Path prediction method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide a path prediction method and device, electronic equipment and storage medium, and relate to the field of data processing. The path prediction method provided by the present application obtains a current path of a current execution instance, calculates a current data matrix corresponding to the current path; obtains a plurality of alternative subsequent paths corresponding to the current execution instance; calculates an alternative data matrix corresponding to each alternative subsequent path; calculates a similarity result of each alternative data matrix and the current data matrix, determines the alternative data matrix with the largest similarity result as a target matrix; and determines the alternative subsequent path corresponding to the target matrix as a target path. The embodiments of the present application only need to calculate the similarity result between matrices to predict the path in the business process, and the calculation method is simple and does not need to train the prediction model with a large amount of event log data.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a path prediction method, apparatus, electronic device, and storage medium. Background Technology

[0002] In production applications, as demand increases, business processes become increasingly complex and diverse, leading to variable execution paths. Therefore, predicting the path of currently executing business processes in advance becomes crucial; this involves predictive analysis of ongoing business process instances to accurately forecast potential execution states over a future period.

[0003] Most current prediction methods are based on deep learning technology, training a process prediction model using event logs and then using the currently executed path as input to output the predicted event. However, deep learning technology itself relies on a large number of training samples, significantly increasing the modeling time of the process prediction model; furthermore, if the process changes, it is necessary to retrain the process prediction model using new training samples. Therefore, existing deep learning-based process execution path prediction methods have significant limitations. Summary of the Invention

[0004] To address the aforementioned technical problems, embodiments of this application provide a path prediction method, apparatus, electronic device, and storage medium.

[0005] In a first aspect, embodiments of this application provide a path prediction method, the method comprising:

[0006] Obtain the current path of the currently executing instance, and calculate the current data matrix corresponding to the current path;

[0007] Obtain multiple alternative subsequent paths corresponding to the currently executing instance;

[0008] Calculate the candidate data matrix corresponding to each of the proposed alternative subsequent paths;

[0009] Calculate the similarity results between each of the candidate data matrices and the current data matrix, and determine the candidate data matrix with the largest similarity result as the target matrix;

[0010] The candidate subsequent paths corresponding to the target matrix are determined as the target path.

[0011] In one embodiment, calculating the current data matrix corresponding to the current path includes:

[0012] Obtain multiple task nodes traversed by the current path and multiple data objects corresponding to each task node;

[0013] The rows of the current data matrix are generated based on the task node, and the columns of the current data matrix are generated based on the data object.

[0014] In one embodiment, before obtaining the multiple task nodes traversed by the current path and the multiple data objects corresponding to each task node, the process includes:

[0015] Obtain the event log data corresponding to the current path. The event log data includes task node information and data object information.

[0016] In one embodiment, calculating the candidate data matrix corresponding to each of the candidate subsequent paths includes:

[0017] Obtain the multiple task nodes traversed by the alternative subsequent path and the multiple data objects corresponding to each task node;

[0018] The rows of the candidate data matrix are generated based on the task node, and the columns of the candidate data matrix are generated based on the data object.

[0019] In one embodiment, calculating the similarity result between each of the candidate data matrices and the current data matrix includes:

[0020] Calculate the similarity between each of the candidate data matrices and the current data matrix;

[0021] Calculate the similarity coefficient between each of the candidate data matrices and the current data matrix;

[0022] Through formula

[0023] s = r × c

[0024] Calculate the similarity results corresponding to each of the candidate data matrices, where s is the similarity result, c is the similarity degree, and r is the similarity coefficient.

[0025] In one embodiment, calculating the similarity between each of the candidate data matrices and the current data matrix includes:

[0026] Through formula

[0027]

[0028] Calculate the similarity between each of the candidate data matrices and the current data matrix, where A is the current data matrix, B is the candidate data matrix, and a i b is the i-th component of the vector obtained by expanding the current data matrix. i The i-th component of the vector obtained by expanding the candidate data matrix.

