Federal data correlation analysis method, data recommendation method and device

Abstract

The present disclosure relates to a federal data association analysis method and device, and a data recommendation method, and relates to the technical field of computers. The method is applied to a first participant in communication connection with at least one second participant. The first participant determines a first frequent item set and a support degree according to obtained sample data. The first participant receives a second frequent item set sent by the second participant, and the second frequent item set is obtained by the second participant according to sample data of the second participant. The confidence of the second frequent item set under the condition that the first frequent item set appears is determined, and an association rule is determined according to the confidence. At this time, the frequent item sets are exchanged between multiple participants while the corresponding support degrees are reserved, a federal association analysis based on privacy protection is realized, the analysis environment can be targeted to data composition diversity and complex association relationship, and without obtaining all source data, the data fragmentation between participants can be avoided, the data privacy security is ensured, the hardware configuration requirement is reduced, the cost is reduced, and the analysis efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular, to a federated data association analysis method, a data recommendation method and device. BACKGROUND

[0002] Association analysis, also known as association mining, is an analysis technique that obtains the association, correlation, causal structure or frequent pattern between item sets through data processing. In the scene of data recommendation, based on the association analysis technique, other items with association can be determined according to the items with which the user has generated interaction behavior, and the other items are recommended to the user, so as to achieve more accurate data recommendation.

[0003] At present, the association analysis is usually limited in the same field and platform, and it is difficult to perform accurate and efficient data association analysis in the case of data fragmentation. Moreover, the businesses in the same field and platform are similar, resulting in single data composition and simple relationship, which makes it difficult to fully represent the complex association state in the network environment, thereby limiting the practicality and accuracy of the association rules. Meanwhile, the association analysis is usually centralized processing, learning and analysis, which requires obtaining all the source data to be analyzed, and has high requirements for hardware configuration, resulting in high analysis cost and great security risks of data privacy.

[0004] It should be noted that the data disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include data that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The purpose of the present disclosure is to provide a federated data association analysis method, a data recommendation method and device, which can discover association rules among multiple fields and platforms, effectively improve the practicality and accuracy of the association rules in the case of diverse data composition and complex association relationship, without obtaining all the source data, ensuring data privacy security, reducing hardware configuration requirements, and reducing analysis cost.

[0006] According to a first aspect of the present disclosure, a federated data association analysis method is provided, which can be applied to a first participant, the first participant being in communication connection with at least one second participant, and the method can include: obtaining sample data; determining a first frequent item set in the sample data and a support degree corresponding to the first frequent item set; receiving a second frequent item set sent by the second participant, the second frequent item set being obtained by the second participant according to sample data of the second participant; determining a confidence degree of occurrence of the second frequent item set in the case of occurrence of the first frequent item set, and determining an association rule according to the confidence degree of occurrence of the second frequent item set.

[0007] Optionally, when the confidence and the federated frequent itemset meet the condition of the association rule, the association rule is determined to include the second frequent itemset appearing in the case of the first frequent itemset appearing in the federated frequent itemset.

[0008] Optionally, when the confidence and the federated frequent itemset meet the condition of the association rule, the association rule is determined to include the second frequent itemset appearing in the case of the first frequent itemset appearing in the federated frequent itemset, and further includes: when the federated frequent itemset includes the combination of each first frequent itemset and second frequent itemset, outputting the association rule.

[0009] Optionally, the condition of the association rule is that the confidence is greater than a confidence threshold, and the length of the federated frequent itemset is less than an itemset length threshold.

[0010] Optionally, the first participant is further communicatively connected with an intermediate party, and the intermediate party is communicatively connected with the second participant, and the federated support corresponding to the federated frequent itemset is determined by: based on the federated frequent itemset, performing secure multi-party computation between the intermediate party and the second participant to determine the federated support corresponding to the federated frequent itemset.

[0011] Optionally, the sample data is obtained by: obtaining original data; and performing column name mapping on the original data to obtain the sample data.

[0012] Optionally, after the first frequent itemset and the support corresponding to the first frequent itemset are determined in the sample data, the method further includes: sending the first frequent itemset to the second participant, so that the second participant determines the confidence of the first frequent itemset appearing in the case of the second frequent itemset appearing, and determines the association rule according to the confidence of the first frequent itemset appearing, and the association rule of the first participant is asymmetric to the association rule of the second participant.

[0013] According to a second aspect of the present disclosure, a data recommendation method is provided, which can include: obtaining historical behavior data of a user; extracting original itemsets from the historical behavior data; matching a recommended itemset in an association rule based on the original itemsets, the association rule being obtained by the federated data association analysis method of the first aspect; and performing data recommendation to the user according to the recommended itemset.

[0014] According to a third aspect of the present disclosure, a federated data association analysis apparatus is provided, which is applied to a first participant, the first participant being communicatively connected with at least one second participant, and the apparatus can include: a sample data acquisition module configured to acquire sample data; a frequent item set determination module configured to determine a first frequent item set in the sample data and a support degree corresponding to the first frequent item set; a frequent item set exchange module configured to receive a second frequent item set sent by the second participant, the second frequent item set being obtained by the second participant based on sample data of the second participant; and an association rule determination module configured to determine a confidence degree of occurrence of the second frequent item set in a case where the first frequent item set occurs, and determine an association rule based on the confidence degree of occurrence of the second frequent item set.

[0015] Optionally, the association rule determination module includes: a federated frequent item set combination sub-module configured to acquire at least one federated frequent item set, each federated frequent item set being formed by combination of one first frequent item set and one second frequent item set, the first frequent item sets and the second frequent item sets being different between the federated frequent item sets; a federated support degree determination sub-module configured to determine a federated support degree corresponding to the federated frequent item set; a confidence degree determination sub-module configured to, in a case where the federated support degree is greater than a federated support degree threshold, calculate, based on the support degree and the federated support degree, the confidence degree of occurrence of the second frequent item set in a case where the first frequent item set in the federated frequent item set occurs; and an association rule determination sub-module configured to, in a case where the confidence degree and the federated frequent item set meet an association rule condition, determine that the association rule includes occurrence of the second frequent item set in a case where the first frequent item set in the federated frequent item set occurs.

[0016] Optionally, the apparatus further includes an association rule output module configured to output the association rule in a case where the federated frequent item set includes a combination of each first frequent item set and second frequent item set.

[0017] Optionally, the association rule condition is that the confidence degree is greater than a confidence degree threshold and a length of the federated frequent item set is less than an item set length threshold.

[0018] Optionally, the first participant is further communicatively connected with an intermediate party, the intermediate party being communicatively connected with the second participant, and the federated support degree determination sub-module is specifically configured to determine the federated support degree corresponding to the federated frequent item set based on the federated frequent item set by performing secure multi-party computation between the intermediate party and the second participant.

