Power distribution network work order classification method, device and equipment and storage medium

Abstract

The invention discloses a power distribution network work order classification method, device and equipment and a storage medium. The method comprises the following steps: acquiring a to-be-classified power distribution network work order; cutting the to-be-classified power distribution network work order to obtain a plurality of vocabularies; calculating an inverse document frequency and a word frequency of each vocabulary through a word frequency-inverse document frequency method; based on the inverse document frequency and the word frequency, combining a convolutional neural network to determine the score of the to-be-classified power distribution network work order in each directory level in different categories of complaint work order classification labels; and analyzing the score of each directory level in the different categories to obtain a classification result of the to-be-classified power distribution network work order. According to the method, the scores of the work orders under different classifications are determined by cutting the work orders and combining the word frequency inverse file frequency method and the convolutional neural network, the classification result of the work orders is obtained, and the classification efficiency of the power distribution network work orders can be improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of text classification, and in particular to a distribution network work order classification method, device, equipment and storage medium. Background Art

[0002] User evaluation is an important channel for enterprises to evaluate their own service status and improve user experience. As a power supply service company, we are particularly sensitive to complaint tickets. In the process of handling customer complaint tickets, there are often problems such as long structured information extraction time, deviation in the focus of information extraction, and inability to accurately locate customer demand hotspots. Therefore, it is particularly important to effectively improve analysis efficiency, strengthen analysis depth and breadth, and accurately control demands, so as to accurately locate risk tickets, achieve the purpose of accurate complaint control, and achieve the goal of intelligent improvement of service and analysis capabilities.

[0003] At present, the bottleneck of manual analysis results in low efficiency in the overall analysis of complaint tickets, and it is impossible to accurately and quickly locate the hot spots of user demands. At the same time, the analysis of tickets is scattered and has little effect. Summary of the invention

[0004] The present invention provides a distribution network work order classification method, device, equipment and storage medium to solve the problem in the prior art that users' complaint work orders against the distribution network cannot be accurately classified.

[0005] According to one aspect of the present invention, a distribution network work order classification method is provided, the method comprising:

[0006] Get the distribution network work order to be classified;

[0007] Slicing the distribution network work order to be classified to obtain multiple words;

[0008] The inverse document frequency and term frequency of each term are calculated by the term frequency-inverse document frequency method;

[0009] Based on the inverse document frequency and word frequency, a convolutional neural network is combined to determine the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label;

[0010] The scores of each directory level in the different categories are analyzed to obtain the classification results of the distribution network work orders to be classified.

[0011] According to another aspect of the present invention, a distribution network work order classification device is provided, the device comprising:

[0012] An acquisition module is used to obtain distribution network work orders to be classified;

[0013] A cutting module, used for cutting the distribution network work order to be classified to obtain multiple words;

[0014] A calculation module, for calculating the inverse document frequency and word frequency of each vocabulary word by a word frequency-inverse document frequency method;

[0015] A determination module, used to determine the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label based on the inverse document frequency and word frequency in combination with a convolutional neural network;

[0016] The analysis module is used to analyze the scores of each directory level in the different categories to obtain the classification results of the distribution network work orders to be classified.

[0017] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising: at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the distribution network work order classification method described in any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the distribution network work order classification method described in any embodiment of the present invention when executed.

[0021] A distribution network work order classification method, device, equipment and storage medium according to an embodiment of the present invention, the method includes: obtaining a distribution network work order to be classified; cutting the distribution network work order to be classified to obtain multiple words; calculating the inverse document frequency and word frequency of each word by the word frequency-inverse file frequency method; based on the inverse document frequency and word frequency, combining a convolutional neural network to determine the score of each directory level in different categories of the complaint work order classification label of the distribution network work order to be classified; analyzing the score of each directory level in the different categories to obtain the classification result of the distribution network work order to be classified. The method cuts the work order, combines the word frequency inverse file frequency method and the convolutional neural network to determine the score of the work order under different categories, and obtains the classification result of the work order, which can improve the efficiency of the classification of the distribution network work order, and solves the problem that the prior art cannot accurately classify the user's complaint work order for the distribution network.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 A schematic diagram of a flow chart of a distribution network work order classification method provided in the first embodiment of the present invention;

[0025] Figure 2 A schematic diagram of the structure of a distribution network work order classification device provided in the second embodiment of the present invention;

[0026] Figure 3 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should belong to the scope of protection of the present invention. It should be understood that the various steps recorded in the method implementation of the present invention can be performed in different orders and / or in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0028] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0029] 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, any variations of the terms "including" and "having" 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 clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0031] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes, and are not used to limit the scope of these messages or information.

