Account evaluation method and device, computer device and storage medium

Abstract

The present disclosure relates to the technical field of big data processing, and discloses an account evaluation method and device, computer equipment, a storage medium and a computer program product. The method comprises: obtaining account resource interaction data of a to-be-evaluated account and a preset sequence mode, wherein the preset sequence mode is obtained based on a non-frequent feature sequence database, the non-frequent feature sequence database is determined according to a non-frequent item set obtained by mining a preset non-frequent mode tree, and the preset non-frequent mode tree and an initial feature database have an association relationship; and evaluating the account resource interaction data according to the preset sequence mode to obtain an evaluation result of the to-be-evaluated account. The present method can improve the accuracy of account evaluation and the security of resource interaction.

Description

Technical Field

[0001] This disclosure relates to the field of big data processing technology, and in particular to an account evaluation method, apparatus, computer equipment, storage medium, and computer program product. Background Technology

[0002] In resource interaction scenarios, to ensure the security and reliability of resource interaction, the parties involved need to assess the risk of the account. Only after the account passes the assessment are access permissions granted. Typically, account risk assessment is combined with risk warning rules for calculation and alerting.

[0003] In traditional techniques, risk warning rules can be determined using decision tree models or pattern recognition methods based on association analysis. However, decision tree models do not support categorical features, resulting in poor rule stability and adaptability. Pattern recognition methods based on association analysis can lead to rule distortion due to imbalanced positive and negative samples. Therefore, risk warning rules obtained through these methods for assessing account risk have poor accuracy and cannot guarantee the security and reliability of resource interaction processes. Summary of the Invention

[0004] Therefore, it is necessary to provide an account assessment method, apparatus, computer equipment, storage medium, and computer program product that can improve the accuracy of account assessment and the security of resource interaction, in order to address the above-mentioned technical problems.

[0005] In a first aspect, embodiments of this disclosure provide a method for evaluating an account. The method includes:

[0006] Obtain account resource interaction data and preset sequence patterns of the account to be evaluated. The preset sequence patterns are obtained by mining a non-frequent feature sequence database. The non-frequent feature sequence database is determined by mining a non-frequent itemset from a preset non-frequent pattern tree. The preset non-frequent pattern tree has an association with the initial feature database.

[0007] The account resource interaction data is evaluated according to the preset sequence pattern to obtain the evaluation result of the account to be evaluated.

[0008] In one embodiment, the method for determining the preset sequence pattern includes:

[0009] Obtain a database of infrequent feature sequences and a support threshold, wherein the database of infrequent feature sequences is determined based on an initial feature database;

[0010] Determine the initial prefix in the infrequent feature sequence database and obtain the corresponding projection database. The length of the initial prefix is ​​equal to a preset length and the support is greater than or equal to the support threshold.

[0011] Feature terms in the projection database with a support greater than or equal to the support threshold are identified as candidate feature terms.

[0012] The candidate feature terms and the initial prefix are merged to obtain a new initial prefix and a corresponding new projection database, until the support of all feature terms in the new projection database is less than or equal to the support threshold, thus obtaining a preset sequence pattern.

[0013] In one embodiment, determining the initial prefix in the infrequent feature sequence database and obtaining the corresponding projection database includes:

[0014] Obtain the preset length corresponding to the infrequent feature sequence database;

[0015] Feature terms with a length equal to the preset length in the infrequent feature sequence database are identified as candidate feature terms;

[0016] Candidate feature terms with support greater than the support threshold are identified as initial prefixes, and corresponding initial suffixes are searched in the candidate infrequent feature sequence database, wherein the support of the feature terms in the candidate infrequent feature sequence database is greater than or equal to the support threshold.

[0017] The set of initial suffixes is determined as the projection database corresponding to the initial prefix.

[0018] In one embodiment, the method for determining the infrequent itemset includes:

[0019] Obtain an initial feature database, wherein the initial feature database includes an initial feature term set consisting of initial feature terms;

[0020] Initial feature terms with support below the target threshold are identified as infrequent feature terms;

[0021] The infrequent feature items are used as feature nodes to construct a frequent pattern tree;

[0022] The frequent pattern tree is mined based on the conditional pattern base corresponding to the feature node to obtain the infrequent itemset.

[0023] In one embodiment, evaluating the account resource interaction data according to the preset sequence pattern includes:

[0024] Determine the feature items and the relationships between the feature items in the preset sequence pattern;

[0025] The evaluation rules are determined based on the account resource interaction data corresponding to the feature items and the correlation between the feature items;

[0026] The account resource interaction data is evaluated using the aforementioned evaluation rules.