[0029] In one embodiment, calculating the similarity coefficient between each of the candidate data matrices and the current data matrix includes:

[0030] Through formula

[0031] r = r³ - p * N zero *r2+q*r1

[0032] Calculate the similarity coefficient between each of the candidate data matrices and the current data matrix;

[0033] in

[0034]

[0035]

[0036]

[0037] N zero The number of 0s in the matrix obtained by subtracting the current data matrix from the candidate data matrix is ​​A′=A·A T B′=B·B T D = A·B T q is the first weight, and p is the second weight. J is the column number of D.

[0038] Secondly, embodiments of this application provide a path prediction device, the path prediction device comprising:

[0039] The first calculation module is used to obtain the current path of the currently executing instance and calculate the current data matrix corresponding to the current path;

[0040] The acquisition module is used to acquire multiple alternative subsequent paths corresponding to the currently executing instance;

[0041] The second calculation module is used to calculate the candidate data matrix corresponding to each of the candidate subsequent paths;

[0042] The third calculation module is used to calculate the similarity results between each of the candidate data matrices and the current data matrix, and to determine the candidate data matrix with the largest similarity result as the target matrix;

[0043] The determination module is used to determine the candidate subsequent paths corresponding to the target matrix as the target path.

[0044] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the computer program executes the path prediction method provided in the first aspect when the processor is running.

[0045] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when run on a processor, executes the path prediction method provided in the first aspect.

[0046] The path prediction method provided in this application obtains the current path of the currently executing instance and calculates the current data matrix corresponding to the current path; obtains multiple alternative subsequent paths corresponding to the currently executing instance; calculates the alternative data matrix corresponding to each alternative subsequent path; calculates the similarity result between each alternative data matrix and the current data matrix, and determines the alternative data matrix with the largest similarity result as the target matrix; and determines the alternative subsequent path corresponding to the target matrix as the target path. It predicts paths in the business process solely by calculating the similarity results between matrices, making the calculation method simple and eliminating the need for training the model with a large amount of event log data. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be considered as a limitation on the scope of protection of this application. In the various drawings, similar components are numbered similarly.

[0048] Figure 1 A flowchart of the path prediction method provided in an embodiment of this application is shown;

[0049] Figure 2 A schematic diagram of a data matrix provided in an embodiment of this application is shown;

[0050] Figure 3 A schematic diagram of multiple process paths provided in an embodiment of this application is shown;

[0051] Figure 4 A schematic diagram of the first process path provided in an embodiment of this application is shown;

[0052] Figure 5 A schematic diagram of the second process path provided in an embodiment of this application is shown;

[0053] Figure 6 A schematic diagram of the third process path provided in an embodiment of this application is shown;

[0054] Figure 7 A schematic diagram of the fourth process path provided in an embodiment of this application is shown;

[0055] Figure 8 A schematic diagram of the path prediction device provided in an embodiment of this application is shown. Detailed Implementation

[0056] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0057] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0058] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0059] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0060] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0061] Example 1

[0062] This disclosure provides a path prediction method.

[0063] In existing technologies, path prediction for process instances mainly falls into two categories: predicting the remaining time of the process instance and predicting subsequent activities of the process instance. Most current prediction methods extract the process prediction model from event logs, then use the trajectory prefix (the currently executed path) as input, and output the predicted event.

[0064] In recent years, deep learning technology has been widely used in path prediction tasks. For example, Long Short-Term Memory (LSTM) recurrent neural networks, bidirectional quasi-recurrent neural networks, two-layer LSTM neural networks, and Temporal Convolutional Networks (TCN), a variant of Convolutional Neural Networks (CNN), have achieved good prediction results. However, traditional deep learning models require a large amount of event log data for training, which leads to long modeling times and significant limitations. Therefore, this application provides a path prediction method with good versatility and fast computation.

[0065] For details, see Figure 1 , Figure 1 A flowchart of a path prediction method provided in an embodiment of this application is shown, the method comprising:

[0066] Step S110: Obtain the current path of the currently executing instance and calculate the current data matrix corresponding to the current path;

[0067] First, obtain the path taken by the instances that have already been executed in the current business process.