[0032] Embodiment 1

[0033] Figure 1 A flow chart of a distribution network work order classification method provided in Embodiment 1 of the present invention is provided. The method can be applicable to the situation of classifying distribution network work orders. The method can be executed by a distribution network work order classification device, wherein the device can be implemented by software and / or hardware and is generally integrated on an electronic device. In this embodiment, the electronic device includes but is not limited to: computers and other devices.

[0034] like Figure 1 As shown, a distribution network work order classification method provided in Embodiment 1 of the present invention includes the following steps:

[0035] S110. Obtain the distribution network work order to be classified.

[0036] Among them, the distribution network work order to be classified may be a received distribution network complaint work order.

[0037] In this embodiment, the distribution network work orders to be classified that need to be classified can be obtained.

[0038] In one embodiment, the obtaining of the distribution network work order to be classified includes: obtaining the received work order content; determining whether there are complaint keywords in the work order content; and if so, treating the work order content as the distribution network work order to be classified.

[0039] The work order content may include multiple work orders. The complaint keywords may be words related to the complaint, and the complaint keywords may be obtained from the classification tags of the work order.

[0040] In this embodiment, after receiving all the work order contents, for each work order, it can be determined whether there are complaint keywords in the work order content. If so, the work order is considered to be a complaint-related work order and is treated as a distribution network work order to be classified.

[0041] S120: Slice the distribution network work order to be classified to obtain a plurality of words.

[0042] The vocabulary may be a phrase or word in the distribution network work order to be classified.

[0043] In this embodiment, the distribution network work orders to be classified may be segmented to obtain multiple words therein.

[0044] S130, calculating the inverse document frequency and word frequency of each vocabulary word by using the word frequency-inverse document frequency method.

[0045] Among them, the Term Frequency-Inverse Document Frequency (TF-IDF) method is a weighted technique commonly used in information retrieval and text mining. The algorithm evaluates the importance of a word in a document set by the term frequency (TF) of each word in the document and the inverse document frequency (IDF) in the entire corpus. A high value of TF-IDF indicates that the word is more important to the document, while a low value indicates that the word is less important.

[0046] In this embodiment, the inverse document frequency and word frequency of each word may be calculated by the word frequency-inverse document frequency method.

[0047] In one embodiment, the inverse document frequency and word frequency of each word are calculated by the word frequency-inverse file frequency method, including: calculating the word frequency of each word by the number of times each word appears in the distribution network work order to be classified; calculating the inverse document frequency of each word by the number of times each word appears in the training sample data set.

[0048] The training sample data set may be a data set used when training a convolutional neural network model.

[0049] In this embodiment, the word frequency of each word can be calculated by the number of times each word appears in the distribution network work order to be classified; and the inverse document frequency of each word can be calculated by the number of times each word appears in the training sample data set.

[0050] TF-IDF is a statistical calculation method that is often used to evaluate the importance of a word in a document. In a given document, the word frequency represents the number of times a specified word appears in the document. The calculation formula is as follows:

[0051]

[0052] where n i，j Indicates the number of times a specified word appears in a file, ∑ k n k，j Represents the sum of all word occurrences in the file.

[0053]

[0054] Among them, IDF is the inverse document frequency, n is the total number of documents, Expressed as the total number of documents containing the term.

[0055] The importance of words is weighted and the calculation formula is:

[0056] W ij =TF i，j ×IDF i ;

[0057] IDF i It is usually normalized to avoid bias towards long files, and the normalized weight calculation formula is:

[0058] W ij =ln(1+TF i，j )×IDF i .

[0059] S140. Based on the inverse document frequency and word frequency, a convolutional neural network is used to determine the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label.

[0060] Among them, convolutional neural network is a deep learning algorithm that can automatically learn and extract useful features from images by simulating the mechanism of the human visual system. It can be used to complete tasks such as image classification, target detection and image segmentation. The complaint ticket classification label can be a custom classification label, and the directory level can be the level of each classification directory in the complaint ticket classification label, which can include first-level directory, second-level directory and third-level directory.