[0027] Secondly, embodiments of this disclosure also provide an account evaluation apparatus. The apparatus includes:

[0028] The acquisition module is used to acquire account resource interaction data and preset sequence patterns of the account to be evaluated. The preset sequence patterns are obtained by mining a non-frequent feature sequence database. The non-frequent feature sequence database is determined by mining a non-frequent itemset from a preset non-frequent pattern tree. There is an association between the preset non-frequent pattern tree and the initial feature database.

[0029] The evaluation module is used to evaluate the account resource interaction data according to the preset sequence pattern to obtain the evaluation result of the account to be evaluated.

[0030] In one embodiment, the module for determining the preset sequence pattern includes:

[0031] The first acquisition submodule is used to acquire a database of infrequent feature sequences and a support threshold, wherein the database of infrequent feature sequences is determined based on an initial feature database;

[0032] The first determining module is used to determine the initial prefix in the infrequent feature sequence database and obtain the corresponding projection database, wherein the length of the initial prefix is ​​equal to the preset length and the support is greater than or equal to the support threshold.

[0033] The second determining module is used to determine feature items in the projection database whose support is greater than or equal to the support threshold as candidate feature items;

[0034] The merging module is used to merge the candidate feature terms and the initial prefix to obtain a new initial prefix and a corresponding new projection database, until the support of all feature terms in the new projection database is less than or equal to the support threshold, thus obtaining a preset sequence pattern.

[0035] In one embodiment, the first determining module includes:

[0036] The acquisition unit is used to acquire a preset length corresponding to the infrequent feature sequence database;

[0037] The first determining submodule is used to determine the feature items in the infrequent feature sequence database whose length is equal to the preset length as candidate feature items;

[0038] The second determining submodule is used to determine candidate feature terms with support greater than the support threshold as initial prefixes, and to search for corresponding initial suffixes in the candidate infrequent feature sequence database, wherein the support of the feature terms in the candidate infrequent feature sequence database is greater than or equal to the support threshold.

[0039] The third determining submodule is used to determine the set of initial suffixes as the projection database corresponding to the initial prefix.

[0040] In one embodiment, the module for determining the infrequent itemsets includes:

[0041] The second acquisition submodule is used to acquire an initial feature database, wherein the initial feature database includes an initial feature item set composed of initial feature items;

[0042] The third determination module is used to determine that initial feature terms with support below the target threshold are infrequent feature terms;

[0043] The construction module is used to construct a frequent pattern tree by using the infrequent feature items as feature nodes;

[0044] The mining module is used to mine the frequent pattern tree based on the conditional pattern base corresponding to the feature node to obtain the infrequent itemset.

[0045] In one embodiment, the evaluation module includes:

[0046] The fourth determining module is used to determine the feature items in the preset sequence pattern and the correlation between the feature items;

[0047] The fifth determining module is used to determine the evaluation rules based on the account resource interaction data corresponding to the feature item and the correlation between the feature item;

[0048] The evaluation submodule is used to evaluate the account resource interaction data using the evaluation rules.

[0049] Thirdly, embodiments of this disclosure also provide a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described in any one of the embodiments of this disclosure.

[0050] Fourthly, embodiments of this disclosure also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the method described in any one of the embodiments of this disclosure.

[0051] Fifthly, embodiments of this disclosure also provide a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in any one of the embodiments of this disclosure.

[0052] In this embodiment of the disclosure, during account evaluation, the account resource interaction data of the account to be evaluated is evaluated using an acquired preset sequence pattern to obtain an evaluation result. Since the preset sequence pattern is mined from a non-frequent feature sequence database, which is mined from a preset non-frequent pattern tree corresponding to the initial feature database, it can effectively determine non-frequent features affected by time series. It can also mine partial order patterns from the partial order phenomenon of non-frequent itemsets, thus enabling the judgment of non-frequent features. In subsequent account evaluations, considering non-frequent features allows for a more comprehensive evaluation of the account resource interaction data, effectively improving the accuracy of the evaluation results and ensuring the security of account resource interactions. Attached Figure Description

[0053] Figure 1 This is a flowchart illustrating an account evaluation method in one embodiment;

[0054] Figure 2 This is a flowchart illustrating how a preset sequence pattern is determined in one embodiment;

[0055] Figure 3 This is a flowchart illustrating how a preset sequence pattern is determined in one embodiment;

[0056] Figure 4 This is a flowchart illustrating how infrequent itemsets are determined in one embodiment.

[0057] Figure 5 This is a flowchart illustrating an account evaluation method in one embodiment;

[0058] Figure 6 This is a structural block diagram of an account evaluation device in one embodiment;

[0059] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the embodiments of this disclosure will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of this disclosure and are not intended to limit the embodiments of this disclosure.

