Electric power overhaul sheet checking method based on multi-view random forest and electronic equipment

Abstract

The invention discloses an electric power overhaul sheet checking method based on a multi-view random forest and electronic equipment. The method comprises the following steps: acquiring a to-be-checked electric power overhaul list; the multi-view random forest is used for determining a normative verification score, the natural language processing module is used for determining an integrity verification score, and the deep learning module is used for determining a risk verification score. The initial verification score is determined based on a normative verification score, an integrity verification score, and a risk verification score; and determining a target verification result based on a comparison result between the initial verification score and a preset score threshold, the target verification result being used for determining whether the electric power maintenance sheet is qualified or not. According to the invention, the technical problem of low safety and efficiency of the electric power overhaul work caused by the fact that the standardization, integrity and risk of the electric power overhaul sheet cannot be comprehensively measured by the electric power overhaul sheet method in the related technology is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution network maintenance, and in particular to a power maintenance single verification method and electronic equipment based on multi-view random forest. Background Art

[0002] With the increasing complexity and scale of power systems, the traditional power maintenance order verification method has been unable to meet the needs of modern power systems. The traditional power maintenance order verification method mainly relies on manual review and verification, which is not only inefficient but also prone to errors. The main problems are as follows:

[0003] (1) Failure to comprehensively measure the standardization, completeness and risk verification of power maintenance orders: Focusing on certain specific information in the maintenance order, while ignoring the comprehensive evaluation of the entire maintenance process. This may lead to potential errors or omissions in the maintenance order, thus affecting the safety of the maintenance work.

[0004] (2) Lack of automation and intelligence: Usually relying on manual operation and judgment, this not only limits the speed and efficiency of verification, but also increases the possibility of error. In complex power systems, manual verification is difficult to quickly and accurately process large amounts of data and information.

[0005] (3) Lack of flexibility and adaptability: When a maintenance order fails the safety check, a lot of manual intervention is required for adjustment and optimization instead of providing an automated solution, which not only increases the workload but also may lead to inconsistency and instability in the check results.

[0006] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0007] The embodiment of the present invention provides a method and electronic device for verifying an electric power inspection form based on a multi-view random forest, so as to at least solve the technical problem that the electric power inspection form method in the related art fails to comprehensively measure the standardization, integrity and risks of the electric power inspection form, resulting in low safety and efficiency of the electric power inspection work.

[0008] According to one embodiment of the present invention, a method for verifying an electric power inspection order based on a multi-view random forest is provided, comprising: obtaining an electric power inspection order to be verified; verifying the electric power inspection order using an inspection order verification model to obtain an initial verification score, wherein the inspection order verification model comprises a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine a normative verification score, the natural language processing module is used to determine an integrity verification score, the deep learning module is used to determine a risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score; based on the comparison result between the initial verification score and a preset score threshold, a target verification result is determined, wherein the target verification result is used to determine whether the electric power inspection order is qualified.

[0009] Optionally, the maintenance order verification model is used to verify the power maintenance order to obtain an initial verification score, including: using a multi-view random forest module to verify the maintenance task information in the power maintenance order to obtain a normative verification score, wherein the maintenance task information includes maintenance time information, maintenance equipment information, maintenance project information, maintenance operation information and safety measures information; using a natural language processing module to detect preset options in the power maintenance order to obtain an integrity verification score; using a deep learning module to predict the risk level of the power maintenance order to obtain a risk verification score; and determining an initial verification score based on the normative verification score, the integrity verification score and the risk verification score.

[0010] Optionally, the multi-view random forest module includes a numerical view random forest, a classification view random forest and a text view random forest. The multi-view random forest module is used to verify the maintenance task information in the power maintenance order to obtain a normative verification score, including: inputting the maintenance time information in the maintenance task information into the numerical view random forest to obtain a first normative verification score; inputting the maintenance equipment information and the maintenance project information into the classification view random forest to obtain a second normative verification score; inputting the maintenance operation information and the safety measures information into the text view random forest to obtain a third normative verification score.

[0011] Optionally, using a natural language processing module to detect preset options in the power maintenance form to obtain an integrity verification score includes: using a preset word segmentation method in the natural language processing module to perform word segmentation on the text data in the power maintenance form to obtain text features; based on the text features, detecting the information integrity in the preset options to obtain an integrity verification score.

[0012] Optionally, a deep learning module is used to predict the risk level of the power maintenance order to obtain a risk verification score, including: standardizing the time data in the power maintenance order to obtain time features; encoding the maintenance equipment information and maintenance project information to obtain label information; inputting the time features and label information into a neural network to obtain a risk level; and scoring the risk according to the risk level to obtain a risk verification score.

[0013] Optionally, determining the initial verification score based on the normative verification score, the integrity verification score and the risk verification score includes: weighted matching the first normative verification score and the integrity verification score to obtain a first verification score; weighted matching the second normative verification score and the third normative verification score to obtain a second verification score; weighted matching the first verification score and the second verification score to obtain a third verification score; weighted matching the third verification score and the risk verification score to obtain a fourth verification score; weighted matching the first verification score, the second verification score and the fourth verification score to obtain the initial verification score.

[0014] Optionally, based on the comparison result between the initial verification score and the preset score threshold, determining the target verification result includes: in response to the initial verification score being less than the preset score threshold, determining the target verification result as the power inspection order being unqualified; in response to the initial verification score being greater than or equal to the preset score threshold, determining the target verification result as the power inspection order being qualified.

[0015] According to one embodiment of the present invention, there is also provided an electric power inspection order verification device based on multi-view random forest, including: an acquisition module, used to obtain the electric power inspection order to be verified; a verification module, used to verify the electric power inspection order using the inspection order verification model to obtain an initial verification score, wherein the inspection order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine the normative verification score, the natural language processing module is used to determine the integrity verification score, the deep learning module is used to determine the risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score; a determination module, used to determine the target verification result based on the comparison result between the initial verification score and a preset score threshold, wherein the target verification result is used to determine whether the electric power inspection order is qualified.