[0061] In this embodiment, the scores of the distribution network work orders to be classified in different categories of complaint work order classification labels can be determined based on the inverse document frequency and word frequency in combination with the convolutional neural network. For example, as shown in Table 1, under the complaint work order classification label classification:

[0062] Table 1 Classification of complaint ticket classification labels

[0063]

[0064] As shown in Table 1, the complaint work order classification labels include 26 categories. For example, the first category "meter reading and billing / meter reading and payment reminder / power outage and restoration for arrears" includes a first-level directory "meter reading and billing", a second-level directory "meter reading and payment reminder" and a third-level directory "power outage and restoration for arrears". The first-level directories and second-level directories of different categories can be the same.

[0065] S150. Analyze the scores of each directory level in the different categories to obtain the classification results of the distribution network work orders to be classified.

[0066] Among them, the classification result can be used to determine which category of the complaint work order classification label the distribution network work order to be classified belongs to.

[0067] In this embodiment, after obtaining the scores at different directory levels in different categories of the complaint ticket classification labels, the scores of each category of each complaint ticket classification label can be calculated according to the scores at the directory levels in different categories, and finally the final classification result can be determined according to the scores of each category.

[0068] In one embodiment, the scores of each directory level in the different categories are analyzed to obtain the classification results of the distribution network work orders to be classified, including: calculating the scores of different categories in the complaint work order classification label based on the weights and corresponding scores of each directory level in the complaint work order classification label; and taking a vocabulary consisting of words corresponding to the highest scores in all categories as the classification results of the distribution network work orders to be classified.

[0069] The weight of each directory level can be customized or determined through training.

[0070] In this embodiment, the scores of different categories in the complaint work order classification label can be calculated based on the weights and corresponding scores of each directory level in the complaint work order classification label, all scores are sorted, and a vocabulary consisting of words corresponding to the highest scores in all categories is used as the classification result of the distribution network work order to be classified. For example, if the classification score of category 1 in the complaint work order classification label is the highest, the words of each directory level in category 1 can be composed of a vocabulary "meter reading and billing / meter reading and urging / power outage and restoration for arrears", and the vocabulary is used as the classification result of the distribution network work order to be classified.

[0071] A distribution network work order classification method provided in a first embodiment of the present invention includes: obtaining a distribution network work order to be classified; cutting the distribution network work order to be classified to obtain multiple words; calculating the inverse document frequency and word frequency of each word by the word frequency-inverse file frequency method; based on the inverse document frequency and word frequency, combining a convolutional neural network to determine the score of each directory level in different categories of the complaint work order classification label of the distribution network work order to be classified; analyzing the score of each directory level in the different categories to obtain the classification result of the distribution network work order to be classified. The method cuts the work order, combines the word frequency inverse file frequency method and the convolutional neural network to determine the score of the work order under different categories, and obtains the classification result of the work order. It can improve the efficiency of classifying distribution network work orders and solve the problem that the prior art cannot accurately classify user complaint work orders for distribution networks.

[0072] Based on the above embodiment, a variant embodiment of the above embodiment is proposed. It should be noted that in order to make the description concise, only the differences from the above embodiment are described in the variant embodiment.

[0073] In one embodiment, the scores of the distribution network work orders to be classified at each directory level in different categories of complaint work order classification labels are determined based on the inverse document frequency and word frequency in combination with a convolutional neural network, including: determining the weight embedding vector of the distribution network work orders to be classified through the complaint work order classification label; inputting the weight embedding vector into the convolution layer of a trained convolutional neural network to obtain an output result; and determining the scores of the distribution network work orders to be classified at each directory level in different categories of complaint work order classification labels based on the output result, the inverse document frequency and word frequency.

[0074] The weight embedding vector may be a vector representing the importance of the directory level in each classification. Each convolutional layer in the convolutional neural network is composed of several convolutional units, and different features of the input may be extracted through convolution operations. The output result may be the result output by the convolutional layer.

[0075] In this embodiment, the weight embedding vector of the distribution network work order to be classified can be determined by the complaint work order classification label, and the weight embedding vector is input into the convolution layer of the trained convolutional neural network to obtain the output result. The output result, inverse document frequency and word frequency can be used to calculate the score of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label.

[0076] In one embodiment, determining the weight embedding vector of the distribution network work order to be classified through the complaint work order classification label includes: based on the same vocabulary in the distribution network work order to be classified as in the complaint work order classification label, searching for the corresponding weight embedding vector from the embedding matrix.