[0068] In one embodiment, calculating the current data matrix corresponding to the current path includes:

[0069] Obtain multiple task nodes traversed by the current path and multiple data objects corresponding to each task node; generate rows of the current data matrix based on the task nodes, and generate columns of the current data matrix based on the data objects.

[0070] A business process includes elements such as start nodes, system task nodes, user task nodes, various gateways, and connecting lines. During the execution of the business process, different execution paths arise because of the existence of gateways, each with its own execution rules. For example, an exclusive gateway executes the first connecting line whose condition is true by judging the conditions behind that gateway. Therefore, the conditional data will differ across execution paths. Consequently, the data will also change during the process. Therefore, from a data perspective, a business process can be viewed as a series of data objects corresponding to several task node machines.

[0071] A data object is defined as a variable that changes along with the execution nodes of a business process, such as time, executor, and parameters. As the process moves from one node to the next, the executor and execution time may change, and the parameters will also change accordingly. In other words, when a task node changes, the data object changes accordingly. Therefore, a variable that changes with the execution nodes of a business process is defined as a data object.

[0072] In fact, changes in task nodes often represent changes in data objects. Therefore, changes in data objects can also indicate changes in the process, allowing for predictions based on task nodes and data objects.

[0073] Please see Figure 2 , Figure 2 A schematic diagram of a data matrix provided in an embodiment of this application is shown. This schematic diagram applies not only to the current data matrix but also to alternative data matrices.

[0074] Each row of the data matrix corresponds to a task node in the process, and each column corresponds to a data object in the process. This data matrix represents the current execution status of the process. When comparing the data matrix with the process instance in the event log, data objects not appearing in the current execution instance are padded with zeros. One-Hot encoding is used for non-numeric variables to resolve discrepancies in attribute values.

[0075] In one embodiment, before obtaining the multiple task nodes traversed by the current path and the multiple data objects corresponding to each task node, the process includes:

[0076] Obtain the event log data corresponding to the current path. The event log data includes task node information and data object information.

[0077] Event log data stores the execution records of business process instances, thus containing task node information and node-related attribute data, i.e., data object information. Therefore, the instance data of process execution can be extracted from the event log.

[0078] Step S120: Obtain multiple alternative subsequent paths corresponding to the currently executing instance;

[0079] Please see Figure 3 , Figure 3This diagram illustrates multiple process paths provided in an embodiment of this application. First, using the current execution instance, each potentially executable path is traversed. A process contains many nodes, including ordinary task nodes (Task 1, Task 2, Task 3, Task 4, Task 5, Task 6, Task 7), each with a unique ID; and various gateways, such as… Figure 3 The exclusive gateway in the process calculates all conditions attached to outgoing migration lines when the process reaches this gateway. The migration line whose first condition resolves to true is selected. That is, when multiple conditions are true, the first decision is executed, and no other outgoing branches are calculated, allowing the process to run along the first selected migration line. There are also multi-path gateways, whose calculation logic is similar to the exclusive gateway, but they calculate all outgoing branches and execute them if the resolution is true. Finally, there are parallel gateways, which directly run all outgoing branches.

[0080] Therefore, when searching for multiple alternative follow-up paths corresponding to the currently executing instance, a new array can be created, starting execution from the beginning node. Specifically, the node ID is stored in the array. When a gateway is encountered, it is determined what kind of gateway it is. If it is an exclusive gateway, the number of outgoing branches is checked, and an array of equal size is created based on the number of outgoing branches. Each outgoing branch is then stored in a different array, and the process continues to traverse separately. When multiple gateways are encountered, all possible combinations are arranged to find the possible outcomes. Finally, the traversal ends at the end node, resulting in every possible execution path, i.e., multiple alternative follow-up paths.

[0081] like Figure 4 , Figure 5 , Figure 6 and Figure 7 As shown. Figure 4 , Figure 5 , Figure 6 and Figure 7 Each node and Figure 3 The same applies, so I won't repeat it here. Figure 4 , Figure 5 , Figure 6 and Figure 7 Showing Figure 3 There are four possible execution paths. Figure 4 A schematic diagram of the first process path provided in an embodiment of this application is shown; Figure 5 A schematic diagram of the second process path provided in an embodiment of this application is shown; Figure 6 A schematic diagram of the third process path provided in an embodiment of this application is shown; Figure 7 A schematic diagram of the fourth process path provided in an embodiment of this application is shown.

