Business data analysis and additional recording system and method thereof

Abstract

The invention provides a business data analysis and additional recording system and method, and relates to the technical field of data processing, and the method comprises the steps: obtaining an additional recording form, and inputting the additional recording form into an additional recording item analysis model, so as to recognize at least one to-be-additionally-recorded power business attribute item in the additional recording form; for each first power service attribute item to be additionally recorded, determining at least one recorded second power service attribute item associated with the first power service attribute item from an additional recording form, and determining additional recording content for the first power service attribute item according to each second power service attribute item; and filling the corresponding first power business attribute item based on each piece of additional recording content information. Therefore, the deep learning model based on the time attention mechanism is used for business data analysis and additional recording, and an efficient, accurate and reliable power grid business data additional recording solution is provided.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a business data analysis and supplementary recording system and method thereof. Background Art

[0002] Grid business data is an important basis for power system operation analysis, but in actual situations, data loss and incorrect entry often occur due to equipment failure, communication barriers or human errors.

[0003] At present, the existing data supplementation methods mainly rely on manual processing or simple rule matching, such as marking the missing business data in the EXCEL (spreadsheet) document through offline supplementation, and supplementing the business data in the EXCEL document. However, these methods are not only time-consuming and labor-intensive, but also inefficient, and often require supplementation personnel to be familiar with different power grid environments or business data specifications in advance, and sometimes introduce new errors, resulting in poor reliability.

[0004] In response to the above problems, the industry has not yet proposed a better technical solution. Summary of the invention

[0005] The present application provides a business data analysis and supplementary recording method and system, which are used to at least solve the problems of low efficiency and poor reliability caused by relying on manual analysis and supplementary data in the prior art.

[0006] The present application provides a business data analysis supplementary recording method, including: obtaining a supplementary recording form, and inputting the supplementary recording form into a supplementary recording item analysis model to identify at least one electric power business attribute item to be supplemented in the supplementary recording form; the supplementary recording item analysis model adopts a deep learning model based on a time attention mechanism; the supplementary recording form includes multiple electric power business attribute items with corresponding timestamps; for each of the first electric power business attribute items to be supplemented, determining at least one entered second electric power business attribute item associated with the first electric power business attribute item from the supplementary recording form, and determining the supplementary recording content information for the first electric power business attribute item according to each of the second electric power business attribute items; and filling in the corresponding first electric power business attribute item based on each of the supplementary recording content information.

[0007] Optionally, the first power service attribute item has a corresponding first timestamp, and each of the second power service attribute items has a corresponding second timestamp; wherein the time difference between each of the second timestamps and the first timestamp is less than a preset time threshold.

[0008] Optionally, the power business attribute items include any one of the following: power grid equipment status, power grid hourly power consumption, power grid equipment fault information, personnel inspection information, power grid voltage fluctuation information, equipment environmental parameters and power grid load mode; the first power business attribute item P1 is equipment fault information, and the second power business attribute item Q1 associated with P1 is: inspection information and power grid voltage fluctuation information; and, the first power business attribute item P2 is power grid load information, and the second power business attribute item Q2 associated with P2 is: power grid hourly power consumption and power grid load mode.

[0009] Optionally, determining at least one entered second electric power business attribute item associated with the first electric power business attribute item from the supplementary recording form includes: based on a preset attribute item association table, determining at least one entered second electric power business attribute item associated with the first electric power business attribute item from the supplementary recording form; the attribute item association table contains multiple attribute item relationships, and the attribute item relationships predefine the relationship between missing attribute items and corresponding entered attribute items.

[0010] Optionally, determining at least one entered second power business attribute item associated with the first power business attribute item from the supplementary entry form includes: inputting the first power business attribute item into the graph convolutional network, so that the graph convolutional network determines at least one entered second power business attribute item associated with the first power business attribute item; each node in the graph convolutional network is defined by a corresponding power business attribute item, and there is an association between the nodes connected by the edges in the graph convolutional network.

[0011] Optionally, determining the supplementary content information for the first power business attribute item based on each of the second power business attribute items includes: displaying the first power business attribute item and each of the second power business attribute items based on a touch interactive screen; obtaining user input information for the first power business attribute item based on the touch interactive screen, and determining the supplementary content information based on the user input information.

[0012] Optionally, determining the supplementary content information for the first power business attribute item based on each of the second power business attribute items includes: inputting each of the second power business attribute items and the first power business attribute item into a supplementary content synthesis model, so that the supplementary content synthesis model synthesizes corresponding supplementary content information for each of the first power business attribute items respectively; the content synthesis model adopts a deep generation model.

