Long-track-oriented business process compliance checking method and system
By filtering event logs and restoring index information, the problems of high computational complexity and insufficient accuracy in compliance inspection of long-track business processes are solved, and efficient compliance inspection is achieved, suitable for industrial-grade applications.
Patent Information
- Application Number
- CN202510552186.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-29
- Publication Date
- 2025-08-15
AI Technical Summary
The prior art faces the problems of high computational complexity and insufficient accuracy when handling long-track business process compliance checks, especially when the event log size and complexity increase, existing alignment methods cannot be effectively processed.
By extracting the activity category collection in the process model, filtering the original event logs, generating new logs that only contain the activity categories of the model, and using index information to restore the optimal alignment result, reducing the alignment search space.
On the premise of ensuring 100% compliance accuracy, the long track alignment efficiency is significantly improved, the calculation complexity and search time are reduced, and the underlying algorithm is not required to be modified for industrial-grade application scenarios.
Smart Images
Figure CN120494464A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of process mining, and in particular to a business process compliance checking method, system, storage medium and computing device for long traces. Background Art
[0002] Compliance checking aims to discover the differences and commonalities between the behavior of business process event logs and model behavior. Alignment is one of the current standard techniques for compliance checking and can accurately locate deviations between log behavior and model behavior. However, with the increase in the scale and complexity of event logs, especially when the logs contain a large number of events outside the set of process model activities, existing alignment methods face performance bottlenecks when processing long trajectories. To address the state space explosion problem faced by A*-based alignment methods when processing long-trajectory event logs, various improvement schemes have been proposed, among which preprocessing of event logs or process models is one of the popular directions. Through preprocessing techniques, researchers aim to simplify the original event logs or process models to remove irrelevant or redundant data, thereby effectively reducing the search space of the alignment algorithm and reducing the computational complexity. However, if the preprocessing method is inappropriate, it may affect the accuracy of the alignment results, thereby reducing the accuracy of compliance checking. To address the above issues, this patent proposes a business process alignment system for long trajectories. First, the system extracts the set of all activity categories in the process model, filters the original log based on this set, and generates a new log containing only the model activity categories to reduce the alignment search space; then, the filtered log is aligned with the process model to generate an incomplete alignment sequence; finally, the optimal alignment result, that is, the complete compliance check result, is restored based on the index information of the filtered activities. Summary of the Invention
[0003] The first objective of this invention is to overcome the shortcomings and deficiencies of existing compliance checking technologies by proposing a compliance checking method for business processes over long traces. The core concept of this method is to filter and optimize event logs to reduce computational complexity and thus improve alignment efficiency. Building on traditional alignment techniques, this invention improves the efficiency of compliance checking while ensuring the accuracy of the results, thus providing a reliable basis for optimizing and improving business processes.
[0004] The second purpose of the present invention is to provide a business process compliance checking system for long trajectories. The system reduces the complexity of aligning event logs with process models by filtering event logs, and obtains the optimal alignment result, that is, the complete compliance checking result, through restoration technology, thereby improving the efficiency of alignment.
[0005] A third object of the present invention is to provide a storage medium.
[0006] A fourth object of the present invention is to provide a computing device.
[0007] The first object of the present invention is achieved by the following technical solution: a business process compliance checking method for long traces, comprising the following steps:
[0008] 1) Obtain basic data, namely the original event log and the corresponding process model;
[0009] 2) Traversing the process model obtained in step 1) to obtain the activity category set therein, filtering the original event log obtained in step 1) based on the activity category set to obtain the corresponding filtered log and index information of the filtered activities;
[0010] 3) performing a compliance check on the filtered log obtained in step 2) and the process model obtained in step 1) to obtain an incomplete alignment sequence;
[0011] 4) The incomplete alignment sequence obtained in step 3) is restored according to the index information obtained in step 2) to obtain the final complete optimal compliance check result.
[0012] Furthermore, in step 1), the event log is a collection of finite event sequences captured and recorded by the information system when executing the business process. Each finite event sequence is called a trace, which records the information of the activity, including execution time, resources and activity name, and is used to describe and analyze the actual execution of the business process; the process model is a Petri net, which is a four-tuple (P, T, F, M0), where P represents the library set, T represents the transition set, F represents the relationship set, and M0 is the initial state, satisfying: P = P L ∪P M , Represents a logical location. Indicates the message repository; F=F L ∪F M , Indicates the flow relationship of the logical library, Indicates the flow relationship of the message library.
