Operation and maintenance work order problem type dual-forest evolution classification algorithm

Through the two forest evolution classification algorithm for operation and maintenance work order problem types, the problems of operation and maintenance work order type errors and insufficient training of large language models are solved, automatic classification and continuous evolution are realized, and operation and maintenance efficiency and prediction accuracy are improved.

CN119066530APending Publication Date: 2024-12-03SHANXI BIG DATA IND DEV CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411004731.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The specification of operation and maintenance work order types in the prior art can easily lead to manual operations duplication and incorrect judgment, resulting in the wrong handling channels for work order flow, resulting in a decrease in operation and maintenance efficiency, and the training degree of operation and maintenance related terms and problems by large language models Not enough, the matching degree is not accurate enough.

Method used

Using the two forest evolution classification algorithm for operation and maintenance work order problems, gradually improve predictions by extracting keyword groups, building initial data sets, training the two forest classification model, automatically sorting work order problems, recording modification results, continuously evolving data sets and models Accuracy.

Benefits of technology

It reduces the workload of users and operation and maintenance personnel to fill in and modify problem types, improves the accuracy and efficiency of work order processing, avoids the invalid flow of work orders, and continuously improves the accuracy of model prediction results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066530A_ABST
    Figure CN119066530A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of operation and maintenance work order management, and solves the problem that the matching degree of operation and maintenance terms and problem conditions is not accurate enough during prediction of a processing channel and a language model of a work order flow error caused by manual operation. An operation and maintenance work order problem type dual-forest evolution classification algorithm is provided, a trained dual-forest classification model is accessed to an operation and maintenance work order system, and automatic classification of problem types is carried out on work orders filled by a user; if the work order with the inaccurate classification result exists, the operation and maintenance personnel modify the problem type of the work order and record the problem type into an evolution data set; and replacing the old dual-forest classification model with the evolved dual-forest classification model, and continuously performing construction of a new evolution data set and continuous evolution of the dual-forest classification model in continuous operation of the operation and maintenance work order system. A user can be prevented from filling question types, and user troubles are avoided; and meanwhile, an evolutionary mechanism is provided, and the accuracy of a prediction result can be continuously improved along with continuous operation of the system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of operation and maintenance work order management, and in particular relates to a dual forest evolutionary classification algorithm for an operation and maintenance work order problem type. Background Art

[0002] Currently, for the work order problem type in the operation and maintenance work order system, the work order problem type is usually specified when the user fills in the work order. When the operation and maintenance personnel process the work order, if they find that the work order type is incorrect, they will correct it.

[0003] The disadvantage of this method is that repetitive manual operations can easily cause fatigue, and misjudgment can cause work orders to flow to the wrong processing channels, resulting in work orders not being resolved in a timely manner, leading to reduced operation and maintenance efficiency.

[0004] In addition, for intelligent differentiation of work order problem types, large language models can be used for prediction, such as general-purpose artificial intelligence models such as ChatGPT. However, they are not sufficiently trained on terms and problem situations related to operation and maintenance, and the matching degree is not accurate enough. Summary of the invention

[0005] In order to solve at least one of the above technical problems existing in the prior art, the present invention provides a dual forest evolutionary classification algorithm for operation and maintenance work order problem type.

[0006] The present invention is implemented by the following technical solution: a dual forest evolutionary classification algorithm for operation and maintenance work order problem type, comprising the following steps:

[0007] S1: Extract key words from different work order problem overviews and work order problem descriptions in the operation and maintenance work order system to generate simplified work order problem overviews and work order problem descriptions;

[0008] S2: Construct an initial data set: including a simplified work order problem overview and work order problem description as model input and the problem types of work orders in the operation and maintenance work order system with a status of resolved as model output; use the initial data set as a training data set to train the dual forest classification model;

[0009] S3: Integrate the trained dual forest classification model into the operation and maintenance work order system to automatically classify the problem types of the work orders submitted by users;

[0010] S4: The operation and maintenance work end receives the work order that has been classified and processes the work order. If there is a work order with inaccurate classification results, the operation and maintenance personnel modify the problem type of the work order and record it in the evolution data set.

