Layered business process model quality evaluation method and device based on alignment

Abstract

The invention provides a hierarchical business process model quality evaluation method and device based on alignment, and belongs to the technical field related to large model quality evaluation, the fitting degree of each level is calculated according to a hierarchical alignment sequence HAT obtained by alignment of a hierarchical process model HPN and a hierarchical event log HL, and the overall fitting degree is obtained through weighted average; constructing a hierarchical prefix accuracy automaton according to the hierarchical alignment sequence HAT, obtaining the accuracy of each level, and obtaining the overall accuracy through weighting; and performing weighted average according to the overall fitting degree and the overall accuracy to obtain a comprehensive score F-value of the hierarchical business process model, and evaluating the quality of the hierarchical business process model according to the comprehensive score F-value, so that not only can a complex hierarchical structure in the hierarchical process model be effectively identified, but also the quality of the hierarchical process model can be comprehensively and accurately evaluated.

Description

Technical Field

[0001] The present invention relates to the technical field related to large model quality evaluation, and in particular to a method and device for evaluating the quality of a hierarchical business process model based on alignment. Background Art

[0002] Process mining is a new research hotspot in the field of Business Process Management (BPM). The main scenarios include: model mining, compliance checking and improvement. BPM usually starts with a process model, which is a graphical representation of the business process within an organization, describing the sequence and interaction of business activities, decisions and information flows. Compliance checking can compare and correlate event logs with process models, diagnose deviations between models and logs, and is an important research direction of process mining. Alignment is one of the most commonly used means of compliance checking and can be used to evaluate the quality of process models. Flat process models can describe event activities at the same level of abstraction, while hierarchical process models can deeply mine the nested relationships between tasks and express multiple complex patterns such as multiple instances and sub-processes, playing an important role in hierarchical business process modeling.

[0003] Hierarchical business process models are used for a variety of purposes in BPM, including analysis, formulation, redesign, and process improvement. Flat process models can describe event activities at the same level of abstraction. Unlike traditional flat process models, hierarchical process models are process models with complex structures such as multiple instances and sub-processes. They can be used to describe the calling relationship between outsourced processes (parent processes) and sub-process activities in cross-organizational outsourcing scenarios, and better reflect the hierarchy and complexity of business processes. The hierarchical structure of the hierarchical process model can improve the comprehensibility and maintainability of the model, while also reducing the complexity of modeling. For hierarchical process models, the corresponding quality evaluation method technology is not yet mature enough.

[0004] Model quality evaluation indicators include: fit, precision, generalization and simplicity, among which: fit evaluates the degree to which the process model can reproduce the traces recorded in the log, and evaluates the similarity between the traces generated by the model and the activity sequences recorded in the real log when simulating or executing the business process. By using this indicator, enterprises can better understand the effectiveness of the process model, find potential improvement points, and optimize the business process to improve the compliance of the model with the actual operation. Fit plays an important role in business process management and optimization. Precision measures the behavior allowed in the process model, which is impossible in the behavior observed in the event log. The purpose of evaluating precision is to improve the flexibility of the model and better adapt to possible changes and abnormal situations in the business process. Through precision, the model can more comprehensively capture potential behavior patterns while maintaining the compliance of the observed behavior. Only focusing on fit may lead to a flower model, while only focusing on precision may lead to overfitting. Therefore, the F-value evaluation method is introduced to measure the quality of the process model, strike a balance between fit and precision, unify fit and precision, and comprehensively evaluate the quality of the hierarchical process model.

[0005] In the field of process mining, compliance checking can compare and correlate event logs with process models and diagnose deviations between models and logs, which is very important for measuring the quality of process models discovered from event logs. In order to measure the effectiveness of process models in describing observed behaviors, different quality indicators can be used. Commonly used quality indicators include: (1) fit, (2) accuracy, (3) generalization, and (4) simplicity.

[0006] The fit describes the proportion of behavior in the event log that can be replayed by the process model. There are several different ways to measure this quality metric. Most of the time, other quality metrics need to be investigated only when the fit between the log and the model is relatively high. Weijters et al. proposed the PM metric to quantify the fit between the event log and the process model discovered from the log. This coarse-grained evaluation method is used to quantify the percentage of traces that can be replayed in the discovered process model. Aalst proposed a goodness of fit check between event logs and process models based on footprint matrices. This method is only meaningful if the log is complete with respect to "directly follows" relationships between activities. De Leoni et al. proposed a method to quantify the fit between the declarative model and the event log, taking into account the cost of executing activities in the log that are not allowed according to the model and vice versa.

[0007] Precision evaluates the extent to which the behavior of the process model is allowed that is not observed in the event log. Since the process model may have infinite behavior, in this case, only precision can be estimated. Carmona et al. proposed several methods to improve precision, constructing a partial state space when replaying the event log on the process model, and calculating the unenabled escape edges from the visited states. The higher the proportion of unused edges, the more additional behaviors are allowed, and the lower the precision of the model. Munoz-Gama et al. proposed the ETCPrecision metric, which takes into account the frequency of escape edges and introduces the concept of escape edges based on prefix automata to measure the precision between event logs and process models. Only focusing on fit may lead to flower models, while only focusing on precision may lead to overfitting. Therefore, the F-value evaluation method is introduced to measure the quality of the process model, strike a balance between fit and precision, and unify fit and precision. However, it is not applicable to models containing sub-process behaviors. Summary of the invention

[0008] The embodiment of the present invention provides a method and device for evaluating the quality of a hierarchical business process model based on alignment, aiming to solve the problem that existing quality evaluation methods cannot efficiently identify and process complex hierarchical structures. The method uses an alignment-based fit evaluation method, an accuracy evaluation method and a weighted average calculation model F-value of the two to comprehensively evaluate the quality of the hierarchical process model on the basis of obtaining an alignment sequence based on the alignment of the hierarchical process model and the event log. The method can not only effectively identify the complex hierarchical structure in the hierarchical process model, but also comprehensively and accurately evaluate the quality of the hierarchical process model.