[0061] It should be noted that the user information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and the process of acquiring information and data involved in the embodiments of this disclosure are all authorized by the user or fully authorized by all parties, and comply with the relevant provisions of national laws and regulations.

[0062] In one embodiment, such as Figure 1 As shown, an account assessment method is provided, the method comprising:

[0063] Step S110: Obtain account resource interaction data and preset sequence pattern of the account to be evaluated. The preset sequence pattern is obtained by mining a non-frequent feature sequence database. The non-frequent feature sequence database is determined by mining a non-frequent itemset from a preset non-frequent pattern tree. The preset non-frequent pattern tree has an association with the initial feature database.

[0064] In this embodiment, account resource interaction data and preset sequence patterns of the account to be evaluated are obtained. Specifically, the account resource interaction data of the account to be evaluated may include, but is not limited to, account attribute data and data generated by the account in resource interactions. In one example, the type of account resource interaction data obtained can be determined according to the actual application scenario, and the type of account resource interaction data may differ in different application scenarios. The preset sequence patterns are obtained through sequence pattern mining, which refers to mining patterns that appear frequently over time or other periods. In this embodiment, the preset sequence patterns may differ for different application scenarios. In one example, this embodiment can be applied to the scenario of account credibility assessment. An initial feature database is obtained based on the account resource interaction data of accounts whose credibility meets preset requirements from historical data. After a series of processing steps on the initial feature database, a preset pattern sequence is obtained. The preset requirements can be determined according to the actual application scenario, such as credibility being lower than a preset threshold, credibility being higher than a preset threshold, etc. In one possible implementation, the data in the initial feature database can be obtained by extracting features from the account resource interaction data.

[0065] In one possible implementation, an initial feature database can be determined based on the actual application scenario. A preset sequence pattern is obtained by processing the initial feature database. This initial feature database corresponds to the application scenario and typically includes itemsets corresponding to account data matching the scenario. The data in the initial feature database can be determined based on data generated by accounts during historical resource interactions. Generally, it can include account resource interaction data from multiple accounts. In one example, the data in the initial feature database has a temporal relationship based on chronological order. When obtaining the preset sequence pattern, infrequent feature items in the initial feature database are identified. An infrequent pattern tree is obtained based on these infrequent feature items. Itemset mining is then performed on the infrequent pattern tree to obtain the corresponding infrequent itemsets. The infrequent feature sequence database is then determined based on these infrequent itemsets. Finally, the preset sequence pattern is obtained by mining the infrequent feature sequence database using a preset sequence pattern mining method. The preset sequence pattern mining method can be determined according to the actual application scenario. The sequence pattern mining method can include, but is not limited to, the mining method based on the AprioriAll algorithm, the mining method based on the GSP algorithm, and the mining method based on the PrefixSpan algorithm. In one example, the mining method based on the PrefixSpan algorithm can be selected as the preset sequence pattern mining method, so that the temporal correlation between data items can be taken into account; and the sequence is traversed in a depth-first manner, without the need to generate candidate sequences, which effectively improves the mining efficiency.

[0066] Step S120: Evaluate the account resource interaction data according to the preset sequence pattern to obtain the evaluation result of the account to be evaluated.

[0067] In this embodiment of the disclosure, after obtaining a preset sequence pattern, the account resource interaction data is evaluated based on the obtained preset sequence pattern to determine the evaluation result corresponding to the account to be evaluated. Specifically, feature points affected by time series can be determined based on the preset sequence pattern. The account resource interaction data is evaluated based on the preset sequence pattern, and the correlation between the account resource interaction data and the features in the preset sequence pattern is determined to obtain the evaluation result of the account to be evaluated. In one example, when evaluating the account resource interaction data according to the preset sequence pattern, evaluation rules can be determined based on the correlation between feature items and feature item sets in the preset sequence pattern. The account resource interaction data is evaluated according to the evaluation rules to obtain the evaluation result corresponding to the account to be evaluated. In some possible implementations, the corresponding evaluation result may include evaluation passed, evaluation failed, or further judgment required. When the account to be evaluated passes the evaluation, the next step can be executed; when the account to be evaluated fails the evaluation, the next step for the account can be rejected; when the account to be evaluated requires further judgment, other data or rules can be combined to further judge the account. In one example, during the evaluation process, other preset rules or evaluation methods can be combined to evaluate the account resource interaction data to obtain the comprehensive evaluation result of the account to be evaluated.