[0077] The embedding matrix can be a matrix constructed by the classification labels of the complaint work order. The embedding matrix can be used to represent the weight of each classification. For example, in the first category, the embedding matrix can be used to find that the weight of "meter reading and billing" is 0.5, the weight of "meter reading and billing" is 0.3, and the weight of "power outage and restoration for arrears" is 0.2.

[0078] In this embodiment, the corresponding weight embedding vector can be found from the embedding matrix based on the same vocabulary in the distribution network work order to be classified as in the classification label of the complaint work order. Exemplarily, if the distribution network work order to be classified includes the vocabulary "meter reading and billing" and "meter reading and urging for payment", it can be determined from the classification label of the complaint work order that the vocabulary in the distribution network work order to be classified has the same vocabulary as the 1st, 9th, 22nd and 25th categories, and the vectors corresponding to the 1st, 9th, 22nd and 25th categories in the embedding matrix are used as weight embedding vectors.

[0079] In one embodiment, the determining the scores of the distribution network work orders to be classified at each directory level in different categories of the complaint work order classification label based on the output result, the inverse document frequency and the word frequency includes: calculating the scores I of the distribution network work orders to be classified at each directory level in different categories of the complaint work order classification label by the following formula:

[0080] I=W ij ×TF k，i ;

[0081]

[0082] W ij =ln(1+TF i，j )×IDF i ;

[0083] Among them, IDF i is the inverse document frequency, TF i，j is the word frequency, C j is the output result, w k，i is the i-th word in the vocabulary corresponding to the k-th work order; C(w k，i ) means counting the frequency of the i-th word in the vocabulary of the k-th work order in the k-th work order.

[0084] In this embodiment, the scores of the distribution network work orders to be classified at each directory level in different categories of the complaint work order classification label can be calculated through output results, inverse document frequency and word frequency.

[0085] The embodiment of the present invention provides a specific implementation method based on the technical solutions of the above embodiments. As a specific implementation method of this embodiment, the training of the convolutional neural network is as follows:

[0086] The input of the convolutional layer is the embedding vector. According to the level of the word or character of the embedding vector, the embedding matrix is ​​divided into a word embedding matrix or a character embedding matrix. According to the directory level, the corresponding embedding vector is found in the embedding matrix and stacked to form the input of the convolutional layer.

[0087] The calculation process of one convolution kernel in the convolution layer is as follows:

[0088] c i =f(X i:i+h-1 ·W h +b i );

[0089] Among them, X i is the i-th row vector in the input matrix X, X i:i+h-1 Represents a matrix consisting of h row vectors from i to i+h-1, which is combined with the convolution kernel W of size h h Perform convolution operation and add offset parameter b i After conversion by activation function f, a convolution result c is obtained i , assuming that the convolution layer input matrix contains n row vectors, after multiple convolution operations, we can get the following expression:

[0090] C j =[c1, c2, c3, …, c n-h+1 ] T ;

[0091] Assuming that there are m convolution kernels of size h, the output of the convolution layer is:

[0092] M=[C1,C2,C3,…,C m ];

[0093] Combined with the above TF-IDF model, the inverse document frequency of each word in the training sample is calculated separately. The calculation formula is as follows:

[0094]

[0095] Where D represents the training sample data set; |D| represents the total number of training samples, D j is the jth sample in the training sample dataset, w i represents the i-th word in the vocabulary determined using the training sample, {j|w i ∈D j} means it contains word w i The set of training samples, |{j|w i ∈D j}| means containing word w i The total number of training samples.

[0096]

[0097] Among them, w k，i is the i-th word in the vocabulary corresponding to the k-th work order, C(w k，i ) means counting the frequency of the i-th word in the vocabulary corresponding to the k-th work order in the k-th work order.

[0098] After the above calculation, we can get the scores of different words in each category I = W ij ×TF k，i , select the words with the highest scores in the category to form the final word list. Such a word list can not only retain the words that contribute to the classification to the greatest extent, but also compress the embedding matrix, reducing training time and memory usage.

[0099] The method of this embodiment, when constructing a convolutional neural network model, in the embedding matrix layer, finds the corresponding embedding vector in the embedding matrix according to the index of the word, and then stacks it to form the input of the convolution layer; uses TF-IDF to evaluate the importance of a word to a single file in the text library, extracts keywords of the category based on TF-IDF to reduce the vocabulary, and further reduces the size of the embedding matrix, thereby improving the efficiency and accuracy of work order classification.