[0009] In a first aspect, the present invention provides a method for evaluating the quality of a hierarchical business process model based on alignment, comprising:

[0010] Hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL HAT Calculate the goodness of fit at each level and obtain the overall goodness of fit by weighted average;

[0011] Hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL HAT Construct a hierarchical prefix accuracy automaton to obtain the accuracy of each level and get the overall accuracy by weighting;

[0012] The comprehensive score F-value of the hierarchical business process model is obtained by taking a weighted average based on the overall fit and the overall accuracy, and the quality of the hierarchical business process model is evaluated based on the comprehensive score F-value.

[0013] Optionally, the comprehensive score F-value of the hierarchical business process model obtained by weighted average of the overall fit and the overall accuracy includes:

[0014] According to the overall fit and overall accuracy, the formula The weighted average is used to obtain the comprehensive score F-value of the hierarchical business process model, where wfit h (HL,HPN i ,orc,sc) for use and w i The weighted goodness of fit between the hierarchical event log HL and the hierarchical process model HPN is calculated; prec(HPA) is the overall accuracy between the hierarchical event log HL and the hierarchical process model HPN.

[0015] Optionally, the step of calculating the degree of fit of each layer according to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, and obtaining the overall degree of fit by weighted average includes:

[0016] According to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, the formula is used Calculate the hierarchical weighted fit wfit for each level hi (L i ,N i ,orc,sc), where mlf hi (L i ,N i ,orc,sc) is L i and N i The hierarchical log fit, L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, mmf hi (L i ,N i ,orc,sc) is L i and N i The moving fit of the hierarchical model;

[0017] Based on hierarchical weighted fitting wfit hi (L i ,N i ,orc,sc), using L i and N i The relative fit between the levels and the weight w of the model fit of the i-th layer of the hierarchical process model i Calculate the weighted fit wfit of the hierarchical event log HL and hierarchical process model HPNh (HL,HPN i ,orc,sc), where the weighted fit wfit h (HL,HPN i ,orc,sc)=∑w i ·wfit hi (L i ,N i ,orc,sc).

[0018] Optionally, use the formula

[0019]

[0020] Calculate L i and N i The hierarchical log fit mlf hi (L i ,N i ,orc,sc), where L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, γ i YesL i and N i The corresponding HAT layer i sub-alignment sequence, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, w i is the weight of the i-th layer process model, σ is a finite set of event sequences in the event space ξ, A is a finite set of activity labels, and T is a finite set of transitions.

[0021] Optionally, use the formula

[0022]

[0023] Calculate L i and N i The moving average fit of the hierarchical model mmf hi (L i, N i, orc,sc), where L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, γ i YesL i and N i The corresponding HAT layer i sub-alignment sequence, sc is the cost function, orc is L i and N iCompared with the Oracle function of sc, w i is the weight of the i-th layer process model, σ is a task sequence with life cycle information, A is a set of activity labels, and T is a transition set.

[0024] Optionally, constructing a hierarchical prefix accuracy automaton based on the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, obtaining the accuracy of each level, and obtaining the overall accuracy by weighting includes:

[0025] According to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, a hierarchical prefix accuracy automaton is constructed using the formula Calculate the accuracy of the i-th layer between the hierarchical process model HPN and the hierarchical event log HL, where is a set of active labels, T is a transition set, HPN=(PN N0, HPN(PN N0 )) is a hierarchical process model on A, Tr={T n0 ,T n1 ,...T nn} is its nested transition set, HPA=(PA N0 ,HPA(PA N0 )) is a hierarchical prefix accuracy automaton of HPN, Sa is the state set, SE is the escape state set, EA is a set of directed arcs with labels between states, and wgt is a weight function that maps states to real numbers;

[0026] According to the formula prec(HPA) = ∑W i ×prec(HPA i ) is weighted to obtain the overall accuracy, where w i is the weight of the i-th layer process model in the hierarchical process model, prec(HPA i ) is the accuracy of the i-th layer between the hierarchical process model HPN and the hierarchical event log HL.

[0027] Optionally, the weight w of the i-th layer process model of the hierarchical process model i Using formula Calculate, where Sa is the set of states, EA is the set of labeled directed arcs between states, and wgt is the weight function that maps states to real numbers.

[0028] Optionally, a hierarchical prefix accuracy automaton PA is constructed based on the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL. Nni Is the nested state sa iThe corresponding prefix-accurate automaton with nested states, where HAN(PA N0 )={(sa i ,PA Nni ,HAN(PA Nni ))∣sa i ∈SA n0}, is a set of active labels, HPN=(PN N0, HPN(PN N0 )) is a hierarchical process model on A, Tr={T n0 ,T n1 ,...T nn} is its nested transition set, HPA=(PA N0, HPA(PA N0 )) is a hierarchical prefix accuracy automaton for HPN, PA N0 is the root node prefix accuracy automaton of HPA, SA N0 is the root node set, HPA(PA N0 ) is PA N0 The corresponding subprefix accuracy automaton.

[0029] In a second aspect, the present invention further provides an alignment-based hierarchical business process model quality evaluation device comprising:

[0030] The fit calculation module is configured to obtain the hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL. HAT Calculate the goodness of fit at each level and obtain the overall goodness of fit by weighted average;

[0031] The accuracy calculation module is configured to obtain a hierarchical alignment sequence based on the alignment of the hierarchical process model HPN with the hierarchical event log HL. HAT Construct a hierarchical prefix accuracy automaton to obtain the accuracy of each level and get the overall accuracy by weighting;

[0032] The model evaluation module is configured to obtain a comprehensive score F-value of the hierarchical business process model based on a weighted average of the overall fit and the overall accuracy, and evaluate the quality of the hierarchical business process model based on the comprehensive score F-value.

