Power supply reliability prediction method based on correlation analysis and deep belief network

Abstract

The invention relates to the technical field of data analysis, and discloses a power supply reliability prediction method based on correlation analysis and a deep belief network, and the method comprises the steps: obtaining power supply service data; selecting characteristics related to the power supply reliability in the power supply service data as influence indexes; performing correlation analysis on the power supply reliability and the influence indexes by adopting a Pearson's correlation coefficient, and selecting positive correlation influence indexes or negative correlation influence indexes as input characteristics; establishing a deep belief network, inputting the power supply service data of the input features into the deep belief network, and optimizing network parameters by the deep belief network by adopting layer-by-layer pre-training and BP reverse fine tuning so as to determine a prediction model structure; and evaluating the prediction model by adopting a root-mean-square error and an average absolute percentage error. The correlation analysis reduces the calculation complexity of the prediction model and enhances the interpretability of the model to a certain extent.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a power supply reliability prediction method based on correlation analysis and a deep belief network. Background Art

[0002] With the development of information technology in power enterprises, the trend of diversified data sources is conducive to better research on data-driven prediction methods. Common prediction algorithms include support vector machines, time series models, and traditional artificial neural networks, and these algorithms are mainly used in load prediction, photovoltaic prediction, etc., and are relatively rarely used in power supply reliability prediction.

[0003] Traditionally, support vector machines are used for prediction. Training large-scale data requires a lot of time and computing resources, and its performance depends largely on the selection of parameters and the design of kernel functions, which increases the complexity and usage requirements of the model. Traditional artificial neural networks need to convert all features into digital form, which may lead to information loss, and the network is difficult to explain the reasoning process and basis, making it difficult to convince people with the prediction results.

[0004] The technical features related to power supply reliability can be obtained based on the analysis of service quality characteristics, and there are many related technical features. Inputting a large number of redundant technical features into support vector machines or traditional artificial neural networks for prediction will not only prolong the training time, but also affect the prediction results. Therefore, there is an urgent need for a prediction model with a convincing reasoning process, more accurate prediction results and short training time. Summary of the invention

[0005] The present invention provides a power supply reliability prediction method based on correlation analysis and deep belief network to solve the technical problems of long training time of existing prediction models and large errors in prediction results.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a power supply reliability prediction method based on correlation analysis and deep belief network, and the method includes:

[0007] Obtain power supply service data;

[0008] Selecting features related to the power supply reliability rate in the power supply service data as influencing indicators;

[0009] The Pearson correlation coefficient is used to perform a correlation analysis between the power supply reliability rate and the influencing index, and a positively correlated influencing index or a negatively correlated influencing index is selected as an input feature;

[0010] Establishing a deep belief network, inputting the power supply service data of the input features into the deep belief network, wherein the deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning to optimize network parameters to determine a prediction model structure;

[0011] The root mean square error and mean absolute percentage error were used to evaluate the prediction model.

[0012] Furthermore, the method further includes preprocessing the power supply service data corresponding to the influencing index, that is, performing normalization and denormalization processing on the power supply service data corresponding to the influencing index.

[0013] Furthermore, the specific steps of using the Pearson correlation coefficient to perform a correlation analysis between the power supply reliability rate and the influencing index and selecting a positively correlated influencing index or a negatively correlated influencing index as an input feature include:

[0014] Select multiple influencing indicators according to actual needs, wherein the influencing indicators include the average price coefficient of user electricity charges, the number of power outages, the fault repair time and the maximum load;

[0015] Calculating the analysis value of the Pearson correlation coefficient between the power supply reliability rate and the influencing index;

[0016] A positively correlated influencing index or a negatively correlated influencing index is selected as an input feature according to the analysis value.

[0017] Furthermore, the calculation formula of the Pearson correlation coefficient is:

[0018]

[0019] Where: ρ is the analysis value corresponding to the Pearson correlation coefficient, X is the influencing indicator, S is the data set composed of the influencing indicators, Y is the power supply reliability index, cov is the covariance of the two indicators, and σ is the standard deviation of the corresponding indicator.