[0019] Optionally, the sample data acquisition module includes: an original data acquisition sub-module configured to acquire original data; and an original data desensitization sub-module configured to perform column name mapping on the original data to obtain the sample data.

[0020] Optionally, the frequent item set exchange module is further configured to send the first frequent item set to the second participant, so that the second participant determines a confidence level of appearance of the first frequent item set in a case where the second frequent item set appears, and determines the association rule according to the confidence level of appearance of the first frequent item set, the association rule of the first participant being asymmetric to the association rule of the second participant.

[0021] According to a fourth aspect of the present disclosure, a data recommendation device is provided, which can include: a historical data acquisition module configured to acquire historical behavior data of a user; an original item set extraction module configured to extract an original item set from the historical behavior data; an association rule matching module configured to match a recommendation item set in an association rule based on the original item set, the association rule being obtained by the federated data association analysis device of the third aspect; and a user data recommendation module configured to perform data recommendation to the user according to the recommendation item set.

[0022] According to a fifth aspect of the present disclosure, a computer readable storage medium is provided, which stores a computer program, the computer program being executed by a processor to implement the federated data association analysis method of the first aspect or the data recommendation method of the second aspect.

[0023] According to a sixth aspect of the present disclosure, an electronic device is provided, which includes:

[0024] a processor; and

[0025] a memory configured to store a computer program of the processor;

[0026] wherein the processor is configured to implement the federated data association analysis method of the first aspect or the data recommendation method of the second aspect by executing the computer program.

[0027] The federated data association analysis method provided by the present disclosure is applied to a first participant, and the first participant is in communication connection with at least one second participant. After the first participant acquires sample data, the first participant determines a first frequent item set in the sample data and a support degree corresponding to the first frequent item set. Then, the first participant receives a second frequent item set sent by the second participant, the second frequent item set being obtained by the second participant according to sample data of the second participant. The first participant determines a confidence level of appearance of the second frequent item set in a case where the first frequent item set appears, and determines an association rule according to the confidence level. At this time, the first participant only receives frequent item sets sent by other participants, without exchanging the corresponding support degrees, thereby realizing federated association analysis based on privacy protection among multiple participants, obtaining high practicability and accuracy of the association rule under the condition of diverse data composition and complex association relationship, breaking the state of data fragmentation among the participants without acquiring all source data, ensuring data privacy security, reducing the requirement of hardware configuration on the processing procedure of the association analysis, thereby reducing the analysis cost and improving the analysis efficiency.

[0028] It is to be understood that the foregoing general description and the following detailed description are only exemplary and explanatory and are not restrictive of the disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0029] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and serve to explain the principles of the disclosure. It is apparent that the accompanying drawings for use in the following description are only some embodiments of the present disclosure, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0030] Figure 1 One of the step flow charts of the federal data correlation analysis method provided for the embodiments of the present disclosure,

[0031] Figure 2 The second step flow chart of the federal data correlation analysis method provided for the embodiments of the present disclosure.

[0032] Figure 3 The architecture flow schematic diagram of a federal data correlation analysis method provided for the embodiments of the present disclosure.

[0033] Figure 4 The step flow chart of a data recommendation method provided for the embodiments of the present disclosure.

[0034] Figure 5 The structural schematic diagram of a federal data correlation analysis device provided for the embodiments of the present disclosure.

[0035] Figure 6 The structural schematic diagram of a data recommendation device provided for the embodiments of the present disclosure.

[0036] Figure 7 The structural schematic diagram of an electronic device according to an embodiment of the present disclosure is illustrated. DETAILED DESCRIPTION

[0037] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations can be implemented in any of various forms, and are not necessarily limited to the examples described herein; rather, examples are provided so that this disclosure will be thorough and complete, and fully convey the scope of example implementations to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more implementations. In the following description, numerous specific details are recited to provide a thorough understanding of implementations of the disclosure. One skilled in the art, however, will recognize that the implementations of the present disclosure can be practiced without one or more of the specific details, or with other methods, components, materials, and so forth. In other instances, well-known structures have not been described in detail so as not to obscure the aspects of the disclosure.

[0038] In addition, the accompanying drawings are included to provide a thorough understanding of the present disclosure and are not intended to be in any way limiting. Like reference numerals in the drawings represent corresponding parts throughout the text in which: the same reference numerals indicate the same or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities, which do not necessarily have to correspond to physically or logically independent entities. These functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0039] It should be noted that the data obtained by the present disclosure, including raw data, sample data, frequent item sets, support, and other data, are all accessed, collected, stored, and applied to subsequent analysis and processing with the consent and authorization of the user or the relevant data owner after informing the user or the relevant data owner of the data collection content, data use, processing method, and other information, and the user or the relevant data owner can be sent the way to access, correct, delete the data, and the method to revoke the consent and authorization.

[0040] Figure 1 A step flowchart of a federated data association analysis method provided by an embodiment of the present disclosure is shown in FIG. 1. The method is applied to a first participant, which is in communication connection with at least one second participant.

[0041] The method can analyze association rules among two or more participants. The participants interact with each other through communication connection to jointly analyze the association rules of their respective local data, thereby obtaining the association rules among multi-source, multi-platform, and multi-field data. In the embodiment of the present disclosure, any participant can be taken as the first participant, and the step flow of the federated data association analysis method is described based on the execution flow of the first participant. At this time, the participants other than the first participant can be taken as the second participant.

[0042] As shown in Figure 1 The method can include steps 101-104. As shown below:

[0043] Step 101, obtaining sample data.

[0044] The sample data can be user behavior data in the first participant to be analyzed for correlation. Depending on the business field corresponding to the first participant and the requirements of the correlation analysis, the types of sample data can be different. For example, if the business field corresponding to the first participant is commodity transaction, the sample data can include user behavior data such as purchase, evaluation, collection or recommendation of commodities; if the business field corresponding to the first participant is information, the sample data can be user behavior data such as browsing, commenting or sharing of information; if the business field corresponding to the first participant is a social platform, the sample data can include user behavior data such as attention, shielding or interaction of social accounts. Those skilled in the art can obtain sample data corresponding to the first participant according to actual requirements and collection conditions, and the present disclosure does not make specific limitations on this.

[0045] Step 102, determining a first frequent item set in the sample data and a support degree corresponding to the first frequent item set.

[0046] The item set refers to a set of items appearing simultaneously in the sample data, the item can be the object of the behavior data, the support degree represents the frequency of the item set appearing in the sample data, and the frequent item set refers to an item set with a support degree greater than or equal to a minimum support degree. For example, taking the sample data as commodity purchase behavior data, item set 1 includes commodity a, which means only purchasing commodity a, and item set 2 includes commodity a and commodity b, which means simultaneously purchasing commodity a and commodity b. At this time, the support degree of item set 1 can be the proportion of the purchase times of commodity a in all commodity purchase times. Further, in the case where the support degree is greater than or equal to the minimum support degree, item set 1 can be taken as the first frequent item set in the sample data. On this basis, other sample data such as information browsing behavior data and account attention behavior data can be similarly determined to obtain corresponding first frequent item sets.

