicator calculation module, configured to determine a business indicator corresponding to the initial threshold based on the intersection potential and the first number of first samples in the first sample set; a threshold determination module, configured to determine at least one target threshold based on the business indicator within the initial threshold; and a data transmission module, configured to provide the target threshold to the data party, so that the data party divides the third sample set into at least two fourth sample sets based on the target threshold and feature values.

[0014] Optionally, the privacy intersection module is specifically used to initiate anonymity privacy intersection with the data party based on each initial threshold, perform random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the first private key on the first sample set to obtain a first encryption result; send the first encryption result to the data party; obtain a second encryption result provided by the data party, which is obtained by the data party performing random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the second private key on the second sample set corresponding to the initial threshold; perform random rearrangement and encryption processing based on the first private key on the second encryption result to obtain a third encryption result; obtain a fourth encryption result provided by the data party, which is obtained by the data party performing random rearrangement and encryption processing based on the second private key on the first encryption result; compare the third encryption result and the fourth encryption result to obtain the intersection potential of the first sample set and the second sample set.

[0015] Optionally, the first sample set includes a positive sample subset and a negative sample subset. The privacy intersection module is specifically used to initiate two anonymous privacy intersections with the data provider based on each initial threshold, respectively calculating the positive sample intersection potential of the positive sample subset and the second sample set corresponding to the initial threshold, and calculating the negative sample intersection potential of the negative sample subset and the second sample set corresponding to the initial threshold. The first sample quantity includes the number of positive samples and the number of negative samples. The indicator calculation module is specifically used to determine the business indicator corresponding to the initial threshold based on the positive sample intersection potential, the negative sample intersection potential, the number of positive samples in the first sample set, and the number of negative samples in the first sample set.

[0016] Optionally, if the number of second samples in the third sample set is greater than the number of first samples, the electronic device further includes a data receiving module and a sample alignment module; wherein, the data receiving module is used to obtain the number of second samples provided by the data provider; and the sample alignment module is used to align the number of samples in the first sample set and the second sample set based on the number of second samples.

[0017] Optionally, the electronic device further includes a data receiving module and an initial threshold module; wherein, the data receiving module is used to acquire the feature value range provided by the data provider; the initial threshold module is used to determine at least one initial threshold based on the feature value range; and the data sending module is also used to provide the initial threshold to the data provider.

[0018] Optionally, business metrics may include at least one of hit rate, accuracy, and improvement.

[0019] Optionally, the data parties include at least two; wherein each data party corresponds to the same feature type; or, at least one data party corresponds to a feature type different from the feature types corresponding to the other data parties.

[0020] According to a fourth aspect of this disclosure, an electronic device is provided, which acts as a requester. The electronic device may include: a query module, used to initiate an anonymous query to a data provider for users to be recommended, and determine the priority corresponding to the users to be recommended; each priority corresponds to a fourth sample set divided by the data provider based on the target threshold and feature values ​​after the electronic device of the third aspect provides a target threshold to the data provider; and a delivery module, used to deliver to users to be recommended based on priority.

[0021] According to a fifth aspect of this disclosure, an electronic device is provided, comprising:

[0022] processor;

[0023] Memory, used to store computer programs for the processor;

[0024] The processor is configured to implement the aforementioned multi-party data processing method of the first aspect or the multi-party joint recommendation method of the second aspect by executing a computer program.

[0025] According to a sixth aspect of this disclosure, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements a multi-party data processing method as described in the first aspect, or a multi-party joint recommendation method as described in the second aspect.

[0026] According to a seventh aspect of this disclosure, a computer program product is provided that, when run on an electronic device, causes the electronic device to perform a multi-party data processing method as described in the first aspect, or a multi-party joint recommendation method as described in the second aspect.

[0027] The multi-party data processing method disclosed herein involves a requesting party forming a first sample set with a first number of user samples of a first sample size, and initiating an anonymous privacy intersection with the data party based on each initial threshold to determine the intersection potential of the first sample set and the second sample set corresponding to the initial threshold. The second sample set is obtained by the data party after filtering its local third sample set according to the initial threshold and feature values. The third sample set includes second user samples, and each second user sample has a corresponding feature value. Based on this, the business indicator corresponding to the initial threshold is determined based on the intersection potential and the first sample size in the first sample set, and a target threshold is determined based on the business indicator within the initial threshold and provided to the data party. Thus, the data party can divide the third sample set into at least two fourth sample sets based on the target threshold and feature values. This method determines the intersection potential between the requester's first sample set and the data provider's second sample set, filtered by the data provider using an initial threshold, without providing or acquiring plaintext information from any party, thus avoiding privacy leaks. Based on this, business metrics are calculated using the intersection potential and the number of samples in the first user sample set. The filtering effect of the initial threshold on the samples is evaluated using the selected intersection quantity, thereby determining the target threshold. This allows for the determination of a feature value distribution that fully meets the requester's business needs without leaking plaintext information. Furthermore, different fourth sample sets are divided by the data provider based on the target threshold and feature values, corresponding to different business metric levels of the requester. This enables the data provider to support the requester's differentiated business needs in subsequent multi-party collaborations.

[0028] The multi-party joint recommendation method disclosed herein can be applied to the requesting party, which initiates an anonymous query to the data party for users to be recommended, thereby determining the priority of the users to be recommended. Each priority corresponds to a fourth sample set partitioned by the data party based on the target threshold and feature values ​​after providing a target threshold to the data party in the aforementioned multi-party data processing method. Recommendations are then delivered to users based on these priorities. In this method, different fourth sample sets are partitioned by the data party based on the target threshold and feature values, and each corresponds to a priority, thus different priorities can correspond to different business indicator levels of the requesting party. Furthermore, the requesting party obtains priorities through anonymous queries without disclosing the information of the users to be recommended, nor does it obtain the feature values ​​of the data party, avoiding privacy leaks for all parties. Moreover, prioritizing these priorities allows for differentiated delivery of recommendations to users by comprehensively considering the feature value distribution of the data party and the recommendation business needs of the requesting party, improving business quality.

[0029] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. Attached Figure Description

[0030] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0031] Figure 1 This is one of the flowcharts of a multi-party data processing method provided in this embodiment of the disclosure.

[0032] Figure 2 This is a schematic diagram of the anonymous privacy intersection process in a multi-party data processing method provided in this embodiment of the disclosure.

[0033] Figure 3 This is a schematic diagram of the interaction flow of a multi-party data processing method provided in an embodiment of this disclosure.

[0034] Figure 4 This is a second flowchart of a multi-party data processing method provided in an embodiment of the present disclosure.

[0035] Figure 5 This is a flowchart illustrating the steps of a multi-party joint recommendation method provided in an embodiment of this disclosure.

[0036] Figure 6 This is one of the structural block diagrams of the electronic device provided in the embodiments of this disclosure.