[0013] Furthermore, in step 2), the original event log obtained in step 1) is filtered, and index information of the filtered activities is obtained. The specific situation is as follows:
[0014] Based on the activity set of Petri net, the filtering criteria are defined. Each trace in the original event log is traversed. For each event in the trace, it is determined whether it belongs to the activity of the Petri net. If it does, it is retained in the filtered trace. Otherwise, the position and name of the event in the trace are recorded and stored in the index. If all events of a trace are filtered, that is, the trace is empty after filtering, the index storage of the trace is skipped. Finally, the traces that are not completely filtered and their index information are stored in the filtering log and index structure respectively.
[0015] Furthermore, the compliance check of the filtered log and the process model in step 3) is as follows:
[0016] The compliance check is based on alignment. Alignment is a sequence of movements. The optimal alignment is the one with the least log movement and model movement. The three types of movements are described below:
[0017] a. Synchronous movement: the events recorded in the trace are consistent with the activities of the process model execution sequence;
[0018] b. Model movement: there is a deviation between the trajectory and the execution sequence of the process model, and the execution activities in the process model are skipped;
[0019] c. Log shift: There is a deviation between the execution sequence of the trace and the process model, and the events recorded in the log are skipped;
[0020] In the model movement, when the model execution activity is the invisible transition τ, that is, (>>,τ), although the cost of this movement is 1, it is not included in the subsequent fitness calculation.
[0021] Furthermore, in step 4), the incomplete alignment sequence is restored, which is divided into two parts: restoring the incomplete alignment sequence obtained in step 3) to obtain the optimal alignment sequence; calculating the goodness of fit value of the optimal alignment sequence to obtain the final complete compliance check result. The specific explanation is as follows:
[0022] A) Based on the index information in step 2), the incomplete alignment sequence obtained by aligning the filtered log with the Petri net in step 3) is restored to the optimal alignment sequence. The specific steps are as follows:
[0023] A.1) Traverse each trace in the original event log from step 1) and determine whether there is a trace where all activities have been filtered. If such a trace exists, construct a corresponding alignment sequence and insert it into the incomplete alignment sequence. Ensure that the incomplete alignment sequence is consistent with the original event log in terms of trajectory position, ensuring that it corresponds to the correct trace when the index is restored;
[0024] A.2) Using the index information from step 2), traverse the trace alignment results in the filtered log one by one, inserting the previously filtered activities into the corresponding alignment sequence according to their positions in the index, to obtain the optimal alignment sequence, i.e., the complete compliance check result;
[0025] B) The calculation of the goodness of fit value requires calculating the sum of the log shift in the aligned sequence and the model shift in step A.2), and the calculation formula is shown in formula (1):
[0026]
[0027] Where Fitness represents the fitness value of the alignment sequence, cost represents the number of moves in the alignment sequence, length(trace) represents the length of the trajectory, and shortlength represents the length of the shortest path in the model.
[0028] The second objective of the present invention is achieved through the following technical solution: a business process compliance checking system for long traces, used to implement the above-mentioned business process compliance checking method for long traces, comprising:
[0029] Data acquisition module, used to obtain basic data, namely the original event log and the corresponding process model;
[0030] The filter log and index acquisition module is used to filter the original event log and obtain the index information of the filtered activities;
[0031] A compliance check module is used to perform compliance checks on the filtered logs and the process model to obtain incomplete alignment sequences;
[0032] The optimal alignment result restoration module is used to restore the incomplete alignment sequence to obtain the optimal alignment result, that is, the complete compliance check result.
[0033] The third purpose of the present invention is achieved through the following technical solution: a storage medium stores a program, and when the program is executed by a processor, it implements the above-mentioned business process compliance checking method for long traces.
[0034] The fourth purpose of the present invention is achieved through the following technical solution: a computing device, comprising a processor and a memory for storing processor executable programs, when the processor executes the program stored in the memory, implements the above-mentioned long-trace oriented business process compliance checking method.
[0035] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0036] 1. This paper removes out-of-model activities from event logs through a pre-filtering mechanism and records their index information, significantly reducing the search space in the core alignment phase. While ensuring 100% compliance check accuracy, it improves the efficiency of long trajectory alignment.