[0047] Step 103, receiving a second frequent item set sent by the second participant, the second frequent item set being obtained by the second participant according to sample data of the second participant.

[0048] In the multi-participant, each participant can determine a local frequent item set and a corresponding support degree. After the first participant determines a first frequent item set corresponding to the sample data, the first participant can receive a second frequent item set sent by the second participant, but does not receive the support degree corresponding to the second frequent item set, so that the partial exchange of the sample data is realized between the first participant and the second participant. Further, the first participant can also send the first frequent item set to the second participant, but does not send the support degree corresponding to the first frequent item set to the second participant, so that each participant can obtain the frequent item set of the other participant, but cannot obtain the support degree corresponding to the frequent item set.

[0049] In the embodiments of the present disclosure, the process of obtaining the second frequent item set by the second participant can correspond to the process of obtaining the first frequent item set by the first participant in the foregoing steps 102 to 103. In the processing process, each participant processes the local sample data in a consistent manner. It should be noted that in the federated association analysis, the frequent item sets provided between the participants can be a collection of behaviors of different users on the same object, such as the item sets of goods purchased by each participant. By expanding the quality and quantity of sample data of similar features, the accuracy of the association analysis can be improved. The frequent item sets provided between the participants can also be a collection of behaviors of the same user on different objects, such as the item set of goods purchased by the user provided by the first participant and the account attention item set of the user provided by the second participant. By combining sample data of different features, the analysis demand of complex association relationship can be met.

[0050] In step 104, the confidence degree of the occurrence of the second frequent item set when the first frequent item set occurs is determined, and the association rule is determined according to the confidence degree of the occurrence of the second frequent item set.

[0051] The confidence degree represents the conditional probability that the second frequent item set occurs when the first frequent item set occurs. The confidence degree can be obtained by dividing the probability that the first frequent item set and the second frequent item set occur simultaneously by the probability that the first frequent item set occurs, and can represent the association degree of the first frequent item set and the second frequent item set. For example, the first frequent item set is extracted based on the goods purchase behavior data and includes {goods a, goods b}, and the second frequent item set is extracted based on the account attention behavior data and includes {account 1}. The confidence degree represents the probability that the user has purchased goods a and goods b and has followed account 1, which can be obtained by dividing the probability that the user has purchased goods a and goods b by the probability that the user has followed account 1. At this time, the confidence degree represents the probability that the user has followed account 1 because of purchasing goods a and goods b.

[0052] In the embodiments of the present disclosure, the association degree of the first frequent item set and the second frequent item set can be determined based on the confidence degree, so that whether the association rule that the second frequent item set occurs when the first frequent item set occurs is established is determined according to the confidence degree, to obtain the corresponding association rule.

[0053] Figure 2 FIG. 2 is a flowchart of a method for federal data association analysis according to an embodiment of the present disclosure, which is applied to a first participant in communication connection with at least one second participant.

[0054] The method can include steps 201-209. As shown below:

[0055] Step 201: Obtain original data.

[0056] The original data can be behavior data directly generated by a user in a business, such as behavior data including the name of a commodity, a user name, a transaction amount, a transaction time, a transaction address, and the like in a commodity purchase behavior, and behavior data including a follower account, a followed account, a following time, and a following frequency, and the like in an account following behavior.

[0057] For example, the first participant obtains original data X A .

[0058] Step 202: Perform column name mapping on the original data to obtain sample data.

[0059] In an embodiment of the present disclosure, the original behavior data can include a large amount of sensitive data in different formats, which can be processed to obtain sample data meeting subsequent processing requirements, such as data cleaning, format unification, and the like. Further, the data can be desensitized, such as mapping the column name of the original data, and obtaining sample data based on the mapping relationship. At this time, only the first participant saves the mapping relationship, thereby ensuring the security of the original data and improving the security protection capability for data privacy.

[0060] For example, the first participant performs column name mapping on the original data X A to obtain sample data (x1, x2, …, x A ).

[0061] Step 203: Determine a first frequent item set in the sample data and a support degree corresponding to the first frequent item set.

[0062] Step 203 can correspond to the description of step 102, and thus will not be repeated here.

[0063] In a method embodiment of the present disclosure, the first participant can employ the FP-growth (Frequent Pattern-growth) algorithm to find the first frequent itemsets in the sample data and determine the support corresponding to the first frequent itemsets. Through the FP-growth algorithm, the first participant can fill the sample data into the FP-tree according to the support of each element item, and find the conditional FP-trees in the FP-tree, each of which includes a set of element items that commonly appear together in the FP-tree, i.e., the first frequent itemsets. Those skilled in the art can also select other methods for finding the first frequent itemsets in the sample data according to the operation conditions and processing requirements, such as the Apriori algorithm, and the present disclosure does not make specific limitations in this regard.

[0064] For example, the first participant finds the first frequent itemsets A_frequent_itemsets=(‘x1’,‘x2’): support=0.5, (‘x1’): support=0.6,··· in the sample data based on the FP-growth algorithm, where frequent_itemsets is the frequent itemsets and support is the support.

[0065] It should be noted that the second participant can also obtain the second frequent itemsets and their support according to the local sample data in this process, such as the second participant finds the second frequent itemsets B_frequent_itemsets=(‘y1’,‘y2’): support=0.43, (‘y1’): support=0.5,··· in the sample data based on the FP-growth algorithm.

[0066] Step 204, receiving the second frequent itemsets sent by the second participant, which are obtained by the second participant according to the sample data of the second participant.

[0067] Wherein, step 204 can correspond to the relevant description of the aforementioned step 103, and to avoid repetition, it will not be described here.

[0068] For example, the first participant can send the first frequent itemsets A to the second participant, including [(‘x1’,‘x2’), (‘x1’),···]; and receive the second frequent itemsets B sent by the second participant, including [(‘y1’,‘y2’), (‘y1’), (‘y3’),···]. In this process, the participants exchange the frequent itemsets, but keep the support corresponding to their own frequent itemsets, thereby ensuring the security and privacy of the complete data.

[0069] In step 205, at least one federal frequent item set is obtained, each federal frequent item set being formed by combining a first frequent item set and a second frequent item set, at least one of the first frequent item sets and the second frequent item sets being different between the federal frequent item sets.

[0070] In the method, the federal frequent item sets can be obtained based on the first frequent item sets and the second frequent item sets, each federal frequent item set being formed by combining a first frequent item set and a second frequent item set, at least one of the first frequent item sets and the second frequent item sets being different between the federal frequent item sets. In this case, the federal frequent item sets including both the first frequent item sets and the second frequent item sets are obtained by combining the randomly selected first frequent item sets and the randomly selected second frequent item sets. Therefore, in subsequent processing, the association between the first frequent item sets and the second frequent item sets can be analyzed based on the federal frequent item sets.