[0013] Optionally, the content synthesis model adopts a variational autoencoder, the variational autoencoder includes an encoder module and a decoder module, and the supplementary content synthesis model synthesizes the supplementary content information by performing the following operations: determining an input time series according to the first power service attribute item and each of the second power service attribute items Among them, g (i) represents the first power business attribute item of the i-th item, and Indicates that g (i) corresponding second power service attribute items;

[0014] Based on the encoder module, construct a latent space parameter distribution corresponding to the input time series:

[0015]

[0016] z=μ φ (g,x)+σ φ (g,x)⊙∈,

[0017] where z represents a random variable in the latent space, φ represents the encoder parameter, θ represents the decoder parameter, and μ φ (g,x) represents the latent space mean of the encoder output, represents the variance of the latent space of the encoder output; ∈ represents the noise sampled from the standard normal distribution, which enables the model to perform backpropagation; q φ (z|g,x) is the output of the encoder, which represents the posterior distribution of z given (g,x);

[0018] Based on the decoder module, sampling and reconstruction are performed from the latent space parameter distribution to determine a corresponding reconstructed data group for the first power service attribute item:

[0019]

[0020] Among them, p θ (g'|z) is the likelihood probability that the decoder generates the reconstructed data set g' for a given z, σ 2 represents the variance distribution of the reconstructed data generated by the decoder, represents normal distribution, I represents the identity matrix;

[0021] Determine each reconstructed data g' in the reconstructed data group based on the decoder (i) The loss function value of the first power service attribute item is determined from the respective reconstructed data according to the loss function value:

[0022]

[0023] Among them, L represents the loss function value; p(z) is the standard normal distribution D KL Represents the reconstructed data g' (i) The potential distribution of q φ (z|g' (i) ) with the standard normal distribution, and E represents the lower bound on maximizing the marginal log-likelihood.

[0024] Optionally, the supplementary item analysis model is an LSTM network with a time attention mechanism, which specifically comprises a cascaded input layer, an LSTM layer, a time attention layer and an output layer; the supplementary item analysis model identifies at least one electric power business attribute item to be supplemented in the supplementary list by performing the following operations: based on the input layer, determining a feature vector sequence corresponding to the supplementary list; the feature vector sequence is determined according to the business attribute items corresponding to each timestamp in the supplementary list; based on the LSTM layer, determining a hidden state vector sequence corresponding to the feature vector sequence:

[0025] h t =LSTM(m t ,h t-1 θ LSTM ),

[0026] Among them, θ LSTM represents LSTM parameters, m t represents the input feature vector at time step t, and h t represents the hidden state vector of LSTM at time step t, which defines the data dependency between time step t-1 and time step t; based on the temporal attention layer and the hidden state vector sequence, the importance of each time step in the feature vector sequence is evaluated to determine the corresponding weighted feature representation:

[0027] e t =v T tanh(W h ·h t +b),

[0028]

[0029] Among them, e t represents the unnormalized attention score at time step t, v,W h ,b represent the corresponding attention layer parameters, α t represents the normalized attention weight at time step t, c represents the attention-weighted context vector; classification prediction is performed based on the weighted feature representation to identify whether the power business attribute item corresponding to the input feature vector at time step t belongs to the first power business attribute item to be supplemented:

[0030]

[0031] in, is the prediction vector, which represents the prediction probability for the missing attribute item; W c ,b c They represent the corresponding output layer parameters respectively.

[0032] The present application also provides a business data analysis and supplementary recording system, including: an acquisition unit, used to acquire a supplementary recording form, and input the supplementary recording form into a supplementary recording item analysis model to identify at least one electric power business attribute item to be supplemented in the supplementary recording form; the supplementary recording item analysis model adopts a deep learning model based on a time attention mechanism; the supplementary recording form contains multiple electric power business attribute items with corresponding timestamps; a determination unit, used to determine at least one entered second electric power business attribute item associated with each first electric power business attribute item to be supplemented from the supplementary recording form, and determine the supplementary recording content information for the first electric power business attribute item according to each second electric power business attribute item; a filling unit, used to fill the corresponding first electric power business attribute item based on each of the supplementary recording content information.

[0033] The present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the business data analysis and supplementary recording method as described above is implemented.

[0034] The present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the business data analysis and supplementary recording methods described above.

[0035] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the business data analysis and supplementary recording methods described above.

[0036] The business data analysis and supplementary recording method, system, electronic device, and non-transitory computer-readable storage medium provided by the present application can produce at least the following technical effects:

[0037] The traditional manual supplementary recording method is inefficient. The deep learning model based on the temporal attention mechanism can automatically identify the missing power business attribute items at important time steps, and then quickly locate the related entered data, which greatly reduces the time required for manual search and matching and improves the supplementary recording efficiency.

[0038] Replacing manual re-entry with automated re-entry methods reduces data inaccuracy caused by manual input errors. Since the model can learn the patterns and rules of historical data, it can effectively avoid human errors and the resulting data inconsistencies.

[0039] Customized deep learning models can infer possible supplementary recording values ​​based on the statistical relationship between historical data and existing data. This inference based on historical data patterns is more accurate than manual guessing, thereby improving the quality and reliability of supplementary power grid business data.