[0082] Furthermore, during the traversal, based on the current process execution path and parameters, and according to the conditional statements of the gateway nodes (including the exclusive gateway, multi-path gateway, and parallel gateway mentioned above), these parameters are evaluated. Instance paths whose current parameters do not meet the evaluation conditions are excluded, while those that do meet the conditions are retained. Paths that cannot be evaluated are also saved. In this way, the final event log data can reflect the execution status of each node in the process and the corresponding data objects.

[0083] In one embodiment, calculating the candidate data matrix corresponding to each of the candidate subsequent paths includes: obtaining multiple task nodes traversed by the candidate subsequent paths and multiple data objects corresponding to each task node; generating rows of the candidate data matrix based on the task nodes, and generating columns of the candidate data matrix based on the data objects.

[0084] Step S130: Calculate the candidate data matrix corresponding to each of the candidate subsequent paths;

[0085] Once the alternative subsequent paths are determined, the alternative data matrix can be calculated using the same method as the current data matrix. To avoid repetition, this will not be elaborated here.

[0086] Step S140: Calculate the similarity results between each of the candidate data matrices and the current data matrix, and determine the candidate data matrix with the largest similarity result as the target matrix;

[0087] In one embodiment, calculating the similarity result between each of the candidate data matrices and the current data matrix includes:

[0088] Calculate the similarity between each of the candidate data matrices and the current data matrix;

[0089] Through formula 1:

[0090] s = r × c

[0091] Calculate the similarity result corresponding to each of the candidate data matrices, where r is the similarity result, c is the similarity degree, and r is the similarity coefficient.

[0092] In one embodiment, calculating the similarity between each of the candidate data matrices and the current data matrix includes:

[0093] Through formula 2:

[0094]

[0095] Calculate the similarity between each of the candidate data matrices and the current data matrix, where A is the current data matrix, B is the candidate data matrix, and a ib is the i-th component of the vector obtained by expanding the current data matrix. i Let i be the i-th component of the vector obtained by expanding the candidate data matrix. Matrix A is represented as A = (a ij ) m×n Matrix B is represented as B = (b ij ) m×n .

[0096] When calculating the similarity of matrices, the matrices are first expanded into vectors, for example, by expanding the matrix into vectors.

[0097]

[0098] Expand the matrix into a vector (a, b, c, d, e, f, g, h, i). Similarly, expand the other matrix into a vector, and then calculate the cosine similarity between the two vectors. i b is the i-th component in the current data matrix expansion. i Let i be the i-th component in the expansion of the candidate data matrix.

[0099] In one embodiment, calculating the similarity coefficient between each of the candidate data matrices and the current data matrix includes:

[0100] Through formula 3:

[0101] r = r³ - p * N zero *r2+q*r1

[0102] Calculate the similarity coefficient between each of the candidate data matrices and the current data matrix;

[0103] Where r3, r2, and r1 are determined by formula 4:

[0104] Formula 4:

[0105]

[0106]

[0107]

[0108] N zero The number of 0s in the matrix obtained by subtracting the current data matrix from each of the candidate data matrices is A′=A·A T B′=B·B T D = A·B T ;

[0109] Through formula 5:

[0110]

[0111] Divide D into three parts: D1, D2, and D3. p is the second weight (which can be 0.36), q is the first weight (which can be 1.3), and J is the column number of D.

[0112] N zero This represents the number of zeros in the subtraction of the matrices. The more zeros there are, the more common elements the two matrices have, and the more similar the two matrices are.

[0113] In one specific embodiment, for example, there are three matrices A1, B1, and C1.

[0114] Matrix A1:

[0115]

[0116] Matrix B1:

[0117]

[0118] Matrix C1:

[0119]

[0120] If A1 is the current data matrix corresponding to the currently executing instance, and B1 and C1 are the alternative data matrices calculated after extracting the event logs of the alternative subsequent paths.