[0068] In this embodiment of the disclosure, during account evaluation, the account resource interaction data of the account to be evaluated is evaluated using an acquired preset sequence pattern to obtain an evaluation result. Since the preset sequence pattern is mined from a non-frequent feature sequence database, which is mined from a preset non-frequent pattern tree corresponding to the initial feature database, it can effectively determine non-frequent features affected by time series. It can also mine partial order patterns from the partial order phenomenon of non-frequent itemsets, thus enabling the judgment of non-frequent features. In subsequent account evaluations, considering non-frequent features allows for a more comprehensive evaluation of the account resource interaction data, effectively improving the accuracy of the evaluation results and ensuring the security of account resource interactions.

[0069] In one embodiment, such as Figure 2 As shown, the method for determining the preset sequence pattern includes:

[0070] Step S210: Obtain a database of infrequent feature sequences and a support threshold, wherein the database of infrequent feature sequences is determined based on the initial feature database;

[0071] Step S220: Determine the initial prefix in the infrequent feature sequence database and obtain the corresponding projection database. The length of the initial prefix is ​​equal to the preset length and the support is greater than or equal to the support threshold.

[0072] Step S230: Determine the feature items in the projection database whose support is greater than or equal to the support threshold as candidate feature items;

[0073] Step S240: Merge the candidate feature terms and the initial prefix to obtain a new initial prefix and a corresponding new projection database, until the support of all feature terms in the new projection database is less than or equal to the support threshold, and obtain a preset sequence pattern.

[0074] In this embodiment, when a preset sequence pattern is obtained based on a non-frequent feature sequence database, a non-frequent feature sequence database and a support threshold are acquired. The non-frequent feature sequence database is determined based on an initial feature database, which and the support threshold can be predetermined according to the actual application scenario. An initial prefix and a corresponding projection database are determined based on the non-frequent feature sequence database. Specifically, itemsets meeting the conditions are determined from the non-frequent feature sequence database according to a preset length and support threshold as initial prefixes. The length of the initial prefix is ​​equal to the length of the corrupted feature sequence and the support is greater than or equal to the support threshold. A corresponding projection database is generated based on the obtained initial prefixes. There can be one or more initial prefixes, each corresponding to a projection data sequence. The projection data sequences construct the projection database. After obtaining the projection database, feature items with support greater than or equal to the support threshold in the projection database are acquired and identified as candidate feature items. The candidate feature items and the initial prefixes are merged to obtain a new initial prefix and a new projection database. This process continues until the support of all feature items in the new projection database is less than or equal to the support threshold, at which point the iteration stops, and the preset sequence pattern is obtained.

[0075] In one possible implementation, infrequent feature itemsets can be mined from an initial feature database. These infrequent feature itemsets are then expanded into infrequent sequences, resulting in an infrequent feature sequence database. Each infrequent sequence is formed by an ordered arrangement of multiple itemsets. The original infrequent sequences are projected based on prefixes. Sequence patterns are mined starting with prefixes of a preset length, and the corresponding prefix projection database is searched to obtain infrequent sequences corresponding to prefixes of the preset length. Then, the preset length is recursively increased to obtain a new preset length, and infrequent sequences corresponding to prefixes of the new preset length are mined. This process is repeated, continuously combining prefixes and outputting infrequent sequences that meet the support threshold, until the prefix projection database for a certain prefix is ​​empty. In one example, the initial preset length can be set to 1, and during the recursive process, the length is incremented by 1 each time to obtain a new preset length. Table 1 shows an infrequent feature sequence database according to one embodiment.

[0076] Table 1

[0077] Identifier (Sid) Sequence C1 <(ab)(ac)> C2 <(cd)(efg)> C3 <(and)hg> C4 <(ad)h>

[0078] When sequence patterns are obtained from the database described in Table 1, the initial length can be set to 1. First, based on all infrequent itemsets that satisfy the minimum support (min_sup), i.e. the support threshold, infrequent itemsets in the infrequent feature sequence database are removed, resulting in Table 2.

[0079] Table 2

[0080] Identifier (Sid) Sequence C1 <a(ac)> C2 <(cd)g> C3 <(and)hg> C4 <(ad)h>

[0081] As shown in Table 3, the infrequent items in Table 2 are used as initial prefixes to generate corresponding projection data sequence libraries. In the projection database, infrequent itemsets that meet the support threshold are searched and combined with the corresponding initial prefixes to form a new initial prefix of length 2.

[0082] Table 3

[0083]

[0084] As shown in Table 4, a new projection database is generated based on the new initial prefix. There are no infrequent itemsets in the new projection database that satisfy the support threshold. The iteration ends, and all sequence patterns of length 2 that satisfy the support threshold are obtained.