[0071] For example, the first frequent item set f t1 =(x1, x2) is provided by the first participant, and the second frequent item set f t2 =(y1) is provided by the second participant, and the federal frequent item set (x1, x2, y1) is obtained by merging.

[0072] In step 206, the federal support degree corresponding to the federal frequent item set is determined.

[0073] In the method, after the federal frequent item set is obtained, the federal support degree of the federal frequent item set can be determined between the first participant and the second participant, which is the probability of simultaneous occurrence of the first frequent item set and the second frequent item set.

[0074] For example, the federal support degree sup of the federal frequent item set (x1, x2, y1) is determined, which represents the probability of simultaneous occurrence of the first frequent item set f t1 =(x1, x2) and the second frequent item set f t2 =(y1).

[0075] In a method embodiment of the disclosure, the first participant is also in communication connection with an intermediate party, and the intermediate party is in communication connection with the second participant. Then, step 206 specifically includes: based on the federal frequent item set, the federal support degree corresponding to the federal frequent item set is determined by performing secure multi-party computation between the intermediate party and the second participant.

[0076] The calculation of the federal support can introduce an intermediate party implementation, the first participant is in communication connection with the intermediate party, and the intermediate party is in communication connection with the second participant. The intermediate party can be used for controlling network transmission, routing addressing, and calculation logic transmission, etc. The first participant and the second participant can perform collaborative calculation based on the federal frequent item set, and obtain the federal support corresponding to the federal frequent item set respectively. Specifically, the collaborative calculation can be initiated by the first participant and the second participant, and the calculation logic of the federal frequent item set is transmitted through the intermediate party. The first participant and the second participant respectively query the corresponding data locally, such as the support corresponding to the first frequent item set and the support corresponding to the second frequent item set, and perform collaborative calculation between the data streams in the secret state based on the data of each participant. In the calculation process, secret sharing, careless transmission, and confusion circuit and other related transmission, encryption and calculation protocol methods can be used to ensure the privacy and security of the data. Finally, in the case of ensuring the data privacy, the intermediate party can send the calculated federal support to the first participant and the second participant, so that each participant obtains the required and correct calculation result.

[0077] In a method embodiment of the present disclosure, the first participant and the second participant can also perform secure multi-party calculation end to end without the intermediate party, and the present disclosure embodiment does not make specific limitation on this.

[0078] In step 207, in the case that the federal support is greater than the federal support threshold, the confidence of the appearance of the second frequent item set in the case of the appearance of the first frequent item set in the federal frequent item set is calculated according to the support and the federal support.

[0079] The first participant can further determine whether the federal support is greater than the federal support threshold after obtaining the federal support. The federal support is the probability of the appearance of the federal frequent item set, and the federal support threshold can be a probability evaluation threshold of the federal frequent item set. When the federal support is less than the federal support threshold, it can be considered that the probability of the simultaneous appearance of the first frequent item set and the second frequent item set is too low, and the possibility of the association between the two frequent item sets is low, and then the generation of the subsequent association rule can be not performed. Conversely, when the federal support is greater than the federal support threshold, the confidence can be further calculated according to the support of the first frequent item set and the federal support of the federal frequent item set. The confidence can represent the probability of the appearance of the second frequent item set in the case of the appearance of the first frequent item set in the federal frequent item set.

[0080] For example, the first participant calculates the ratio of the federal support to the support corresponding to the first frequent item set in the case that the sup is greater than the min_support, and obtains the confidence of (′x1′,′x2′)->(′y1′).

[0081] It should be noted that after the second participant obtains the federated support, the second participant can also calculate the ratio of the federated support to the support corresponding to the second frequent itemset, and obtain the confidence of (′y1′)->(′x1′,′x2′), provided that sup is greater than min_support.

[0082] Step 208: If the confidence level and the federated frequent itemset meet the association rule conditions, determine that the association rule includes the occurrence of the second frequent itemset when the first frequent itemset appears in the federated frequent itemset.

[0083] In the first participant, the association rule condition can refer to the condition that makes "given the occurrence of the first frequent itemset, the occurrence of the second frequent itemset" meet the rule generation requirements. These requirements can be determined based on the analysis conditions and practical applications of the association rule. Since confidence level can characterize the probability of the second frequent itemset occurring given the occurrence of the first frequent itemset in the federated frequent itemset, and the federated frequent itemset includes all elements of both the first and second frequent itemsets, this rule generation requirement can be based on the definitions of confidence level and federated frequent itemsets to ensure that the probability of occurrence and specific content of the association rule meet the application requirements.

[0084] In one embodiment of the method disclosed herein, the association rule condition is that the confidence level is greater than the confidence level threshold and the length of the federated frequent itemset is less than the itemset length threshold.

[0085] The association rule conditions can be that the confidence level is greater than a confidence threshold and the length of the federated frequent itemset is less than an itemset length threshold. The confidence threshold can be a threshold for assessing the probability of "the second frequent itemset appearing when the first frequent itemset appears." If the confidence level is less than the confidence threshold, the probability of occurrence is considered too low and does not meet the rule generation requirements. The itemset length threshold can be a threshold for assessing the length and number of items contained in the federated frequent itemset. If the length of the items in the federated frequent itemset is too long or the number of items is too large, the length of the federated frequent itemset may exceed the itemset length threshold. In this case, the federated frequent itemset can be considered not to meet the rule generation requirements. Therefore, if the confidence level is greater than the confidence threshold and the length of the federated frequent itemset is less than the itemset length threshold, it can be determined that the association rule for the first participant includes "the second frequent itemset appearing when the first frequent itemset appears."

[0086] For example, in the first participant, if the confidence of (′x1′,′x2′)->(′y1′) is greater than min_confidence, and len((′x1′,′x2′,′y1′))<=max_len, then the association rule of the first participant includes (′x1′,′x2′)->(′y1′).

[0087] In a method embodiment of the present disclosure, after step 203, step 209 is further included. As shown below:

[0088] Step 209, sending the first frequent item set to the second participant, so that the second participant determines the confidence of the first frequent item set appearing in the case of the second frequent item set appearing, and determines the association rule based on the confidence, the association rule of the first participant and the association rule of the second participant are asymmetric.

[0089] It should be noted that the embodiments of the present disclosure take the first participant as the main body to explain the step flow of the federated data association analysis method. However, in each participant, the step flow performed by each participant can correspond to the relevant description of the foregoing steps 101 to 104 or steps 201 to 208.