[0011] S5: When the amount of data in the evolutionary data set exceeds the cumulative threshold, the evolutionary data set and the previous training data set are simultaneously used as new training data sets to train and generate the evolved dual forest classification model;

[0012] S6: Replace the old dual forest classification model with the evolved dual forest classification model. While the operation and maintenance work order system is running continuously, continue to build new evolutionary data sets and continuously evolve the dual forest classification model.

[0013] Preferably, keyword groups in different work order problem overviews and work order problem descriptions in the operation and maintenance work order system are extracted based on jieba and NLTK libraries; in the training data set, the proportion of work order problem overviews and work order problem descriptions in the model input is 60% and 40% respectively.

[0014] Preferably, the dual forest classification model is connected to a user work order submission response interface in the operation and maintenance work order system.

[0015] Preferably, a model correction trigger is connected to the operation and maintenance personnel work order modification function interface in the operation and maintenance work order system, and the modification of the problem type is recorded in the evolution data set.

[0016] Preferably, the data in the evolutionary data set includes a work order problem overview and a work order problem description with inaccurate classification results, and a modified problem type of the work order.

[0017] Preferably, the cumulative threshold of the amount of data in the evolution data set is 1000.

[0018] Preferably, the dual forest classification model is verified by a verified data set, and when the accuracy of the problem type of the work order reaches a preset value, the evolved dual forest classification model replaces the old dual forest classification model.

[0019] Preferably, the verified data set is a simplified work order problem overview and work order problem description and the problem types of the work orders with the correctly corresponding status of resolved; the preset value of the correctness of the problem types of the work orders is 70%.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] The present invention can save users from having to fill in the problem type, thus avoiding user trouble. For ordinary users, it is difficult to correctly determine the type of problem they encounter, thus reducing their workload. Operation and maintenance personnel usually do not need to modify the problem type, thus reducing their workload. At the same time, it can avoid the abnormal circulation of work orders. For different problem types, corresponding functional departments will handle the work orders, thus improving operation and maintenance efficiency. Previously, incorrect filling of the problem type will lead to invalid circulation of work orders between different departments.

[0022] In the dual forest classification model, different weights are used for the work order problem overview and the work order problem description, highlighting the impact of the problem overview on the prediction results, fitting the meaning of the data itself, and greatly improving the accuracy of the model's prediction results.

[0023] Conventional classification models do not have an evolutionary mechanism and are fixed after model training. This algorithm uses various triggers embedded in the operation and maintenance system to manually correct the model. As the system continues to run, the accuracy of the prediction results will continue to improve. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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 paying creative work.

[0025] Figure 1 It is the overall flow chart of the present invention. DETAILED DESCRIPTION

[0026] In conjunction with the drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so that people familiar with this technology can understand and read them. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention. It should be noted that in this specification, relational terms such as first and second are only used to distinguish one entity from several other entities, and do not necessarily require or imply any actual relationship or order between these entities.

[0028] The present invention provides an embodiment:

[0029] like Figure 1 As shown, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type includes the following steps:

[0030] S1: Extract key words from different work order problem overviews and work order problem descriptions in the operation and maintenance work order system to generate simplified work order problem overviews and work order problem descriptions;

[0031] S2: Construct an initial data set: including a simplified work order problem overview and work order problem description as model input and the problem types of work orders in the operation and maintenance work order system with a status of resolved as model output; use the initial data set as a training data set to train the dual forest classification model;

[0032] S3: Integrate the trained dual forest classification model into the operation and maintenance work order system to automatically classify the problem types of the work orders submitted by users;

[0033] S4: The operation and maintenance work end receives the work order that has been classified and processes the work order. If there is a work order with inaccurate classification results, the operation and maintenance personnel modify the problem type of the work order and record it in the evolution data set.

[0034] S5: When the amount of data in the evolutionary data set exceeds the cumulative threshold, the evolutionary data set and the previous training data set are simultaneously used as new training data sets to train and generate the evolved dual forest classification model;

[0035] S6: Replace the old dual forest classification model with the evolved dual forest classification model. While the operation and maintenance work order system is running continuously, continue to build new evolutionary data sets and continuously evolve the dual forest classification model.