[0033] In a third aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer execution instructions, and when the computer execution instructions are executed by a processor, they are used to implement the above-mentioned alignment-based hierarchical business process model quality evaluation method.

[0034] Beneficial technical effects of the present invention:

[0035] The present invention provides a method and device for evaluating the quality of a hierarchical business process model based on alignment, which is based on the hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL. HAT Calculate the fit of each level and get the overall fit by weighted average; align the hierarchical process model HPN with the hierarchical event log HL to get the hierarchical alignment sequence HAT A hierarchical prefix accuracy automaton is constructed to obtain the accuracy of each level, and the overall accuracy is obtained by weighting; the comprehensive score F-value of the hierarchical business process model is obtained by weighted average of the overall fit and the overall accuracy, and the quality of the hierarchical business process model is evaluated based on the comprehensive score F-value. Based on the alignment sequence obtained by aligning the hierarchical process model with the event log, the alignment-based fit evaluation method, the accuracy evaluation method and the weighted average of the two are used to calculate the F-value of the model to comprehensively evaluate the quality of the hierarchical process model. This method can not only effectively identify the complex hierarchical structure in the hierarchical process model, but also comprehensively and accurately evaluate the quality of the hierarchical process model. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 A schematic diagram of the basic framework of a method for evaluating the quality of a hierarchical business process model based on alignment provided by the present invention;

[0038] Figure 2 A schematic diagram of a hierarchical process model and a hierarchical event log provided by the present invention;

[0039] Figure 3 A schematic diagram of another hierarchical process model and hierarchical event log provided by the present invention;

[0040] Figure 4 A schematic diagram of an optimally aligned hierarchical alignment sequence provided by the present invention;

[0041] Figure 5 A schematic diagram of an optimally aligned hierarchical prefix accuracy automaton provided by the present invention;

[0042] Figure 6 Schematic diagram of all optimal alignments between the trajectories σ1= and σ6=<g> provided by the present invention and the hierarchical process model;

[0043] Figure 7Hierarchical prefix accuracy automata for all optimal alignments provided by the present invention;

[0044] Figure 8 All optimal alignment representatives between the trajectories σ1= and σ6=<g> provided by the present invention and the hierarchical process model;

[0045] Fig. 9 All optimal alignments of hierarchical process models with hierarchical event logs provided for the present invention represent hierarchical prefix precision automata. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] It should be understood that in various embodiments of the present invention, the size of the sequence number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0048] It should be understood that in the present invention, "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0049] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0050] With the development of information technology, many business processes and systems have become more and more complex, usually showing multi-level and multi-dimensional characteristics. These complex hierarchical structures often make it difficult for traditional quality evaluation methods to cope with them in practical applications, resulting in inaccurate and incomplete model evaluation results, which affects the reliability and application effect of the hierarchical process model. When facing a complex hierarchical structure, the existing quality evaluation method usually adopts the method of converting the hierarchical process model into a flat process model for quality evaluation. This method may be effective when dealing with simple or medium-complexity models, but it is inefficient when facing large and complex models. In order to solve this technical problem, an embodiment of the present invention provides a hierarchical business process model quality evaluation method and device based on alignment, based on the alignment sequence obtained by aligning the hierarchical process model with the event log, the hierarchical process model fit, precision and F-value quality indicators are formally defined, and a hierarchical process model quality evaluation method is proposed, that is, a fit evaluation method based on alignment, a precision evaluation method and a weighted average calculation of the F-value of the model are performed on the two, so as to comprehensively evaluate the quality of the hierarchical process model. This method can not only effectively identify the complex hierarchical structure in the hierarchical process model, but also comprehensively and accurately evaluate the quality of the hierarchical process model. It can be widely used in the quality assessment of hierarchical process models such as timestamps, resources, and life cycle information.

[0051] refer to Figure 1 , Figure 2 , Figure 3 and Figure 4 As shown, the embodiment of the present invention provides a method for evaluating the quality of a hierarchical business process model based on alignment, comprising the following steps:

[0052] Step 100: Calculate the fit of each level according to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, and obtain the overall fit by weighted average.

[0053] refer to Figure 1 As shown, a quality evaluation method (HQE) for the hierarchical process model is defined based on the alignment between the hierarchical process model and the event log, including an alignment-based fit evaluation method, an alignment-based accuracy evaluation method and an alignment-based F-value evaluation method.

[0054] The hierarchical alignment sequence between the event log trajectory and the hierarchical process model network provides an explicit mapping of each level between the activities in the trajectory and the transitions in the hierarchical process model, so that the activities in the event log are mapped to the transitions in the hierarchical process model in a deterministic way. It is also based on this that the embodiment of the present invention uses this relationship to quantify the fit, accuracy and F-value of the hierarchical process model. Among them, the hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL is HATAs input, the corresponding fitting degree of each level is obtained, and the overall fitting degree is obtained by weighted average.

[0055] L i ∈B(A i * ) is the layered event log HL, the i-th layer sub-log, A is a finite set of active labels. i =(P,T,F;M) is L i The corresponding HPN layer i sub-model. fit i is the fit of the i-th layer of the hierarchical process model HPN, as shown in formula (1)

[0056]

[0057] Among them, sum(σ i ) represents the number of trajectories that can be reproduced by the i-th layer sub-model of the hierarchical process model HPN, sum(L i ) represents the number of trajectories in the i-th sub-log of the hierarchical event log HL, and sum(L) represents the number of trajectories in the hierarchical event log HL. i is the weight of the i-th layer model of the hierarchical process model, and the weight is determined according to the proportion of the trajectory in the event log corresponding to the layer model in the trajectory in the event log corresponding to the overall hierarchical process model, and ∑w i = 1. As shown in formula (2).

