Artificial intelligence grouping method and system based on power customer complaints

Abstract

The invention discloses an artificial intelligence grouping method and system based on power customer complaints, and the method is characterized in that the method comprises the following steps: obtaining power customer complaint information, and extracting complaint features from the power customer complaint information through employing a neural network; respectively inputting the complaint features into a plurality of hierarchical memory models according to information dimensions, and acquiring outputs of the plurality of hierarchical memory models; connecting the outputs of the plurality of hierarchical memory models, and estimating joint distribution of connection results; and judging the event emergency of the power customer complaint information based on the joint distribution, and issuing a power customer complaint work order according to the event emergency. The method is accurate and effective, the priority ranking of the automatic order sending service is realized, the response efficiency of the system is improved, and the effective response to the emergency complaint is ensured.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and more specifically, to an artificial intelligence grouping method and system based on power customer complaints. Background Art

[0002] Power supply service complaints refer to power customers within the power company's business scope expressing dissatisfaction with the damage to their rights and interests caused by the power supply company's responsibility in terms of power supply services, business operations, power outages, power supply quality, and grid construction, and demanding that their rights and interests be protected.

[0003] Customer complaints mainly include service complaints, business complaints, power outage complaints, power supply quality complaints, power grid construction complaints and other types. In service complaints, users complain about the irregular service behavior of power supply company employees and the blocked service channels. In business complaints, power supply companies have problems such as overtime, negligence, and errors in business expansion and registration, electricity consumption changes, meter reading and collection, electricity charges and prices, and energy metering. Therefore, it is necessary to make business improvements based on the content of the complaints. In power outage complaints, there are service deviations in power outage information announcements, power outage plan execution, emergency repair quality, and the opening of value-added services. Therefore, it is necessary to improve the service content based on complaints. In addition, in power supply quality complaints, customers may raise opinions on power supply quality problems such as voltage deviation, frequency deviation, voltage imbalance, voltage fluctuation or flicker in the transmitted electricity. Therefore, in order to optimize the power supply capacity in response to power supply quality complaints, it is necessary to make substantial improvements to transmission parameters such as precise voltage deviation and frequency deviation.

[0004] At present, power companies mainly provide customer service through WeChat business halls, online business halls, 95598 call hotlines, physical business halls, etc. Customer service for complaints mainly relies on manual customer service for classification and dispatch, and the power industry has not yet truly realized intelligent power customer service.

[0005] In the traditional complaint handling process, operators usually answer user complaints over the phone, then compile them into text messages for recording, and then assign them to relevant departments and personnel for processing. However, due to the wide variety of complaints and the involvement of many departments and personnel, qualified operators often need a lot of training before they can take up their posts, and they are unable to handle emergency issues during special hours at night. Therefore, there is an urgent need for automated dispatching applications to improve the efficiency of customer complaint handling.

[0006] Secondly, the categories of complaint cases are stored and classified in the database in a tree-like hierarchical structure. When accepting a complaint, it is necessary to first correctly classify it at each level of the tree, and finally accurately trace the complaint. If a category in an upper layer is classified incorrectly, the classification of the next level will also be wrong, which leads to the accumulation of error rates layer by layer.

[0007] On the other hand, with the upgrade of the customer service platform, various forms of complaints such as voice, text, pictures and videos can be obtained during the complaint service process. The complaint contents of various modes are interrelated and complementary, but it is difficult to provide complete complaint information depending on the source and access method of the complaint. These problems lead to the accumulation of classification model error rates.

[0008] When the error rate of the classification model is high, the automatic real-time dispatching service cannot be implemented accurately and effectively.

[0009] In response to the above problems, there is an urgent need for an artificial intelligence grouping method and system based on power customer complaints. Summary of the invention

[0010] In order to solve the deficiencies in the prior art, the present invention provides an artificial intelligence grouping method and system based on power customer complaints, which extracts complaint features from power customer complaint information through a neural network, inputs the complaint information into a hierarchical memory model in a dimensionally divided manner, and uses the feedforward features of the memory model to obtain the joint distribution, thereby obtaining the event urgency of the complaint information.

[0011] The present invention adopts the following technical solution.