[0085] Table 4

[0086]

[0087] This embodiment of the disclosure mines a non-frequent feature sequence database based on a support threshold to obtain a preset sequence pattern. This realizes the mining of frequently occurring patterns from non-frequent itemsets, and can identify the commonalities of features with certain time patterns and low occurrence frequencies. This improves the comprehensiveness of the evaluation process in the subsequent evaluation process, effectively enhances the accuracy of the evaluation results, and thus ensures the security and reliability of resource interaction. In this embodiment, sequence pattern mining is performed through a prefix sum projection database, eliminating the need to generate a candidate sequence database, which greatly reduces the amount of computation. Moreover, the mining object is a non-frequent itemset database, reducing the amount of data to be processed, thereby improving the efficiency of sequence pattern mining, obtaining the preset sequence pattern more quickly, which is beneficial for subsequent iterative optimization and improves resource utilization.

[0088] In one embodiment, such as Figure 3 As shown, determining the initial prefix in the infrequent feature sequence database and obtaining the corresponding projection database includes:

[0089] Step S221: Obtain the preset length corresponding to the infrequent feature sequence database;

[0090] Step S222: Determine the feature items in the infrequent feature sequence database whose length is equal to the preset length as candidate feature items;

[0091] Step S223: Determine candidate feature terms with support greater than the support threshold as initial prefixes, and search for corresponding initial suffixes in the candidate infrequent feature sequence database, wherein the support of feature terms in the candidate infrequent feature sequence database is greater than or equal to the support threshold;

[0092] Step S224: Determine the set of initial suffixes as the projection database corresponding to the initial prefix.

[0093] In this embodiment, when determining the initial prefix and corresponding projection database, a preset length corresponding to the infrequent feature sequence database is obtained. This preset length can be pre-set according to the actual application scenario. Based on the preset length, feature items with a length equal to the preset length are selected as candidate feature items from the infrequent feature sequence database. The initial prefix is ​​obtained based on the support of the candidate feature items. When the support of a candidate feature item is greater than or equal to a support threshold, the candidate feature item is determined as the initial prefix. Based on the support of the feature items in the infrequent feature sequence database, a candidate infrequent feature sequence database is obtained. Specifically, feature items with support less than the support threshold are filtered out from the infrequent feature sequence database to obtain the candidate infrequent feature sequence database. The support of all feature items in the candidate infrequent feature sequence database is greater than or equal to the support threshold. An initial suffix corresponding to the initial prefix is ​​searched in the infrequent feature sequence database, and the set of initial suffixes obtained is determined as the projection database corresponding to the initial prefix.

[0094] In this embodiment of the disclosure, an initial prefix and projection database is determined based on a preset length and a threshold value, thereby enabling the subsequent mining of preset sequence patterns. This method eliminates the need to generate candidate sequences, and the projection database shrinks quickly with a small data processing volume. The method of obtaining preset sequence patterns through iterative processing of the prefix and suffix projection database can effectively improve the efficiency of sequence pattern mining and increase resource utilization.

[0095] In one embodiment, such as Figure 4 As shown, the method for determining the infrequent itemsets includes:

[0096] Step S410: Obtain an initial feature database, wherein the initial feature database includes an initial feature item set composed of initial feature items;

[0097] Step S420: Determine the initial feature terms whose support is lower than the target threshold as infrequent feature terms;

[0098] Step S430: Construct a frequent pattern tree by using the infrequent feature items as feature nodes;

[0099] Step S440: Mining the frequent pattern tree based on the conditional pattern base corresponding to the feature node to obtain the infrequent itemset.

[0100] In this embodiment, infrequent itemsets are determined using a frequent pattern tree. Specifically, an initial feature database is obtained, which includes an initial feature itemset composed of initial feature items. Typically, the initial feature items are determined based on the actual application scenario. In one example, the initial feature items may be determined after feature extraction based on historical resource interaction data. Initial feature items in the initial feature database with support below a target threshold are identified as infrequent feature items. The target threshold can be determined based on the actual application scenario and may be the same as the support threshold. A frequent pattern tree is constructed based on the infrequent feature items, and the infrequent feature items in the obtained frequent pattern tree correspond to feature nodes. Based on the conditional patterns corresponding to the feature nodes and mining the frequent pattern tree, the resulting itemset is the infrequent itemset. In one example, the conditional pattern base can be determined based on the actual application scenario. In some possible implementations, the process of obtaining infrequent itemsets based on the frequent pattern tree can be performed using association analysis algorithms for data processing and mining. These association analysis algorithms may include, but are not limited to, the FP-Growth algorithm. In one example, when using the FP-Growth algorithm for infrequent itemset mining, the data can be compressed by constructing an infrequent itemset tree. The database is then scanned twice to determine the given pattern. In the second scan, the feature set containing infrequent feature items is retained, while other features are filtered out. Based on the path corresponding to the infrequent feature items, the corresponding conditional basis is found, and a conditional FP tree is constructed. The infrequent feature set is then recursively obtained using the FP-growth algorithm.