[0037] Figure 7This is a second structural block diagram of an electronic device provided in an embodiment of this disclosure.

[0038] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0039] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0040] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0041] It should be noted that the data obtained by the requesting party and the data provider in this disclosure are accessed, collected, stored, and used for subsequent analysis and processing only after clearly informing the user or relevant data owner of the content of the data collection, the purpose of the data, the processing method, and other information. Furthermore, the requesting party and the data owner may send the user or relevant data owner the means to access, correct, or delete the data, as well as the method to revoke consent or authorization.

[0042] In multi-party collaborative scenarios, due to the existence of data silos among the parties, it is usually necessary for each party to exchange and analyze at least some plaintext data. This leads to the risk of privacy leakage and low data security in multi-party collaborative data processing and analysis.

[0043] In the embodiments of this disclosure, in the joint data processing of the requester and one or more data parties, the requester and data parties perform anonymity and privacy intersection to determine the intersection potential of the data party samples and the requester samples screened with different initial thresholds. Based on this, business indicators are calculated using the intersection potential to evaluate whether the feature value distribution under the initial threshold division meets the requester's business needs, so as to determine the target threshold in the initial threshold.

[0044] Furthermore, the embodiments of this disclosure also provide a joint recommendation method between the requester and one or more data providers. The data providers can divide their samples into different priorities based on the aforementioned target threshold and feature values. Thus, the requester can obtain the priority of the user to be recommended by initiating an anonymous query to the data provider, thereby carrying out differentiated delivery based on different priorities.

[0045] The joint recommendation method of this disclosure can be applied to various scenarios involving a requesting party jointly delivering data to one or more data providers.

[0046] For example, in a credit marketing scenario, the requesting party can possess credit marketing characteristics, while the data provider can possess credit conversion characteristics. The requesting party can build a highly discriminative and accurate marketing response model based on these characteristics to identify high-intent users for credit business. However, there may be underestimation issues regarding credit conversion. In this case, a data silo problem exists between the requesting party and the data provider.

[0047] Currently, when requesting parties collaborate with data providers for credit marketing, they typically need to provide a small batch of user samples to the data provider in advance to obtain the data provider's credit conversion characteristic distribution. They then conduct joint evaluations by querying the corresponding characteristic distribution information of the users to be marketed to the data provider. Providing a small batch of user samples may result in the leakage of user information to the data provider. Furthermore, the requesting party may infer additional information based on the credit conversion characteristics returned by the data provider, such as the correlation between the variables held by the requesting party and the credit conversion characteristics.

[0048] In this embodiment, the requesting party can perform anonymity and privacy intersection with the data provider to determine the intersection potential of the data provider samples and the requesting party samples filtered by different initial thresholds. Based on this intersection potential, business indicators are calculated to evaluate whether the credit feature distribution under the initial threshold division meets the requesting party's business needs, thereby determining the target threshold within the initial threshold. Furthermore, the data provider can divide the data provider samples into different priorities based on the aforementioned target threshold and credit features. Thus, the requesting party can obtain the priority of the users to be marketed by initiating an anonymity query with the data provider, thereby conducting differentiated marketing campaigns based on different priorities.

[0049] In another credit marketing scenario, the requester can hold credit marketing characteristics, while the data provider can hold post-loan risk characteristics.

[0050] The embodiments disclosed herein can also be applied to other marketing scenarios, such as product recommendation scenarios in e-commerce businesses, where the requesting party can hold product recommendation features and the data party can hold revenue and expenditure conversion features; or music recommendation scenarios in multimedia businesses, where the requesting party can hold music recommendation features and the data party can hold music product consumption features, etc.

[0051] Those skilled in the art can implement the multi-party data processing method and multi-party joint recommendation method provided in this disclosure in different application scenarios according to actual needs.

[0052] The following section, in conjunction with the above application scenarios, provides reference... Figure 1 This describes a multi-party data processing method according to exemplary embodiments of the present disclosure.

[0053] Figure 1 This is one of the flowcharts of a multi-party data processing method provided in this disclosure. The method can be applied to a requesting party, which forms a first sample set with a first number of first user samples.

[0054] In this embodiment of the disclosure, the requesting party may be an initiator that jointly processes and analyzes data with one or more data parties. The requesting party may hold user information and analyze the user information by comprehensively considering the feature information held by the data parties, supporting user-related classification, reasoning, recommendation, and other business processes. In the multi-party data processing process initiated by the requesting party to the data parties, the requesting party may form a first sample set with a first number of first user samples. Each first user sample may correspond to one piece of user information, such as each first user sample corresponding to a user identifier. The first sample set may include all or part of the user information held by the requesting party.

[0055] In this embodiment of the disclosure, the requesting party may hold local feature information in addition to user information. The feature information held by the requesting party and the feature information held by the data party may be of the same feature type or different feature types; this embodiment of the disclosure does not impose specific limitations in this regard.

[0056] like Figure 1 As shown, the method may include steps 101 to 104.

[0057] In step 101, an anonymous privacy intersection is initiated with the data party based on each initial threshold to calculate the intersection potential of the first sample set and the second sample set corresponding to the initial threshold. The second sample set is obtained by the data party after filtering the local third sample set according to the initial threshold and feature values. The third sample set includes second user samples, and each second user sample has a corresponding feature value.

[0058] In this embodiment, the data provider holds a third sample set locally. This third sample set consists of a second user sample set held locally by the data provider, and each second user sample has a corresponding feature value. This feature value is the feature information held by the data provider. Based on this, the data provider filters the local third sample set according to an initial threshold and the feature value to obtain a second sample set. Thus, each initial threshold has a corresponding second sample set, and the feature values ​​corresponding to all second samples in the second sample set corresponding to the initial threshold are consistent with the comparison relationship of the initial threshold. This comparison relationship can be selected according to actual calculation conditions, business requirements, etc. For example, the comparison relationship can be that the feature value is greater than or equal to the initial threshold, then the feature values ​​corresponding to all second samples in the second sample set corresponding to each initial threshold are greater than or equal to the initial threshold; and so on. The comparison relationship can also be less than or equal to, greater than, less than, equal to, etc. For each initial threshold, the data provider can determine a second sample set. The initial threshold can be determined by the data provider and provided to the requesting party, or determined by the requesting party and provided to the data provider, or jointly determined by the requesting party and the data provider through negotiation.

[0059] In this embodiment of the disclosure, the requesting party can initiate anonymity and privacy intersection with the data party based on each initial threshold, and initiate one round of anonymity and privacy set intersection for each initial threshold. In each round of anonymity and privacy intersection, the requesting party and the data party calculate the intersection potential of the first sample set and the second sample set corresponding to the initial threshold of that round. The intersection potential reflects the number of intersections between the second user samples obtained by the data party based on the initial threshold and the first user samples in the requesting party.