[0012] The first aspect of the present invention relates to an artificial intelligence grouping method based on electricity customer complaints, the method comprising the following steps: obtaining electricity customer complaint information, and using a neural network to extract complaint features from the electricity customer complaint information; inputting the complaint features into multiple hierarchical memory models according to information dimensions, and obtaining outputs of multiple hierarchical memory models; connecting the outputs of multiple hierarchical memory models, and estimating the joint distribution of the connection results; judging the event urgency of the electricity customer complaint information based on the joint distribution, and issuing an electricity customer complaint work order according to the event urgency.

[0013] Preferably, obtaining the power customer complaint information includes: obtaining customer complaint images, complaint videos, complaint voices and complaint texts; and mapping the power customer complaint information of different dimensions into the same feature space.

[0014] Preferably, a neural network is used to extract complaint features from the electricity customer complaint information, including: extracting complaint features from the electricity customer complaint information based on a spatiotemporal data mining method; the complaint features are represented in vector form, and the spatiotemporal data mining method is implemented using CNN or RNN.

[0015] Preferably, the complaint features include complaint image features, complaint video features, complaint voice features and complaint text features; wherein, the complaint features are input into multiple hierarchical memory models according to information dimensions, including: inputting the complaint image features, complaint video features, complaint voice features and complaint text features into the first to fourth hierarchical memory models respectively.

[0016] Preferably, the complaint features are input into multiple hierarchical memory models according to information dimensions, including: using the outputs of the first to fourth hierarchical memory models to calculate the first to fourth complaint urgency parameters respectively as the outputs of the multiple hierarchical memory models.

[0017] Preferably, the outputs of multiple hierarchical memory models are connected and the joint distribution of the connection results is estimated, including: connecting the first to fourth complaint urgency parameters and calculating the joint distribution probability of the connection results.

[0018] Preferably, judging the event urgency of the power customer complaint information based on the joint distribution includes: judging the event urgency of the power customer complaint information using the joint distribution probability.

[0019] Preferably, the urgency of the incident is:

[0020]

[0021] The sample mean is The variance is

[0022] Where k is the number of feedforward optimal matching sequences,

[0023] W is the window length of the event urgency score,

[0024] The value range of j is between 1 and W.

[0025] s t-j is the urgency score of the tjth event,

[0026] l is the number of the data processing unit of the current column in the hierarchical memory model,

[0027] is a probability function.

[0028] The second aspect of the present invention relates to an artificial intelligence grouping system based on power customer complaints using the method in the first aspect of the present invention, the system comprising an acquisition module, a stratification module, a connection module and a judgment module; wherein the acquisition module is used to acquire power customer complaint information and extract complaint features from the power customer complaint information using a neural network; the stratification module is used to input the complaint features into multiple hierarchical memory models according to information dimensions, and acquire outputs of multiple hierarchical memory models; the connection module is used to connect the outputs of multiple hierarchical memory models and estimate the joint distribution of the connection results; the judgment module is used to judge the event urgency of the power customer complaint information based on the joint distribution, and issue a power customer complaint work order according to the event urgency.

[0029] Preferably, the acquisition module is used to obtain power customer complaint information, including: obtaining customer complaint images, complaint videos, complaint voices and complaint texts; and mapping the power customer complaint information of different dimensions into the same feature space.

[0030] Preferably, an acquisition module is used to extract complaint features from the electricity customer complaint information using a neural network, including: extracting complaint features from the electricity customer complaint information based on a spatiotemporal data mining method; the complaint features are represented in vector form, and the spatiotemporal data mining method is implemented using CNN or RNN.

[0031] Preferably, the complaint features acquired by the acquisition module include complaint image features, complaint video features, complaint voice features, and complaint text features;

[0032] A hierarchical module is used to input the complaint features into multiple hierarchical memory models according to information dimensions, including: inputting complaint image features, complaint video features, complaint voice features and complaint text features into the first to fourth hierarchical memory models respectively.

[0033] Preferably, the hierarchical module inputs the complaint features into multiple hierarchical memory models according to information dimensions, including: using the outputs of the first to fourth hierarchical memory models to calculate the first to fourth complaint urgency parameters respectively as the outputs of the multiple hierarchical memory models.