[0016] Optionally, the power maintenance order verification device based on multi-view random forest also includes a processing module, which is used to use the multi-view random forest module to verify the maintenance task information in the power maintenance order to obtain a normative verification score, wherein the maintenance task information includes maintenance time information, maintenance equipment information, maintenance project information, maintenance operation information and safety measures information; the power maintenance order verification device based on multi-view random forest also includes a detection module, which is used to use the natural language processing module to detect preset options in the power maintenance order to obtain an integrity verification score; the power maintenance order verification device based on multi-view random forest also includes a prediction module, which is used to use the deep learning module to predict the risk level of the power maintenance order to obtain a risk verification score; the determination module is also used to determine the initial verification score based on the normative verification score, the integrity verification score and the risk verification score.

[0017] Optionally, the processing module is also used to: input the maintenance time information in the maintenance task information into the numerical view random forest to obtain a first normative verification score; input the maintenance equipment information and the maintenance project information into the classification view random forest to obtain a second normative verification score; input the maintenance operation information and safety measures information into the text view random forest to obtain a third normative verification score.

[0018] Optionally, the detection module is also used to: perform word segmentation processing on the text data in the power maintenance order using a preset word segmentation method in the natural language processing module to obtain text features; detect the information integrity in the preset options based on the text features to obtain an integrity verification score.

[0019] Optionally, the prediction module is also used to: standardize the time data in the power maintenance order to obtain time characteristics; encode the maintenance equipment information and maintenance project information to obtain label information; input the time characteristics and label information into the neural network to obtain the risk level; and perform risk scoring according to the risk level to obtain a risk verification score.

[0020] Optionally, the prediction module is also used to: weightedly match the first normative verification score and the integrity verification score to obtain a first verification score; weightedly match the second normative verification score and the third normative verification score to obtain a second verification score; weightedly match the first verification score and the second verification score to obtain a third verification score; weightedly match the third verification score and the risk verification score to obtain a fourth verification score; weightedly match the first verification score, the second verification score, and the fourth verification score to obtain an initial verification score.

[0021] Optionally, the determination module is also used to: in response to the initial verification score being less than a preset score threshold, determine that the target verification result is that the power inspection order is unqualified; in response to the initial verification score being greater than or equal to a preset score threshold, determine that the target verification result is that the power inspection order is qualified.

[0022] According to one embodiment of the present invention, there is also provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the above-mentioned power maintenance single verification method based on multi-view random forest is executed when the program is running.

[0023] According to one embodiment of the present invention, a non-volatile storage medium is further provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned power maintenance single verification method based on multi-view random forest when running.

[0024] According to one embodiment of the present invention, a computer program product is also provided, including a computer program, which implements the above-mentioned power maintenance single verification method based on multi-view random forest when executed by a processor.

[0025] In an embodiment of the present invention, a method is adopted in which an electric power inspection form to be verified is obtained, and an inspection form verification model is used to verify the electric power inspection form to obtain an initial verification score. Based on the comparison result between the initial verification score and a preset score threshold, a target verification result is determined, thereby achieving the purpose of comprehensively measuring the standardization, integrity and risks of the electric power inspection form, thereby achieving the technical effect of improving the safety and efficiency of the electric power inspection work, and further solving the technical problem that the electric power inspection form method in the related art fails to comprehensively measure the standardization, integrity and risks of the electric power inspection form, resulting in low safety and efficiency of the electric power inspection work. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0027] Figure 1 It is a hardware structure block diagram of a power maintenance single verification method based on multi-view random forest according to one embodiment of the present invention;

[0028] Figure 2 is a flow chart of a power maintenance single verification method based on multi-view random forest according to one embodiment of the present invention;

[0029] Figure 3 is a flow chart of another electric power maintenance single verification method based on multi-view random forest according to one embodiment of the present invention;

[0030] Figure 4 It is a structural block diagram of a power maintenance single verification device based on multi-view random forest according to one embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. 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 embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] The embodiment of the electric power maintenance single verification method based on multi-view random forest provided in the embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 1 is a hardware structure block diagram of a power maintenance single verification method based on multi-view random forest according to one embodiment of the present invention. Figure 1 As shown, the computer terminal 10 (or electronic device 10) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission module 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0034] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or electronic device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0035] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the electric power maintenance single verification method based on multi-view random forest in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the above-mentioned electric power maintenance single verification method based on multi-view random forest is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0036] The transmission module 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission module 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission module 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0037] The display may be, for example, a touch screen type liquid crystal display (LCD), which may enable a user to interact with a user interface of the computer terminal 10 (or electronic device).

[0038] It should be noted that, in some optional embodiments, the above Figure 1 The computer device (or electronic device) shown may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of hardware elements and software elements. It should be noted that Figure 1This is merely one example of a particular embodiment and is intended to illustrate the types of components that may be present in the above-described computer device (or electronic device).

[0039] According to an embodiment of the present invention, a method embodiment of an electric power maintenance single verification method based on multi-view random forest is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0040] Figure 2 is a flow chart of a power maintenance single verification method based on multi-view random forest according to one embodiment of the present invention, such as Figure 2 As shown, the method comprises the following steps:

[0041] Step S20, obtaining the power maintenance order to be checked;

[0042] In step S20, the above-mentioned power maintenance order includes but is not limited to: maintenance task information and safety measures information, and obtains a normative verification score, wherein the maintenance task information includes maintenance time information, maintenance equipment information, maintenance project information, maintenance operation information and safety measures information;

[0043] Step S22, using the maintenance order verification model to verify the power maintenance order to obtain an initial verification score, wherein the maintenance order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine the normative verification score, the natural language processing module is used to determine the integrity verification score, the deep learning module is used to determine the risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score;

[0044] Specifically, the power maintenance order is input into the maintenance order verification model, and the maintenance order verification model is used to verify the power maintenance order, wherein the maintenance order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module.