[0040] Traditional manual supplementary recording usually requires operators to have a certain understanding of the power grid environment or business data specifications, but the automated supplementary recording model supplements the recording by learning the existing data relationships, thereby reducing the requirements for the professional skills of operators, while reducing professional requirements and labor costs;

[0041] In the technical solution provided in this application, the business data analysis and supplementary recording method brought by the deep learning model based on the temporal attention mechanism can provide an efficient, accurate and reliable power grid business data supplementary recording solution, meet the growing demand for automated processing of power data, and help promote the modernization of the power industry in data management and analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 A flowchart of an example of a method for supplementary recording of business data analysis according to an embodiment of the present application is shown;

[0044] Figure 2 A structural block diagram of an example of a supplementary item analysis model according to an embodiment of the present application is shown;

[0045] Figure 3 A structural block diagram of an example of a supplementary content synthesis model according to an embodiment of the present application is shown;

[0046] Figure 4 A structural block diagram of an example of a business data analysis and supplementary recording system according to an embodiment of the present application is shown;

[0047] Figure 5 It is a structural schematic diagram of the electronic device provided by this application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] Figure 1 A flowchart of an example of a business data analysis supplementary recording method according to an embodiment of the present application is shown.

[0050] Regarding the executor of the method of the embodiment of the present application, it can be any electronic device with processing and computing capabilities, such as a computer, mobile phone, etc., to automatically identify missing business data items in the supplementary entry form and provide an efficient, accurate and reliable power grid business data supplementary entry solution.

[0051] like Figure 1 As shown, in step S110, a supplementary entry form is obtained, and the supplementary entry form is input into a supplementary entry item analysis model to identify at least one power business attribute item to be supplemented in the supplementary entry form.

[0052] Here, the supplementary entry form includes multiple power business attribute items with corresponding timestamps. In some embodiments, the power business attribute items include any one of the following: power grid equipment status, power grid hourly power consumption, power grid equipment failure information, personnel inspection information, power grid voltage fluctuation information, equipment environmental parameters, and power grid load mode. In addition, the power business attribute items may also include other items, such as equipment operating status, environmental monitoring information, and the like.

[0053] In some examples, the supplementary recording personnel input the supplementary recording form into the business data analysis supplementary recording system, or the business data analysis supplementary recording system obtains the supplementary recording form from the power business system to perform preprocessing on the obtained supplementary recording form, including missing item identification, outlier processing, etc.

[0054] In some examples, the supplementary item analysis model uses a deep learning model based on the temporal attention mechanism. By introducing the temporal attention mechanism to capture the importance of different time points in the time series data, the supplementary item analysis model can focus on the time steps that have an important impact on the supplementary prediction, and ensure the comprehensive extraction of various related business attribute items that are missing at the same time. In addition, the use of the temporal attention mechanism in the deep learning model not only helps the comprehensiveness of the missing item identification results in the data supplementation task, but also improves the interpretability of the model prediction process because it can intuitively see which parts of the time series the model focuses on.

[0055] In step S120, for each first power service attribute item to be supplemented, at least one already-entered second power service attribute item associated with the first power service attribute item is determined from the supplementary recording form.

[0056] It should be noted that there are numerous power business attribute items, and there are correlations between different power business attribute items. In particular, a missing attribute item can be derived from other known attribute items.

[0057] In some embodiments, based on a preset attribute item association table, at least one entered second power business attribute item associated with the first power business attribute item is determined from the supplementary entry form. Here, the attribute item association table includes a plurality of attribute item relationships, and the attribute item relationship predefines the relationship between the missing attribute item and the corresponding entered attribute item. Thus, by using the attribute item association table in which the various association relationships are pre-stored, at least one entered attribute item associated with the missing attribute item can be obtained explicitly and accurately.

[0058] For example, the first power business attribute item P1 is equipment failure information, and the second power business attribute item Q1 associated with P1 is: inspection information and grid voltage fluctuation information; and the first power business attribute item P2 is grid load information, and the second power business attribute item Q2 associated with P2 is: grid hourly power consumption and grid load mode. In addition, more associations can be recorded in the attribute item association table, for example: the missing first power business attribute item is the real-time power generation efficiency of the power plant, and the corresponding second power business attribute item is the coal consumption, ambient temperature, and unit equipment status information of the power plant.

[0059] More preferably, the first power service attribute item has a corresponding first timestamp, each second power service attribute item has a corresponding second timestamp, and the time difference between each second timestamp and the first timestamp is less than a preset time threshold. Thus, by limiting the time between the timestamp of the entered associated service attribute item and the timestamp of the missing attribute item, service attribute items with weak time correlation are filtered, which can reduce the resource occupation of the system and improve the system response performance.

[0060] It should be understood that, in addition to the above-mentioned business attribute items that are obviously derived from physical formulas, the correlation between some business attribute items may not be expressed by explicit relational expressions. For example, the hidden relationship between different business attribute items can be learned through a machine learning model.

[0061] In step S130, for each first power service attribute item to be supplemented, the supplementary recording content for the first power service attribute item is determined according to each second power service attribute item.