[0058]

[0059] Using w i and fit i Calculate the fit of the hierarchical process model fit h , as shown in formula (3).

[0060] fit h =∑w i ×fit i (3)

[0061] L i ∈B(A i * ) is the layered event log HL, the i-th layer sub-log, A is a finite set of active labels. i =(P,T,F;M) is L i The corresponding HPN layer i sub-model. i YesL i and N i The corresponding HAT layer i sub-alignment sequence. sc:(A i ＞＞ ×Ti ＞＞ )→IR is the cost function, orc:{σ∈L i}→((A i ＞＞ ×T i ＞＞ ) * →IP) is L i and N i Compared with the optimal Oracle function of sc, w i is the weight of the i-th layer process model. Use orc and sc to calculate L i and N i The relative fit between the levels As shown in formula (4).

[0062]

[0063] The relative fitting degree of each layer of the hierarchical process model is obtained by taking full account of the hierarchical structure and nested structure of the hierarchical process model. and w i The relative fit between the hierarchical event log HL and the hierarchical process model HPN is calculated as shown in formula (5).

[0064]

[0065] w i Represents the weight of the fit of the i-th layer. The overall relative fit of the hierarchical process model is between 0 and 1. Moreover, ∑w i =1, we can get

[0066] L i ∈B(A i * ) is the HL layer i sub-log, A is a finite set of active labels. i =(P,T,F;M) is L i The corresponding HPN layer i sub-model. i YesL i and N i The corresponding HAT layer i sub-alignment sequence. sc:(A i ＞＞ ×T i ＞＞ )→IR is the cost function, orc:{σ∈L i}→((A i ＞＞ ×T i ＞＞ ) * →IP) is Li and N i Compared with the optimal Oracle function of sc, w i is the weight of the i-th layer process model. Use orc and sc to calculate L i and N i The hierarchical log fit mlf hi (L i ,N i ,orc,sc), as shown in formula (6).

[0067]

[0068] Among them, L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, γ i YesL i and N i The corresponding HAT layer i sub-alignment sequence, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, w i is the weight of the i-th layer process model, σ is a finite set of event sequences in the event space ξ, A is a finite set of activity labels, and T is a finite set of transitions.

[0069] For the obtained hierarchical log moving goodness of fit, on the basis of fully considering the hierarchical structure and nested structure of the hierarchical process model, and w i The log moving goodness of fit of the hierarchical event log HL and the hierarchical process model HPN is calculated as shown in formula (7).

[0070]

[0071] Similarly, using orc sc Calculate L with sc i and N i The moving average fit of the hierarchical model mmf hi (L i ,N i ,orc,sc), as shown in formula (8).

[0072]

[0073] For the obtained moving fit of the hierarchical model, on the basis of fully considering the hierarchical structure and nested structure of the hierarchical process model, and w iThe model shift fit of the hierarchical event log HL and the hierarchical process model HPN is calculated as shown in formula (9).

[0074]

[0075] N i =(P,T,F;M) is L i The corresponding HPN layer i sub-model. i YesL i and N i The corresponding HAT layer i sub-alignment sequence. sc:(A i ＞＞ ×T i ＞＞ )→IR is the cost function, orc:{σ∈L i}→((A i >>×T i >>) * →IP) is L i and N i Compared with the optimal Oracle function of sc, w i is the weight of the i-th layer process model. Use orc and sc to calculate L i and N i The hierarchical weighted fitting degree of is shown in formula (10).

[0076]

[0077] That is, the hierarchical alignment sequence HAT is obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, using the formula Calculate the hierarchical weighted fit wfit for each level hi (L i ,N i ,orc,sc), where wfit hi (L i ,N i ,orc,sc) is L i and N i The hierarchical log fit, L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, mmf hi (L i ,N i ,orc,sc) is L i and N iThe moving goodness of fit of the hierarchical model.

[0078] Based on hierarchical weighted fitting wfit hi (L i ,N i ,orc,sc), using L i and N i The relative fit between the levels and the weight w of the model fit of the i-th layer of the hierarchical process model i Calculate the weighted fit wfit of the hierarchical event log HL and hierarchical process model HPN h (HL,HPN i ,orc,sc), where the weighted fit wfit h (HL,HPN i ,orc,sc)=∑w i ·wfit hi (L i ,N i ,orc,sc).

[0079] That is, for the obtained hierarchical weighted fitting degree, use and w i The weighted fit of the hierarchical event log HL and the hierarchical process model HPN is calculated as shown in formula (11).

[0080]

[0081] by Figure 2 Take the hierarchical event log and hierarchical process model as an example. Use the Oracle function to map each trajectory in the log to the optimal alignment between the trajectory and the model. Use the likelihood cost function to construct the Oracle function and measure the degree of deviation. Using the method shown in step 100, obtain Figure 2 The top model PN in the hierarchical process model N0 The fit degree with the root log rootLog is 0.98, PN N1 and Nlog a The fit is 0.94, PN N2 and Nlog b The fitting degree is 1, and the overall relative fitting degree is rfit h =0.98×2 / 4+0.94×1 / 4+1×1 / 4=0.975. mlf at each level between log and model hi and mmf hi They are 0.98 and 0.96, 1 ​​and 0.97, and 0.94 and 1 respectively. Figure 2 The overall mlf between the hierarchical event log and the hierarchical process model shown h and mmfh They are 0.98 and 0.97 respectively. Figure 2 The hierarchical wfit between the hierarchical event log and the hierarchical process model shown hi They are 0.97, 0.95 and 1 respectively. Figure 2 The overall wfit between the hierarchical event log and the hierarchical process model shown h It is 0.97. From the calculation results, it can be seen that the top model PN has a relatively low degree of fit N0 and PN N1 The evaluation method of the embodiment of the present invention can provide key clues for the optimization and improvement of the hierarchical process model and provide a direction for model optimization.