[0090] Among them, after obtaining the first frequent item set, the first participant can send the first frequent item set to the second participant. Similarly, the second participant can obtain the second frequent item set and the support degree corresponding to the second frequent item set according to its sample data, and then provide the second frequent item set to the first participant and receive the first frequent item set provided by the first participant; Then the second participant can determine the confidence of the first frequent item set appearing in the case of the second frequent item set appearing, and determine whether the association rule required by the second participant is established based on the confidence of the first frequent item set appearing.

[0091] For example, the first frequent item set is extracted based on commodity purchase behavior data, including {commodity a, commodity b}, and the second frequent item set is extracted based on account attention behavior data, including {account 1}. The confidence of the second participant indicates the probability that the user purchases commodity a and commodity b because of following account 1. It can be seen that the association rules of the first participant and the second participant are asymmetric. For example, in the second participant, the association rule condition can refer to the condition that makes "in the case of the second frequent item set appearing, the first frequent item set appearing" meet the rule generation requirement. Referring to the related description of the first participant, in the second participant, the confidence of (′y1′)->(′x1′,′x2′) can be calculated in the case of the federated support being greater than the federated support threshold, and the association rule of the second participant includes (′y1′)->(′x1′,′x2′) in the case of the confidence of (′y1′)->(′x1′,′x2′) being greater than min_confidence, and len((′x1′,′x2′,′y1′))<=max_len. Wherein, the federated support threshold, the confidence threshold and the item set length threshold adopted by the first participant and the second participant can be the same or different, in order to adapt to the association analysis requirements of different fields and different platforms.

[0092] Step 210: output the association rule when the federal frequent item set includes each combination of the first frequent item set and the second frequent item set.

[0093] The federal frequent item set can include different combinations of the first frequent item set and the second frequent item set. When the federal frequent item set includes each combination of the first frequent item set and the second frequent item set, it can be determined that the evaluation of each combination of the first frequent item set and the second frequent item set has been completed in the generation of the association rule, and all association relationships that meet the association rule condition based on the first frequent item set and the second frequent item set have been included in the association rule. At this time, the association rule can be output to end the data association analysis process.

[0094] In an embodiment of the method of the present disclosure, when the federal frequent item set does not include each combination of the first frequent item set and the second frequent item set, steps 205 to 208 can be executed in a loop, and each time the loop is executed, one federal frequent item set is calculated and evaluated. The federal frequent item set processed in different loops is different, so that the loop ends when the federal frequent item set includes each combination of the first frequent item set and the second frequent item set, and the output association rule is obtained.

[0095] The federal data association analysis method provided by the present disclosure is applied to a first participant, and the first participant is in communication connection with at least one second participant. After the first participant obtains sample data, the first participant determines a first frequent item set in the sample data and a support degree corresponding to the first frequent item set. Then, the first participant receives a second frequent item set sent by the second participant, and the second frequent item set is obtained by the second participant according to sample data of the second participant. The first participant determines a confidence degree of occurrence of the second frequent item set under the condition of occurrence of the first frequent item set, and determines an association rule according to the confidence degree. At this time, the first participant only receives the frequent item set sent by other participants, and does not exchange the corresponding support degree, thereby realizing the federal association analysis based on privacy protection among multiple participants. The federal association analysis can obtain high practicality and accuracy of the association rule under the condition that the data is diverse and the association relationship is complex. In addition, the federal association analysis can break the data fragmentation among the participants without obtaining all the source data, thereby ensuring the data privacy and security, reducing the requirement of the hardware configuration in the processing process of the association analysis, and thus reducing the analysis cost and improving the analysis efficiency.

[0096] Figure 3 An architecture flow diagram of a federal data association analysis method provided by an embodiment of the present disclosure is provided. The architecture includes a first participant A, a second participant B, and an intermediate party C. The first participant A is configured with a support degree threshold 1, a confidence degree threshold 1, and an item set length threshold 1. The second participant B is configured with a support degree threshold 2, a confidence degree threshold 2, and an item set length threshold 2.

[0097] It should be noted that the support threshold 1, confidence threshold 1, and itemset length threshold 1, or support threshold 2, confidence threshold 2, and itemset length threshold 2, are only used to indicate that the objects configured for the thresholds are different, and do not restrict the specific values ​​of each threshold. In practical applications, the values ​​of each threshold can be the same or different between the first participant and different second participants, or between different second participants, depending on the application requirements.

[0098] like Figure 3 As shown, this federal data correlation analysis method includes steps 301A to 306A, and steps 301B to 306B. See below:

[0099] Step 301A: The first participant A determines the first frequent itemset in the sample data and the support corresponding to the first frequent itemset based on the FP-growth algorithm.

[0100] Step 301B: The second participant B determines the local second frequent itemset and the support corresponding to the second frequent itemset in the sample data based on the FP-growth algorithm.

[0101] Step 302A: First participant A sends a first frequent itemset to second participant B and receives a second frequent itemset sent by second participant B.

[0102] Step 302B: The second participant B sends the second frequent itemset to the first participant A and receives the first frequent itemset sent by the first participant A.

[0103] Step 303: First participant A provides the first frequent itemset, and second participant B provides the second frequent itemset to obtain the federated frequent itemset. Secure multi-party computation is then performed through intermediary C to obtain the federated support of the federated frequent itemset.

[0104] Step 304: Intermediate party C sends federation support to first participant A and to second participant B.

[0105] Step 305A: When the federated support is greater than the federated support threshold 1, the first participant A calculates the confidence that the second frequent itemset will appear when the first frequent itemset appears, based on the federated support and the support of the first frequent itemset. When the confidence is greater than the confidence threshold 1 and the length of the federated frequent itemset is less than the itemset length threshold 1, the association rule is determined to include the occurrence of the second frequent itemset when the first frequent itemset appears.

[0106] Step 305B, when the federation support is greater than the federation support threshold 2, the second participant B calculates the confidence of the appearance of the first frequent item set in the case of the appearance of the second frequent item set according to the federation support and the support of the second frequent item set, and when the confidence is greater than the confidence threshold 2 and the length of the federated frequent item set is less than the item set length threshold 2, determines that the association rule includes the appearance of the first frequent item set in the case of the appearance of the second frequent item set.

[0107] Step 306A, when the federated frequent item set includes each combination of the first frequent item set and the second frequent item set, the first participant A outputs the association rule; or when the federated frequent item set does not include each combination of the first frequent item set and the second frequent item set, the steps 303 to 305A are executed in a loop.

[0108] Step 306B, when the federated frequent item set includes each combination of the first frequent item set and the second frequent item set, the second participant B outputs the association rule; or when the federated frequent item set does not include each combination of the first frequent item set and the second frequent item set, the steps 303 to 305B are executed in a loop.