[0060] In step 102, the business indicators corresponding to the initial threshold are determined based on the intersection potential and the number of the first samples in the first sample set.

[0061] In this embodiment of the disclosure, for one or more initial thresholds, each initial threshold can be calculated using anonymity and privacy intersection calculation. Based on this, the business metrics corresponding to the initial threshold can be determined based on the intersection potential and the number of first samples in the first sample set. Since the initial threshold filters the feature value range of the second user samples, the intersection potential can represent the number of first user samples that satisfy the feature value range. Therefore, based on the intersection potential and the number of first samples, the distribution of first user samples satisfying the feature range in the first sample set can be reflected. The business metrics can be set based on the requester's business needs. The filtering effect of the initial threshold on the first user samples can be represented by describing the comparison relationship between the intersection potential and the number of first samples. For example, the business metrics can be the ratio between the intersection potential and the number of first samples. The business metrics can also be the quantitative relationship between the number of first user samples corresponding to different types of first user samples and the intersection potential. The type of first user sample can be set according to business needs, such as gender, region, age range, occupation, consumption behavior, etc. The number of business metrics can be one or more. Those skilled in the art can set the content and number of business metrics according to actual needs; this embodiment of the disclosure does not impose specific limitations on this.

[0062] In step 103, at least one target threshold is determined from the initial threshold based on business metrics.

[0063] In this embodiment of the disclosure, for one or more initial thresholds, it can be determined whether they meet business requirements based on their respective business indicators, and then the target threshold can be determined.

[0064] For example, taking the ratio of the intersection potential to the number of first samples as a business indicator, it can be that when the ratio of the intersection potential to the number of first samples is greater than or equal to a preset ratio, the initial threshold is considered to meet the business requirements, and then the initial threshold is determined as the target threshold.

[0065] Furthermore, when the requesting party has differentiated business needs, multiple target thresholds can be determined. In this case, multiple preset ratios can be set based on different business needs to confirm the target thresholds that meet different levels of business needs.

[0066] For example, a first preset ratio and a second preset ratio are set, and the first preset ratio is greater than the second preset ratio. Then, when the ratio of the intersection potential to the number of first samples is greater than or equal to the first preset ratio, the initial threshold is determined as the first target threshold; when the ratio of the intersection potential to the number of first samples is less than or equal to the second preset ratio, the initial threshold is determined as the second target threshold.

[0067] In step 104, a target threshold is provided to the data provider so that the data provider can divide the third sample set into at least two fourth sample sets based on the target threshold and feature values.

[0068] In this embodiment of the disclosure, the requesting party provides a definite target threshold to the data party. Based on this, the data party divides the third sample set into at least two fourth sample sets according to the target threshold. For example, when there is one target threshold, the third sample set can be divided into two fourth sample sets, each including second user samples with feature values ​​greater than or equal to the target threshold, and second user samples with feature values ​​less than the target threshold. When there are two target thresholds, the third sample set can be divided into three fourth sample sets, each including second user samples with feature values ​​greater than or equal to the first target threshold, second user samples with feature values ​​less than the first target threshold but greater than the second target threshold, and second user samples with feature values ​​less than or equal to the second target threshold. This process continues when there are three target thresholds. Based on this, the data party's second user samples are divided into different fourth sample sets according to the feature value distribution, a division method that meets the requesting party's business requirements.

[0069] In this embodiment, the requesting party initiates a Private Set Intersection (PSI) to the data party to calculate the intersection potential, enabling the determination of the number of elements in the intersection among multiple parties without disclosing their respective data content. In practical applications, homomorphic encryption, ECDH (Elliptic Curve Diffie-Hellman key Exchange) protocol, and other methods can be used to calculate the intersection potential. Taking the ECDH protocol as an example... Figure 2 This is a schematic diagram illustrating the anonymous privacy intersection process in a multi-party data processing method provided in this embodiment of the disclosure. Figure 2 As shown, on the requesting side, the anonymous privacy request in step 101 can include steps 201 to 206 as follows.

[0070] In step 201, an anonymous privacy intersection is initiated with the data party based on each initial threshold, and the first sample set is randomly rearranged, mapped to a predetermined elliptic curve, and encrypted based on the first private key to obtain the first encryption result.

[0071] In this embodiment of the disclosure, when the requesting party initiates anonymity-privacy intersection with the data party for each initial threshold, in each round of anonymity-privacy intersection, the requesting party can prepare a first private key and randomly rearrange its local first sample set, making the rearranged first sample set more randomly ordered than before. Further, the requesting party can perform a predetermined elliptic curve mapping on the randomly rearranged first sample set. The predetermined elliptic curve can be an elliptic curve over a finite field pre-negotiated by the requesting party and the data party, used for ECDH encryption calculations. The requesting party can hash each first user sample in the first sample set and map it to a corresponding point on the predetermined elliptic curve. Based on this, encryption is performed again using the first private key to obtain a first encryption result of the first sample set after ECDH encryption.

[0072] In step 202, the first encryption result is sent to the data provider.

[0073] In this embodiment of the disclosure, the requesting party can send the first encryption result to the data party. When there are multiple data parties, the requesting party can use the same first private key and predetermined elliptic curve for multiple data parties; or it can use different predetermined elliptic curves and prepare different first private keys according to the agreement, and then send the first encryption result obtained by ECDH encryption based on the corresponding predetermined elliptic curve and first private key to the corresponding data party.

[0074] In step 203, the second encryption result provided by the data provider is obtained. The second encryption result is obtained by the data provider through random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the second private key on the second sample set corresponding to the initial threshold.

[0075] In this embodiment of the disclosure, the requesting party can obtain a second encrypted result provided by the data provider. When the requesting party initiates anonymity-privacy intersection with the data provider for each initial threshold, the data provider can prepare a second private key in each round of anonymity-privacy intersection and randomly rearrange the second sample set corresponding to that initial threshold locally, so that the rearranged second sample set is randomly ordered compared to the original. Further, the data provider can perform a predetermined elliptic curve mapping on the randomly rearranged second sample set. The data provider can hash each second user sample in the second sample set and map it to the corresponding point on the predetermined elliptic curve. Based on this, encryption is performed again based on the second private key to obtain a second encrypted result of the second sample set encrypted with ECDH, and the second encrypted result is provided to the requesting party.

[0076] In step 204, the second encryption result is randomly rearranged and encrypted based on the first private key to obtain the third encryption result.

[0077] In this embodiment of the disclosure, the requesting party can randomly rearrange the obtained second encryption result and then encrypt the second encryption result again based on the first private key to obtain a third encryption result. Based on this, the third encryption result undergoes two rearrangements between the requesting party and the data party, and two encryptions based on the first and second private keys respectively.

[0078] In step 205, the fourth encryption result provided by the data provider is obtained. The fourth encryption result is obtained by the data provider randomly rearranging the first encryption result and encrypting it based on the second private key.