[0101] In this embodiment, infrequent feature items are obtained by judging the support of feature items in the initial feature database, and a frequent pattern tree is constructed. This enables the mining of infrequent itemsets, which is beneficial for obtaining preset sequence patterns with low support and high confidence. Itemset mining through the frequent pattern tree ensures the efficiency of the mining and the accuracy of the final infrequent itemsets, further improving the accuracy of account evaluation and ensuring the security and reliability of resource interaction.

[0102] In one embodiment, such as Figure 5 As shown, the evaluation of the account resource interaction data according to the preset sequence pattern includes:

[0103] Step S121: Determine the feature items in the preset sequence pattern and the correlation between the feature items;

[0104] Step S122: Determine the evaluation rules based on the account resource interaction data corresponding to the feature items and the correlation between the feature items;

[0105] Step S123: Evaluate the account resource interaction data using the evaluation rules.

[0106] In this embodiment, when evaluating account resource interaction data using a preset sequence pattern, the feature items and their relationships are determined based on the preset sequence pattern. The preset sequence pattern is determined based on an initial feature database, which is typically pre-determined according to the actual application scenario. In this embodiment, since account resource interaction data needs to be evaluated, the initial feature database can be determined based on historical resource interaction data, and there is a correspondence between feature items and resource interaction data. Based on the feature items in the preset sequence pattern, the corresponding account resource interaction data is determined, and the evaluation rules are determined based on the corresponding account resource interaction data and the relationships between feature items. The obtained evaluation rules are used to evaluate the account resource interaction data, where the evaluation rules can be reflected through threshold and relevance settings. In one example, when evaluating account resource interaction data, other data, rules, models, etc., can also be combined for comprehensive evaluation to obtain the evaluation result; this disclosure does not impose any limitations on this.

[0107] In this embodiment of the disclosure, evaluation rules are determined by a preset sequence pattern, and the evaluation rules are used to evaluate account resource interaction data, which can make account evaluation more efficient and accurate. The evaluation rules determined by the preset sequence pattern are more in line with the application scenario, which can effectively reduce the complexity of the evaluation and reduce the amount of data processing. Thus, while ensuring accuracy, the efficiency of account evaluation is improved, and the user experience and resource utilization are enhanced.

[0108] In one possible implementation, the mining of preset sequence patterns can be performed through the following process: Construct a conditional FP-tree based on infrequent feature nodes. Using a divide-and-conquer strategy, process the subproblems of their associated nodes to discover all infrequent itemsets ending with a specific suffix, resulting in the final candidate sequence feature set. This can be achieved by pruning the FP-tree to retain paths containing infrequent features. Infrequent itemsets are mined by finding the conditional basis of element items and constructing the conditional FP-tree. A conditional FP-tree is constructed for each infrequent feature node. The node and the conditional FP-tree are input as parameters into the FP-Growth algorithm to recursively obtain the infrequent feature set, thus obtaining the final candidate sequence feature set. Scan the final candidate sequence feature library to generate all sequence patterns that meet the support threshold and have a length of 1. Generate a corresponding projection (suffix) database based on the sequence patterns of length 1. Repeat the above steps on the corresponding projection database until no sequence patterns of length 1 can be generated on the corresponding projection database. Repeat the above process on different projection databases until no new sequence patterns of length 1 are generated, thus obtaining the final sequence pattern.

[0109] This embodiment uses the FP-Growth algorithm to mine infrequent itemsets and the PrefixSpan algorithm to consider time-series factors. By discovering the time series of infrequent feature sets, it obtains infrequent feature sets with low support but high confidence. When mining infrequent itemsets, if a node in the FP-tree belongs to an infrequent feature, a conditional FP-tree based on that node is constructed and input as a parameter into the FP-Growth algorithm to obtain the infrequent feature set. It only requires compressing and storing the dataset by constructing the infrequent itemset tree, and performing two scans of the database to determine the given pattern. In the second scan, only feature sets containing infrequent features are considered, and other features are pruned. The algorithm runs quickly and has high performance. The PrefixSpan algorithm, based on the divide-and-conquer approach, expands the infrequent feature itemset into infrequent sequences for mining preset sequence patterns. Through depth-first traversal, it improves pattern mining efficiency and can uncover some low-frequency features affected by time series. In subsequent evaluation processes, it ensures the depth and breadth of evaluation, improving the accuracy of account evaluation.