[0062] In one example of the embodiment of the present application, after the system outputs each first power business attribute item to be supplemented and the associated second power business attribute item, the supplementary recording content input by the user is received. In another example of the embodiment of the present application, the system can perform analysis and calculation through the second power business attribute item to obtain the supplementary recording content of the corresponding first power business attribute item, for example, using a deep generation model to perform content synthesis operations.

[0063] In step S140, the corresponding first power service attribute items are filled based on each supplementary content information.

[0064] Thus, the generated supplementary content information is filled into the corresponding first power business attribute item, filling the missing items in the original data, improving the integrity of the power business data, and providing a data basis for subsequent power system operation analysis. In addition, the use of a deep learning model based on the temporal attention mechanism can automatically identify power business attribute items that are missing at important time steps, and then quickly locate the associated entered data, greatly reducing the time required for manual search and matching, and improving the efficiency of supplementary recording.

[0065] Regarding the details of step S110, in some embodiments, a deep learning network based on a temporal attention mechanism can better complete the analysis of supplementary items that are strongly related to business timestamps. The use of a temporal attention mechanism can help the model focus on the most relevant time steps when processing time series data, thereby improving prediction accuracy.

[0066] Figure 2 A structural block diagram of an example of a supplementary item analysis model according to an embodiment of the present application is shown.

[0067] like Figure 2 As shown, the supplementary item analysis model 200 adopts an LSTM (long short-term memory) network with a temporal attention mechanism, specifically comprising a cascaded input layer 210, an LSTM layer 220, a temporal attention layer 230, and an output layer 240. Specifically, the supplementary item analysis model 200 identifies at least one power business attribute item to be supplemented in the supplementary item form by performing the following operations:

[0068] Based on the input layer 210, a feature vector sequence corresponding to the supplementary recording form is determined, and the feature vector sequence is determined according to the business attribute items corresponding to each time stamp in the supplementary recording form.

[0069] The hidden state vector sequence corresponding to the feature vector sequence is determined based on the LSTM layer 220:

[0070] h t =LSTM(m t ,ht-1 θ LSTM ),

[0071] Among them, θ LSTM represents LSTM parameters, m t represents the input feature vector at time step t, and h t represents the hidden state vector of LSTM at time step t, which defines the data dependency between time step t-1 and time step t.

[0072] Based on the temporal attention layer 230 and the hidden state vector sequence, the importance of each time step in the feature vector sequence is evaluated to determine the corresponding weighted feature representation:

[0073] e t =v T tanh(W h ·h t +b),

[0074]

[0075] Among them, e t represents the unnormalized attention score at time step t, v,W h ,b represent the corresponding attention layer parameters, α t represents the normalized attention weight at time step t, and c represents the attention-weighted context vector.

[0076] It should be noted that the key to the temporal attention mechanism is to learn a weight distribution α t , which indicates the importance of each time step for the attribute item to be supplemented, and is related to the hidden state sequence h of the sequence model t are used together to create a weighted representation c.

[0077] Based on the output layer 240, classification prediction is performed using weighted feature representation to identify whether the power service attribute item corresponding to the input feature vector at time step t belongs to the first power service attribute item to be supplemented:

[0078]

[0079] in, is the prediction vector, which represents the prediction probability for the missing attribute item; W c ,b c They represent the corresponding output layer parameters respectively.

[0080] For the training process of the supplementary item analysis model 200, the model parameters can be learned by optimizing the loss function (e.g., cross entropy loss) and using the back propagation algorithm. Exemplarily, by collecting and preprocessing data, such as normalization of timestamps, feature encoding of business attribute items, etc. Then, the network parameters such as θ are initialized. LSTM ,W h ,v,b,W c ,b c . Then, iterative training is performed on the training set, the loss function is calculated, and the parameters are updated by gradient descent. Finally, the model performance is evaluated on the validation set and fine-tuned as needed. Furthermore, once the model training converges, it is possible to identify the power business attribute items that need to be supplemented in each supplementary entry form.

[0081] Regarding the details of the temporal attention layer 230, in some embodiments, the purpose of the design of the temporal attention layer is to enable the model to focus on the moments that are most important for the supplementary recording task. The temporal attention layer 230 is generally implemented by assigning different weights to the feature vectors of each time step output by the convolutional layer, so that the model can focus on processing those time steps that are more likely to contain information that is helpful for the supplementary recording task.

[0082] Exemplarily, the design of the temporal attention layer is carried out through the following steps. First, the time-related query matrix (Query Matrix) needs to be learned. The model learns a query matrix Q, which maps the feature vector of each time step of the time series data to produce a query vector. This is essentially a transformation of the features of a given time step.

[0083] Q=W q +X conv +b q ,

[0084] Among them, X conv is the output of the convolutional layer, W q is the learnable weight matrix, b q is the bias term.

[0085] Then, the attention score is calculated for each time step. Specifically, the model uses the query vector and a training parameter W k Generate an attention score to determine how much the model pays attention to each time step.

[0086] A=W k Q,

[0087] Among them, W k is the learnable time-step key weight matrix.