[0079] In this embodiment of the disclosure, the requesting party can obtain the fourth encryption result provided by the data party. After obtaining the first encryption result provided by the requesting party in step 203, the data party can randomly rearrange the first encryption result and encrypt it based on the second private key to obtain the fourth encryption result and provide it to the requesting party. On this basis, the fourth encryption result undergoes two rearrangements between the requesting party and the data party, and two encryptions based on the first private key and the second private key respectively.

[0080] In step 206, the third encryption result is compared with the fourth encryption result to obtain the intersection potential of the first sample set and the second sample set.

[0081] In this embodiment, the requester can obtain a third encryption result and a fourth encryption result. The third encryption result is obtained by rearranging and encrypting the second sample set twice, and the fourth encryption result is obtained by rearranging and encrypting the first sample set twice. Since ECDH encryption is commutative, when elements of the first and second sample sets overlap, corresponding elements of the fourth and third encryption results also overlap. Therefore, the intersection potential of the first and second sample sets can be obtained by comparing the third and fourth encryption results. In the anonymous and privacy-preserving intersection process, the requester and the data provider can each randomly sort their data, preventing the requester from inferring the correspondence between the first and fourth encryption results based on bit order when obtaining the fourth encryption result. This ensures that the plaintext of the intersection is not leaked, and only the intersection potential is output.

[0082] In an optional embodiment of this disclosure, the first sample set includes a positive sample subset and a negative sample subset. Based on this, step 101 may include step A1.

[0083] In step A1, two anonymous privacy intersections are initiated with the data party based on each initial threshold. The positive sample intersection potential between the positive sample subset and the second sample set corresponding to the initial threshold is calculated, and the negative sample intersection potential between the negative sample subset and the second sample set corresponding to the initial threshold is calculated.

[0084] In this embodiment, the first sample set can be divided into a positive sample subset and a negative sample subset. The setting of positive and negative samples can be determined based on the data provider's feature type and the requester's business needs. For example, in a credit marketing scenario, when the data provider holds credit conversion characteristics, the positive sample subset may include first user samples that have completed credit conversion, and the negative sample subset may include first user samples that have not completed credit conversion; when the data provider holds post-loan risk characteristics, the positive sample subset may include first user samples that have not defaulted after the loan, and the negative sample subset may include first user samples that have defaulted after the loan. Based on this, when initiating an anonymous privacy intersection with the data provider based on each initial threshold, two anonymous privacy intersections can be performed in each round to calculate the positive sample intersection potential between the positive sample subset and the second sample set corresponding to the initial threshold, and to calculate the negative sample intersection potential between the negative sample subset and the second sample set corresponding to the initial threshold. The calculation of each anonymous privacy intersection can be referred to the relevant descriptions of steps 201 to 206 above; to avoid repetition, they will not be repeated here.

[0085] Furthermore, the first sample size includes the number of positive samples and the number of negative samples. Based on the aforementioned step A1, the aforementioned step 102 includes the following step A2.

[0086] In step A2, the business indicators corresponding to the initial threshold are determined based on the intersection potential of positive samples, the intersection potential of negative samples, the number of positive samples in the first sample set, and the number of negative samples in the first sample set.

[0087] In this embodiment, the number of first samples in the first sample set can be further divided into the number of positive samples and the number of negative samples. The number of positive samples corresponds to the number of first user samples in the positive sample subset, and the number of negative samples corresponds to the number of first user samples in the negative sample subset. Based on this, the positive sample intersection potential represents the number of elements in the first user samples within the feature value range after initial threshold filtering that overlap with the positive sample subset, and the negative sample intersection potential represents the number of elements in the first user samples within the feature value range after initial threshold filtering that overlap with the negative sample subset. At this point, based on the positive sample intersection potential, negative sample intersection potential, number of positive samples, and number of negative samples, business indicators can be designed more precisely to evaluate the distribution of the data party's feature values ​​on the first sample set.

[0088] For example, taking three initial thresholds T1, T2, and T3 from small to large as an example, Figure 3 This is a schematic diagram of the interaction flow of a multi-party data processing method provided in an embodiment of this disclosure. Figure 3 As shown, the interaction process between the requester and the data provider is as follows, from steps 301 to 309.

[0089] Step 301: The requesting party initiates two anonymous privacy intersections with the data party based on the initial threshold T1, respectively calculating the positive sample intersection potential 1 between the positive sample subset and the second sample set corresponding to the initial threshold T1, and calculating the negative sample intersection potential 1 between the negative sample subset and the second sample set corresponding to the initial threshold T1.

[0090] Step 302: The requester calculates business indicator 1 based on the intersection potential of positive samples 1, the intersection potential of negative samples 1, the number of positive samples, and the number of negative samples.

[0091] Step 303: The requesting party initiates two anonymous privacy intersections with the data party based on the initial threshold T2, respectively calculating the positive sample intersection potential 2 of the positive sample subset and the second sample set corresponding to the initial threshold T2, and calculating the negative sample intersection potential 2 of the negative sample subset and the second sample set corresponding to the initial threshold T2.

[0092] Step 304: The requester calculates business indicator 2 based on the intersection potential of positive samples 2, the intersection potential of negative samples 2, the number of positive samples, and the number of negative samples.

[0093] Step 305: The requesting party initiates two anonymous privacy intersections with the data party based on the initial threshold T3, respectively calculating the positive sample intersection potential 3 between the positive sample subset and the second sample set corresponding to the initial threshold T3, and calculating the negative sample intersection potential 3 between the negative sample subset and the second sample set corresponding to the initial threshold T3.

[0094] Step 306: The requester calculates business indicator 3 based on the intersection potential of positive samples 3, the intersection potential of negative samples 3, the number of positive samples, and the number of negative samples.

[0095] Step 307: Based on business metric 1, business metric 2 and business metric 3, the requester determines the initial thresholds T1 and T3 as target thresholds from the initial thresholds T1, T2 and T3.

[0096] Step 308: Request the data provider to provide target thresholds T1 and T3.

[0097] Step 309: The data is divided based on the target thresholds T1 and T3 to obtain the fourth sample set corresponding to the feature values ​​[T3,max], [T1,T3) and [min,T1).

[0098] Figure 4 The second flowchart of a multi-party data processing method provided in this embodiment of the present disclosure is applicable to a requesting party, which forms a first sample set with a first number of first user samples.

[0099] In this embodiment of the disclosure, the requesting party, the first user sample, and the first sample set can be referred to as described above. Figure 1To avoid repetition, the relevant descriptions will not be repeated here.

[0100] For example, the requester prepares a small batch of first user samples in the local user sample database. x Forming the first sample set At this point, the number of the first sample is... m .

[0101] In addition, the data provider holds a second user sample locally. y and the corresponding eigenvalues Then there is a third sample set. At this point, the number of the second sample is... n .