[0037] 2. This paper proposes an innovative architecture of "model activity set filtering + index restoration", which decouples off-model activities from core alignment calculations, avoids redundant state space traversal, and effectively copes with resource-constrained scenarios for long-trajectory event logs.
[0038] 3. The present invention is compatible with the existing process mining tool chain through standardized preprocessing, and can adapt to industrial-grade application scenarios without modifying the underlying alignment algorithm, thereby reducing deployment costs.
[0039] 4. The present invention has a wide range of applications in improving and enhancing compliance inspections, and has broad prospects for improving the efficiency of compliance inspections, providing a reference basis for the improvement and optimization of subsequent compliance inspection methods. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 Schematic diagram of the logic flow of the method of the present invention.
[0041] Figure 2 This is a schematic diagram of the case process model.
[0042] Figure 3 This is an architectural diagram of the system of the present invention. DETAILED DESCRIPTION
[0043] The present invention will be further described below with reference to specific embodiments.
[0044] Example 1
[0045] like Figure 1 As shown, this embodiment discloses a business process compliance check method for long traces. First, basic data, namely, original event logs and corresponding process models, are obtained. Then, the process model is traversed to obtain a set of activity categories therein. The original event logs are filtered based on the set of activity categories to obtain corresponding filtered logs and index information of the filtered activities. The filtered logs are subjected to a compliance check with the process model to obtain an incomplete alignment sequence. Finally, the incomplete alignment sequence is restored based on the obtained index information to obtain an optimal alignment result, namely, a complete compliance check result. The method includes the following steps:
[0046] 1) Obtain basic data, namely the original event log and the corresponding process model; the event log is a collection of finite event sequences captured and recorded by the information system when executing the business process. Each finite event sequence is called a trace, which records information such as the execution time, resources, and activity name of the activity, and is used to describe and analyze the actual execution of the business process. The process model is a Petri net, which is a four-tuple (P, T, F, M0), where P represents the library set, T represents the transition set, F represents the relationship set, and M0 is the initial state, satisfying: P = P L ∪P M , Represents a logical location. Indicates the message repository; F=F L ∪F M , Indicates the flow relationship of the logical library, Indicates the flow relationship of the message library.
[0047] Using the above steps, input the sample event log as L=[<a,x,b,y,c> , <a,c,e,z<,<a,b,c,e> ,<x,y,z,t> ,<x,a,c,e,t> ], the corresponding example process model is a Petri net, such as Figure 2 As shown, the four-tuple of the Petri net is represented as: P = {Start, p1, p2, p3, Final}, T = {a, b, c, d, e}, F = {(Start, a), (Start, b), (a, p1), (b, p1), (p1, c), (c, p2), (c, p3), (p2, d), (p3, e), (d, Final), (e, Final)}, M0 = {Start}.
[0048] 2) Obtaining filtered logs and indexes: Traverse the process model obtained in step 1) to obtain the activity category set therein. Based on the activity category set, filter the original event log obtained in step 1) to obtain the corresponding filtered logs and index information of the filtered activities. The specific explanation is as follows:
[0049] Filtering criteria are defined based on the Petri net's activity set, and each trace in the original event log is traversed. For each event in the trace, it is determined whether it belongs to a Petri net activity. If so, it is retained in the filtered trace; otherwise, the event's position and name in the trace are recorded and stored in the index. If all events in a trace are filtered (i.e., empty after filtering), the index storage for that trace is skipped. Finally, traces that have not been fully filtered and their index information are stored in the filter log and index structure, respectively.
[0050] Using the above steps, we traverse the transitions of the input Petri net and obtain its activity set Set = [a, b, c, d, e]. After determining the activity set of the model, we traverse the trajectory of the event log L and determine whether the activity in the trajectory is in the activity set Set. If so, we add it to the filtered log trajectory, otherwise, we record the activity information in the index. For example, if we traverse to the trajectory<a,x,b,y,c> When it is determined that activities x and y in the trajectory are not in the activity set Set, then these two activities are recorded in the index {0: [(1,x), (3,y)]}, and the trajectory of the filtered log is<a,b,c> Traverse to the last trace of the log and get the filtered log as filtered_log=[<a,b,c> ,<a,c,e> ,<a,b,c,e> ,<a,c,e> ], the index information is index = {0: [(1, x), (3, y)], 1: [(3, z)], 4: [(0, x), (4, t)]}. It should be noted that when all activities in the trajectory are not in the activity set, there is no need to record the index information. For example, the trajectory in event log L<x,y,z,t> .