[0088] Then, the attention scores are normalized by the softmax function to ensure that the sum of the attention weights of all time steps is 1.

[0089] A softmax =softmax(A),

[0090] Among them, A softmax The shape of is the same as A, but all elements have values ​​between 0 and 1, and all values ​​sum to 1, so that the weight of each time step represents its relative importance.

[0091] Finally, the features output by the convolutional layer are weighted using the normalized attention weights to obtain the features adjusted by temporal attention.

[0092] X attn =A softmax ⊙X conv ,

[0093] In this formula, ⊙ represents an element-wise broadcast multiplication, applying the attention weights to the corresponding feature vectors.

[0094] Through the above deployment operation for the time attention layer 230, the time attention layer 230 can automatically learn which time steps to pay more attention to, so as to more effectively process sequence data and capture the temporal dependencies related to the task. When training the model, the parameters of this attention part are automatically adjusted to achieve the best effect of the supplementary recording task. In the implementation process, the attention mechanism not only helps the data supplementary recording task, but also improves the interpretability, because it can be intuitively seen which parts of the time series the model focuses more on.

[0095] As for the details of step S120, in some cases, a graph convolutional network (GCN) is used to associate different power business attribute items. Specifically, the first power business attribute item is input into the graph convolutional network, so that the graph convolutional network determines at least one entered second power business attribute item associated with the first power business attribute item. Here, each node in the graph convolutional network is defined by a corresponding power business attribute item, and there is an association between the nodes connected by the edges in the graph convolutional network. In combination with an example of practical application, the first power business attribute item is used as the input node of the graph convolutional network, so that the graph convolutional network infers the corresponding second power business attribute items through the neighboring nodes connected to each edge of the input node. Thus, the information transmission mechanism of the graph convolutional network is used to aggregate the nodes and extract the association features between the nodes. Therefore, through the GCN model, the association relationship between the first power business attribute item and the second power business attribute item can be mined in the graph structure, the accuracy and comprehensiveness of the association can be improved, the implicit association between different business attribute items can be learned, and the graph structure of the graph convolutional network can be used to better model the complex relationship between the power business attribute items, which helps to capture the potential rules in the business data.

[0096] In some embodiments, the deployment process for the graph neural network mainly includes the following operations: first, by collecting various types of power business attribute items, including equipment status, power consumption, fault information, inspection information, voltage fluctuations, environmental parameters, and load mode attributes. Then, the known associations between the business attribute items are constructed as edges, and their feature vectors are initialized; if there is no explicit relationship, a complete graph can be constructed, and the uncertainty of the association can be reduced through subsequent learning. Then, the constructed graph data is input into the GCN model for training, so that the model can learn the complex interactions and associations between nodes (i.e., power business attribute items).

[0097] Exemplarily, a graph convolutional network layer can be defined in the following way:

[0098]

[0099] Among them, H (l) is the node feature representation matrix of the lth layer, H (l+1) is the node feature representation matrix of the l+1th layer, is the adjacency matrix with self-connection added (A is the original adjacency matrix and I is the identity matrix), yes The degree matrix of W (l) is the weight matrix of the lth layer, and σ uses the ReLU nonlinear activation function.

[0100] Combined with the examples of the embodiments of this application, H(l) It can contain the characteristics of each power business attribute item. Represents the association relationship between attribute items, and W (l) Represents the weight matrix corresponding to these relationships. Through a multi-layer graph convolutional network, the GCN model can capture the high-order neighborhood aggregation of node features and update the representation of each node to integrate neighborhood information.

[0101] Regarding the implementation details of the above-mentioned step S130, in an example of an embodiment of the present application, based on the touch interactive screen, the first power business attribute item and each of the second power business attribute items are displayed. Then, based on the touch interactive screen, the user input information for the first power business attribute item is obtained, and the supplementary content information is determined according to the user input information. Specifically, the supplementary recording system displays the first business attribute item to be supplemented and the associated second business attribute item that has been entered through the screen, so that the supplementary recording personnel can deduce and calculate through the entered second business attribute item to obtain the supplementary content information of the first business attribute item, and then input the supplementary content information into the system through the touch interactive screen. Thus, the confirmation and information entry of the power business attribute items to be supplemented are realized through user interaction, and there is no need for the supplementary recording personnel to search for the supplementary items and the associated recorded items, thereby improving the efficiency of content supplementation.

[0102] With respect to the implementation details of the above-mentioned step S130, in another example of the embodiment of the present application, each second power business attribute item and the first power business attribute item are input into the supplementary content synthesis model, so that the supplementary content synthesis model synthesizes the corresponding supplementary content information for each first power business attribute item, and the content synthesis model adopts a deep generation model. In the implementation of this example, each first power business attribute item and its associated second power business attribute item are input as inputs to a deep generation model, such as a generative adversarial network (GAN) or a variational autoencoder (VAE), to synthesize the supplementary content information related to the first power business attribute item. Thus, through the deep generation model, the supplementary content related to the first power business attribute item can be synthesized more accurately, the accuracy of the supplementary content is improved, it is helpful to generate diverse supplementary content, increase the diversity of data, and help to better reflect the complexity of real business scenarios. In addition, based on the deep generation model, potential business features can be learned, so as to better adapt to different business scenarios and improve the generalization ability of the model.