[0109] The architecture flow of the federated data association analysis method provided by the present disclosure is applied to a first participant, a second participant and an intermediate party, and the first participant, the second participant and the intermediate party are in communication connection. Wherein, each participant in the first participant and the second participant extracts frequent item sets and corresponding supports from sample data locally, and exchanges the frequent item sets respectively, and then performs secure multi-party computation among the multiple participants through the intermediate party to obtain the federation support of the federated frequent item set, and the federated frequent item set includes a frequent item set of each participant, and then each participant can determine the association of the federated item set among the participants based on the federation support of the federated frequent item set and the support of the local frequent item set of each participant. At this time, the frequent item sets are exchanged among the multiple participants, and the corresponding supports are not exchanged, thereby realizing the federated association analysis based on privacy protection among the multiple participants, obtaining high practicability and accuracy of the association rules under the condition of diverse data composition and complex association relationship, breaking the data fragmentation state among the participants without obtaining all the source data, ensuring data privacy and security, and reducing the requirement of the processing process of the association analysis on the hardware configuration, thereby reducing the analysis cost and improving the analysis efficiency.

[0110] Figure 4 A step flowchart of a data recommendation method provided by an embodiment of the present disclosure, which can include steps 401 to 404. As follows:

[0111] Step 401, obtaining historical behavior data of a user.

[0112] Wherein, the user is the target of data recommendation, and can be an individual, an institution or an enterprise, etc., and the historical behavior data can be historical interaction data of the user in the business, such as historical purchase data, historical collection data, historical evaluation data of commodities of the user on a shopping platform, etc., or historical browsing data, historical sharing data, historical collection data of information of the user on an information platform, etc., or historical attention data, historical shielding data, historical comment data of accounts of the user on a social platform, etc. Different historical behavior data of the user can be obtained according to the user type, the recommendation demand, etc., and the embodiments of the present disclosure do not make specific limitation on this.

[0113] Step 402, extracting an original item set from the historical behavior data.

[0114] Wherein, the original item set can be a collection of element items in the historical behavior data, and the element item can refer to an entity of user behavior interaction, such as a commodity, information or an account, etc. Taking the historical purchase data of commodities of the user on a shopping platform as an example, the original item set (daily commodity a, daily commodity b) can be extracted from a piece of historical purchase data of commodities "purchased daily commodities a and b on X month X day". Taking the account attention data of the user on a social platform as an example, the original item set (account 1) can be extracted from a piece of data "attended account 1 on X month X day".

[0115] Step 403, matching a recommendation item set in the association rule based on the original item set, and the association rule is described based on Figures 1 to 3 Any federal data association analysis method.

[0116] Wherein, the association relationship between the occurrence of the first frequent item set and the occurrence of the second frequent item set is described based on the association rule, and the association rule can be matched based on the original item set, and the recommendation item set that can occur in the case of the occurrence of the original item set can be obtained. For example, the original item set is (daily commodity a, daily commodity b), and the corresponding association rule in the case of purchasing daily commodity a and daily commodity b can be matched, such as association rule 1 is that in the case of purchasing daily commodity a and daily commodity b, electronic product commodity c is purchased; and association rule 2 is that in the case of purchasing daily commodity a and daily commodity b, account 1 is attended. On this basis, the recommendation item set matched by the user based on the original item set can be (electronic product commodity c), (account 1), etc.

[0117] Step 404, data recommendation is performed to the user according to the recommendation item set.

[0118] The data recommendation to the user according to the recommended item set can be directly pushing the recommended item set to the user, or recommending different recommended item sets to the user based on the business type of the platform, the field and the like, such as recommending (an electronic product commodity c) to the user on a shopping platform, recommending (an account 1) to the user on a social platform, and providing differentiated data recommendation to the user based on the association rule.

[0119] The data recommendation method provided by the present disclosure can obtain historical behavior data of a user, and extract an original item set from the historical behavior data; then, a recommended item set is matched in an association rule based on the original item set, and data recommendation is performed to the user according to the recommended item set, wherein the association rule is obtained by the foregoing Figures 1 to 3 The federal data association analysis method can obtain any of the foregoing. Based on the foregoing federal data association analysis method, the association rule is extracted from more diversified data features and a wider sample size, and then the user can be provided with differentiated, cross-platform and cross-field data recommendation more accurately, thereby improving the accuracy and practicability of data recommendation.

[0120] Figure 5 A structural schematic diagram of a federal data association analysis device provided by an embodiment of the present disclosure is provided. The device is applied to a first participant, the first participant is in communication connection with at least one second participant, and the device can include:

[0121] The sample data acquisition module 501 is configured to acquire sample data.

[0122] The frequent item set determination module 502 is configured to determine a first frequent item set and a support degree corresponding to the first frequent item set in the sample data.

[0123] The frequent item set exchange module 503 is configured to receive a second frequent item set sent by the second participant, the second frequent item set being obtained by the second participant based on sample data of the second participant.

[0124] The association rule determination module 504 is configured to determine a confidence degree of the second frequent item set appearing under the condition that the first frequent item set appears, and determine an association rule according to the confidence degree of the second frequent item set appearing.

[0125] In a device embodiment of the present disclosure, the association rule determination module 504 includes:

[0126] The federal frequent item set combination submodule is configured to acquire at least one federal frequent item set, each federal frequent item set being formed by combining one first frequent item set and one second frequent item set, and the first frequent item set and the second frequent item set being different between the federal frequent item sets at least in one.

[0127] The federal support degree determination submodule is configured to determine a federal support degree corresponding to the federal frequent item set.

[0128] a confidence degree determination submodule, configured to, in a case where the federal support degree is greater than the federal support degree threshold, calculate, according to the support degree and the federal support degree, a confidence degree of appearance of the second frequent item set in a case where the first frequent item set in the federal frequent item set appears;

[0129] an association rule determination submodule, configured to, in a case where the confidence degree and the federal frequent item set meet an association rule condition, determine that the association rule includes appearance of the second frequent item set in a case where the first frequent item set in the federal frequent item set appears.

[0130] In an apparatus embodiment of the present disclosure, the apparatus further includes an association rule output module, configured to, in a case where the federal frequent item set includes a combination of each first frequent item set and second frequent item set, output the association rule.

[0131] In an apparatus embodiment of the present disclosure, the association rule condition is that the confidence degree is greater than a confidence degree threshold, and the length of the federal frequent item set is less than an item set length threshold.

[0132] In an apparatus embodiment of the present disclosure, the first participant is further communicatively connected with an intermediate party, the intermediate party is communicatively connected with the second participant, and the federal support degree determination submodule is specifically configured to determine the federal support degree corresponding to the federal frequent item set by performing secure multi-party computation with the second participant through the intermediate party based on the federal frequent item set.

[0133] In an apparatus embodiment of the present disclosure, the sample data acquisition module 501 includes:

[0134] an original data acquisition submodule, configured to acquire original data;

[0135] an original data desensitization submodule, configured to perform column name mapping on the original data to obtain the sample data.