[0020] Furthermore, the specific steps of establishing a deep belief network, inputting the power supply service data of the input features into the deep belief network, and optimizing network parameters by layer-by-layer pre-training and BP reverse fine-tuning to determine the prediction model structure include:

[0021] The deep belief network includes a plurality of stacked restricted Boltzmann machines and a logistic regression layer stacked on top;

[0022] Obtaining the power supply service data corresponding to the input feature, and inputting it into the deep belief network;

[0023] The network parameters are trained layer by layer using the unsupervised learning method of restricted Boltzmann machine;

[0024] The logistic regression layer includes a BP neural network, which uses a supervised learning method of the BP neural network to receive the output of the last restricted Boltzmann machine, and uses a back propagation algorithm of the BP neural network to fine-tune the entire network to determine the prediction model structure.

[0025] Furthermore, the specific steps of using the unsupervised learning method of the restricted Boltzmann machine to train the network parameters layer by layer include:

[0026] The restricted Boltzmann machine includes a visible layer and a hidden layer;

[0027] Training the visible layer of the first restricted Boltzmann machine;

[0028] After the visible layer of the first restricted Boltzmann machine is trained, its hidden layer is used as the visible layer of the next restricted Boltzmann machine to perform parameter training on the layer, and the training is performed layer by layer until completion.

[0029] Furthermore, the calculation formulas of the root mean square error (RMSE) and the mean absolute percentage error (MAPE) are respectively:

[0030]

[0031] In the formula, N is the total number of observed data, Y i is the actual load value, is the load forecast value.

[0032] The embodiment of the present invention further provides a power supply reliability prediction system based on correlation analysis and deep belief network, the system comprising:

[0033] A data acquisition module, used to acquire power supply service data;

[0034] A feature extraction module, used for selecting features related to the power supply reliability rate in the power supply service data as influencing indicators;

[0035] A correlation analysis module, used for performing a correlation analysis between the power supply reliability rate and the influencing index by using the Pearson correlation coefficient, and selecting the influencing index with a positive correlation or the influencing index with a negative correlation as an input feature;

[0036] A model structure determination module is used to establish a deep belief network, input the power supply service data of the input features into the deep belief network, and the deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning to optimize network parameters to determine the prediction model structure;

[0037] The model evaluation module is used to evaluate the prediction model using root mean square error and mean absolute percentage error.

[0038] An embodiment of the present invention further provides a computer device, including:

[0039] one or more processors;

[0040] A storage device for storing one or more programs;

[0041] When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement any of the above methods.

[0042] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which implements any of the above-mentioned methods when executed by a processor.

[0043] Beneficial effects of the embodiments of the present invention:

[0044] This embodiment uses a deep belief network DBN to predict the power supply reliability rate, and overcomes the limitations of traditional artificial neural networks through a deep nonlinear network structure. First, the Person coefficient is introduced to analyze the relevant features of the power supply reliability rate. The correlation analysis reduces the computational complexity of the prediction model and enhances the interpretability of the model to a certain extent. Then the selected features are input into the prediction model, and the model parameters are optimized using a layer-by-layer unsupervised training method and a back propagation training method. Finally, the model is evaluated using MAPE and RMSE. The simulation results show that the deep belief network DBN prediction model is superior to other models, with a maximum increase of 0.0419% in MAPE and a maximum increase of 0.0508% in RMSE, verifying the effectiveness of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flow chart of an embodiment of the present invention;

[0046] Figure 2 is a flow chart of an embodiment of the present invention;

[0047] Figure 3 is a flowchart of a specific implementation of S30 of the present invention;

[0048] Figure 4 is a flow chart of a specific implementation of S40 of the present invention;

[0049] Figure 5 It is a schematic diagram of the structure of the deep belief network;

[0050] Figure 6It is a graph showing the changing trend of the power supply reliability curve before and after normalization;

[0051] Figure 7 It is the correlation coefficient diagram among the indicators;

[0052] Figure 8 This is a comparison chart of the prediction results of five models;

[0053] Fig. 9 is a schematic block diagram of an example computer device of an embodiment of the present invention.