[0082] Step 200: Obtain a hierarchical alignment sequence by aligning the hierarchical process model HPN with the hierarchical event log HL HAT A hierarchical prefix accuracy automaton is constructed to obtain the accuracy of each level and the overall accuracy is obtained by weighting.

[0083] Among them, the hierarchical prefix accuracy automaton is constructed based on the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, using the formula Calculate the accuracy of the i-th layer between the hierarchical process model HPN and the hierarchical event log HL, where is a set of active labels, T is a transition set, HPN=(PN N0 ,HPN(PN N0 )) is a hierarchical process model on A, Tr={T n0 ,T n1 ,...T nn} is its nested transition set, HPA=(PA N0 ,HPA(PA N0 )) is a hierarchical prefix accuracy automaton of HPN, Sa is the state set, SE is the escape state set, EA is a set of directed arcs with labels between states, and wgt is a weight function that maps states to real numbers; according to the formula prec(HPA) = ∑W i ×prec(HPA i ) is weighted to obtain the overall accuracy, where w i is the weight of the i-th layer process model in the hierarchical process model, prec(HPA i ) is the accuracy of the i-th layer between the hierarchical process model HPN and the hierarchical event log HL.

[0084] An alignment-based hierarchical process model quality assessment method provided by an embodiment of the present invention first constructs a hierarchical prefix accuracy automaton from a log, compares the log record behavior with the model allowed behavior, identifies the point where the model allowed behavior exceeds the log record behavior, i.e., the escape state, and calculates the accuracy by considering the log record behavior and all identified escape state sets.

[0085] The embodiment of the present invention provides a method for evaluating the quality of a hierarchical process model based on alignment, PA L =(SA,SE,EA,T,ss,wgt,l SA ) is a labeled prefix accuracy automaton where:

[0086] 1) That is, the sequence is triggered to align all prefix sets;

[0087] 2) That is, the set of enabling transitions that have never been triggered,

[0088] 3)EA={(ρ,t,ρ')∈SA×T×(SA∪SE)|ρ'=ρ·t};

[0089] 4)ss=<>;

[0090] 5) The weight of a state takes into account both the number of trajectories that “enter” that state and the alignment probability of the trajectories.

[0091] 6)l SA :SA→Γ is a state labeling function, where Γ is a labeling set.

[0092] The embodiment of the present invention provides a method for evaluating the quality of a hierarchical process model based on alignment, PA N =(PA L, N L ) is a nested prefix accuracy automaton where:

[0093] 1) PA L =(SA,SE,EA,T,ss,wgt,l SA ) is a labeled prefix accuracy automaton;

[0094] 2)N L :SA→{A L ,N L} is a nested state function that satisfies Indicates that sa is in normal state, N L (sa) = N L Indicates that sa is in a nested state.

[0095] The embodiment of the present invention provides a method for evaluating the quality of a hierarchical process model based on alignment, which constructs a hierarchical prefix accuracy automaton PA according to a hierarchical alignment sequence HAT obtained by aligning a hierarchical process model HPN with a hierarchical event log HL. Nni Is the nested state sa i The corresponding prefix-accurate automaton with nested states, where HAN(PA N0 )={(sa i ,PA Nni ,HAN(PA Nni ))∣sa i ∈SA n0}, is a set of active labels, HPN=(PN N0, HPN(PN N0 )) is a hierarchical process model on A, Tr={T n0 ,T n1 ,...T nn} is its nested transition set, HPA=(PA N0, HPA(PA N0 )) is a hierarchical prefix accuracy automaton for HPN, PA N0 is the root node prefix accuracy automaton of HPA, SA N0 Is the root node set, HPA(PA N0 ) is PA N0 The corresponding sub-prefix accuracy automaton. Among them:

[0096] 1) PA N0 It is the root node prefix accuracy automaton of HPA;

[0097] 2)SA N0 Is the root node set, that is, the top-level PA N0 The nested state collection in the network;

[0098] 3) HPA (PA N0 ) is PA N0 The corresponding subprefix accuracy automaton, where:

[0099] (a) If

[0100] (b) Otherwise HAN(PA N0 )={(sa i ,PA Nni ,HAN(PA Nni ))∣sa i ∈SA n0}, where PA Nni Is the nested state sa iThe corresponding prefix-accurate automaton with nested states.

[0101] The embodiment of the present invention provides a method for evaluating the quality of a hierarchical process model based on alignment, wherein the accuracy of the i-th layer between the HL and the HPN is calculated by HPA, as shown in formula (12).

[0102]

[0103] W i Represents the weight of the i-th layer process model of the hierarchical process model. Weight W i It is determined according to the proportion of the set of directed arcs with labels between the states of this layer in the set of directed arcs of the entire model, and ∑W i =1 as shown in formula (13).

[0104]

[0105] According to prec(HPA i ), the accuracy contribution of each layer is integrated to calculate the overall accuracy between HL and HPN, as shown in formula (14).