[0110] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the accompanying drawings may include multiple steps or stages, which are not necessarily completed at the same time, but may be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but may be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0111] Based on the same inventive concept, this disclosure also provides an account evaluation apparatus for implementing the account evaluation method described above. The solution provided by this apparatus is similar to the implementation described in the above method; therefore, the specific limitations of the one or more account evaluation apparatus embodiments provided below can be found in the limitations of the account evaluation method described above, and will not be repeated here.

[0112] In one embodiment, such as Figure 6 As shown, an account evaluation device 600 is provided, comprising:

[0113] The acquisition module 610 is used to acquire account resource interaction data and preset sequence patterns of the account to be evaluated. The preset sequence patterns are obtained by mining a non-frequent feature sequence database. The non-frequent feature sequence database is determined by mining a non-frequent itemset from a preset non-frequent pattern tree. The preset non-frequent pattern tree has an association with the initial feature database.

[0114] The evaluation module 620 is used to evaluate the account resource interaction data according to the preset sequence pattern to obtain the evaluation result of the account to be evaluated.

[0115] In one embodiment, the module for determining the preset sequence pattern includes:

[0116] The first acquisition submodule is used to acquire a database of infrequent feature sequences and a support threshold, wherein the database of infrequent feature sequences is determined based on an initial feature database;

[0117] The first determining module is used to determine the initial prefix in the infrequent feature sequence database and obtain the corresponding projection database, wherein the length of the initial prefix is ​​equal to the preset length and the support is greater than or equal to the support threshold.

[0118] The second determining module is used to determine feature items in the projection database whose support is greater than or equal to the support threshold as candidate feature items;

[0119] The merging module is used to merge the candidate feature terms and the initial prefix to obtain a new initial prefix and a corresponding new projection database, until the support of all feature terms in the new projection database is less than or equal to the support threshold, thus obtaining a preset sequence pattern.

[0120] In one embodiment, the first determining module includes:

[0121] The acquisition unit is used to acquire a preset length corresponding to the infrequent feature sequence database;

[0122] The first determining submodule is used to determine the feature items in the infrequent feature sequence database whose length is equal to the preset length as candidate feature items;

[0123] The second determining submodule is used to determine candidate feature terms with support greater than the support threshold as initial prefixes, and to search for corresponding initial suffixes in the candidate infrequent feature sequence database, wherein the support of the feature terms in the candidate infrequent feature sequence database is greater than or equal to the support threshold.

[0124] The third determining submodule is used to determine the set of initial suffixes as the projection database corresponding to the initial prefix.

[0125] In one embodiment, the module for determining the infrequent itemsets includes:

[0126] The second acquisition submodule is used to acquire an initial feature database, wherein the initial feature database includes an initial feature item set composed of initial feature items;

[0127] The third determination module is used to determine that initial feature terms with support below the target threshold are infrequent feature terms;

[0128] The construction module is used to construct a frequent pattern tree by using the infrequent feature items as feature nodes;

[0129] The mining module is used to mine the frequent pattern tree based on the conditional pattern base corresponding to the feature node to obtain the infrequent itemset.

[0130] In one embodiment, the evaluation module includes:

[0131] The fourth determining module is used to determine the feature items in the preset sequence pattern and the correlation between the feature items;

[0132] The fifth determining module is used to determine the evaluation rules based on the account resource interaction data corresponding to the feature item and the correlation between the feature item;

[0133] The evaluation submodule is used to evaluate the account resource interaction data using the evaluation rules.

[0134] The modules in the aforementioned account evaluation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can invoke and execute the operations corresponding to each module.

[0135] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data such as account resource interaction data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements an account evaluation method.

[0136] Those skilled in the art will understand that Figure 7 The structures shown are merely block diagrams of some structures related to the embodiments of this disclosure and do not constitute a limitation on the computer devices on which the embodiments of this disclosure are applied. Specific computer devices may include more or fewer components than those shown in the figures, or combine certain components, or have different component arrangements.

[0137] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0138] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0139] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0140] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0141] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this disclosure can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this disclosure may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this disclosure may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0142] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0143] The above-described embodiments are merely illustrative of several implementation methods of the present disclosure, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the patent for the embodiments of the present disclosure. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of the present disclosure, and these all fall within the protection scope of the embodiments of the present disclosure. Therefore, the protection scope of the embodiments of the present disclosure should be determined by the appended claims.