[0102] like Figure 4 As shown, in an optional embodiment of this disclosure, the number of second samples of the second user samples in the third sample set is greater than the number of first samples, and the method may include the following steps 401 to 402.

[0103] In step 401, the second number of samples provided by the data provider is obtained.

[0104] In this embodiment of the disclosure, when the second sample size is greater than the first sample size, the requesting party can perform sample size alignment during data preparation. This sample size alignment is based on the second sample size provided by the data provider, and the data provider can synchronize the second sample size to the requesting party.

[0105] For example, in At the same time, the data receiving party synchronizes... n .

[0106] In step 402, the sample counts of the first sample set and the second sample set are aligned based on the second sample count.

[0107] In this embodiment of the disclosure, sample number alignment ensures that the number of elements in the first sample set and the second sample set are the same. This can be achieved by filling empty spaces in the first sample set with invalid data. The first sample set after sample number alignment includes, in addition to the first user samples of the first sample number, invalid data whose quantity satisfies the difference between the first and second sample numbers.

[0108] For example, based on n Fill in the first sample set ( n - m () invalid data points, obtain the first sample set .

[0109] Furthermore, if the first sample set includes a positive sample subset and a negative sample subset, invalid data can be filled into the corresponding set according to the ratio of the number of positive samples to the number of negative samples to align the sample numbers.

[0110] For example, if positive samples account for 80% of the first sample size and negative samples account for 20%, then based on... n In the positive sample set Fill Invalid data, and in the negative sample set Fill Invalid data.

[0111] like Figure 4 As shown, in an optional embodiment of this disclosure, the method may include steps 403 to 405.

[0112] In step 403, the range of feature values ​​provided by the data provider is obtained.

[0113] In step 404, at least one initial threshold is determined based on the feature value range.

[0114] In step 405, an initial threshold is provided to the data provider.

[0115] In this embodiment of the disclosure, the feature value range can be the total range of feature values ​​corresponding to the second user sample held by the data provider. Based on this, the requesting party can determine one or more initial thresholds according to business needs, historical experience, and other factors. The specific number of initial thresholds can be determined through negotiation between the requesting party and the data provider.

[0116] In an optional embodiment of this disclosure, when providing feature values, the data provider may also provide the requesting party with one or more suggested initial thresholds based on feature value distribution, business experience, etc. The requesting party may adopt, adjust, or revise the initial thresholds, thus improving the efficiency and accuracy of subsequent target threshold determination by integrating the experience and needs of both the requesting party and the data provider during the preparation phase.

[0117] For example, the requesting party obtains a range of feature values ​​provided by the data provider, and then synchronizes data with the data provider based on this range of feature values. k Initial thresholds .in, k It can take the value [3, 6].

[0118] Based on this, the data provider can prepare k A second sample set .in, That is, each second sample set includes second user samples whose feature values ​​are greater than the corresponding initial threshold, and each second user sample includes the corresponding feature value. On this basis, invalid data is used to align the number of samples in each second sample set.

[0119] For example, with an initial threshold For example, the second sample set Y 1 include score Greater than or equal to of The second user sample, and Invalid data.

[0120] like Figure 4 As shown, the method may further include steps 406 to 409.

[0121] In step 406, an anonymous privacy intersection is initiated with the data party based on each initial threshold to calculate the intersection potential of the first sample set and the second sample set corresponding to the initial threshold; the second sample set is obtained by the data party after filtering the local third sample set according to the initial threshold and feature values; the third sample set includes second user samples, and each second user sample has a corresponding feature value.

[0122] In this embodiment of the disclosure, step 406 can be referred to the relevant description of step 101 above. To avoid repetition, it will not be repeated here.

[0123] In this embodiment of the disclosure, based on k Each initial threshold requester initiates k rounds of anonymous privacy intersection calculations with the data provider. The first sample set includes... and For example, in the first i In the process of seeking privacy and engaging in clandestine activities, The specific process for executing the ECDH protocol is shown in Table 1 below:

[0124] Table 1

[0125]

[0126] And so on, in the... i In the process of seeking privacy and engaging in clandestine activities, When executing the ECDH protocol, the requester outputs the positive sample intersection potential. .

[0127] In step 407, the business indicators corresponding to the initial threshold are determined based on the intersection potential and the number of the first samples in the first sample set.

[0128] In this embodiment of the disclosure, step 407 can be referred to the relevant description of step 102 above. To avoid repetition, it will not be repeated here.

[0129] In an optional embodiment of this disclosure, the business metrics include at least one of hit rate, accuracy, and lift.

[0130] In this embodiment of the disclosure, the business metrics may include hit rate, which represents the proportion of hits in the first user sample after initial threshold screening, such as the proportion of hits of intersection elements in the total number of samples, or the proportion of hits of intersection elements of positive samples in the number of positive samples, etc.; it may also include accuracy, which represents the proportion of samples that meet the requester's business requirements after initial threshold screening, such as the proportion of positive sample intersection potential in the total intersection potential; it may also include lift, which represents the improvement effect of user samples that meet business requirements after referencing data party feature values ​​versus user samples that meet business requirements without referencing data party feature values, such as the ratio between the proportion of positive sample intersection potential in the total intersection potential and the proportion of the number of positive samples in the first number of samples.

[0131] For example, hit rate hit_rate It can be calculated using the following formula (1):

[0132] (1)

[0133] in, This is the number of positive samples in the first sample count of the requester. In this case, the hit rate is expressed as the proportion of the positive sample intersection potential in the total number of positive samples.

[0134] For example, accuracy precision It can be calculated using the following formula (2):

[0135] (2)

[0136] At this point, accuracy is expressed as the proportion of the positive sample intersection potential in the sum of the positive sample intersection potential and the negative sample intersection potential.

[0137] For example, lift degree lift It can be calculated using the following formula (3):

[0138] (3)

[0139] At this point, the lift is expressed as the ratio between the proportion of the positive sample intersection potential in the sum of the positive sample intersection potential and the negative sample intersection potential, and the proportion of the number of positive samples in the first sample quantity.

[0140] In step 408, at least one target threshold is determined from the initial threshold based on business metrics.

[0141] In this embodiment, step 408 can be referred to the relevant description of step 103 above, and will not be repeated here to avoid repetition. Specifically, multi-dimensional analysis can be performed based on one business indicator or a combination of multiple business indicators to determine the target threshold, thereby improving the accuracy of the target threshold determination.

[0142] For example, at the initial threshold When the improvement is >1.5, accuracy >20%, and hit rate >90%, it is considered that the initial threshold is met. By selecting user samples that meet the requester's business needs, this test threshold can be set. This is determined as the target threshold.

[0143] In step 409, a target threshold is provided to the data provider so that the data provider can divide the third sample set into at least two fourth sample sets based on the target threshold and feature values.