[0103] Figure 3 A structural block diagram of an example of a supplementary content synthesis model according to an embodiment of the present application is shown.

[0104] like Figure 3As shown, the supplementary content synthesis model 300 adopts VAE, and the variational autoencoder includes an encoder module 310 and a decoder module 320. Thus, the encoder module 310 converts the input data into a representation of the latent space, and the decoder module 320 restores the original data from the representation of the latent space, and synthesizes the supplementary content by learning the latent representation of the input data.

[0105] Specifically, the supplementary content synthesis model synthesizes the supplementary content information by performing the following operations:

[0106] Determine the input time sequence according to the first power service attribute item and each second power service attribute item Among them, g (i) represents the first power business attribute item of the i-th item, and Indicates that g (i) corresponding second power service attribute items;

[0107] Based on the encoder module, construct the latent space parameter distribution corresponding to the input time series:

[0108]

[0109] z=μ φ (g,x)+σ φ (g,x)⊙∈,

[0110] where z represents a random variable in the latent space, φ represents the encoder parameter, θ represents the decoder parameter, and μ φ (g,x) represents the latent space mean of the encoder output, represents the variance of the latent space of the encoder output; ∈ represents the noise sampled from the standard normal distribution, which enables the model to perform backpropagation; q φ (z|g,x) is the output of the encoder, which represents the posterior distribution of z given (g,x);

[0111] Based on the decoder module, sampling and reconstruction are performed from the latent space parameter distribution to determine a corresponding reconstructed data group for the first power service attribute item:

[0112]

[0113] Among them, p θ (g'|z) is the likelihood probability that the decoder generates the reconstructed data set g' for a given z, σ 2 represents the variance distribution of the reconstructed data generated by the decoder, represents normal distribution, I represents the identity matrix;

[0114] Determine the reconstructed data g' in the reconstructed data group based on the decoder (i)The loss function value of the first power business attribute item is determined from each reconstructed data according to the loss function value:

[0115]

[0116] Among them, L represents the loss function value; p(z) is the standard normal distribution D KL Represents the reconstructed data g' (i) The potential distribution q φ (z|g' (i) ) with the standard normal distribution, and E represents the lower bound on maximizing the marginal log-likelihood.

[0117] Regarding the training deployment details of VAE, in some embodiments, in order to find parameters φ and θ to maximize the ELBO (Evidence Lower Bound) under the dataset X, the parameters φ and θ are updated by stochastic gradient descent (SGD) so that the posterior distribution q is minimized while the reconstruction error is minimized. φ Minimize the divergence between (z|x) and the prior distribution p(z).

[0118] For example, a VAE can be trained as follows: for each instance x in the dataset (i) , the encoder module generates the mean μ φ (x (i) ) and variance Sample z using the reparameterization trick and use the decoder module to try to recover x from the sampled random variables z (i) , calculate the ELBO and update φ and θ through the optimizer.

[0119] After the VAE model is trained, the decoder p can be used θ By generating new supplementary content and sampling new z values ​​through the latent space and passing them to the decoder, the distribution of possible supplementary data x can be obtained, and the supplementary content information corresponding to the input can be automatically synthesized.

[0120] In the implementation process of this application, multiple artificial intelligence algorithm models are integrated, and an LSTM network based on the time attention mechanism is used in the supplementary item analysis model, which can effectively extract the time series characteristics of the power business attribute items in the supplementary form and capture the temporal dependencies of the business data. Based on the time attention mechanism, the model can evaluate the importance of different time steps, improve the focus on key time points, and help to more accurately identify the power business attribute items to be supplemented. Using the graph convolutional network, the association between the first power business attribute item and the second power business attribute item can be more accurately determined, and the correlation of the graph structure can be fully utilized. Through node representation learning, the model can better capture the explicit or implicit relationship between the power business attribute items and has a strong generalization ability. Using a deep generative model to automatically generate supplementary content information can more accurately synthesize the supplementary content related to the first power business attribute item, improve the accuracy and fidelity of the synthesized content, and help to better simulate the complexity of the actual business scenario. Through the variational autoencoder, the model can learn potential business features and improve the generalization ability and adaptability of the model. Therefore, the advantages of time series characteristics, correlation information and generation models are fully utilized to improve the accuracy and efficiency of supplementary recording of power business data and enhance the adaptability of the system.

[0121] The business data analysis and supplementary recording system provided by the present application is described below. The business data analysis and supplementary recording system described below and the business data analysis and supplementary recording method described above can be referenced to each other.

[0122] Figure 4 A structural block diagram of an example of a business data analysis and supplementary recording system according to an embodiment of the present application is shown.