Claims

1. A method of assessing an account, characterized by, The method comprises: obtaining account resource interaction data of an account to be evaluated and a preset sequence mode, wherein the preset sequence mode is obtained based on non-frequent feature sequence database mining, the non-frequent feature sequence database is determined according to non-frequent item set mining of a preset non-frequent mode tree, and the preset non-frequent mode tree and an initial feature database have an association relationship; evaluating the account resource interaction data according to the preset sequence mode to obtain an evaluation result of the account to be evaluated; wherein the determination manner of the preset sequence mode comprises: obtaining a non-frequent feature sequence database and a support threshold, wherein the non-frequent feature sequence database is determined based on an initial feature database; determining an initial prefix in the non-frequent feature sequence database and obtaining a corresponding projection database, wherein the length of the initial prefix is equal to a preset length and the support is greater than or equal to the support threshold; determining a feature item in the projection database with a support greater than or equal to the support threshold as a candidate feature item; merging the candidate feature item and the initial prefix to obtain a new initial prefix and a corresponding new projection database, until the support of the feature items in the new projection database is less than or equal to the support threshold, and a preset sequence mode is obtained.

2. The method of claim 1, wherein, The determination of the initial prefix in the non-frequent feature sequence database and the corresponding projection database comprises: obtaining a preset length corresponding to the non-frequent feature sequence database; determining a feature item in the non-frequent feature sequence database with a length equal to the preset length as a candidate feature item; determining a candidate feature item with a support greater than the support threshold as an initial prefix, and searching for a corresponding initial suffix in a candidate non-frequent feature sequence database, wherein the support of the feature items in the candidate non-frequent feature sequence database is greater than or equal to the support threshold; determining a set of the initial suffixes as the projection database corresponding to the initial prefix.

3. The method of claim 1, wherein, The determination manner of the non-frequent item set comprises: obtaining an initial feature database, wherein the initial feature database includes an initial feature item set composed of initial feature items; determining an initial feature item with a support lower than a target threshold as a non-frequent feature item; constructing a frequent mode tree by taking the non-frequent feature item as a feature node; mining the frequent mode tree according to a conditional mode base corresponding to the feature node to obtain a non-frequent item set.

4. The method of claim 1, wherein, The evaluation of the account resource interaction data according to the preset sequence mode comprises: determining feature items in the preset sequence mode and an association relationship between the feature items; determining an evaluation rule based on the account resource interaction data corresponding to the feature items and the association relationship between the feature items; evaluating the account resource interaction data by using the evaluation rule.

5. An account evaluation apparatus, characterized by, The device comprises: The acquisition module is configured to acquire account resource interaction data of a to-be-evaluated account and a preset sequence mode, wherein the preset sequence mode is obtained based on a non-frequent feature sequence database, and the non-frequent feature sequence database is determined according to a non-frequent item set obtained by mining a preset non-frequent mode tree, and the preset non-frequent mode tree has an association relationship with an initial feature database; The evaluation module is configured to evaluate the account resource interaction data according to the preset sequence mode, to obtain an evaluation result of the to-be-evaluated account. The determination module of the preset sequence mode includes: The first acquisition submodule is configured to acquire a non-frequent feature sequence database and a support threshold, wherein the non-frequent feature sequence database is determined based on an initial feature database; The first determination module is configured to determine an initial prefix in the non-frequent feature sequence database, and obtain a corresponding projection database, wherein the initial prefix has a length equal to a preset length and a support equal to or greater than the support threshold; The second determination module is configured to determine a feature item in the projection database as a candidate feature item, wherein the feature item has a support equal to or greater than the support threshold; The merging module is configured to merge the candidate feature item and the initial prefix, to obtain a new initial prefix and a corresponding new projection database, until the supports of the feature items in the new projection database are all less than or equal to the support threshold, to obtain a preset sequence mode.

6. The apparatus of claim 5, wherein, The first determination module includes: The acquisition unit is configured to acquire a preset length corresponding to the non-frequent feature sequence database; The first determination submodule is configured to determine a feature item in the non-frequent feature sequence database as a candidate feature item, wherein the feature item has a length equal to the preset length; The second determination submodule is configured to determine a candidate feature item having a support greater than the support threshold as an initial prefix, and find a corresponding initial suffix in a candidate non-frequent feature sequence database, wherein the feature items in the candidate non-frequent feature sequence database have supports equal to or greater than the support threshold; The third determination submodule is configured to determine a set of the initial suffixes as a projection database corresponding to the initial prefix.

7. The apparatus of claim 5, wherein, The determination module of the non-frequent item set includes: The second acquisition submodule is configured to acquire an initial feature database, wherein the initial feature database includes an initial feature item set composed of initial feature items; The third determination module is configured to determine an initial feature item having a support lower than a target threshold as a non-frequent feature item; The construction module is configured to construct a frequent mode tree by taking the non-frequent feature item as a feature node; The mining module is configured to mine the frequent mode tree according to a conditional pattern base corresponding to the feature node, to obtain a non-frequent item set.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the account evaluation method in any one of claims 1 to 4.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the account evaluation method in any one of claims 1 to 4.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the account evaluation method in any one of claims 1 to 4.

Images