[0045] Furthermore, the multi-view random forest is used to determine the normative verification score of the power maintenance order, the natural language processing module is used to determine the integrity verification score of the power maintenance order, and the deep learning module is used to determine the risk verification score of the power maintenance order. The initial verification score of the power maintenance order is determined based on the normative verification score, the integrity verification score and the risk verification score.

[0046] Step S24, based on the comparison result between the initial verification score and the preset score threshold, determine the target verification result, wherein the target verification result is used to determine whether the power maintenance order is qualified.

[0047] Specifically, the initial verification score is compared with the preset score threshold to obtain a comparison result, and the target verification result is determined based on the comparison result, wherein the target verification result indicates that the power inspection order is qualified or the power inspection order is unqualified. Unqualified power inspection orders are intelligently returned, and the entire process is recorded. The target verification results and feedback process are stored in the power inspection order database, and the error modes and improvements of the power inspection orders are regularly analyzed. Therefore, the target verification results are not only used for the evaluation of the current power inspection order, but also used as feedback to optimize the subsequent power inspection order verification process. For example, the person who reported the power inspection order and the relevant responsible persons are automatically notified through SMS, email, etc., and the specific errors and improvement suggestions are pointed out.

[0048] In the above steps S20 to S24, the power inspection form to be verified is obtained, and the power inspection form verification model is used to verify the power inspection form to obtain an initial verification score. Based on the comparison result between the initial verification score and the preset score threshold, the target verification result is determined, thereby achieving the purpose of comprehensively measuring the standardization, integrity and risks of the power inspection form, thereby achieving the technical effect of improving the safety and efficiency of power inspection work, and further solving the technical problem of low safety and efficiency of power inspection work caused by the power inspection form method in the related art due to the failure to comprehensively measure the standardization, integrity and risks of the power inspection form.

[0049] Optionally, in step S22, the power maintenance order is verified using the maintenance order verification model to obtain an initial verification score including:

[0050] Step S221, using the multi-view random forest module to check the maintenance task information in the power maintenance order to obtain a normative verification score, wherein the maintenance task information includes maintenance time information, maintenance equipment information, maintenance item information, maintenance operation information and safety measures information;

[0051] In step S221, the above maintenance time information is used to characterize the equipment maintenance time and equipment operation time. For example, a transformer needs to undergo a routine inspection every 6 months, and the maintenance time information will record the date of the next maintenance, such as "2023-12-15". At the same time, if the transformer has been running for 180 days since the last maintenance, the equipment operation time will also be recorded.

[0052] The above maintenance equipment information is used to represent the equipment name information and the unique identifier information of the equipment.

[0053] The above maintenance project information is used to characterize the maintenance work content that needs to be performed, such as inspection, cleaning, replacement of parts, etc. For example, the maintenance project information may include "check whether the insulator has cracks", "clean the transformer oil", "replace the aged coil", etc.

[0054] The above maintenance operation information is used to characterize the maintenance operation steps, operators, required tools and other information. For example, the operation information includes: Operation steps: disconnect the power supply of the transformer, ensure that the transformer is in a power-free state, and hang a warning sign of "Do not close the switch, someone is working"; use a screwdriver and a wrench to disassemble the transformer casing according to the established procedures; install the three new coils into the transformer in the correct way; use a multimeter and other testing tools to test the electrical performance of the transformer to ensure that all indicators meet the standards; inject 1 barrel of transformer oil to ensure that the oil level reaches the specified position; perform no-load and load tests, observe the operation of the transformer and record the data. Operator: Electrician A is responsible. Required tools: screwdriver, wrench, 3 new coils and 1 barrel of transformer oil.

[0055] Step S222, using a natural language processing module to detect preset options in the power maintenance order to obtain a completeness verification score;

[0056] In step S222, the above-mentioned preset options are used to represent the mandatory options in the power maintenance form, such as maintenance operation steps, maintenance work content, safety measures, etc.

[0057] Specifically, the natural language processing module is used to detect whether the required options in the power maintenance form have been completed, and obtain the completeness verification score S required .

[0058] Step S223, using the deep learning module to predict the risk level of the power maintenance order and obtain a risk verification score;

[0059] Specifically, the deep learning module uses deep neural networks (DNN) to predict the risk level of power maintenance orders. During the training process of DNN, the safety event records in the historical power maintenance order data are used as labels. The labels can be binary categories. For example, risky and non-risk) The labels can also be different levels of risk. For example, low risk, medium-low risk, medium risk, medium-high risk, high risk; the loss function uses the cross entropy loss function; the optimization method selects Adam's gradient descent algorithm. The power maintenance order is input into the trained DNN, and the risk level corresponding to the power maintenance order is output. Finally, the risk verification score S is obtained according to the risk level. risk .

[0060] Step S224, determining an initial verification score based on the normative verification score, the integrity verification score and the risk verification score.

[0061] Specifically, the initial verification score S is obtained by weighting the normative verification score, integrity verification score and risk verification score. total .

[0062] Based on the above steps S221 to S224, the multi-view random forest module is used to verify the maintenance task information in the power maintenance order to obtain a standardization verification score; the natural language processing module is used to detect the preset options in the power maintenance order to obtain an integrity verification score; the deep learning module is used to predict the risk level of the power maintenance order to obtain a risk verification score; the initial verification score is determined based on the standardization verification score, the integrity verification score and the risk verification score, and the standardization, integrity and risk verification scores are weighted and matched to obtain a comprehensive initial verification score, which can comprehensively reflect the quality of the power maintenance order and achieve the purpose of comprehensively measuring the standardization, integrity and risk of the power maintenance order, thereby achieving the technical effect of improving the safety and efficiency of the power maintenance work, and further solving the technical problem that the power maintenance order method in the related technology fails to comprehensively measure the standardization, integrity and risk of the power maintenance order, resulting in low safety and efficiency of the power maintenance work.