[0034] Preferably, a connection module is used to connect the outputs of multiple hierarchical memory models and estimate the joint distribution of the connection results, including: connecting the first to fourth complaint urgency parameters and calculating the joint distribution probability of the connection results.

[0035] Preferably, the judgment module is used to judge the event urgency of the power customer complaint information based on the joint distribution, including: judging the event urgency of the power customer complaint information by using the joint distribution probability.

[0036] Preferably, the urgency of the event determined by the determination module is:

[0037]

[0038] The sample mean is The variance is

[0039] Where k is the number of feedforward optimal matching sequences,

[0040] W is the window length of the event urgency score,

[0041] The value range of j is between 1 and W.

[0042] s t-j is the urgency score of the tjth event,

[0043] l is the number of the data processing unit of the current column in the hierarchical memory model,

[0044] is a probability function.

[0045] A third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method described in the first aspect of the present invention.

[0046] A fourth aspect of the present invention relates to a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the method in the first aspect of the present invention.

[0047] The beneficial effect of the present invention is that, compared with the prior art, the artificial intelligence grouping method and system based on power customer complaints in the present invention extracts complaint features from power customer complaint information through a neural network, inputs the complaint information into a hierarchical memory model in dimensions, and obtains a joint distribution using the feedforward features of the memory model, thereby obtaining the event urgency of the complaint information. The present invention is accurate and effective, realizes the priority sorting of automatic dispatching services, improves the response efficiency of the system, and ensures an effective response to emergency complaints.

[0048] The beneficial effects of the present invention also include:

[0049] 1. The method is able to integrate and process multiple types of data, understand and analyze complaint information from different sources, and effectively combine them to produce more accurate and comprehensive results. Therefore, the method ensures that the system effectively responds to customer complaints and suggestions through intelligent sorting.

[0050] 2. The present invention can process image information and video information at the same time, and supports analyzing the substantive content of the complaint provided in the image and video information in combination with natural language understanding, confirming the cause of the complaint, the root cause and type of the fault, thereby ensuring an accurate and reasonable response to the complaint.

[0051] 3. The method combines convolutional neural networks or recursive neural networks to support the processing of large amounts of unstructured data and extract useful features from them, so as to simultaneously analyze the image and voice information in the video, organically link the image and voice information, and provide a deeper and more comprehensive understanding of the complaint information.

[0052] 4. In order to obtain the urgency of the complaint content, the method extracts information such as event subject information, complaint sentiment attitude information, complaint probability information, and complaint peak period from the information features when identifying multi-dimensional information. Based on the automatic identification of the above information, the analysis process of time urgency is ensured to be accurate and effective, and the order dispatch is scientifically implemented accordingly.

[0053] 5. The analysis results of the present invention can also be used to objectively and truthfully conduct other scientific analyses of complaint information according to the actual needs of the system, and apply the analysis results to each link of data management. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The present invention is a flowchart of an artificial intelligence grouping method based on power customer complaints. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in the present invention are only embodiments of a part of the present invention, not all embodiments. Based on the spirit of the present invention, all other embodiments not recorded in the present invention obtained by ordinary technicians in this field according to the embodiments recorded in the present invention without making creative work should belong to the protection scope of the present invention.

[0056] Figure 1 The figure is a flow chart of an artificial intelligence grouping method based on power customer complaints in the present invention. Figure 1 As shown, the first aspect of the present invention relates to an artificial intelligence grouping method based on electricity customer complaints, and the method includes steps 1 to 4.

[0057] Step 1: Obtain the power customer complaint information and use a neural network to extract complaint features from the power customer complaint information.

[0058] In the present invention, based on the main affairs, customer service content, and emotional state extracted from the customer's complaints or suggestions, deep learning and other technologies in the field of artificial intelligence are used to classify the customer service types, and the urgency and complaint probability of this business are obtained, so that the relevant departments can reasonably arrange the processing order according to the different urgency of the customers.

[0059] Customer service information is the complaints and suggestions of customers about various events in the real world. The random occurrence of real-world events makes the generation of customer service information random in time and quantity. Sometimes only a small amount of data is generated in a long period of time, and sometimes a large amount of data is generated in a short period of time.