[0054] The reference numerals of the drawings in the specification are as follows:

[0055] 600, computer device; 601, computing unit; 602, ROM; 603, RAM; 604, bus; 605, I / O interface; 606, input unit; 607, output unit; 608, storage unit; 609, communication unit. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Embodiment 1

[0058] Please refer to Figure 1-2 , a power supply reliability prediction method based on correlation analysis and deep belief network provided by an embodiment of the present invention, the method comprising:

[0059] S10: Obtain power supply service data.

[0060] Power supply companies can use intelligent data collection and monitoring equipment to collect customer electricity usage information and provide data support for services such as data analysis and optimization management.

[0061] S20: Select the features related to power supply reliability in the power supply service data as influencing indicators.

[0062] Specifically, the service quality characteristics analysis provides more relevant indicators for the power supply reliability prediction. Power supply reliability is an important indicator of power supply quality. Power supply reliability is directly reflected in the power supply reliability rate. According to the actual application, the influencing indicators are considered: user electricity price coefficient, power outage times, fault repair time and maximum load, etc.

[0063] These impact indicators constitute the impact indicator data set S, S = [X price ,X outage,X repair ,X load ...], where X is the impact index, X price is the average price coefficient of user electricity charges, X outage is the number of power outages, X repair is the fault repair time, X load For maximum load.

[0064] Specifically, the method further includes preprocessing the power supply service data corresponding to the influencing index, that is, performing normalization and denormalization processing on the power supply service data corresponding to the influencing index.

[0065] Since different influencing indicators have different dimensions, in order to simplify the calculation and eliminate the differences in dimensions and value ranges between the data, they are converted into dimensionless quantities. Specifically, they can be normalized, which is conducive to improving the prediction accuracy of technical indicators. Among them, the normalized calculation formula is: In the formula, x is the technical indicator, x' is the normalized value of the technical indicator, the data interval is [0,1], x min and x max It is the minimum and maximum value of the technical indicator before normalization.

[0066] Denormalization is also performed to restore the normalized data to the range of the original data, so that the normalized data can be better understood and interpreted during the data analysis and model application stages. Denormalization ensures the interpretability and practicality of the data, making the processed data more in line with actual business needs.

[0067] The calculation formula for denormalization is: x = (x max -x min )x'+x min , where x is the technical indicator, x' is the normalized value of the technical indicator, the data interval is [0,1], x min and x max It is the minimum and maximum value of the technical indicator before normalization.

[0068] S30: The Pearson correlation coefficient is used to perform correlation analysis between power supply reliability and influencing indicators, and positively correlated influencing indicators or negatively correlated influencing indicators are selected as input features.

[0069] If all influencing indicators are used as input features of the prediction model, it will inevitably increase redundant information, which may make the calculation of the model more complicated and may also lead to the occurrence of dimensionality disaster. Selecting a suitable feature set is conducive to enhancing the interpretability of the model and reducing the computational complexity. Therefore, this embodiment adopts the Pearson correlation coefficient for correlation analysis and selects influencing indicators with correlation as input features.

[0070] like Figure 3 As shown, S31: multiple influencing indicators are selected according to actual needs, and the influencing indicators include user electricity price average coefficient, power outage times, fault repair time and maximum load.

[0071] The influencing indicators can be selected by professional technicians according to actual needs. This step is to first select some features from all the features that may have an impact as influencing indicators, and then select the relevant features through the following steps.

[0072] S32: Calculate the analysis value of the Pearson correlation coefficient between the power supply reliability and the influencing index.

[0073] S33: Select a positively correlated influencing indicator or a negatively correlated influencing indicator as an input feature according to the analysis value.