[0063] Optionally, in step S221, the multi-view random forest module includes a numerical view random forest, a classification view random forest, and a text view random forest. The multi-view random forest module is used to check the maintenance task information in the power maintenance order, and the normative verification score obtained includes:

[0064] Step S2211, inputting the maintenance time information in the maintenance task information into the numerical view random forest to obtain a first normative verification score;

[0065] Step S2212, input the maintenance equipment information and maintenance item information into the classification view random forest to obtain a second normative verification score;

[0066] Step S2213, input the maintenance operation information and safety measures information into the text view random forest to obtain the third normative verification score.

[0067] Specifically, the multi-view random forest module includes numerical view random forest, classification view random forest and text view random forest. The establishment process of the multi-view random forest module is as follows: 1. Data collection and preprocessing: (1) Data collection: Extract maintenance order data from the power system maintenance management platform. The data fields include maintenance time information, maintenance equipment information, maintenance project information, maintenance operation information and safety measures information. The data source may include databases, spreadsheet files, CSV files, etc. (2) Data preprocessing: Remove duplicate records and delete duplicate items in the maintenance order; check and remove abnormal data, such as invalid records of equipment names or maintenance time; for numerical data, linear interpolation is used to fill missing values. Jieba, a word segmentation tool in natural language processing (NLP) technology, is used to segment the text content of the maintenance order, including the removal of stop words and the extraction of stems.

[0068] 2. Multi-feature extraction: (1) Numerical data feature extraction: Numerical data includes various numerical fields in the maintenance list, such as equipment maintenance time and equipment operation time. The numerical data is standardized to have the same dimension. The standardization formula is shown in expression (1):

[0069]

[0070] Among them, x i represents the i-th original numerical data; N represents the format of the original data; x 标准化 Represents normalized numeric data.

[0071] (2) Categorical data feature extraction: Categorical data includes maintenance equipment information and maintenance project information. Categorical data needs to be encoded. For ordinary unordered categorical data, one-hot encoding is used for encoding; for sequential categorical data, label encoding is used.

[0072] (3) Text data feature extraction: The text data is the maintenance operation information and safety measures information in the maintenance list. The TF-IDF vectorization method is used to convert the text data into a feature vector. The TF-IDF vectorization process is shown in expression (2).

[0073]

[0074] Among them, TF(t,d) represents the frequency of word t in document d; N represents the total number of documents; DF(t) represents the number of documents containing word t.

[0075] (4) Word embedding: Using the pre-trained Word2Vec word embedding model, the text data of historical maintenance orders is collected as the training corpus of the word embedding model to learn the semantic representation of words. Then, the text data in each maintenance order is converted into a word vector, and then the word vectors in each maintenance order are averaged to generate a document-level feature vector.

[0076] 3. Multi-view random forest model establishment: In order to be able to properly handle different features of numerical, classification and text types, a multi-view random forest model is constructed. The multi-view random forest model consists of multiple sub-random forests, each of which processes a type of feature view, namely numerical view, classification view and text view. Then, the results of these sub-random forests are integrated to obtain the final verification result.

[0077] (1) Feature combination: construct feature views of numerical data features, categorical data features, and text data features respectively. That is, the numerical feature is T 运行 and T 检修 , the classification features are the unique hot encoding of the maintenance equipment information and maintenance project information, and the text features are the document-level feature vectors generated by word embedding.

[0078] (2) View definition: Numerical view V num :V num =[T 运行 ,T 检修 ]; Classification view V cat :V cat =[unique hot encoding of maintenance equipment information, unique hot encoding of maintenance project information]; text view V Test :V Test =[document level feature vector].

[0079] (3) Construction of sub-random forests: A separate random forest is constructed for each view, namely, random forest for numerical view, random forest for classification view, and random forest for text view. The construction process of sub-random forests is as follows. 1) Random sampling: Multiple subsets are generated by random sampling with replacement from the training data set of the corresponding view. Each subset is used to train a decision tree. Assume that M decision trees are generated, and the training data subset of each tree is Where view is the view category. 2) Decision tree generation: For each training data subset Select the optimal features and split points to maximize information gain to build a decision tree. Then set the maximum depth of a tree or the minimum number of samples of a leaf node to avoid overfitting. The calculation process of information gain is shown in expression (3).

[0080]

[0081] Among them, IG(T,X) represents the information gain of feature X to data set T; H(T) represents the entropy of data set T; T i It is the data subset when feature X takes the i-th value; n represents the number of values ​​of feature X.

[0082] (4) Tree integration: The integration method is majority voting. By integrating the prediction results of all decision trees in each view, the accuracy and stability of the multi-view random forest model are improved. The specific process of majority voting is shown in expression (4).

[0083]

[0084] in, Indicates the prediction result of the mth tree for sample x; mode() means taking the mode; In particular, they are the final prediction results of the random forest for the corresponding views.

[0085] Then, the weighted voting method is used to assign different weights according to the prediction accuracy of the random forest of each view and fuse the prediction results of each view. The specific process is shown in expression (5).

[0086]

[0087] Among them, w view Represents the weight of the view; Indicates the prediction result of the corresponding view; mode() means taking the mode.

[0088] Finally, we use the stacking ensemble method to take the prediction results of each view random forest as new features and train a meta-classifier for the final classification:

[0089] (5) Model optimization: After the model training is completed, feature importance analysis is performed to evaluate the contribution of each feature to the model prediction results. The feature importance calculation process is shown in expression (6). According to the results of feature importance analysis, features with higher importance are selected for modeling, reducing feature dimensions and improving the model training speed and prediction performance.

[0090] Finally, the cross-validation method is used to tune the hyperparameters of the multi-view random forest model.

[0091]

[0092] Among them, FI j represents the importance of feature j; T m represents the node set of the mth tree; ΔG t is the Gini gain of node t; represents the indicator function, where the value is 1 when feature j is in node t and 0 otherwise.