[0144] In this embodiment of the disclosure, step 409 can be referred to the relevant description of step 104 above. To avoid repetition, it will not be repeated here.

[0145] In an optional embodiment of this disclosure, the data parties include at least two; wherein each data party corresponds to the same feature type; or, at least one data party corresponds to a feature type different from the feature types corresponding to other data parties.

[0146] In this embodiment of the disclosure, the requesting party may perform joint data processing with one or more data parties. When the number of data parties is greater than or equal to two, each data party may have the same feature type, or there may be at least one data party with a different feature type. For example, in a credit marketing scenario, the requesting party may perform joint data processing with multiple data parties that provide credit conversion features, or the requesting party may perform joint data processing with data parties that provide credit conversion features and data parties that provide post-loan risk features, respectively. The above feature types are only examples, and those skilled in the art can select the data parties to be combined with the requesting party according to actual business needs and data conditions. This embodiment of the disclosure does not impose specific limitations in this regard.

[0147] The multi-party data processing method disclosed herein involves a requesting party forming a first sample set with a first number of user samples of a first sample size, and initiating an anonymous privacy intersection with the data party based on each initial threshold to determine the intersection potential of the first sample set and the second sample set corresponding to the initial threshold. The second sample set is obtained by the data party after filtering its local third sample set according to the initial threshold and feature values. The third sample set includes second user samples, and each second user sample has a corresponding feature value. Based on this, the business indicator corresponding to the initial threshold is determined based on the intersection potential and the first sample size in the first sample set, and a target threshold is determined based on the business indicator within the initial threshold and provided to the data party. Thus, the data party can divide the third sample set into at least two fourth sample sets based on the target threshold and feature values. This method determines the intersection potential between the requester's first sample set and the data provider's second sample set, filtered by the data provider using an initial threshold, without providing or acquiring plaintext information from any party, thus avoiding privacy leaks. Based on this, business metrics are calculated using the intersection potential and the number of samples in the first user sample set. The filtering effect of the initial threshold on the samples is evaluated using the selected intersection quantity, thereby determining the target threshold. This allows for the determination of a feature value distribution that fully meets the requester's business needs without leaking plaintext information. Furthermore, different fourth sample sets are divided by the data provider based on the target threshold and feature values, corresponding to different business metric levels of the requester. This enables the data provider to support the requester's differentiated business needs in subsequent multi-party collaborations.

[0148] Figure 5 This disclosure provides a flowchart of a multi-party joint recommendation method, which can be applied to the requesting party. The method may include steps 501 to 502.

[0149] In step 501, an anonymous query is initiated to the data provider for the user to be recommended to determine the priority of the user to be recommended; each priority corresponds to a fourth sample set divided by the data provider based on the target threshold and feature value after the target threshold is provided to the data provider in the aforementioned multi-party data processing method.

[0150] In this embodiment of the disclosure, the requesting party and the data party are based on the foregoing Figure 1 , Figure 4The multi-party data processing method determines a target threshold, and after the data party divides the data into fourth sample sets based on the target threshold and feature values, it can set a corresponding priority for each fourth sample set. The priority setting can be based on the requester's differentiated business needs; for example, the fourth sample set whose business metrics better match the requester's business needs has a higher priority. For instance, taking a target threshold as an example, if the data party divides the data into two fourth sample sets, and the business metrics better match the requester's business needs when the feature value is greater than or equal to the target threshold, the fourth sample set containing the second user sample whose feature value is greater than or equal to the target threshold has a higher priority than the other fourth sample set.

[0151] Building upon this foundation, the requesting party can collaborate with the data provider for multi-party joint recommendations. The requesting party can initiate an anonymous query to the data provider regarding local users to be recommended, obtaining priority data from the data provider based on the fourth sample set matched by the users to be recommended, without disclosing the users themselves. This anonymous query can employ a keyword-matching-based scheme, implemented through homomorphic encryption. Communication is conducted using homomorphic ciphertext, preventing the data provider from determining the specific information of the users to be recommended, thus protecting the requesting party's privacy.

[0152] In step 502, the recommended users are delivered based on priority.

[0153] In this embodiment of the disclosure, the requesting party can target users to be recommended based on priority. For example, it can remove low-priority user groups from multiple users to be recommended and only target high-priority user groups, or target a small number of low-priority user groups and a large number of high-priority user groups. Alternatively, it can combine the requesting party's local feature information, weight the local feature information according to priority, and then target it differently. It can also remove low-priority user groups and target high-priority user groups after evaluating them based on local feature information.

[0154] For example, in a credit marketing scenario, when the requester holds marketing response characteristics and the data provider holds credit conversion characteristics, the requester can anonymously query the data provider for the priority of the users to be recommended, thereby eliminating users with relatively low priority, and determining the periodic marketing volume for different users to be recommended based on the marketing response characteristics in the user group with relatively high priority.

[0155] The multi-party joint recommendation method disclosed herein can be applied to the requesting party, which initiates an anonymous query to the data party for users to be recommended, thereby determining the priority of the users to be recommended. Each priority corresponds to a fourth sample set partitioned by the data party based on the target threshold and feature values ​​after providing a target threshold to the data party in the aforementioned multi-party data processing method. Recommendations are then delivered to users based on these priorities. In this method, different fourth sample sets are partitioned by the data party based on the target threshold and feature values, and each corresponds to a priority, thus different priorities can correspond to different business indicator levels of the requesting party. Furthermore, the requesting party obtains priorities through anonymous queries without disclosing the information of the users to be recommended, nor does it obtain the feature values ​​of the data party, avoiding privacy leaks for all parties. Moreover, prioritizing these priorities allows for differentiated delivery of recommendations to users by comprehensively considering the feature value distribution of the data party and the recommendation business needs of the requesting party, improving business quality.

[0156] Figure 6 This is one of the structural block diagrams of an electronic device 600 provided in an embodiment of this disclosure. The electronic device 600 acts as a requester, forming a first sample set with a first number of user samples. The electronic device 600 may include: a privacy intersection module 601, used to initiate anonymity and privacy intersection with the data provider based on each initial threshold, calculating the intersection potential of the first sample set and the second sample set corresponding to the initial threshold; the second sample set is obtained by the data provider after filtering a local third sample set according to the initial threshold and feature values; the third sample set includes second user samples, each second user sample having a corresponding feature value; an indicator calculation module 602, used to determine the business indicator corresponding to the initial threshold based on the intersection potential and the first number of samples in the first sample set; a threshold determination module 603, used to determine at least one target threshold based on the business indicator within the initial threshold; and a data sending module 604, used to provide the target threshold to the data provider, enabling the data provider to divide the third sample set into at least two fourth sample sets based on the target threshold and feature values.