[0100] After the above training, the work orders can be classified by the trained model:

[0101] Step 1: Enter the received work order content;

[0102] Step 2: Filter out the processing text information from all the acceptance text information in the work order, and cut the processing text information into single words;

[0103] Step 3: Find the corresponding embedding vector according to the directory level of the vocabulary in the embedding matrix and stack them to form the input of the convolutional layer;

[0104] Step 4: Perform convolution calculation on the convolution kernel to obtain the output result after convolution;

[0105] Step 5: Use all the words in the training sample to calculate the inverse document frequency of each word;

[0106] Step 6: Calculate the frequency of words in each category;

[0107] Step 7: For each category of vocabulary, calculate the TF-IDF value of each word based on the output results;

[0108] Step 8: Get the scores of different words in each category, and select the words with the highest scores in the category to form the final vocabulary.

[0109] Embodiment 2

[0110] Figure 2 This is a structural schematic diagram of a distribution network work order classification device provided in Embodiment 2 of the present invention. The device can be applicable to the situation of classifying distribution network work orders, wherein the device can be implemented by software and / or hardware and is generally integrated on an electronic device.

[0111] like Figure 2 As shown, the device comprises:

[0112] An acquisition module 210 is used to acquire a distribution network work order to be classified;

[0113] A cutting module 220, used for cutting the distribution network work order to be classified to obtain a plurality of words;

[0114] A calculation module 230, for calculating the inverse document frequency and the word frequency of each word by a word frequency-inverse document frequency method;

[0115] A determination module 240 is used to determine the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label based on the inverse document frequency and word frequency in combination with a convolutional neural network;

[0116] The analysis module 250 is used to analyze the scores of each directory level in the different categories to obtain the classification results of the distribution network work orders to be classified.

[0117] The present embodiment provides a distribution network work order classification device, including: an acquisition module, used to acquire a distribution network work order to be classified; a cutting module, used to cut the distribution network work order to be classified to obtain multiple words; a calculation module, used to calculate the inverse document frequency and word frequency of each word by the word frequency-inverse file frequency method; a determination module, used to determine the score of each directory level of the distribution network work order to be classified in different categories of complaint work order classification labels based on the inverse document frequency and word frequency in combination with a convolutional neural network; an analysis module, used to analyze the score of each directory level in the different categories to obtain the classification result of the distribution network work order to be classified. By cutting the work order, combining the word frequency inverse file frequency method and the convolutional neural network, the scores of the work order under different categories are determined, and the classification result of the work order is obtained, which can improve the efficiency of the classification of the distribution network work order, and solve the problem that the prior art cannot accurately classify the user's complaint work order for the distribution network.

[0118] Furthermore, the acquisition module 210 is specifically used for:

[0119] Get the content of the received work order;

[0120] Determine whether there are complaint keywords in the work order content;

[0121] If so, the work order content is used as the distribution network work order to be classified.

[0122] Furthermore, the calculation module 230 is specifically configured to:

[0123] Calculate the word frequency of each word by the number of times each word appears in the distribution network work order to be classified;

[0124] The inverse document frequency of each word is calculated by the number of times each word appears in the training sample dataset.

[0125] Furthermore, the determination module 240 is specifically configured to:

[0126] Determine the weight embedding vector of the distribution network work order to be classified by the complaint work order classification label;

[0127] Inputting the weight embedding vector into the convolutional layer of the trained convolutional neural network to obtain an output result;

[0128] Based on the output result, the inverse document frequency and the word frequency, the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label are determined.

[0129] Furthermore, determining the weight embedding vector of the distribution network work order to be classified by the complaint work order classification label includes:

[0130] Based on the same vocabulary in the distribution network work order to be classified as in the classification label of the complaint work order, the corresponding weight embedding vector is found from the embedding matrix.

[0131] Further, the determining of the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label based on the output result, the inverse document frequency and the word frequency includes:

[0132] The score I of each directory level in different categories of the complaint work order classification label is calculated by the following formula:

[0133] I=W ij ×TF k，i ;

[0134]

[0135] W ij =ln(1+TF i，j )×IDF i ;

[0136] Among them, IDF i is the inverse document frequency, TF i，j is the word frequency, C j is the output result, wk，i is the i-th word in the vocabulary corresponding to the k-th work order; C(w k，i ) means counting the frequency of the i-th word in the vocabulary of the k-th work order in the k-th work order.