[0093] Furthermore, after data preprocessing and feature extraction of the power maintenance order, the feature vector x=

[0094] [T 运行 ,T 检修 , unique hot encoding of maintenance equipment information, unique hot encoding of maintenance project information, document level feature vector]. The multi-view random forest module will process different views in the feature vector according to the feature vector, so as to verify the standardization score of the power maintenance order.

[0095] Specifically, the multi-view random forest module inputs the maintenance time information in the maintenance task information into the numerical view random forest to obtain the first normative verification score S fomat , where the first normative verification score S fomat As shown in expression (7).

[0096]

[0097] in, Represents the verification result of the i-th time item, where 0 indicates that the maintenance time information does not meet the specifications of the power maintenance order, and 1 indicates that the maintenance time information meets the specifications of the power maintenance order; N is the total number of time items.

[0098] The multi-view random forest module inputs the maintenance equipment information and maintenance project information into the classification view random forest to obtain the second normative verification score S concent , where the second normative verification score S concent As shown in expression (8).

[0099]

[0100] in, Represents the verification result of the i-th maintenance equipment and maintenance project item, where 0 means that both the maintenance equipment and the maintenance project do not meet the specifications of the power maintenance order, 1 means that the maintenance equipment or maintenance project meets the specifications of the power maintenance order, and 2 means that both the maintenance equipment and maintenance project meet the specifications of the power maintenance order; N is the total number of maintenance equipment and maintenance project items.

[0101] The maintenance operation information and safety measures information of the multi-view random forest module are input into the text view random forest to obtain the third normative verification score S safety , among which, the third normative verification score S safety As shown in expression (9).

[0102]

[0103] in, Indicates the verification result of the i-th maintenance operation information and safety measure information item, where 0 indicates that both the maintenance operation information and the safety measure information do not conform to the specifications of the power maintenance order, 1 indicates that either the maintenance operation information or the safety measure information conforms to the specifications of the power maintenance order, and 2 indicates that both the maintenance operation information and the safety measure information conform to the specifications of the power maintenance order; N is the total number of maintenance operation information and safety measure information items.

[0104] Based on the above steps S2211 to S2213, the maintenance time information in the maintenance task information is input into the numerical view random forest to obtain the first normative verification score; the maintenance equipment information and the maintenance project information are input into the classification view random forest to obtain the second normative verification score; the maintenance operation information and the safety measure information are input into the text view random forest to obtain the third normative verification score. Through the random forest models of the numerical view, the classification view, and the text view, different types of data in the power maintenance order can be specifically processed and analyzed, and the characteristics of the data can be captured more comprehensively, improving the accuracy of verification.

[0105] Optionally, in step S222, using the natural language processing module to detect the preset options in the power maintenance order, the integrity verification score includes:

[0106] Step S2221, using the preset word segmentation method in the natural language processing module to perform word segmentation processing on the text data in the power maintenance order to obtain text features;

[0107] Step S2222, based on the text features, detect the information integrity in the preset options to obtain the integrity verification score.

[0108] Specifically, use the word segmentation tool Jieba in the NLP module to perform word segmentation processing on the information in the power maintenance order to obtain text features, where the word segmentation processing includes removing stop words, extracting stems, etc. Use the text features obtained after word segmentation processing to identify the required fields and check whether they are filled in completely. For each required field f i , if it is filled in completely, the score is 1, otherwise the score is 0. The process of checking the required fields is shown in expression (10).

[0109]

[0110] Among them, R i represents the verification result of the i-th required field, filled in as 1, not filled in as 0; K is the total number of required fields. If S required <K, it indicates that the integrity verification is unqualified, and S required =0.

[0111] Based on the above steps S2221 to S2222, the text data in the power maintenance form is segmented using the preset word segmentation method in the natural language processing module to obtain text features; the information integrity in the preset options is detected based on the text features to obtain an integrity verification score. The natural language processing module can accurately identify key information in the text and improve the accuracy of verification. The automated word segmentation and verification process can significantly improve the processing speed and reduce the time required for manual review. The use of preset word segmentation rules and verification standards can ensure that all maintenance forms are reviewed according to unified standards, reducing errors caused by human factors, and the natural language processing module can be easily integrated into the existing power maintenance management system without the need for large-scale system reconstruction.

[0112] Optionally, in step S223, the risk level of the power maintenance order is predicted by using a deep learning module, and the risk verification score obtained includes:

[0113] Step S2231, standardize the time data in the power maintenance list to obtain time characteristics;

[0114] Step S2232, encoding the maintenance equipment information and the maintenance item information to obtain label information;

[0115] Step S2233, input the time feature and label information into the neural network to obtain the risk level;

[0116] Step S2234, perform risk scoring according to the risk level to obtain a risk verification score.

[0117] Specifically, the time data in the power maintenance list is standardized according to expression (1) to obtain the time feature: V num =[T 运行 ,T 检修 ], use the unique hot encoding method for the maintenance equipment information and maintenance project information to obtain the label information V cat = [Unique hot encoding of maintenance equipment information, unique hot encoding of maintenance project information]. Combine the time feature and label information to obtain the combined vector x = [T 运行, T 检修 , unique hot encoding of maintenance equipment information, unique hot encoding of maintenance project information]. Input the combined vector x corresponding to the power maintenance order into the trained DNN, output the risk level corresponding to the power maintenance order, and finally obtain the risk verification score S according to the risk level. risk .

[0118] From step S2231 to step S2234, the time data in the power maintenance form is standardized to obtain time features; the maintenance equipment information and maintenance project information are encoded to obtain label information; the time features and label information are input into the neural network to obtain the risk level; the risk is scored according to the risk level to obtain the risk verification score. Through the automated feature extraction and risk assessment process, human intervention can be reduced and the accuracy of the assessment can be improved. The deep learning model can quickly process large amounts of data, thereby providing risk assessment results in a short time, thereby improving the efficiency of power maintenance work.