[0123] like Figure 4 As shown, the business data analysis supplementary recording system 400 includes an acquisition unit 410, a determination unit 420 and a filling unit 430.

[0124] The acquisition unit 410 is used to obtain a supplementary recording form and input the supplementary recording form into a supplementary recording item analysis model to identify at least one power business attribute item to be supplemented in the supplementary recording form; the supplementary recording item analysis model adopts a deep learning model based on a temporal attention mechanism; the supplementary recording form includes multiple power business attribute items with corresponding timestamps.

[0125] The determination unit 420 is used to determine, for each first power business attribute item to be supplemented, at least one already entered second power business attribute item associated with the first power business attribute item from the supplementary recording form, and determine the supplementary recording content for the first power business attribute item based on each second power business attribute item.

[0126] The filling unit 430 is used to fill in the corresponding first power service attribute item based on each of the supplementary content information.

[0127] In some embodiments, an embodiment of the present application provides a non-volatile computer-readable storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute the above-mentioned business data analysis and supplement method of the present application.

[0128] In some embodiments, the embodiments of the present application also provide a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the above-mentioned business data analysis and supplement method.

[0129] In some embodiments, the embodiments of the present application also provide an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform a business data analysis supplement method.

[0130] Figure 5 is a schematic diagram of the hardware structure of an electronic device for executing a business data analysis and supplementary recording method provided by another embodiment of the present application, such as Figure 5 As shown, the device includes:

[0131] One or more processors 510 and memory 520, Figure 5 A processor 510 is taken as an example.

[0132] The device for executing the business data analysis supplementary recording method may also include: an input device 530 and an output device 540.

[0133] The processor 510, the memory 520, the input device 530 and the output device 540 may be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.

[0134] The memory 520 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the business data analysis and supplementary recording method in the embodiment of the present application. The processor 510 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 520, that is, the business data analysis and supplementary recording method in the above method embodiment is implemented.

[0135] The memory 520 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 520 may optionally include a memory remotely arranged relative to the processor 510, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0136] The input device 530 may receive input digital or character information and generate signals related to user settings and function control of the electronic device. The output device 540 may include a display device such as a display screen.

[0137] The one or more modules are stored in the memory 520, and when executed by the one or more processors 510, the business data analysis and supplementary recording method in any of the above method embodiments is executed.

[0138] The above product can execute the business data analysis and supplementary recording method provided in the embodiment of this application, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not described in detail in this embodiment, please refer to the method provided in the embodiment of this application.

[0139] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:

[0140] (1) Mobile communication equipment: This type of equipment is characterized by having mobile communication functions and its main purpose is to provide voice and data communications. This type of terminal includes: smart phones, multimedia phones, functional phones, and low-end phones.

[0141] (2) Ultra-mobile personal computer devices: These devices belong to the category of personal computers, have computing and processing functions, and generally also have mobile Internet access features. These terminals include: PDA, MID and UMPC devices, etc.

[0142] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players, handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0143] (4) Other onboard electronic devices with data interaction functions, such as on-board devices installed in vehicles.

[0144] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0145] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0146] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A business data analysis supplementary recording method, comprising: Obtaining a supplementary entry form, and inputting the supplementary entry form into a supplementary entry item analysis model to identify at least one power business attribute item to be supplemented in the supplementary entry form; the supplementary entry analysis model adopts a deep learning model based on a temporal attention mechanism; The supplementary entry form includes a plurality of power business attribute items with corresponding timestamps; For each of the first power business attribute items to be supplemented, determining at least one entered second power service attribute item associated with the first power service attribute item from the supplementary entry form, Determine, according to each of the second power service attribute items, supplementary content information for the first power service attribute item; Based on each of the supplementary content information, fill in the corresponding first power service attribute item.

2. The method according to claim 1, wherein: The first power service attribute item has a corresponding first timestamp, and each of the second power service attribute items has a corresponding second timestamp; wherein the time difference between each of the second timestamps and the first timestamp is less than a preset time threshold.

3. The method according to claim 1, wherein: The power service attribute items include any one of the following: power grid equipment status, power grid hourly power consumption, power grid equipment failure information, personnel inspection information, power grid voltage fluctuation information, equipment environment parameters and power grid load mode; The first power service attribute item P1 is equipment failure information, and the second power service attribute item Q1 associated with P1 is: inspection information and grid voltage fluctuation information; Furthermore, the first power service attribute item P2 is the power grid load information, and the second power service attribute item Q2 associated with P2 is: power grid hourly power consumption and power grid load mode.

4. The method according to claim 1, wherein: The determining, from the supplementary entry form, at least one entered second power service attribute item associated with the first power service attribute item comprises: Based on a preset attribute item association table, at least one entered second power business attribute item associated with the first power business attribute item is determined from the supplementary entry form; the attribute item association table contains multiple attribute item relationships, and the attribute item relationships predefine the relationship between missing attribute items and corresponding entered attribute items.