[0157] In an optional embodiment of this disclosure, the privacy intersection module 601 is specifically configured to initiate anonymity privacy intersection with the data party based on each initial threshold, perform random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the first private key on the first sample set to obtain a first encryption result; send the first encryption result to the data party; obtain a second encryption result provided by the data party, the second encryption result being obtained by the data party performing random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the second private key on the second sample set corresponding to the initial threshold; perform random rearrangement and encryption processing based on the first private key on the second encryption result to obtain a third encryption result; obtain a fourth encryption result provided by the data party, the fourth encryption result being obtained by the data party performing random rearrangement and encryption processing based on the second private key on the first encryption result; and compare the third encryption result with the fourth encryption result to obtain the intersection potential of the first sample set and the second sample set.

[0158] In an optional embodiment of this disclosure, the first sample set includes a positive sample subset and a negative sample subset. The privacy intersection module 601 is specifically used to initiate two anonymous privacy intersections with the data provider based on each initial threshold, respectively calculating the positive sample intersection potential of the positive sample subset and the second sample set corresponding to the initial threshold, and calculating the negative sample intersection potential of the negative sample subset and the second sample set corresponding to the initial threshold. The first sample quantity includes the number of positive samples and the number of negative samples. The indicator calculation module 602 is specifically used to determine the business indicator corresponding to the initial threshold based on the positive sample intersection potential, the negative sample intersection potential, the number of positive samples in the first sample set, and the number of negative samples in the first sample set.

[0159] In an optional embodiment of this disclosure, the number of second samples in the third sample set is greater than the number of first samples. The electronic device further includes a data receiving module and a sample alignment module. The data receiving module is used to obtain the number of second samples provided by the data provider. The sample alignment module is used to align the number of samples in the first sample set and the second sample set based on the number of second samples.

[0160] In an optional embodiment of this disclosure, the electronic device 600 further includes a data receiving module and an initial threshold module; wherein, the data receiving module is used to acquire a range of feature values ​​provided by the data provider; the initial threshold module is used to determine at least one initial threshold based on the range of feature values; and the data sending module 604 is further used to provide the initial threshold to the data provider.

[0161] In an optional embodiment of this disclosure, the business metrics include at least one of hit rate, accuracy, and lift.

[0162] In an optional embodiment of this disclosure, the data parties include at least two; wherein each data party corresponds to the same feature type; or, at least one data party corresponds to a feature type different from the feature types corresponding to other data parties.

[0163] The electronic device 600 provided in this disclosure can, as a requester, form a first sample set with a first number of user samples of a first sample size, and initiate an anonymous privacy intersection with the data provider based on each initial threshold to determine the intersection potential of the first sample set and the second sample set corresponding to the initial threshold. The second sample set is obtained by the data provider after filtering the local third sample set according to the initial threshold and feature values. The third sample set includes second user samples, and each second user sample has a corresponding feature value. Based on this, the business indicator corresponding to the initial threshold is determined based on the intersection potential and the first sample size in the first sample set, and a target threshold is determined based on the business indicator in the initial threshold and provided to the data provider. Thus, the data provider can divide the third sample set into at least two fourth sample sets based on the target threshold and feature values. This method determines the intersection potential between the requester's first sample set and the data provider's second sample set, filtered by the data provider using an initial threshold, without providing or acquiring plaintext information from any party, thus avoiding privacy leaks. Based on this, business metrics are calculated using the intersection potential and the number of samples in the first user sample set. The filtering effect of the initial threshold on the samples is evaluated using the selected intersection quantity, thereby determining the target threshold. This allows for the determination of a feature value distribution that fully meets the requester's business needs without leaking plaintext information. Furthermore, different fourth sample sets are divided by the data provider based on the target threshold and feature values, corresponding to different business metric levels of the requester. This enables the data provider to support the requester's differentiated business needs in subsequent multi-party collaborations.

[0164] Figure 7 The second structural block diagram of the electronic device 700 provided in this embodiment of the present disclosure is shown. The electronic device 700 acts as a requester and may include: a query module 701, used to initiate an anonymous query to the data provider for the user to be recommended and determine the priority corresponding to the user to be recommended; after the electronic device in the third aspect of the above provides a target threshold to the data provider, the data provider divides a fourth sample set based on the target threshold and feature values; and a delivery module 702, used to deliver to the user to be recommended based on the priority.

[0165] The electronic device 700 provided in this disclosure can act as a requester to initiate an anonymous query to a data provider for users to be recommended, thereby determining the priority of each user. Each priority corresponds to a fourth sample set partitioned by the data provider based on the target threshold and feature values ​​after providing a target threshold to the data provider in the aforementioned multi-party data processing method. Users to be recommended are then delivered based on their priorities. In this method, different fourth sample sets are partitioned by the data provider based on the target threshold and feature values, each corresponding to a priority, thus allowing different priorities to correspond to different business indicator levels of the requester. Furthermore, the requester obtains priorities through anonymous queries without disclosing information about the users to be recommended, and also avoids obtaining the data provider's feature values, preventing privacy leaks for all parties. Moreover, prioritizing priorities allows for differentiated delivery of users to be recommended by combining the feature value distribution of the data provider with the requester's recommendation business needs, improving business quality.

[0166] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0167] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0168] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0169] In an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method is also provided.

[0170] Those skilled in the art will understand that various aspects of this disclosure can be implemented as a system, method, or program product. Therefore, various aspects of this disclosure can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software aspects, collectively referred to herein as a "circuit," "module," or "system."

[0171] The following reference Figure 8 To describe an electronic device 800 according to such an embodiment of the present disclosure. Figure 8 The electronic device 800 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0172] like Figure 8 As shown, the electronic device 800 is manifested in the form of a general-purpose computing device. The components of the electronic device 800 may include, but are not limited to: at least one processing unit 810, at least one storage unit 820, and a bus 830 connecting different system components (including storage unit 820 and processing unit 810).

[0173] The storage unit stores program code, which can be executed by the processing unit 810, causing the processing unit 810 to perform the steps described in the "Exemplary Methods" section above according to various exemplary embodiments of this disclosure.

[0174] Storage unit 820 may include readable media in the form of volatile storage units, such as random access memory (RAM) 8201 and / or cache 8202, and may further include read-only memory (ROM) 8203.

[0175] The storage unit 820 may also include a program / utility 8204 having a set (at least one) of program modules 8205, such program modules 8205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0176] Bus 830 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0177] Electronic device 800 can also communicate with one or more external devices (e.g., keyboard, pointing device, Bluetooth device, etc.), one or more devices that enable a user to interact with electronic device 800, and / or any device that enables electronic device 800 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed through display unit 840 and input / output (I / O) interface 850 connected to display unit 840. Furthermore, electronic device 800 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 860. As shown, network adapter 860 communicates with other modules of electronic device 800 via bus 830. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0178] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0179] In exemplary embodiments of this disclosure, a computer-readable medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible implementations, various aspects of this disclosure may also be implemented as a program product including program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure.