[0121] In fact, direct observation also reveals that A1 and B1 are more similar. Therefore, the calculation is performed according to the method provided in step S140.

[0122] Similarity between A1 and B1

[0123] c1 is approximately equal to 0.68688. Similarly, the similarity c2 between A1 and C1 is approximately equal to 0.55221.

[0124] Then calculate the similarity coefficient: r = r³ - 0.36 * N zero *r2+1.3r1

[0125] The values ​​of r1 for A1 and B1 are 0.38289, r2 is 0.34122289, and r3 is 0.70004, so r1 is -0.64483.

[0126] For A1 and C1, r1 equals 0.33500, r2 equals 0.383704, and r3 equals 0.6968745. Therefore, r2 equals -0.525225. Hence, r1 is less than r2.

[0127] s1 = c1 * r1 = -0.44292; s2 = c2 * r2 = -0.29003. Therefore, s1 is less than s2, and the similarity between A1 and B1 is less than that between A1 and C1.

[0128] Step S150: The alternative subsequent paths corresponding to the target matrix are determined as the target path.

[0129] The path prediction method provided in this embodiment has the following beneficial effects: First, it uses data objects within the process for process prediction, reflecting changes in the process through these data objects and executing path transformations. Second, it reduces the cycle time of the prediction process; constructing the data object matrix of each instance in the process event log is significantly less time-consuming than building a prediction model using deep learning. Third, it requires less event log data, only one normal execution data from each instance in the process is needed. Fourth, it supports more flexible process modifications; when a process node changes, only the instance's data object matrix needs to be updated to achieve subsequent process path prediction. Fifth, it calculates matrix similarity using a similarity algorithm, directly reflecting the degree of similarity between paths, thereby obtaining the prediction result.

[0130] In summary, the path prediction method provided in this embodiment is simple and fast to calculate, does not require a lot of time and a large number of samples to train the model, and has a wide range of applications, which can be flexibly changed as business processes change.

[0131] Example 2

[0132] Furthermore, embodiments of this disclosure provide a path prediction device.

[0133] Specifically, such as Figure 8 As shown, the path prediction device 800 includes:

[0134] The first calculation module 810 is used to obtain the current path of the currently executing instance and calculate the current data matrix corresponding to the current path;

[0135] The acquisition module 820 is used to acquire multiple alternative subsequent paths corresponding to the currently executing instance;

[0136] The second calculation module 830 is used to calculate the candidate data matrix corresponding to each of the candidate subsequent paths;

[0137] The third calculation module 840 is used to calculate the similarity results between each of the candidate data matrices and the current data matrix, and to determine the candidate data matrix with the largest similarity result as the target matrix;

[0138] The determination module 850 is used to determine the alternative subsequent paths corresponding to the target matrix as the target path.

[0139] The path prediction device 800 provided in this embodiment can implement the path prediction method provided in Embodiment 1. To avoid repetition, it will not be described again here.

[0140] The path prediction device provided in this embodiment is simple and fast to calculate, does not require a lot of time and a large number of samples to train the model, and has a wide range of applications, which can be flexibly changed as business processes change.

[0141] Example 3

[0142] Furthermore, this disclosure provides an electronic device including a memory and a processor, wherein the memory stores a computer program that executes the path prediction method provided in Embodiment 1 when the computer program is run on the processor.

[0143] The electronic device provided in this embodiment of the invention can implement the path prediction method provided in Embodiment 1. To avoid repetition, it will not be described again here.

[0144] The electronic device provided in this embodiment is simple and fast to calculate, does not require a lot of time and a large number of samples to train the model, and has a wide range of applications, and can be flexibly changed as business processes change.

[0145] Example 4

[0146] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the path prediction method provided in Embodiment 1.

[0147] In this embodiment, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0148] The computer-readable storage medium provided in this embodiment can implement the path prediction method provided in Embodiment 1. To avoid repetition, it will not be described again here.