5. The method according to claim 1, wherein: The determining, from the supplementary entry form, at least one entered second power service attribute item associated with the first power service attribute item comprises: The first power business attribute item is input into the graph convolutional network so that the graph convolutional network determines at least one entered second power business attribute item associated with the first power business attribute item; each node in the graph convolutional network is defined by a corresponding power business attribute item, and there is an association between the nodes connected by the edges in the graph convolutional network.

6. The method according to claim 1, wherein: The determining, according to each of the second power service attribute items, the supplementary content information for the first power service attribute item includes: Based on the touch interactive screen, displaying the first power service attribute item and each of the second power service attribute items; Based on the touch interactive screen, user input information for the first power service attribute item is obtained, and supplementary content information is determined according to the user input information.

7. The method according to claim 1, wherein: The determining, according to each of the second power service attribute items, the supplementary content information for the first power service attribute item includes: Each of the second power business attribute items and the first power business attribute items is input into a supplementary content synthesis model, so that the supplementary content synthesis model synthesizes corresponding supplementary content information for each of the first power business attribute items respectively; the content synthesis model adopts a deep generation model.

8. The method according to claim 7, wherein: The content synthesis model adopts a variational autoencoder, which includes an encoder module and a decoder module. The supplementary content synthesis model synthesizes the supplementary content information by performing the following operations: Determine an input time sequence according to the first power service attribute item and each of the second power service attribute items Among them, g (i) represents the first power business attribute item of the i-th item, and Indicates that g (i) corresponding second power service attribute items; Based on the encoder module, construct a latent space parameter distribution corresponding to the input time series: z=μ φ (g,x)+σ φ (g,x)⊙∈, where z represents a random variable in the latent space, φ represents the encoder parameter, θ represents the decoder parameter, and μ φ (g,x) represents the latent space mean of the encoder output, represents the latent space variance of the encoder output; ∈ represents the noise sampled from the standard normal distribution, which enables the model to perform backpropagation; q φ (z|g,x) is the output of the encoder, which represents the posterior distribution of z given (g,x); Based on the decoder module, sampling and reconstruction are performed from the latent space parameter distribution to determine a corresponding reconstructed data group for the first power service attribute item: Among them, p θ (g'|z) is the likelihood probability that the decoder generates the reconstructed data set g' for a given z, σ 2 represents the variance distribution of the reconstructed data generated by the decoder, represents normal distribution, I represents the identity matrix; Determine each reconstructed data g' in the reconstructed data group based on the decoder (i) The loss function value of the first power service attribute item is determined from the respective reconstructed data according to the loss function value: Among them, L represents the loss function value; p(z) is the standard normal distribution D KL Represents the reconstructed data g' (i) The potential distribution q φ (z|g' (i) ) with the standard normal distribution, and E represents the lower bound on maximizing the marginal log-likelihood.

9. The method according to claim 1, wherein: The supplementary item analysis model is an LSTM network with a temporal attention mechanism, which specifically comprises a cascaded input layer, an LSTM layer, a temporal attention layer and an output layer; The supplementary entry analysis model identifies at least one power service attribute item to be supplemented in the supplementary entry form by performing the following operations: Based on the input layer, determining a feature vector sequence corresponding to the supplementary recording form; The feature vector sequence is determined according to the service attribute items corresponding to each time stamp in the supplementary recording table; Determine the hidden state vector sequence corresponding to the feature vector sequence based on the LSTM layer: h t =LSTM(m t ,h t-1 ;θ LSTM ), Among them, θ LSTM represents LSTM parameters, m t represents the input feature vector at time step t, and h t represents the hidden state vector of LSTM at time step t, which defines the data dependency between time step t-1 and time step t; Based on the temporal attention layer and the hidden state vector sequence, the importance of each time step in the feature vector sequence is evaluated to determine the corresponding weighted feature representation: e t =v T ·tanh(W h ·h t +b), Among them, e t represents the unnormalized attention score at time step t, v,W h ,b represent the corresponding attention layer parameters, α t represents the normalized attention weight at time step t, and c represents the attention-weighted context vector; Classification prediction is performed based on the weighted feature representation to identify whether the power service attribute item corresponding to the input feature vector at time step t belongs to the first power service attribute item to be supplemented: in, is the prediction vector, which represents the prediction probability for the missing attribute item; W c ,b c They represent the corresponding output layer parameters respectively.

10. A business data analysis and supplementary recording system, comprising: An acquisition unit, used for acquiring a supplementary recording form, and inputting the supplementary recording form into a supplementary recording item analysis model to identify at least one power business attribute item to be supplemented in the supplementary recording form; the supplementary recording item analysis model adopts a deep learning model based on a temporal attention mechanism; The supplementary entry form includes a plurality of power business attribute items with corresponding timestamps; a determining unit, configured to determine, for each of the first power service attribute items to be supplemented, from the supplementary recording form, at least one already recorded second power service attribute item associated with the first power service attribute item, and determine supplementary recording content information for the first power service attribute item according to each of the second power service attribute items; A filling unit is used to fill in the corresponding first power service attribute item based on each of the supplementary content information.

Images