[0136] In an apparatus embodiment of the present disclosure, the frequent item set exchange module 503 is further configured to send the first frequent item set to the second participant, so that the second participant determines a confidence degree of appearance of the first frequent item set in a case where the second frequent item set appears, and determines an association rule according to the confidence degree of appearance of the first frequent item set, and the association rule of the first participant is asymmetric to the association rule of the second participant.

[0137] The federal data association analysis apparatus provided by the present disclosure is applied to a first participant, and the first participant is communicatively connected with at least one second participant. After the first participant acquires sample data, the first participant determines a first frequent item set and a support degree corresponding to the first frequent item set in the sample data; then receives a second frequent item set sent by a second participant, the second frequent item set is obtained by the second participant according to sample data of the second participant; and determines a confidence degree of appearance of the second frequent item set in a case where the first frequent item set appears, so as to determine an association rule according to the confidence degree.

[0138] At this time, the first participant only receives the frequent item set sent by the other participants without exchanging the corresponding support, thereby realizing the privacy protection based federated association analysis among multiple participants, and obtaining high practicability and accuracy of the association rules under the condition of diversified data composition and complex association relationship; and without obtaining all source data, the data fragmentation state among the participants is broken, thereby ensuring the data privacy security, reducing the requirement of the processing procedure of the association analysis on the hardware configuration, thereby reducing the analysis cost and improving the analysis efficiency.

[0139] Figure 6 A structural schematic diagram of a data recommendation device provided by the embodiments of the present disclosure, which can include:

[0140] The historical data acquisition module 601 is configured to acquire historical behavior data of a user.

[0141] The original item set extraction module 602 is configured to extract an original item set from the historical behavior data.

[0142] The association rule matching module 603 is configured to match a recommendation item set in an association rule based on the original item set, and the association rule is obtained by the federated data association analysis device of the third aspect; and the user data recommendation module 604 is configured to recommend data to the user according to the recommendation item set.

[0143] The data recommendation device provided by the present disclosure can acquire historical behavior data of a user and extract an original item set from the historical behavior data; then match a recommendation item set in an association rule based on the original item set, and recommend data to the user according to the recommendation item set, wherein the association rule is obtained by the aforementioned Figures 1 to 3 The association rule is obtained by the federated data association analysis method. Based on the aforementioned federated data association analysis method, the association rules are extracted from more diversified data features and more extensive sample quantities, and then the user can be provided with differentiated, cross-platform and cross-domain data recommendation more accurately, thereby improving the accuracy and practicability of the data recommendation.

[0144] It should be noted that although several modules or units of the device for action execution are mentioned in the foregoing detailed description, such division is not mandatory. In fact, according to the embodiments of the present 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 into multiple modules or units.

[0145] Moreover, although individual steps of the methods in the present disclosure are described in a particular order in the figures, this is not required or implied as to the order of execution of the steps, nor is it required that all of the steps be executed to achieve the desired result. Additionally or alternatively, certain steps can be omitted, multiple steps can be combined into a single step, a single step can be broken into multiple steps, etc.

[0146] From the above description of the embodiments, those skilled in the art will easily understand that the example embodiments described herein can be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash disk, a mobile hard disk, etc.) or a network, and includes a number of instructions to make a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) execute the method according to the embodiments of the present disclosure.

[0147] In the example embodiments of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0148] Those skilled in the art can understand that each aspect of the present disclosure can be implemented as a system, a method or a program product. Therefore, each aspect of the present disclosure can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, which can be collectively referred to as "circuitry", "module" or "system".

[0149] The electronic device 700 according to this embodiment of the present disclosure will be described below with reference to Figure 7 Figure 7 The electronic device 700 shown is merely an example, and should not bring any limitation to the function and use range of the embodiments of the present disclosure.

[0150] As shown in Figure 7 The components of the electronic device 700 can include, but are not limited to, the above-mentioned at least one processing unit 710, the above-mentioned at least one storage unit 920, and a bus 930 connecting different system components, including the storage unit 920 and the processing unit 710.

[0151] The storage unit stores program code, which can be executed by the processing unit 710, so that the processing unit 710 performs the steps according to various example embodiments of the present disclosure described in the above "example method" section of the present specification.

[0152] ​The storage unit 920 can include a readable medium in the form of volatile storage such as random access memory (RAM) 9201 and / or cache memory 9202, and also include a non-volatile storage such as read only memory (ROM) 9203.

[0153] The storage unit 920 can also include a program / utility 9204 having a set of program modules 9205 such as an operating system, one or more application programs, other program modules, and program data, each of which can give the electronic device 700 its functionality, at least in part. The program modules 9205 can include, but are not limited to, a network environment implementation, either alone or in some combination.

[0154] The bus 930 can represent one or more of several types of bus structures, including a storage unit bus or bus for storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit bus, or a local bus using any of a variety of bus architectures.

[0155] The electronic device 700 can also communicate with one or more external devices 700 such as a keyboard or pointing device, a Bluetooth device, etc.; other devices that enable a user to interact with the electronic device 700; and / or one or more devices that enable the electronic device 700 to communicate with one or more other computing devices. Such communication can be via the display unit 940 and an input / output (I / O) interface 950 that is connected to the display unit 940. The electronic device 700 can also communicate with one or more networks such as a local area network (LAN), a wide area network (WAN), and / or the Internet through a network adapter 960. As depicted, the network adapter 960 is communicatively coupled to the other components of the electronic device 700 through the bus 930. It should be appreciated that the electronic device 700 can be implemented using other hardware and / or software components in addition to or in place of those depicted in FIG. 9. For example, the electronic device 700 can include a microcode implementation, a device driver implementation, a redundant processing unit implementation, an external disk drive array, a RAID system, a tape drive, and a data backup storage system, etc.

[0156] Those skilled in the art will readily appreciate that the example embodiments described herein can be implemented by software and / or firmware in addition to or instead of hardware. The application embodiments thus can be implemented in software and / or firmware code to be executed by one or more of the processing units 910 as "machine- readable code" and / or in hardware. Thus, the technology described herein can be embodied in a machine- readable code, which can be implemented in software, firmware, and / or hardware. Accordingly, the term "processor", as used herein can refer to one or more of custom circuitry and / or one or more microprocessors or microcontrollers that respond to and process program code instructions (which can be in the form of software or firmware) to control and perform various aspects of the technology described herein.

[0157] In the example embodiments of the present disclosure, a computer readable storage medium is also provided, on which a program product capable of implementing the above-mentioned method of the present specification is stored. In some possible implementations, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program codes for causing a terminal device to perform the steps described in the above-mentioned “example method” section of the present specification according to various example embodiments of the present disclosure when the program product is run on the terminal device.