[0119] Optionally, in step S224, determining the initial verification score based on the normative verification score, the integrity verification score, and the risk verification score includes:

[0120] Step S2241, weighting and matching the first normativeness verification score and the integrity verification score to obtain a first verification score;

[0121] Specifically, the first normative verification score and the integrity verification score are weighted and matched to obtain the first verification score, wherein the calculation process of the first verification score is shown in expression (11).

[0122] S norm =αS required +βS fomat (11)

[0123] Among them, α and β represent weight coefficients respectively; S required Indicates the integrity verification score; S fomat Represents the first normative verification score.

[0124] Step S2242, weighting and matching the second normative verification score and the third normative verification score to obtain a second verification score;

[0125] Specifically, the second normative verification score and the third normative verification score are weighted and matched to obtain the second verification score, wherein the calculation process of the second verification score is shown in expression (12).

[0126] S correct =γS content +δS safety (12)

[0127] Among them, γ and δ represent weight coefficients respectively; S concent Indicates the second normative verification score; S safety Represents the third normative verification score.

[0128] Step S2243, weighting and matching the first verification score and the second verification score to obtain a third verification score;

[0129] Specifically, the first verification score and the second verification score are weighted and matched to obtain a third verification score, wherein the calculation process of the third verification score is shown in expression (13).

[0130] S initialsafety =λS norm +μS correct (13)

[0131] Among them, λ and μ represent weight coefficients respectively; S norm Indicates the first verification score; S correct Indicates the second verification score.

[0132] Step S2244, weighting and matching the third verification score and the risk verification score to obtain a fourth verification score;

[0133] Specifically, the third verification score and the risk verification score are weighted and matched to obtain the fourth verification score, wherein the calculation process of the fourth verification score is shown in expression (14).

[0134] S finalsafety =νS initialsafety +ξS risk (14)

[0135] Among them, ν, ξ represent weight coefficients respectively; S initial safety Indicates the third verification score; S risk Indicates the risk verification score.

[0136] Step S2245, weighted matching is performed on the first verification score, the second verification score and the fourth verification score to obtain an initial verification score.

[0137] Specifically, the first verification score, the second verification score and the fourth verification score are weighted and matched to obtain an initial verification score, wherein the calculation process of the initial verification score is shown in expression (15).

[0138] S total =αS norm +βS correct +γS final safety (15)

[0139] Among them, α, β, and γ represent weight coefficients respectively; S norm Indicates the first verification score; S correct Indicates the second verification score; S final safety Indicates the fourth verification score.

[0140] Based on the above steps S2241 to S2245, the first normative verification score and the integrity verification score are weighted and matched to obtain the first verification score; the second normative verification score and the third normative verification score are weighted and matched to obtain the second verification score; the first verification score and the second verification score are weighted and matched to obtain the third verification score; the third verification score and the risk verification score are weighted and matched to obtain the fourth verification score; the first verification score, the second verification score and the fourth verification score are weighted and matched to obtain the initial verification score. By combining the scores of normativeness, integrity, risk and other aspects, a more comprehensive verification score can be obtained. The compliance and risks of the power maintenance order can be evaluated accurately, and the focus of the evaluation can be adjusted according to the specific situation by adjusting the weight coefficients of different scores, so that the score is more in line with the actual needs. The weight coefficient can be adjusted according to the feedback of the actual evaluation to adapt to the ever-changing evaluation standards and environment, so as to achieve the purpose of comprehensively measuring the standardization, integrity and risks of the power maintenance order, thereby achieving the technical effect of improving the safety and efficiency of the power maintenance work, and further solving the technical problem of low safety and efficiency of the power maintenance work caused by the power maintenance order method in the related technology due to the failure to comprehensively measure the standardization, integrity and risks of the power maintenance order.

[0141] Optionally, in step S24, based on the comparison result between the initial verification score and the preset score threshold, determining the target verification result includes:

[0142] Step S241, in response to the initial verification score being less than a preset score threshold, determining that the target verification result is that the power maintenance order is unqualified;

[0143] Step S242, in response to the initial verification score being greater than or equal to the preset score threshold, determining that the target verification result is that the power maintenance order is qualified.

[0144] Specifically, set a preset scoring threshold T 0 , compare the initial verification score with the preset score threshold T 0 The comparison is performed to obtain a comparison result, and the target verification result is determined according to the comparison result, wherein the target verification result indicates that the power inspection order is qualified or the power inspection order is unqualified. When the initial verification score is less than the preset score threshold T 0 When the target verification result is determined to be unqualified for the power maintenance order; when the initial verification score is greater than or equal to the preset score threshold T 0, determine that the target verification result is qualified for the power maintenance order. Intelligently return unqualified power maintenance orders and record the entire process. Store the target verification results and feedback process in the power maintenance order database, and regularly analyze the error modes and improvements of the power maintenance order. Therefore, the target verification results are not only used to evaluate the current power maintenance order, but also used as feedback to optimize the subsequent power maintenance order verification process. For example, the person who reported the power maintenance order and the relevant responsible persons are automatically notified through SMS, email, etc., and specific errors and improvement suggestions are pointed out.

[0145] Furthermore, a preset scoring threshold T is set 0 It is a process that comprehensively considers data analysis, model evaluation, business needs and security strategies, and needs to be further determined based on actual conditions. The final threshold can be determined from the following aspects: (1) Based on data analysis and the use of historical data: Use the data of historical maintenance orders to analyze the distribution of their comprehensive scores. By statistically analyzing the scores of known qualified and unqualified maintenance orders, the boundaries of the score distribution can be identified. Risk preference analysis can also be set to determine the system's tolerance for risk. For example, if the system wants to be more conservative to minimize any potential risks, the threshold can be set higher. (2) Based on business needs and security strategies: Set different preset score thresholds T according to the content of the maintenance order and its importance to the operation of the power system. 0 For critical equipment or high-risk areas, the preset scoring threshold T 0 It should be stricter. For example, multiple levels of low, medium, and high thresholds should be set, and the corresponding preset scoring thresholds T 0 The trigger response can be to trigger the most basic alarm, trigger a higher level alarm or require further review, and trigger the highest warning to automatically return the repair order and require resubmission.