[0180] In embodiments of this disclosure, a program product for implementing the above-described methods is also provided. This product may be a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of this disclosure is not limited thereto. In this document, the readable medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0181] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0182] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0183] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0184] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0185] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0186] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

Claims

1. A multi-party data processing method, characterized in that, The method is applied to the requesting party, which forms a first sample set with a first number of first user samples. The method includes: Based on each initial threshold, an anonymous privacy intersection is initiated with the data provider to calculate the intersection potential of the first sample set and the second sample set corresponding to the initial threshold; the second sample set is obtained by the data provider after filtering the local third sample set according to the initial threshold and feature value; the third sample set includes second user samples, and each second user sample has a corresponding feature value; Based on the intersection potential and the number of first samples in the first sample set, the business indicator corresponding to the initial threshold is determined; the business indicator is determined by the ratio between the intersection potential and the number of first samples in the first sample set. Based on the business metrics, at least one target threshold is determined from the initial threshold; The target threshold is provided to the data provider so that the data provider divides the third sample set into at least two fourth sample sets based on the target threshold and the feature value. The first sample set includes a positive sample subset and a negative sample subset. The step of initiating anonymity and privacy intersection with the data provider based on each initial threshold, and calculating the intersection potential between the first sample set and the second sample set corresponding to the initial threshold, includes: Based on each initial threshold, two anonymous privacy intersections are initiated with the data party to calculate the positive sample intersection potential of the positive sample subset and the second sample set corresponding to the initial threshold, and to calculate the negative sample intersection potential of the negative sample subset and the second sample set corresponding to the initial threshold. The first sample count includes the number of positive samples and the number of negative samples. The step of determining the business metric corresponding to the initial threshold based on the intersection potential and the first sample count in the first sample set includes: Based on the intersection potential of the positive samples, the intersection potential of the negative samples, the number of positive samples in the first sample set, and the number of negative samples in the first sample set, the business indicators corresponding to the initial threshold are determined.

2. The multi-party data processing method according to claim 1, characterized in that, The step of initiating anonymity and privacy intersection with the data party based on each initial threshold, and calculating the intersection potential of the first sample set and the second sample set corresponding to the initial threshold, includes: Based on each of the initial thresholds, an anonymous privacy intersection is initiated with the data party to perform random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the first private key on the first sample set to obtain the first encryption result; Send the first encryption result to the data provider; Obtain the second encryption result provided by the data provider. The second encryption result is obtained by the data provider through random rearrangement, predetermined elliptic curve mapping, and encryption processing based on the second private key on the second sample set corresponding to the initial threshold. The second encryption result is randomly rearranged and encrypted based on the first private key to obtain the third encryption result; Obtain the fourth encryption result provided by the data provider, which is obtained by the data provider through random rearrangement of the first encryption result and encryption based on the second private key; By comparing the third encryption result with the fourth encryption result, the intersection potential of the first sample set and the second sample set is obtained.

3. The multi-party data processing method according to claim 1, characterized in that, The number of second samples in the third sample set of the second user samples is greater than the number of first samples. Before initiating anonymity and privacy intersection with the data party based on each initial threshold, and calculating the positive and negative sample intersection potentials of the first sample set and the second sample set corresponding to the initial threshold, the method further includes: Obtain the number of the second samples provided by the data provider; The sample counts of the first sample set and the second sample set are aligned based on the second sample count.

4. The multi-party data processing method according to claim 1, characterized in that, Before initiating anonymity and privacy intersection with the data party based on each initial threshold, and calculating the positive and negative sample intersection potentials of the first sample set and the second sample set corresponding to the initial threshold, the method further includes: Obtain the feature value range provided by the data provider; At least one initial threshold is determined based on the range of the feature values; The initial threshold is provided to the data provider.

5. The multi-party data processing method according to claim 1, characterized in that, The business metrics include at least one of hit rate, accuracy, and improvement.

6. The multi-party data processing method according to claim 1, characterized in that, The data parties include at least two; wherein each data party corresponds to the same feature type; or, at least one data party corresponds to a feature type that is different from the feature types corresponding to the other data parties.

7. A multi-party joint recommendation method, characterized in that, The method is applied to the requesting party, and the method includes: An anonymous query is initiated to the data provider for the user to be recommended, and the priority of the user to be recommended is determined; each priority corresponds to a fourth sample set divided by the data provider based on the target threshold and feature value after the data provider provides the target threshold to the data provider in any of the multi-party data processing methods described in any of the preceding claims 1 to 6. The recommended users are then targeted based on the aforementioned priority.

8. An electronic device, characterized in that, The electronic device acts as the requester, and the requester comprises a first sample set consisting of a first number of user samples. The electronic device includes: The privacy intersection module is used to initiate anonymity and privacy intersection with the data party based on each initial threshold, and calculate the intersection potential of the first sample set and the second sample set corresponding to the initial threshold; the second sample set is obtained by the data party after filtering the local third sample set according to the initial threshold and feature value; the third sample set includes second user samples, and each second user sample has a corresponding feature value; The indicator calculation module is used to determine the business indicator corresponding to the initial threshold based on the intersection potential and the number of first samples in the first sample set; the business indicator is determined by the ratio between the intersection potential and the number of first samples in the first sample set. The threshold determination module is used to determine at least one target threshold based on the business indicators from the initial threshold; The data sending module is used to provide the target threshold to the data party, so that the data party can divide the third sample set into at least two fourth sample sets based on the target threshold and the feature value; The first sample set includes a positive sample subset and a negative sample subset. The privacy intersection module is specifically used to initiate two anonymous privacy intersections with the data party based on each initial threshold, respectively calculating the positive sample intersection potential of the positive sample subset and the second sample set corresponding to the initial threshold, and calculating the negative sample intersection potential of the negative sample subset and the second sample set corresponding to the initial threshold. The first sample quantity includes the number of positive samples and the number of negative samples. The indicator calculation module is specifically used to determine the business indicator corresponding to the initial threshold based on the intersection potential of the positive samples, the intersection potential of the negative samples, the number of positive samples in the first sample set, and the number of negative samples in the first sample set.

9. An electronic device, characterized in that, The electronic device acts as the requester, and the electronic device includes: The query module is used to initiate an anonymous query to the data provider for the user to be recommended, and determine the priority of the user to be recommended; each priority corresponds to a fourth sample set divided by the data provider based on the target threshold and feature value after the electronic device of claim 8 provides the target threshold to the data provider. The delivery module is used to deliver recommendations to the users to be recommended based on the priority.

10. An electronic device, characterized in that, include: processor; Memory for storing the computer program of the processor; The processor is configured to execute the multi-party data processing method of any one of claims 1 to 6, or the multi-party joint recommendation method of claim 7, by executing the computer program.

11. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the multi-party data processing method as described in any one of claims 1 to 6, or the multi-party joint recommendation method as described in claim 7.

Images