[0060] The randomness of customer service information places requirements on the processing capacity and service stability of the power intelligent customer service under various traffic and emergency situations. When the amount of customer service information exceeds the system processing capacity, the customer service information belonging to the suggestion can be temporarily backlogged to reduce the computing pressure, allowing the system to smoothly process the data flow and flatten the traffic peak.

[0061] To this end, the present invention supports automatic identification of customer service types and customer emotions based on customer service information, and realizes business classification, urgency, and customer complaint probability assessment.

[0062] In order to achieve the above process, the method first obtains customer complaint information. The information can be obtained from different sources. For example, customer complaints and suggestions are collected from WeChat business halls, online business halls, 95598 call hotlines, and physical business halls, and semantic extraction is performed on the customer complaints and suggestions.

[0063] With the widespread application of multimedia technology and the enrichment of data description methods, a large amount of data content in different forms is widely available, and the description of a description object is jointly realized from multiple angles, in multiple languages, and in multiple visual forms. In the complaint information of the power system, the same complaint content may include multiple different forms such as audio, image, text, and video. Pictures and videos from different sources may have different pixel intensity, grayscale, and texture information. How to integrate various types of information and realize decision-making and prediction based on multi-modal data is the key issue considered in this invention.

[0064] To this end, the method collects complaint voice, text, picture and video information through a variety of complaint channels, and constructs a data set by using the different dimensions of the collection and the dimensions of the data content.

[0065] Preferably, obtaining power customer complaint information includes: obtaining customer complaint images, complaint videos, complaint voices and complaint texts; and mapping power customer complaint information of different dimensions into the same feature space.

[0066] In one embodiment of the present invention, semantic extraction is performed on image information, video information, sound information and text information respectively, including: preprocessing the video information and image information through a computer vision algorithm; preprocessing the sound information through a speech recognition algorithm; and semantic understanding and keyword extraction of text information, preprocessed image information, video information, and sound information through a natural language processing algorithm.

[0067] Obtaining key semantics from the extracted semantic information needs to be implemented through semantic information mapping. Specifically, the method constructs a semantic data set based on the extracted image information semantics, video information semantics, sound information semantics and text information semantics, and uses the semantic data set as the key semantics.

[0068] Through the above methods, the method can obtain a preliminary semantic understanding of the above multi-type information and extract keywords. Therefore, in order to achieve unified mapping, the method needs to construct a corresponding data set. For example, the method can effectively extract instances with multiple different dimensional information, thereby obtaining the distance between instances based on these instances, constructing the multi-dimensional space where the instances are located, ensuring the validity of the data set, and providing a basis for the subsequent classification of data instance types.

[0069] In another embodiment, a neural network is used to extract complaint features from power customer complaint information, including: extracting complaint features from power customer complaint information based on a spatiotemporal data mining method; the complaint features are represented in vector form, and the spatiotemporal data mining method is implemented using CNN or RNN.

[0070] In spatiotemporal data mining methods, different application scenarios and spatiotemporal data types lead to different categories of data mining tasks and problem statements. Different deep learning models usually have different preferences for the types of spatiotemporal data and different requirements for the input data format, so spatiotemporal data needs to be classified. Through classification, the method obtains data instances. The instances include descriptions of events, trajectories, time series, spatial graphs, and spatiotemporal folk songs. By describing different types of spatiotemporal data instances, the method can obtain deep information of different contents, such as points, sequences, graphs, two-dimensional matrices, and three-dimensional vectors.

[0071] In one embodiment of the present invention, the convolutional neural network is supported to analyze the information content of visual images, videos, etc., and accurate information output is obtained through a feedforward method. In addition, the method also supports the method of recursive neural network, through the connection between nodes, a directed graph is formed along the sequence, so as to identify the sequential features and achieve reasonable prediction. In RNN, the method also supports the use of long short-term memory network to achieve long-term dependency of learning input data.

[0072] Step 2: Input the complaint features into multiple hierarchical memory models according to the information dimensions, and obtain the outputs of multiple hierarchical memory models.

[0073] After passing through the neural network, the method obtains outputs of multiple different dimensions, such as complaint features including complaint image features, complaint video features, complaint voice features and complaint text features.