[0036] In this embodiment, the keyword groups in the work order problem overview and work order problem description in the operation and maintenance work order system are extracted based on Jieba and NLTK libraries; in the training data set, the work order problem overview and work order problem description account for 60% and 40% of the model input respectively. In the dual forest classification model, different weights are used for the work order problem overview and work order problem description, highlighting the impact of the problem overview on the prediction results, fitting the meaning of the data itself, and greatly improving the accuracy of the model's prediction results.

[0037] The dual forest classification model is connected to the user work order submission response interface public_ticket in the operation and maintenance work order system to classify the problem type of the user's work order.

[0038] The model correction trigger is connected to the ticket modification function interface of the operation and maintenance work order system, and the modification of the problem type is recorded in the evolutionary data set. The data in the evolutionary data set includes the problem overview and description of the work order with inaccurate classification results, as well as the problem type of the modified work order.

[0039] The cumulative threshold of the amount of data in the evolutionary data set is 1000. Set a scheduled task to obtain the number of data in the evolutionary data set. When the threshold of 1000 is met, retrain the model.

[0040] The dual forest classification model is verified by the verified data set. When the accuracy of the problem type of the work order reaches the preset value of 70%, the evolved dual forest classification model is used to replace the old dual forest classification model to complete one evolution. The verified data set is the simplified work order problem overview and work order problem description and the problem type of the work order with the correct corresponding status of resolved.

[0041] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A dual forest evolutionary classification algorithm for operation and maintenance work order problem type, characterized by: The following steps are involved: S1: Extract key words from different work order problem overviews and work order problem descriptions in the operation and maintenance work order system to generate simplified work order problem overviews and work order problem descriptions; S2: Construct an initial data set: including a simplified work order problem overview and work order problem description as model input and the problem types of work orders in the operation and maintenance work order system with a status of resolved as model output; use the initial data set as a training data set to train the dual forest classification model; S3: Integrate the trained dual forest classification model into the operation and maintenance work order system to automatically classify the problem types of the work orders submitted by users; S4: The operation and maintenance work end receives the work order that has been classified and processes the work order; If there is a work order with inaccurate classification results, the operation and maintenance personnel will modify the problem type of the work order and record it in the evolutionary data set; S5: When the amount of data in the evolutionary data set exceeds the cumulative threshold, the evolutionary data set and the previous training data set are simultaneously used as new training data sets to train and generate the evolved dual forest classification model; S6: Replace the old dual forest classification model with the evolved dual forest classification model. While the operation and maintenance work order system is running continuously, continue to build new evolutionary data sets and continuously evolve the dual forest classification model.

2. According to claim 1, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: Based on jieba and NLTK libraries, keyword phrases in different work order problem overviews and work order problem descriptions in the operation and maintenance work order system are extracted; in the training data set, the proportion of work order problem overviews and work order problem descriptions in the model input is 60% and 40% respectively.

3. According to claim 1, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: The dual forest classification model is connected to the user work order submission response interface in the operation and maintenance work order system.

4. According to claim 1, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: The model correction trigger is connected to the operation and maintenance personnel work order modification function interface in the operation and maintenance work order system, and the modification of the problem type is recorded in the evolutionary data set.

5. According to claim 4, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: The data in the evolutionary dataset include the problem overview and description of the work order with inaccurate classification results, as well as the modified problem type of the work order.

6. According to claim 5, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: The cumulative threshold of the amount of data in the evolutionary dataset is 1000.

7. According to claim 1, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: The dual forest classification model is verified through the verified data set. When the accuracy of the problem type of the work order reaches the preset value, the evolved dual forest classification model replaces the old dual forest classification model.

8. According to claim 7, a dual forest evolutionary classification algorithm for operation and maintenance work order problem type is characterized by: The verified data set is the simplified work order problem overview and work order problem description and the problem type of the work order with the correctly corresponding status of resolved; The default value for the correctness of the problem type of a ticket is 70%.