[0106]

[0107] According to the method of step 200, the hierarchical prefix precision automaton is constructed using an Oracle function. Three methods can be used to construct the Oracle function of the model and the log, including: calculating an optimal alignment for each trajectory, calculating all optimal alignments for each trajectory, and calculating the representative of all optimal alignments for each trajectory. Figure 3 Taking the hierarchical event log and hierarchical process model as an example, the event log

[0108] hl3={<a s ,a c >,<a s ,a c ,b s ,b c ,c s ,g s ,g c ,c c ,d s ,d c >,<a s ,a c ,c s ,g s ,g c ,

[0109] h s ,h c ,i s ,i c ,js ,j c ,c c ,b s ,b c ,e s ,e c >,<a s ,a c ,b s ,b c ,f s ,f c ,b s ,b c ,c s ,g s ,g c ,i s ,i c ,h s ,h c ,j s ,j c ,c c ,d s ,d c >}

[0110] Figure 4 A hierarchical alignment sequence with an optimal alignment method for each track in the hierarchical event log is given. The hierarchical prefix accuracy automaton of the hierarchical process model is constructed using the Oracle function, as Figure 5 As shown. Using the constructed hierarchical prefix accuracy automaton, according to the above algorithm, we can get Figure 3 The hierarchical accuracy of the hierarchical event log and hierarchical process model are 0.7377 and 0.8529 respectively, and the overall accuracy of the hierarchical event log and hierarchical process model is 0.7953.

[0111] like Figure 4 As shown, γ1, γ2, γ3, γ4, γ6, γ7 and γ8 are the trajectories σ1=, σ2=<a,b,c,d>, σ3=<a,c,b,e>, σ4=<a,b,f,b,c,d> and

[0112] σ6=<g>,σ7=<g,h,i,j>,σ8=<g,i,h,j>corresponding to an optimal alignment. Figure 3 As shown,

[0113] γ1,γ2,γ3,γ4,γ6,γ7 are the root alignment sequences of the hierarchical alignment sequence rootA, γ6,γ7,γ8 are the sub-alignment sequences of the hierarchical alignment sequence NAlign nb .

[0114] like Figure 5As shown, there are 5 states and several escape states. Figure 5 Take the state ρ=<t1,t2,t6> in as an example. This sequence is the projection prefix of the γ2 model movement, that is, ρ≤π2(γ4) ↓T , and it is not Figure 4 The prefix of any other projection in the alignment model. Therefore, the weight of this state is the same as the trajectory π2(γ4) ↓A Occurs the same number of times in the hierarchical event log. Figure 3 The initiation of the model initial identity [p1] leads to the identities [p2, p3].

[0115] Two transitions are enabled from the markup: t2 and t3. Figure 6 The only transition induced after induced ρ is t2. Therefore, the state <t1,t2,t6,t3> is the escape state of <t1,t2,t6>. Figure 5 The state <t1, t3> shown in is a nested state. Figure 3 Neutron Model PN N1 The corresponding subprefix accuracy automaton.

[0116] by Figure 3 Taking the hierarchical event log and hierarchical process model as an example, Figure 6 The hierarchical alignment sequence of all optimal alignments of the traces σ1= and σ6=<g> in the log is shown. The hierarchical prefix accuracy automaton constructed using the Oracle function is as follows: Figure 7 As shown in , the Oracle function generates all optimal alignments for each trajectory, such as Figure 6 shown.

[0117] refer to Figure 6 As shown, γ 1,1 ,γ 1,2 ,γ 1,3 ,γ 1,4 are all optimal alignments between the hierarchical process model and the trajectory σ1=, γ 6,1 ,γ 6,2 ,γ 6,3 ,γ 6,4 are all optimal alignments between the hierarchical process model and the trajectory and σ6=<g>.

[0118] like Figure 7 As shown in the figure, there are 7 states and several escape states. The state <t1, t3> shown in the figure is a nested state. Figure 3 Neutron Model PN N1 The corresponding subprefix accuracy automaton. Use Figure 7The hierarchical prefix accuracy automaton shown in the figure calculates the hierarchical accuracy of the hierarchical process model and the hierarchical event log to be 0.7184 and 0.8182 respectively, and the overall accuracy of the hierarchical process model and the hierarchical event log is 0.7591. Please note that the optimal alignment of trajectories σ1 and σ6 includes 4 possible cases, so the alignment weights of σ1 and σ6 are evenly divided into 4 parts, that is, the weight of each alignment is 1 / 4 = 0.25.

[0119] The Oracle function is constructed by calculating all optimal alignment representatives. Figure 8 Given the trajectory σ1 = and σ6= <g> A set of representatives of all optimal alignments between the hierarchical process model.< / g> Figure 8 As shown, γ 1,1 ,γ 1,3 is the optimal alignment representative of all possible alignments between the hierarchical process model and the trajectory σ1=, γ 6,1 ,γ 6,2 are all optimal alignment representatives between the hierarchical process model and the trajectory and σ6=<g>. For each group of representatives, the number of optimal alignments represented by the group of representatives is calculated, and Fig. 9 As shown in , the weight is obtained by calculating the ratio of the optimal alignment number to the total optimal alignment number. The higher the weight, the higher the probability. Figure 8 As shown, trajectory σ1 has 4 optimal alignments. 1,1 represents the 2 best alignments with a probability of 2 / 4 = 0.5.

[0120] like Fig. 9 As shown in the figure, there are 7 states and several escape states. The state <t1, t3> shown in the figure is a nested state. Figure 3 Neutron Model PN N1 The corresponding subprefix accuracy automaton. Fig. 9 The hierarchical prefix accuracy automaton constructed using all optimal alignment representatives is shown. Using this automaton, the hierarchical accuracy of the calculated log hl3 and model hpn3 are 0.7333 and 0.8955 respectively, and the overall accuracy of the hierarchical process model and the hierarchical event log is 0.8144.

[0121] According to the above results, different Oracle functions may produce different accuracy results. Among the three solutions proposed in the embodiment of the present invention, only the Oracle function that produces non-zero probability for all optimal alignments of each trajectory provides the highest accuracy and can minimize the number of escape states.

[0122] Step 300: Obtain a comprehensive score F-value of the hierarchical business process model by taking a weighted average based on the overall fit and the overall accuracy, and evaluate the quality of the hierarchical business process model based on the comprehensive score F-value.