[0074] Preferably, the complaint features are input into multiple hierarchical memory models according to information dimensions, including: inputting complaint image features, complaint video features, complaint voice features and complaint text features into the first to fourth hierarchical memory models respectively.

[0075] There are multiple hierarchical memory models in the present invention, and each hierarchical memory model is used to process data information of one dimension. After each hierarchical memory model is trained, it can effectively predict the urgency of the complaint event. Before training, the method can select different parameters according to the different characteristics of the information to be input, such as the number of layers of the model, the number of sequences in each layer, the probability of the activation state of the data processing unit in each sequence, the number of data processing units, etc.

[0076] In addition, the method also supports the adjustment of model parameters by inputting different training sets into each model.

[0077] Preferably, complaint features are input into multiple hierarchical memory models according to information dimensions, including: using the outputs of the first to fourth hierarchical memory models to calculate the first to fourth complaint urgency parameters respectively as the outputs of the multiple hierarchical memory models.

[0078] Therefore, the present invention supports processing data using the simplest and most effective analysis method and the most reasonable analysis tool based on business logic.

[0079] Through the above model, each input data can obtain the sample mean and variance. The sample mean is The variance is In addition, the method also calculates

[0080] On this basis, the method obtains the event urgency score s for each sequence information t-j Specifically, there is

[0081]

[0082] In the above formula, x t and x t-1 is the input content of the information sequence. π(·) is the binary set of each independent prediction result of the model, and a(·) is the sparse binary code of the input information.

[0083] Step 3, join the outputs of multiple hierarchical memory models and estimate the joint distribution of the join results.

[0084] In the above calculation process, the method needs to be optimized. Therefore, judging the event urgency of the power customer complaint information based on the joint distribution includes: judging the event urgency of the power customer complaint information using the joint distribution probability.

[0085] The formula for calculating the joint distribution probability is:

[0086]

[0087] The value of the calculated joint distribution probability function is limited to a certain range, and the method obtains a threshold of the probability and applies it to the analysis of the urgency of the event.

[0088] Step 4: determine the event urgency of the power customer complaint information based on the joint distribution, and issue a power customer complaint work order according to the event urgency.

[0089] Preferably, the urgency of the incident is:

[0090]

[0091] The sample mean is The variance is

[0092] Where k is the number of feedforward optimal matching sequences,

[0093] W is the window length of the event urgency score,

[0094] The value range of j is between 1 and W.

[0095] s t-j is the urgency score of the tjth event,

[0096] l is the number of the data processing unit of the current column in the hierarchical memory model,

[0097] is a probability function.

[0098] The value range of the probability function is limited by the aforementioned threshold. Only when the value of the probability function falls within the threshold range can the output of the model be judged to be accurate, thereby achieving model training.

[0099] Specifically, the expression of formula Q(x) is:

[0100]

[0101] The present invention can automatically determine the urgency of the complaint event through the above method. In addition, combined with the natural language processing method, the method can also obtain the content of the main event to which the complaint belongs, as well as the emotional attitude at the time of the complaint. In addition, combined with the analysis of the peak time period of historical complaints in the customer service system, it is analyzed whether the current complaint is in the peak period, so as to intelligently realize the priority of dispatching orders. For example, if the current complaint is in an emergency state, it is directly classified and dispatched. If the current state is not urgent, it is determined whether it is in the peak period of complaints. If so, the current complaint is temporarily backlogged. If it is not in the peak period, the system operation and maintenance personnel or the automated operation and maintenance equipment can handle it, and the classified dispatch is carried out in real time.

[0102] In addition, the classified dispatch here can obtain customer service content based on the classification of the analyzed main events, thereby performing classification. It is easy to imagine that according to the emotional state at the time of the complaint, the method also supports obtaining the complaint probability of the current fault or problem through intelligent analysis, which can be used for subsequent improvement of the model.