[0149] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0150] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0151] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A path prediction method, characterized in that, The method includes: Obtain the current path of the currently executing instance, and calculate the current data matrix corresponding to the current path; Obtain multiple alternative subsequent paths corresponding to the currently executing instance; By traversing each possible subsequent execution path through the current execution instance, during the traversal, the parameters are judged according to the condition judgment formula of the gateway node based on the current process execution path and parameters, and paths whose current parameters do not meet the judgment conditions are excluded, resulting in multiple candidate subsequent paths. Calculate the candidate data matrix corresponding to each of the proposed alternative subsequent paths; Calculate the similarity results between each of the candidate data matrices and the current data matrix, and determine the candidate data matrix with the largest similarity result as the target matrix; The candidate subsequent paths corresponding to the target matrix are determined as the target path; The calculation of the current data matrix corresponding to the current path includes: Obtain multiple task nodes traversed by the current path and multiple data objects corresponding to each task node; The rows of the current data matrix are generated based on the task node, and the columns of the current data matrix are generated based on the data object.

2. The path prediction method according to claim 1, characterized in that, Before obtaining the multiple task nodes traversed by the current path and the multiple data objects corresponding to each task node, the process includes: Obtain the event log data corresponding to the current path. The event log data includes task node information and data object information.

3. The path prediction method according to claim 1, characterized in that, The calculation of the candidate data matrix corresponding to each of the candidate subsequent paths includes: Obtain the multiple task nodes traversed by the alternative subsequent path and the multiple data objects corresponding to each task node; The rows of the candidate data matrix are generated based on the task node, and the columns of the candidate data matrix are generated based on the data object.

4. The path prediction method according to claim 1, characterized in that, The calculation of the similarity results between each of the candidate data matrices and the current data matrix includes: Calculate the similarity between each of the candidate data matrices and the current data matrix; Calculate the similarity coefficient between each of the candidate data matrices and the current data matrix; Through formula Calculate the similarity results corresponding to each of the candidate data matrices, where, s For the similar results, c For the aforementioned similarity, r The similarity coefficient is denoted as .

5. The path prediction method according to claim 4, characterized in that, The calculation of the similarity between each of the candidate data matrices and the current data matrix includes: Through formula Calculate the similarity between each of the candidate data matrices and the current data matrix, where, A For the current data matrix, B The candidate data matrix, a i The first vector obtained by expanding the current data matrix i One portion, b i The first vector obtained by expanding the candidate data matrix i Each component.

6. The path prediction method according to claim 4, characterized in that, The calculation of the similarity coefficient between each of the candidate data matrices and the current data matrix includes: Through formula Calculate the similarity coefficient between each of the candidate data matrices and the current data matrix; in N zero The number of 0s in the matrix obtained by subtracting the current data matrix from the candidate data matrix. , , q As the first weight, p As the second weight, , , , J for D The number of columns.

7. A path prediction device, characterized in that, The device includes: The first calculation module is used to obtain the current path of the currently executing instance and calculate the current data matrix corresponding to the current path; The first calculation module is also used to obtain multiple task nodes traversed by the current path and multiple data objects corresponding to each task node; generate rows of the current data matrix based on the task nodes, and generate columns of the current data matrix based on the data objects; The acquisition module is used to acquire multiple alternative subsequent paths corresponding to the current execution instance; it traverses each possible subsequent execution path through the current execution instance; during the traversal, it judges the parameters based on the condition judgment formula of the gateway node through the current process execution path and parameters, and excludes paths whose current parameters do not meet the judgment conditions, thereby obtaining multiple alternative subsequent paths. The second calculation module is used to calculate the candidate data matrix corresponding to each of the candidate subsequent paths; The third calculation module is used to calculate the similarity results between each of the candidate data matrices and the current data matrix, and to determine the candidate data matrix with the largest similarity result as the target matrix; The determination module is used to determine the candidate subsequent paths corresponding to the target matrix as the target path.

8. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program that executes the path prediction method according to any one of claims 1 to 6 when the processor is running.

9. A computer-readable storage medium, characterized in that, It stores a computer program that, when run on a processor, executes the path prediction method according to any one of claims 1 to 6.

Images