[0158] In the example embodiments of the present disclosure, a computer readable storage medium is also provided, on which a program product capable of implementing the above-mentioned method of the present specification is stored. In some possible implementations, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program codes for causing a terminal device to perform the steps described in the above-mentioned “example method” section of the present specification according to various example embodiments of the present disclosure when the program product is run on the terminal device.

[0159] The program product can take any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium, for example, can be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any suitable combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0160] The computer readable signal medium can include a data signal propagated in baseband or propagated as a carrier wave in a propagated data signal, in which the readable program code is embodied. Such propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable signal medium can also be any readable medium other than the readable storage medium, which can send, propagate or transmit the program for use by or in connection with an instruction execution system, apparatus or device.

[0161] The program code contained on the readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0162] The program code may be executed by one or more programmable processing devices, which can include processors, microprocessor, microcomputer or microcontrollers, as well as other electronic circuits that include elements that can execute a program of instructions written in any one of a variety of programming languages, including an object oriented programming language such as Java, C++, or the like, as well as conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's computing device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, such as through the Internet using an Internet Service Provider (ISP).

[0163] In addition, the above-described flowcharts are merely illustrative of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to limit the present disclosure. It is readily understood that the processes shown in the above-described flowcharts do not indicate or limit the time sequence of the processes. In addition, it is readily understood that the processes can be executed synchronously or asynchronously, for example, in a plurality of modules.

[0164] Other embodiments of the present disclosure will be readily apparent to those skilled in the art upon considering the description hereof, prefaced by the Summary and supported in detail through the disclosure contained herein. The application is intended to cover any adaptations or variations of the present disclosure followed in the general principles of the present disclosure and including such modifications as would be readily apparent to those skilled in the art and desired to be protected by the claims that follow. The application is to be limited only as specified in the claims.

Claims

1. A method for federal data correlation analysis, characterized in that, The method is applied to a first participant, which is communicatively connected to at least one second participant. The method includes: Obtain sample data; the sample data is user behavior data of the first participant whose correlation is to be analyzed; wherein, the user behavior data includes product purchase, evaluation, collection or recommendation behavior data, or information browsing, commenting or sharing behavior data; In the sample data, a first frequent itemset and the support corresponding to the first frequent itemset are determined; The system receives a second frequent itemset sent by the second participant, which is obtained by the second participant based on its sample data; the sample data of the second participant consists of user behavior data from the second participant whose correlations are to be analyzed. The confidence level of the occurrence of the second frequent itemset is determined when the first frequent itemset occurs, and an association rule is determined based on the confidence level of the occurrence of the second frequent itemset; the confidence level represents the degree of association between the first frequent itemset and the second frequent itemset; the association rule is valid when the second frequent itemset occurs when the first frequent itemset occurs based on the confidence level.

2. The method according to claim 1, characterized in that, The step of determining the confidence level of the occurrence of the second frequent itemset when the first frequent itemset occurs, and determining the association rule based on the confidence level of the occurrence of the second frequent itemset, includes: Obtain at least one federated frequent itemset, each federated frequent itemset being formed by combining a first frequent itemset and a second frequent itemset, wherein at least one of the first frequent itemset and the second frequent itemset is different among the federated frequent itemsets; Determine the federal support degree corresponding to the federal frequent itemset; If the federated support is greater than the federated support threshold, the confidence level of the occurrence of the second frequent itemset is calculated based on the support and the federated support, given that the first frequent itemset appears in the federated frequent itemset. If the confidence level and the federated frequent itemset meet the association rule conditions, the association rule is determined to include the occurrence of the second frequent itemset when the first frequent itemset appears in the federated frequent itemset.

3. The method according to claim 2, characterized in that, The step of determining that the association rule includes the occurrence of the second frequent itemset when the first frequent itemset appears in the federated frequent itemset, given that the confidence level and the federated frequent itemset meet the association rule conditions, further includes: The association rule is output when the federated frequent itemset includes each combination of the first frequent itemset and the second frequent itemset.

4. The method according to claim 2, characterized in that, The association rule conditions are that the confidence level is greater than the confidence level threshold, and the length of the federated frequent itemset is less than the itemset length threshold.

5. The method according to claim 2, characterized in that, The first participant also communicates with an intermediary, and the intermediary communicates with the second participant. Determining the federated support degree corresponding to the federated frequent itemset includes: Based on the federated frequent itemset, the federated support degree corresponding to the federated frequent itemset is determined by the secure multi-party computation between the intermediary and the second participant.

6. The method according to claim 1, characterized in that, The acquisition of sample data includes: Obtain the raw data; The original data is mapped by column names to obtain the sample data.

7. The method according to any one of claims 1-6, characterized in that, After determining the first frequent itemset and the support corresponding to the first frequent itemset in the sample data, the method further includes: The first frequent itemset is sent to the second participant so that the second participant can determine the confidence level of the occurrence of the first frequent itemset when the second frequent itemset occurs, and determine the association rule based on the confidence level of the occurrence of the first frequent itemset. The association rule of the first participant is asymmetric with the association rule of the second participant.

8. A data recommendation method, characterized in that, The method includes: Obtain users' historical behavior data; Extract the original itemset from the historical behavior data; Based on the original itemset, a recommended itemset is matched in the association rule, wherein the association rule is obtained by the federated data association analysis method according to any one of claims 1 to 7; Data recommendations are made to the user based on the set of recommended items.

9. A federal data correlation analysis device, characterized in that, The device is applied to a first participant, which is communicatively connected to at least one second participant. The device includes: The sample data acquisition module is used to acquire sample data; the sample data is user behavior data of the first participant whose correlation is to be analyzed; wherein, the user behavior data includes product purchase, evaluation, collection or recommendation behavior data, or information browsing, commenting or sharing behavior data; A frequent itemset determination module is used to determine a first frequent itemset in the sample data, and the support corresponding to the first frequent itemset; The frequent itemset exchange module is used to receive a second frequent itemset sent by the second participant. The second frequent itemset is obtained by the second participant based on the sample data of the second participant. The sample data of the second participant is user behavior data of the second participant whose correlation is to be analyzed. The association rule determination module is used to determine the confidence level of the occurrence of the second frequent itemset when the first frequent itemset occurs, and to determine the association rule based on the confidence level of the occurrence of the second frequent itemset; the confidence level represents the degree of association between the first frequent itemset and the second frequent itemset; the association rule is valid when the second frequent itemset occurs when the first frequent itemset occurs based on the confidence level.

10. A data recommendation device, characterized in that, The device includes: The historical data acquisition module is used to acquire users' historical behavior data; The original itemset extraction module is used to extract original itemsets from the historical behavior data; The association rule matching module is used to match recommended itemsets in association rules based on the original itemset, wherein the association rules are implemented by the federated data association analysis device as described in claim 9; The user data recommendation module is used to recommend data to the user based on the set of recommendation items.

Images