[0123] When measuring the quality of a model, the fit and precision are two crucial evaluation indicators for the hierarchical process model based on it. However, the two cannot always reach the optimum at the same time. In order to evaluate the quality of the model more comprehensively and objectively, the embodiment of the present invention proposes an F-value that combines the two indicators, as shown in formula (15). The F-value is a comprehensive score obtained by calculating the harmonic mean of the fit and precision. The higher the F-value, the better the hierarchical process model performs in terms of both fit and precision.

[0124] According to the overall fit and overall accuracy, the formula The weighted average is used to obtain the comprehensive score F-value of the hierarchical business process model, where wfit h (HL,HPN i ,orc,sc) for use and w i The weighted goodness of fit between the hierarchical event log HL and the hierarchical process model HPN is calculated; prec(HPA) is the overall accuracy between the hierarchical event log HL and the hierarchical process model HPN.

[0125] for Figure 2 The hierarchical process model and the corresponding hierarchical event log data shown in the figure are comprehensively evaluated by applying a hierarchical business process model quality evaluation method based on alignment according to an embodiment of the present invention, and the calculated F-value is 0.887. Figure 3 The hierarchical process model and the corresponding hierarchical event log data in the embodiment of the present invention are also comprehensively evaluated by applying a hierarchical business process model quality evaluation method based on alignment according to an embodiment of the present invention, and the calculated F-value is 0.976, which is consistent with Figure 3 The quality of the hierarchical process model shown is better than Figure 2 This is consistent with the facts shown for the quality of the hierarchical process model.

[0126] The present invention provides a method for evaluating the quality of a hierarchical business process model based on alignment, which is based on the hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL. HAT Calculate the fit of each level and get the overall fit by weighted average; align the hierarchical process model HPN with the hierarchical event log HL to get the hierarchical alignment sequence HATA hierarchical prefix accuracy automaton is constructed to obtain the accuracy of each level, and the overall accuracy is obtained by weighting; the comprehensive score F-value of the hierarchical business process model is obtained by weighted average of the overall fit and the overall accuracy, and the quality of the hierarchical business process model is evaluated based on the comprehensive score F-value. Based on the alignment sequence obtained by aligning the hierarchical process model with the event log, the alignment-based fit evaluation method, the accuracy evaluation method and the weighted average of the two are used to calculate the F-value of the model to comprehensively evaluate the quality of the hierarchical process model. This method can not only effectively identify the complex hierarchical structure in the hierarchical process model, but also comprehensively and accurately evaluate the quality of the hierarchical process model.

[0127] The present invention provides a method for evaluating the quality of a hierarchical business process model based on alignment. By aligning the hierarchical process model with the hierarchical event log and fully analyzing the complex hierarchical structure relationship, a set of quality evaluation methods applicable to various hierarchical process model hierarchical structures is proposed. First, the quality evaluation method of the embodiment of the present invention can evaluate the quality of the hierarchical process model more comprehensively and accurately. Through the hierarchical alignment method of the hierarchical process model and the event log, the changes and abnormal situations in the hierarchical process model can be quickly identified and processed, thereby improving the efficiency of quality evaluation. Secondly, the hierarchical business process model quality evaluation method proposed in the embodiment of the present invention is applicable to various hierarchical process model hierarchical structures, has strong versatility and adaptability, and has important guiding significance and application value in practical applications. By more comprehensively and accurately evaluating the quality of the hierarchical process model, a scientific basis can be provided for business process optimization, system design and management decision-making, thereby improving the efficiency and effect of the business process. At the same time, the method of the present invention can also provide a reference for research in related fields and promote the development and progress of hierarchical process model quality evaluation technology.

[0128] Based on the same inventive concept, the embodiment of the present application also provides a hierarchical business process model quality evaluation device based on alignment, which can execute the processing flow provided by a hierarchical business process model quality evaluation method embodiment based on alignment. The hierarchical business process model quality evaluation device based on alignment provided by the present application includes:

[0129] The fit calculation module is configured to obtain the hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL. HAT Calculate the goodness of fit at each level and obtain the overall goodness of fit by weighted average;

[0130] The accuracy calculation module is configured to obtain a hierarchical alignment sequence based on the alignment of the hierarchical process model HPN with the hierarchical event log HL. HAT Construct a hierarchical prefix accuracy automaton to obtain the accuracy of each level and get the overall accuracy by weighting;

[0131] The model evaluation module is configured to obtain a comprehensive score F-value of the hierarchical business process model based on a weighted average of the overall fit and the overall accuracy, and evaluate the quality of the hierarchical business process model based on the comprehensive score F-value.

[0132] The device provided in the embodiment of the present application can be specifically used to execute the solution provided in the corresponding method embodiment, and the specific functions and technical effects that can be achieved are not repeated here.

[0133] An embodiment of the present application also provides a computer-readable storage medium, in which computer execution instructions are stored. When the computer execution instructions are executed by a processor, they are used to implement the solution provided by any of the above method embodiments. The specific functions and technical effects that can be achieved are not repeated here.

[0134] An embodiment of the present application also provides a computer program product, which includes: a computer program, which is stored in a readable storage medium. At least one processor of an electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device executes the solution provided by any of the above method embodiments. The specific functions and technical effects that can be achieved are not repeated here.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating the quality of a hierarchical business process model based on alignment, characterized in that: The hierarchical business process model quality evaluation method comprises: Hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL HAT Calculate the goodness of fit at each level and obtain the overall goodness of fit by weighted average; Hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL HAT Construct a hierarchical prefix accuracy automaton to obtain the accuracy of each level and get the overall accuracy by weighting; The comprehensive score F-value of the hierarchical business process model is obtained by taking a weighted average based on the overall fit and the overall accuracy, and the quality of the hierarchical business process model is evaluated based on the comprehensive score F-value.