[0146] Based on the above steps S241 to S242, in response to the initial verification score being less than the preset score threshold, the target verification result is determined to be that the power maintenance order is unqualified; in response to the initial verification score being greater than or equal to the preset score threshold, the target verification result is determined to be that the power maintenance order is qualified. Objectively evaluating the maintenance order through the preset score threshold can reduce judgment errors caused by human factors, and by setting up risk preference analysis, risks can be better managed to ensure that maintenance orders for key equipment and high-risk areas receive higher attention, and the verification results and feedback process are stored in the database, which can improve the transparency of the process and facilitate tracking and auditing.

[0147] Figure 3 is a flowchart of another power maintenance single verification method based on multi-view random forest according to one embodiment of the present invention, such as Figure 3 As shown, the method comprises the following steps:

[0148] Step S301, obtaining the power maintenance order to be checked;

[0149] Step S302, using a multi-view random forest module to check the maintenance task information in the power maintenance order to obtain a standardization verification score;

[0150] Step S303, using a natural language processing module to detect preset options in the power maintenance order to obtain a completeness verification score;

[0151] Step S304, using the deep learning module to predict the risk level of the power maintenance order and obtain a risk verification score;

[0152] Step S305, determining an initial verification score based on the compliance verification score, the integrity verification score, and the risk verification score;

[0153] Step S306, in response to the initial verification score being less than a preset score threshold, determining that the target verification result is that the power maintenance order is unqualified;

[0154] Step S307, in response to the initial verification score being greater than or equal to the preset score threshold, determining that the target verification result is that the power maintenance order is qualified.

[0155] Based on the above steps S301 to S307, the method of obtaining the power maintenance order to be verified and verifying the power maintenance order using the maintenance order verification model to obtain an initial verification score is adopted. Based on the comparison result between the initial verification score and the preset score threshold, the target verification result is determined, thereby achieving the purpose of comprehensively measuring the standardization, integrity and risks of the power maintenance order, thereby achieving the technical effect of improving the safety and efficiency of power maintenance work, and further solving the technical problem of low safety and efficiency of power maintenance work caused by the power maintenance order method in the related technology due to the failure to comprehensively measure the standardization, integrity and risks of the power maintenance order.

[0156] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0157] In an embodiment of the present invention, there is also provided an electric power maintenance single verification device based on multi-view random forest, which is used to implement the above-mentioned embodiments and preferred implementation modes, and will not be repeated hereafter. As used below, the term "module" may be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0158] Figure 4 is a structural block diagram of a power maintenance single verification device based on multi-view random forest according to one embodiment of the present invention, such as Figure 4 As shown, the device comprises:

[0159] The acquisition module 401 is used to acquire the power maintenance order to be checked;

[0160] Verification module 402, used to verify the power maintenance order using the maintenance order verification model to obtain an initial verification score, wherein the maintenance order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine the normative verification score, the natural language processing module is used to determine the integrity verification score, the deep learning module is used to determine the risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score;

[0161] The determination module 403 is used to determine a target verification result based on a comparison result between the initial verification score and a preset score threshold, wherein the target verification result is used to determine whether the power maintenance order is qualified.

[0162] Optionally, the electric power maintenance order verification device based on multi-view random forest also includes a processing module 404, which is used to use the multi-view random forest module to verify the maintenance task information in the electric power maintenance order to obtain a normative verification score, wherein the maintenance task information includes maintenance time information, maintenance equipment information, maintenance project information, maintenance operation information and safety measures information; the electric power maintenance order verification device based on multi-view random forest also includes a detection module 405, which is used to use the natural language processing module to detect preset options in the electric power maintenance order to obtain an integrity verification score; the electric power maintenance order verification device based on multi-view random forest also includes a prediction module 406, which is used to use the deep learning module to predict the risk level of the electric power maintenance order to obtain a risk verification score; the determination module 403 is also used to determine the initial verification score based on the normative verification score, the integrity verification score and the risk verification score.

[0163] Optionally, the processing module 404 is also used to: input the maintenance time information in the maintenance task information into the numerical view random forest to obtain a first normative verification score; input the maintenance equipment information and the maintenance project information into the classification view random forest to obtain a second normative verification score; input the maintenance operation information and safety measures information into the text view random forest to obtain a third normative verification score.

[0164] Optionally, the detection module 405 is also used to: perform word segmentation processing on the text data in the power maintenance order using the preset word segmentation method in the natural language processing module to obtain text features; detect the information integrity in the preset options based on the text features to obtain an integrity verification score.

[0165] Optionally, the prediction module 406 is also used to: standardize the time data in the power maintenance order to obtain time characteristics; encode the maintenance equipment information and the maintenance project information to obtain label information; input the time characteristics and label information into the neural network to obtain the risk level; and perform risk scoring according to the risk level to obtain a risk verification score.

[0166] Optionally, the prediction module 406 is also used to: weight the first normative verification score and the integrity verification score to obtain a first verification score; weight the second normative verification score and the third normative verification score to obtain a second verification score; weight the first verification score and the second verification score to obtain a third verification score; weight the third verification score and the risk verification score to obtain a fourth verification score; weight the first verification score, the second verification score, and the fourth verification score to obtain an initial verification score.

[0167] Optionally, the determination module 403 is also used to: in response to the initial verification score being less than a preset score threshold, determine that the target verification result is that the power inspection order is unqualified; in response to the initial verification score being greater than or equal to a preset score threshold, determine that the target verification result is that the power inspection order is qualified.

[0168] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0169] According to one embodiment of the present invention, there is also provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the above-mentioned power maintenance single verification method based on multi-view random forest is executed when the program is running.

[0170] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0171] Step S1, obtaining the power maintenance order to be checked;

[0172] Step S2, using the maintenance order verification model to verify the power maintenance order to obtain an initial verification score, wherein the maintenance order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine the normative verification score, the natural language processing module is used to determine the integrity verification score, the deep learning module is used to determine the risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score;

[0173] Step S3, based on the comparison result between the initial verification score and the preset score threshold, determine the target verification result, wherein the target verification result is used to determine whether the power maintenance order is qualified.