[0103] The second aspect of the present invention relates to an artificial intelligence grouping system based on electricity customer complaints. The system is used to implement the method in the first aspect of the present invention, and the system includes an acquisition module, a hierarchical module, a connection module and a judgment module; wherein the acquisition module is used to obtain electricity customer complaint information and extract complaint features from the electricity customer complaint information using a neural network; the hierarchical module is used to input the complaint features into multiple hierarchical memory models according to the information dimensions, and obtain the outputs of multiple hierarchical memory models; the connection module is used to connect the outputs of multiple hierarchical memory models and estimate the joint distribution of the connection results; the judgment module is used to judge the event urgency of the electricity customer complaint information based on the joint distribution, and issue an electricity customer complaint work order according to the event urgency.

[0104] Preferably, the acquisition module is used to obtain power customer complaint information, including: obtaining customer complaint images, complaint videos, complaint voices and complaint texts; and mapping the power customer complaint information of different dimensions into the same feature space.

[0105] Preferably, an acquisition module is used to extract complaint features from the electricity customer complaint information using a neural network, including: extracting complaint features from the electricity customer complaint information based on a spatiotemporal data mining method; the complaint features are represented in vector form, and the spatiotemporal data mining method is implemented using CNN or RNN.

[0106] Preferably, the complaint features acquired by the acquisition module include complaint image features, complaint video features, complaint voice features, and complaint text features;

[0107] A hierarchical module is used to input the complaint features into multiple hierarchical memory models according to information dimensions, including: inputting complaint image features, complaint video features, complaint voice features and complaint text features into the first to fourth hierarchical memory models respectively.

[0108] Preferably, the hierarchical module inputs the complaint features into multiple hierarchical memory models according to information dimensions, including: using the outputs of the first to fourth hierarchical memory models to calculate the first to fourth complaint urgency parameters respectively as the outputs of the multiple hierarchical memory models.

[0109] Preferably, a connection module is used to connect the outputs of multiple hierarchical memory models and estimate the joint distribution of the connection results, including: connecting the first to fourth complaint urgency parameters and calculating the joint distribution probability of the connection results.

[0110] Preferably, the judgment module is used to judge the event urgency of the power customer complaint information based on the joint distribution, including: judging the event urgency of the power customer complaint information by using the joint distribution probability.

[0111] Preferably, the urgency of the event determined by the determination module is:

[0112]

[0113] The sample mean is The variance is

[0114] Where k is the number of feedforward optimal matching sequences,

[0115] W is the window length of the event urgency score,

[0116] The value range of j is between 1 and W.

[0117] s t-j is the urgency score of the tjth event,

[0118] l is the number of the data processing unit of the current column in the hierarchical memory model,

[0119] is a probability function.

[0120] The third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method in the first aspect of the present invention.

[0121] A fourth aspect of the present invention relates to a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the method in the first aspect of the present invention.

[0122] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0123] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0124] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0126] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0127] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An artificial intelligence grouping method based on power customer complaints, characterized in that: include: Obtaining power customer complaint information, and extracting complaint features from the power customer complaint information using a neural network; Inputting the complaint features into multiple hierarchical memory models according to information dimensions, and obtaining outputs of multiple hierarchical memory models; Connecting the outputs of multiple hierarchical memory models and estimating the joint distribution of the connection results; The event urgency of the power customer complaint information is determined based on the joint distribution, and a power customer complaint work order is issued according to the event urgency.

2. The artificial intelligence grouping method based on power customer complaints according to claim 1 is characterized by: The obtaining of power customer complaint information includes: Obtain customer complaint images, complaint videos, complaint voices and complaint texts; The power customer complaint information of different dimensions is mapped into the same feature space.

3. The artificial intelligence grouping method based on power customer complaints according to claim 2 is characterized by: The extracting complaint features from the power customer complaint information using a neural network includes: Based on the spatiotemporal data mining method, the complaint features are extracted from the power customer complaint information; The complaint features are represented in vector form, and the spatiotemporal data mining method is implemented using CNN or RNN.

4. The artificial intelligence grouping method based on power customer complaints according to claim 3 is characterized by: The complaint features include complaint image features, complaint video features, complaint voice features and complaint text features; The step of inputting the complaint features into multiple hierarchical memory models according to information dimensions includes: The complaint image features, complaint video features, complaint voice features and complaint text features are input into the first to fourth hierarchical memory models respectively.