2. The hierarchical business process model quality evaluation method according to claim 1, characterized in that: The comprehensive score F-value of the hierarchical business process model obtained by weighted average of the overall fit and the overall accuracy includes: According to the overall fit and overall accuracy, the formula The weighted average is used to obtain the comprehensive score F-value of the hierarchical business process model, where wfit h (HL,HPN i ,orc,sc) for use and w i The weighted goodness of fit between the hierarchical event log HL and the hierarchical process model HPN is calculated; prec(HPA) is the overall accuracy between the hierarchical event log HL and the hierarchical process model HPN.

3. The hierarchical business process model quality evaluation method according to claim 1, characterized in that: The hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL HAT Calculate the fit of each level and get the overall fit by weighted average, including: According to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, the formula is used Calculate the hierarchical weighted fit wfit for each level hi (L i ,N i ,orc,sc), where mlf hi (L i ,N i ,orc,sc) is L i and N i The hierarchical log fit, L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, mmf hi (L i ,N i ,orc,sc) is L i and N i The moving fit of the hierarchical model; Based on hierarchical weighted fitting wfit hi (L i ,N i ,orc,sc), using L i and N i The relative fit between the levels and the weight w of the model fit of the i-th layer of the hierarchical process model i Calculate the weighted fit wfit of the hierarchical event log HL and hierarchical process model HPN h (HL,HPN i ,orc,sc), where the weighted fit wfit h (HL,HPN i ,orc,sc)=∑w i ·wfit hi (L i ,N i ,orc,sc).

4. The hierarchical business process model quality evaluation method according to claim 3 is characterized in that: Using formula Calculate L i and N i The hierarchical log fit mlf hi (L i ,N i ,orc,sc), where L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, γ i YesL i and N i The corresponding HAT layer i sub-alignment sequence, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, w i is the weight of the i-th layer process model, σ is a finite set of event sequences in the event space ξ, A is a finite set of activity labels, and T is a finite set of transitions.

5. The hierarchical business process model quality evaluation method according to claim 3, characterized in that: Using formula Calculate L i and N i The moving average fit of the hierarchical model mmf hi (L i ,N i ,orc,sc), where L i is the i-th level sub-log of the hierarchical event log HL, N i For L i The corresponding hierarchical process model HPN layer i sub-model, γ i YesL i and N i The corresponding HAT layer i sub-alignment sequence, sc is the cost function, orc is L i and N i Compared with the Oracle function of sc, w i is the weight of the i-th layer process model, σ is a task sequence with life cycle information, A is a set of activity labels, and T is a transition set.

6. The hierarchical business process model quality evaluation method according to claim 1, characterized in that: The hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL HAT Construct a hierarchical prefix accuracy automaton to obtain the accuracy of each level, and obtain the overall accuracy by weighting, including: According to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, a hierarchical prefix accuracy automaton is constructed using the formula Calculate the accuracy of the i-th layer between the hierarchical process model HPN and the hierarchical event log HL, where is a set of active labels, T is a transition set, HPN=(PN N0 ,HPN(PN N0 )) is a hierarchical process model on A, Tr={T n0, T n1 ,...T nn } is its nested transition set, HPA=(PA N0, HPA(PA N0 )) is a hierarchical prefix accuracy automaton of HPN, Sa is the state set, SE is the escape state set, EA is a set of directed arcs with labels between states, and wgt is a weight function that maps states to real numbers; According to the formula prec(HPA) = ∑W i ×prec(HPA i ) is weighted to obtain the overall accuracy, where w i is the weight of the i-th layer process model in the hierarchical process model, prec(HPA i ) is the accuracy of the i-th layer between the hierarchical process model HPN and the hierarchical event log HL.

7. The hierarchical business process model quality evaluation method according to claim 6, characterized in that: The weight w of the i-th layer process model of the hierarchical process model i Using formula Calculate, where Sa is the set of states, EA is the set of labeled directed arcs between states, and wgt is the weight function that maps states to real numbers.

8. The hierarchical business process model quality evaluation method according to claim 6, characterized in that: According to the hierarchical alignment sequence HAT obtained by aligning the hierarchical process model HPN with the hierarchical event log HL, a hierarchical prefix accuracy automaton PA is constructed. Nni Is the nested state sa i The corresponding prefix-accurate automaton with nested states, where HAN(PA N0 )={(sa i ,PA Nni ,HAN(PA Nni ))∣sa i ∈SA n0 }, is a set of active labels, HPN=(PN N0 ,HPN(PN N0 )) is a hierarchical process model on A, Tr={T n0 ,T n1 ,...T nn } is its nested transition set, HPA=(PA N0 ,HPA(PA N0 )) is a hierarchical prefix accuracy automaton for HPN, PA N0 is the root node prefix accuracy automaton of HPA, SA N0 is the root node set, HPA(PA N0 ) is PA N0 The corresponding subprefix accuracy automaton.

9. A device for evaluating the quality of a hierarchical business process model based on alignment, characterized in that: The hierarchical business process model quality evaluation device comprises: The fit calculation module is configured to obtain the hierarchical alignment sequence obtained by aligning the hierarchical process model HPN with the hierarchical event log HL. HAT Calculate the goodness of fit at each level and obtain the overall goodness of fit by weighted average; The accuracy calculation module is configured to obtain a hierarchical alignment sequence based on the alignment of the hierarchical process model HPN with the hierarchical event log HL. HAT Construct a hierarchical prefix accuracy automaton to obtain the accuracy of each level and get the overall accuracy by weighting; The model evaluation module is configured to obtain a comprehensive score F-value of the hierarchical business process model based on a weighted average of the overall fit and the overall accuracy, and evaluate the quality of the hierarchical business process model based on the comprehensive score F-value.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the alignment-based hierarchical business process model quality evaluation method according to any one of claims 1 to 8 when executed by a processor.

Images