[0174] According to one embodiment of the present invention, a non-volatile storage medium is further provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned power maintenance single verification method based on multi-view random forest when running.

[0175] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0176] Step S1, obtaining the power maintenance order to be checked;

[0177] Step S2, using the maintenance order verification model to verify the power maintenance order to obtain an initial verification score, wherein the maintenance order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine the normative verification score, the natural language processing module is used to determine the integrity verification score, the deep learning module is used to determine the risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score;

[0178] Step S3, based on the comparison result between the initial verification score and the preset score threshold, determine the target verification result, wherein the target verification result is used to determine whether the power maintenance order is qualified.

[0179] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0180] According to one embodiment of the present invention, a computer program product is also provided, including a computer program, which implements the above-mentioned power maintenance single verification method based on multi-view random forest when executed by a processor.

[0181] Optionally, in this embodiment, the computer program product may be configured to store computer program instructions for performing the following steps:

[0182] Step S1, obtaining the power maintenance order to be checked;

[0183] Step S2, using the maintenance order verification model to verify the power maintenance order to obtain an initial verification score, wherein the maintenance order verification model includes a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine the normative verification score, the natural language processing module is used to determine the integrity verification score, the deep learning module is used to determine the risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score;

[0184] Step S3, based on the comparison result between the initial verification score and the preset score threshold, determine the target verification result, wherein the target verification result is used to determine whether the power maintenance order is qualified.

[0185] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0186] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0187] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0188] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0189] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0190] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0191] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0192] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A power maintenance single verification method based on multi-view random forest, characterized in that: include: Obtain the power maintenance order to be verified; The power maintenance order is verified by using a maintenance order verification model to obtain an initial verification score, wherein the maintenance order verification model comprises a multi-view random forest module, a natural language processing module and a deep learning module, the multi-view random forest is used to determine a normative verification score, the natural language processing module is used to determine an integrity verification score, the deep learning module is used to determine a risk verification score, and the initial verification score is determined based on the normative verification score, the integrity verification score and the risk verification score; Based on the comparison result between the initial verification score and the preset score threshold, a target verification result is determined, wherein the target verification result is used to determine whether the power inspection order is qualified.

2. The electric power maintenance single verification method based on multi-view random forest according to claim 1 is characterized in that: The power maintenance order is verified using the maintenance order verification model, and the initial verification score obtained includes: Using the multi-view random forest module to check the maintenance task information in the power maintenance order to obtain the normative verification score, wherein the maintenance task information includes maintenance time information, maintenance equipment information, maintenance item information, maintenance operation information and safety measures information; Using the natural language processing module to detect preset options in the power maintenance order, and obtain the integrity verification score; Using the deep learning module to predict the risk level of the power maintenance order, and obtain the risk verification score; The initial verification score is determined based on the compliance verification score, the integrity verification score, and the risk verification score.

3. The electric power maintenance single verification method based on multi-view random forest according to claim 2 is characterized in that: The multi-view random forest module includes a numerical view random forest, a classification view random forest and a text view random forest. The multi-view random forest module is used to check the maintenance task information in the power maintenance order to obtain the normative verification score, including: Inputting the maintenance time information in the maintenance task information into the numerical view random forest to obtain a first normative verification score; Inputting the maintenance equipment information and the maintenance item information into the classification view random forest to obtain a second normative verification score; The maintenance operation information and safety measures information are input into the text view random forest to obtain a third normative verification score.

4. The electric power maintenance single verification method based on multi-view random forest according to claim 2 is characterized in that: Utilizing the natural language processing module to detect preset options in the power maintenance order and obtain the integrity verification score includes: Using the preset word segmentation method in the natural language processing module to perform word segmentation processing on the text data in the power maintenance order to obtain text features; The information integrity in the preset options is detected based on the text features to obtain the integrity verification score.

5. The electric power maintenance single verification method based on multi-view random forest according to claim 2 is characterized in that: Using the deep learning module to predict the risk level of the power maintenance order to obtain the risk verification score includes: Standardizing the time data in the power maintenance list to obtain time characteristics; Encoding the maintenance equipment information and the maintenance item information to obtain label information; Inputting the time feature and the label information into a neural network to obtain the risk level; A risk score is performed according to the risk level to obtain the risk verification score.

6. The electric power maintenance single verification method based on multi-view random forest according to claim 3 is characterized in that: Determining the initial verification score based on the compliance verification score, the integrity verification score, and the risk verification score includes: Performing a weighted ratio on the first compliance verification score and the integrity verification score to obtain a first verification score; Performing a weighted ratio on the second normative verification score and the third normative verification score to obtain a second verification score; Performing a weighted ratio on the first verification score and the second verification score to obtain a third verification score; Performing weighted matching on the third verification score and the risk verification score to obtain a fourth verification score; The first verification score, the second verification score, and the fourth verification score are weighted and matched to obtain the initial verification score.

7. The electric power maintenance single verification method based on multi-view random forest according to claim 1 is characterized in that: Based on the comparison result between the initial verification score and the preset score threshold, determining the target verification result includes: In response to the initial verification score being less than the preset score threshold, determining that the target verification result is that the power maintenance order is unqualified; In response to the initial verification score being greater than or equal to the preset score threshold, the target verification result is determined to be that the power maintenance order is qualified.

8. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein when the program is run, the power maintenance single verification method based on multi-view random forest as described in any one of claims 1 to 7 is executed.

9. A non-volatile storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the electric power maintenance single verification method based on multi-view random forest as described in any one of claims 1 to 7 when running.

10. A computer program product, characterized in that It comprises a computer program, which, when executed by a processor, implements the electric power maintenance single verification method based on multi-view random forest according to any one of claims 1 to 7.

Images