5. The artificial intelligence grouping method based on power customer complaints according to claim 3 is characterized by: The complaint features are input into multiple hierarchical memory models according to information dimensions, including: The outputs of the first to fourth hierarchical memory models are used to calculate the first to fourth complaint urgency parameters respectively as the outputs of the multiple hierarchical memory models.

6. The artificial intelligence grouping method based on power customer complaints according to claim 5 is characterized by: The step of connecting the outputs of the plurality of hierarchical memory models and estimating the joint distribution of the connection results comprises: The first to fourth complaint urgency parameters are connected, and the joint distribution probability of the connection results is calculated.

7. The artificial intelligence grouping method based on power customer complaints according to claim 6 is characterized by: The determining the event urgency of the power customer complaint information based on the joint distribution includes: The joint distribution probability is used to determine the event urgency of the power customer complaint information.

8. The artificial intelligence grouping method based on power customer complaints according to claim 7 is characterized by: The urgency of the incident is: The sample mean is The variance is Where k is the number of feedforward optimal matching sequences, W is the window length of the event urgency score, The value range of j is between 1 and W. s t-j is the urgency score of the tjth event, l is the number of the data processing unit of the current column in the hierarchical memory model, is a probability function.

9. An artificial intelligence grouping system based on power customer complaints, characterized by: The system is used to implement the method described in any one of claims 1 to 8; The system includes an acquisition module, a stratification module, a connection module and a judgment module; wherein, The acquisition module is used to acquire power customer complaint information and extract complaint features from the power customer complaint information using a neural network; The hierarchical module is used to input the complaint features into multiple hierarchical memory models according to information dimensions, and obtain outputs of multiple hierarchical memory models; The connection module is used to connect the outputs of multiple hierarchical memory models and estimate the joint distribution of the connection results; The judgment module is used to judge the event urgency of the power customer complaint information based on the joint distribution, and issue a power customer complaint work order according to the event urgency.

10. The artificial intelligence grouping system based on power customer complaints according to claim 9 is characterized by: The acquisition module is used to obtain power customer complaint information, including: Obtain customer complaint images, complaint videos, complaint voices and complaint texts; The power customer complaint information of different dimensions is mapped into the same feature space.

11. The artificial intelligence grouping system based on power customer complaints according to claim 10 is characterized by: The acquisition module is used to extract complaint features from the power customer complaint information using a neural network, including: Based on the spatiotemporal data mining method, the complaint features are extracted from the power customer complaint information; The complaint features are represented in vector form, and the spatiotemporal data mining method is implemented using CNN or RNN.

12. The artificial intelligence grouping system based on power customer complaints according to claim 11 is characterized in that: In the acquisition module, the complaint features include complaint image features, complaint video features, complaint voice features and complaint text features; The step of inputting the complaint features into multiple hierarchical memory models according to information dimensions includes: The complaint image features, complaint video features, complaint voice features and complaint text features are input into the first to fourth hierarchical memory models respectively.

13. The artificial intelligence grouping system based on power customer complaints according to claim 12 is characterized by: The hierarchical module is used to input the complaint features into multiple hierarchical memory models according to information dimensions, including: The outputs of the first to fourth hierarchical memory models are used to calculate the first to fourth complaint urgency parameters respectively as the outputs of the multiple hierarchical memory models.

14. The artificial intelligence grouping system based on power customer complaints according to claim 13 is characterized by: The connection module is used to connect the outputs of multiple hierarchical memory models and estimate the joint distribution of the connection results, including: The first to fourth complaint urgency parameters are connected, and the joint distribution probability of the connection results is calculated.

15. The artificial intelligence grouping system based on power customer complaints according to claim 14 is characterized in that: The judgment module is used to judge the event urgency of the power customer complaint information based on the joint distribution, including: The joint distribution probability is used to determine the event urgency of the power customer complaint information.

16. The artificial intelligence grouping system based on power customer complaints according to claim 15 is characterized by: In the judgment module, The urgency of the incident is: The sample mean is The variance is Where k is the number of feedforward optimal matching sequences, W is the window length of the event urgency score, The value range of j is between 1 and W. s t-j is the urgency score of the tjth event, l is the number of the data processing unit of the current column in the hierarchical memory model, is a probability function.

17. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-8.

Images