[0137] Furthermore, the analysis module 250 is specifically used for:

[0138] Based on the weights and corresponding scores of each directory level in the complaint ticket classification label, the scores of different categories in the complaint ticket classification label are calculated;

[0139] A vocabulary consisting of words corresponding to the highest scores in all categories is used as the classification result of the distribution network work order to be classified.

[0140] The above-mentioned distribution network work order classification device can execute the distribution network work order classification method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0141] Embodiment 3

[0142] Figure 3 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0143] like Figure 3 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0144] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0145] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a distribution network work order classification method.

[0146] In some embodiments, the distribution network work order classification method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the distribution network work order classification method described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the distribution network work order classification method in any other appropriate manner (for example, by means of firmware).

[0147] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0148] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0149] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0150] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0151] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0152] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0153] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0154] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A distribution network work order classification method, characterized in that: The method comprises: Get the distribution network work order to be classified; Slicing the distribution network work order to be classified to obtain multiple words; The inverse document frequency and term frequency of each term are calculated by the term frequency-inverse document frequency method; Based on the inverse document frequency and word frequency, a convolutional neural network is combined to determine the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label; The scores of each directory level in the different categories are analyzed to obtain the classification results of the distribution network work orders to be classified.

2. The method according to claim 1, characterized in that The step of obtaining the distribution network work order to be classified includes: Get the content of the received work order; Determine whether there are complaint keywords in the work order content; If so, the work order content is used as the distribution network work order to be classified.

3. The method according to claim 1, characterized in that The method of calculating the inverse document frequency and the word frequency of each word by the word frequency-inverse document frequency method includes: Calculate the word frequency of each word by the number of times each word appears in the distribution network work order to be classified; The inverse document frequency of each word is calculated by the number of times each word appears in the training sample dataset.

4. The method according to claim 1, characterized in that: The method of determining the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label based on the inverse document frequency and word frequency in combination with a convolutional neural network includes: Determine the weight embedding vector of the distribution network work order to be classified by the complaint work order classification label; Inputting the weight embedding vector into the convolutional layer of the trained convolutional neural network to obtain an output result; Based on the output result, the inverse document frequency and the word frequency, the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label are determined.

5. The method according to claim 4, characterized in that The step of determining the weight embedding vector of the distribution network work order to be classified by using the complaint work order classification label includes: Based on the same vocabulary in the distribution network work order to be classified as in the classification label of the complaint work order, the corresponding weight embedding vector is found from the embedding matrix.

6. The method according to claim 4, characterized in that The step of determining the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label based on the output result, the inverse document frequency and the word frequency includes: The score I of each directory level in different categories of the complaint work order classification label is calculated by the following formula: I=W ij ×TF k，i ; W ij =ln(1+TF i，j )×IDF i ; Among them, IDF i is the inverse document frequency, TF i，j is the word frequency, C j is the output result, w k，i is the i-th word in the vocabulary corresponding to the k-th work order; C(w k , i ) means counting the frequency of the i-th word in the vocabulary of the k-th work order in the k-th work order.

7. The method according to claim 1, characterized in that The analyzing the scores of each directory level in the different categories to obtain the classification results of the distribution network work orders to be classified includes: Based on the weights and corresponding scores of each directory level in the complaint ticket classification label, the scores of different categories in the complaint ticket classification label are calculated; A vocabulary consisting of words corresponding to the highest scores in all categories is used as the classification result of the distribution network work order to be classified.

8. A distribution network work order classification device, characterized in that: The device comprises: An acquisition module is used to obtain distribution network work orders to be classified; A cutting module, used for cutting the distribution network work order to be classified to obtain multiple words; A calculation module, for calculating the inverse document frequency and word frequency of each vocabulary word by a word frequency-inverse document frequency method; A determination module, used to determine the scores of the distribution network work order to be classified at each directory level in different categories of the complaint work order classification label based on the inverse document frequency and word frequency in combination with a convolutional neural network; The analysis module is used to analyze the scores of each directory level in the different categories to obtain the classification results of the distribution network work orders to be classified.

9. An electronic device, characterized in that: The device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the distribution network work order classification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the distribution network work order classification method according to any one of claims 1 to 7 when executed.

Images