[0051] 3) Performing a compliance check on the filtered log obtained in step 2) and the process model obtained in step 1) yields an incomplete alignment sequence, as explained below:
[0052] The compliance check between the log and the process model is based on alignment. The alignment is a sequence of movements; the optimal alignment is the one that minimizes log and model movements. The three types of movements are described below:
[0053] a. Synchronous movement: the events recorded in the trace are consistent with the activities of the process model execution sequence;
[0054] b. Model movement: there is a deviation between the trajectory and the execution sequence of the process model, and the execution activities in the process model are skipped;
[0055] c. Log shift: There is a deviation between the execution sequence of the trace and the process model, and the recorded events in the log are skipped;
[0056] In the model movement, when the model execution activity is the invisible transition τ, that is, (>>,τ), although the cost of this movement is 1, it is not included in the subsequent fitness calculation.
[0057] Using the above steps, the filtered log obtained in step 2) is filtered_log=[<a,b,c> ,<a,c,e> ,<a,b,c,e> ,<a,c,e> ], enter the Petri net such as Figure 2The filtered log is aligned with the Petri net to obtain an incomplete alignment sequence incomplete_result = [<(a,a),(b,>>),(c,c),(>>,d),(>>,e)>,<(a,a),(c,c),(>>,d),(e,e)>,<(a,a),(b,>>),(c,c),(d,>>),(e,e)>,<(a,a),(c,c),(>>,d),(e,e)>].
[0058] 4) Restoring the incomplete alignment sequence is divided into two parts: restoring the incomplete alignment sequence obtained in step 3) to obtain the optimal alignment sequence; calculating the goodness of fit value of the optimal alignment sequence to obtain the final complete compliance check result. The specific explanation is as follows:
[0059] A) Based on the index information obtained in step 2), the incomplete alignment sequence obtained in step 3) is restored to obtain a complete optimal alignment result, i.e., a complete compliance check result, which is specifically explained as follows:
[0060] A.1) Traverse each trace in the original event log from step 1) and determine whether there is a trace where all activities have been filtered. If such a trace exists, construct a corresponding alignment sequence and insert it into the incomplete alignment sequence. This ensures that the incomplete alignment sequence is consistent with the original event log in terms of trajectory position, ensuring that it corresponds to the correct trace when the index is restored.
[0061] A.2) Use the index information in step 2) to traverse the trajectory alignment results in the filtered log one by one, and insert the previously filtered activities into the corresponding alignment sequence according to the position in the index to obtain the optimal alignment sequence.
[0062] B) The calculation of the goodness of fit requires calculating the sum of the log shift in the aligned sequence in step A.2) and the model shift, and the calculation formula is shown in formula (1):
[0063]
[0064] Where Fitness represents the fitness value of the alignment sequence, cost represents the number of moves in the alignment sequence, length(trace) represents the length of the trajectory, and shortlength represents the length of the shortest path in the model.
[0065] Using the above steps, first determine whether the entire track in the original event log has been filtered. If so, directly insert the alignment result corresponding to the track into the incomplete alignment result. Traverse the track in the event log L and determine whether the track is found.<x,y,z,t> All activities in are filtered out, so the alignment result <(x,>>),(y,>>),(z,>>),(t,>>),(>>,a),(>>,c),(>>,d),(>>,e)> corresponding to this trajectory is inserted into the incomplete alignment result, and the insertion position is consistent with the position of this trajectory in the log L.
[0066] The incomplete alignment sequence at this time is incomplete_result = [<(a,a),(b,>>),(c,c),(>>,d),(>>,e)>,<(a,a),(c,c),(>>,d),(e,e)>,<(a,a),(b,>>),(c,c),(>>,d),(e,e)>,<(x,>>),(y,>>),(z,>>),(t,>>),(>>,a),(>>,c),(>>,d),(>>,e)>,<(a,a),(c,c),(>>,d),(e,e)>]. Then, we start traversing the information in the index obtained in step 2). For example, if the first index traversal is {0:[(1,x),(3,y)]}, then (x,>>) is inserted at the second position of the first alignment sequence, and (y,>>) is inserted at the fourth position. During these insertions, it is necessary to confirm whether the incomplete alignment sequence has model shifts. If so, the insertion position needs to be shifted back. The restored result of the first alignment sequence is: <(a,a),(x,>>),(b,>>),(y,>>),(c,c),(>>,d),(>>,e)>. Traversing the index information in sequence, the final restored complete alignment result is final_result=[<(a,a),(x,>>),(b,>>),(y,>>),(c,c),(>>,d),(>>,e)>,<(a,a),(c,c),(>>,d),(e,e),(z,>>)>,<(a,a),(b,>>),(c,c),(>>,d),(e,e)>,<(x,>>),(y,>>),(z,>>),(t,>>),(>>,a),(>>,c),(>>,d),(>>,e)>,<(x,>>),(a,a),(c,c),(>>,d),(e,e),(t,>>)>]
[0067] The final restored complete alignment result final_result=[<(a,a),(x,>>),(b,>>),(y,>>),(c,c),(>>,d),(>>,e)>,<(a,a),(c,c),(>>,d),(e,e),(z,>>)>,<(a,a),(b,>>),(c,c),(>>,d),(e,e)>,<(x,>>) Take the first alignment sequence <(a,a),(x,>>),(b,>>),(y,>>),(c,c),(>>,d),(>>,e)> in [ ,(y,>>),(z,>>),(t,>>),(>>,a),(>>,c),(>>,d),(>>,e)>,<(x,>>),(a,a),(c,c),(>>,d),(e,e),(t,>>)>] as an example. The number of log moves in this alignment sequence is 3, the number of model moves is 2, and the trajectory length is 5. The corresponding model is Figure 2 The Petri net shown in the figure has a shortest path length of 4. Therefore, using formula (1), the fitting formula of this trajectory in the model is:
[0068] Example 2
[0069] This embodiment discloses a business process compliance checking system for long traces, which is used to implement the business process compliance checking method for long traces described in Example 1. Figure 3 As shown, the system includes the following functional modules:
[0070] Data acquisition module, used to obtain basic data, namely the original event log and the corresponding process model;
[0071] The filter log and index acquisition module is used to filter the original event log and obtain the index information of the filtered activities;
[0072] A compliance check module is used to perform compliance checks on the filtered logs and the process model to obtain incomplete alignment sequences;
[0073] The optimal alignment result restoration module is used to restore the incomplete alignment sequence to obtain the optimal alignment result, that is, the complete compliance check result.
[0074] Example 3
[0075] This embodiment discloses a storage medium storing a program. When the program is executed by a processor, the long-trace-oriented business process compliance checking method described in Embodiment 1 is implemented.
[0076] The storage medium in this embodiment can be a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), a USB flash drive, a mobile hard disk, or the like.
[0077] Example 4
[0078] This embodiment discloses a computing device, including a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the long-trace business process compliance checking method described in Example 1 is implemented.
[0079] The computing device described in this embodiment may be a desktop computer, a laptop computer, a smart phone, a PDA handheld terminal, a tablet computer, a programmable logic controller (PLC), or other terminal devices with a processor function.
[0080] In summary, by employing the above scheme, this invention provides a new method and system for compliance checking. By filtering and optimizing event logs, this method reduces computational complexity while ensuring 100% compliance accuracy, thereby improving alignment efficiency. In particular, when checking compliance on long-trajectory event logs, the size of the cumulative network state space is successfully reduced, significantly reducing search time. This feature not only enhances the algorithm's practical application value but also demonstrates its potential adaptability and scalability in addressing complex business processes.
[0081] The above-described embodiments are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, any changes made based on the shape and principle of the present invention should be included in the scope of protection of the present invention.
Claims
1. A business process compliance checking method for long traces, characterized in that: The following steps are involved: 1) Obtain basic data, namely the original event log and the corresponding process model; 2) Traversing the process model obtained in step 1) to obtain the activity category set therein, filtering the original event log obtained in step 1) based on the activity category set to obtain the corresponding filtered log and index information of the filtered activities; 3) performing a compliance check on the filtered log obtained in step 2) and the process model obtained in step 1) to obtain an incomplete alignment sequence; 4) The incomplete alignment sequence obtained in step 3) is restored according to the index information obtained in step 2) to obtain the final complete optimal compliance check result.
2. A business process compliance checking method for long traces according to claim 1, characterized in that: In step 1), the event log is a collection of finite event sequences captured and recorded by the information system when executing the business process. Each finite event sequence is called a trace, which records activity information, including execution time, resources, and activity name, and is used to describe and analyze the actual execution of the business process. The process model is a Petri net, which is a four-tuple (P, T, F, M0), where P represents the set of libraries, T represents the set of transitions, F represents the set of relations, and M0 is the initial state, satisfying: Represents a logical location. Indicates the message repository; Indicates the flow relationship of the logical library, Indicates the flow relationship of the message library.
3. A business process compliance checking method for long traces according to claim 2, characterized in that: In step 2), the original event log obtained in step 1) is filtered, and index information of the filtered activities is obtained. The details are as follows: Define the filtering criteria based on the activity set of the Petri net, traverse each trace in the original event log, and for each event in the trace, determine whether it belongs to the activity of the Petri net. If it does, keep it in the filtered trace. Otherwise, record the position and name of the event in the trace and store it in the index. If all events of a trajectory are filtered, that is, empty after filtering, the index storage of the trajectory is skipped; finally, the trajectory that is not completely filtered and its index information are stored in the filtering log and index structure respectively.
4. A business process compliance checking method for long traces according to claim 3, characterized in that: The compliance check between the filtered log and the process model in step 3) is as follows: The compliance check is based on alignment. Alignment is a sequence of movements. The optimal alignment is the one with the least log movement and model movement. The three types of movements are described below: a. Synchronous movement: the events recorded in the trace are consistent with the activities of the process model execution sequence; b. Model movement: there is a deviation between the trajectory and the execution sequence of the process model, and the execution activities in the process model are skipped; c. Log shift: There is a deviation between the execution sequence of the trace and the process model, and the events recorded in the log are skipped; In the model movement, when the model execution activity is the invisible transition τ, that is, (>>,τ), although the cost of this movement is 1, it is not included in the subsequent fitness calculation.
5. A long-track business process compliance checking method according to claim 4, characterized in that: In step 4), the incomplete alignment sequence is restored, which is divided into two parts: restoring the incomplete alignment sequence obtained in step 3) to obtain the optimal alignment sequence; calculating the goodness of fit value of the optimal alignment sequence to obtain the final complete compliance check result. The specific explanation is as follows: A) Based on the index information in step 2), the incomplete alignment sequence obtained by aligning the filtered log with the Petri net in step 3) is restored to the optimal alignment sequence. The specific steps are as follows: A.1) Traverse each trace in the original event log from step 1) and determine whether there is a trace where all activities have been filtered. If such a trace exists, construct a corresponding alignment sequence and insert it into the incomplete alignment sequence. Ensure that the incomplete alignment sequence is consistent with the original event log in terms of trajectory position, ensuring that it corresponds to the correct trace when the index is restored; A.2) Using the index information from step 2), traverse the trace alignment results in the filtered log one by one, inserting the previously filtered activities into the corresponding alignment sequence according to their positions in the index, to obtain the optimal alignment sequence, i.e., the complete compliance check result; B) The calculation of the goodness of fit value requires calculating the sum of the log shift in the aligned sequence and the model shift in step A.2), and the calculation formula is shown in formula (1): Where Fitness represents the fitness value of the alignment sequence, cost represents the number of moves in the alignment sequence, length(trace) represents the length of the trajectory, and shortlength represents the length of the shortest path in the model.
6. A business process compliance checking system for long traces, characterized by: A long-trace business process compliance checking method for implementing any one of claims 1 to 5, comprising: Data acquisition module, used to obtain basic data, namely the original event log and the corresponding process model; The filter log and index acquisition module is used to filter the original event log and obtain the index information of the filtered activities; A compliance check module is used to perform compliance checks on the filtered logs and the process model to obtain incomplete alignment sequences; The optimal alignment result restoration module is used to restore the incomplete alignment sequence to obtain the optimal alignment result, that is, the complete compliance check result.
7. A storage medium storing a program, characterized in that: When the program is executed by a processor, the business process compliance checking method for long traces according to any one of claims 1 to 5 is implemented.
8. A computing device comprising a processor and a memory for storing a program executable by the processor, characterized in that: When the processor executes the program stored in the memory, the long-trace business process compliance checking method according to any one of claims 1 to 5 is implemented.