[0074] The Pearson correlation coefficient has been widely recognized in measuring the correlation between two continuous variables. Therefore, this embodiment uses the Pearson correlation coefficient as a scoring function to select indicators. The influencing indicators selected according to the analysis value of the Pearson correlation coefficient can improve the calculation efficiency of the model and help the model better avoid the problem of overfitting.

[0075] The calculation formula of Pearson correlation coefficient is:

[0076]

[0077] Where ρ is the analysis value corresponding to the Pearson correlation coefficient, X is the influencing indicator, S is the data set composed of the influencing indicators, Y is the power supply reliability index, cov is the covariance of the two indicators, and σ is the standard deviation of the corresponding indicator.

[0078] In order to verify the effectiveness of the method, this embodiment selected power supply service data from a certain area in Anhui from January 2020 to December 2021 for relevant research and analysis, and used Matlab to perform normalization and denormalization simulation calculations on the power supply service data. Figure 6 The fluctuation of power supply reliability rate in two years is shown, and it can be seen that the curve change trend is consistent before and after normalization. Among them, data from some time periods are selected for correlation analysis, and the specific data can be seen in Table 1.

[0079]

[0080]

[0081] Table 1

[0082] The data in Table 1 were analyzed by Pearson's correlation coefficient, where Figure 7The correlation coefficients between the indicators are shown. As can be seen from the figure, there is a strong negative correlation between the power supply reliability and the number of power outages, and a weak negative correlation with the fault repair time and the maximum load. There is a strong positive correlation between the maximum load and the number of power outages, because the greater the load, the more likely it is that the system will experience a power outage. There is a strong positive correlation between the number of power outages and the fault repair time, and the more power outages there are, the more likely it is that the fault repair time will be. Therefore, the number of power outages, the fault repair time and the maximum load are selected as the input features of the prediction model.

[0083] S40: A deep belief network (DBN) is established, and the power supply service data with input features is input into the deep belief network. The deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning to optimize network parameters to determine the prediction model structure.

[0084] like Figure 4 As shown, specifically, S41: the deep belief network (DBN) includes several stacked restricted Boltzmann machines and a logistic regression layer stacked on top.

[0085] Figure 5 This is a schematic diagram of the structure of a deep belief network (DBN). As can be seen from the figure, the hidden layer is stacked on top of the visible layer, and the logistic regression layer is stacked on top of the hidden layer.

[0086] S42: Obtain power supply service data corresponding to the input features, and input them into the deep belief network.

[0087] S43: Use the unsupervised learning method of restricted Boltzmann machine to train the network parameters layer by layer. This method can effectively avoid the overfitting problem of deep learning.

[0088] More specifically, the restricted Boltzmann machine includes a visible layer and a hidden layer; the visible layer of the first restricted Boltzmann machine is trained; after the visible layer of the first restricted Boltzmann machine is trained, its hidden layer is used as the visible layer of the next restricted Boltzmann machine to train the parameters of the layer, and the training is performed layer by layer until the end.

[0089] S44: The logistic regression layer includes a BP neural network, which receives the output of the last restricted Boltzmann machine using the supervised learning method of the BP neural network, and fine-tunes the entire network using the back propagation algorithm of the BP neural network to determine the prediction model structure.

[0090] A Deep Belief Network (DBN) is a feedforward neural network algorithm that can find the best solution in the hidden layer relatively quickly. Compared with traditional shallow neural networks, DBN can represent complex functions with fewer parameters. DBN can reconstruct its input in a probabilistic way and consists of multiple simple learning modules, namely the restricted Boltzmann machine unsupervised learning part and the logistic regression layer part for prediction, which helps to avoid overfitting problems.

[0091] The restricted Boltzmann machine is a two-layer neural network that can learn the probability distribution of an input set. The restricted Boltzmann machine (RBM) consists of a visible layer v (input layer) and a hidden layer h. In the RBM, there is no connection between neurons in the layer, and there is full connection between neurons in different layers, which makes the hidden units independent under given visible state conditions. The learning process of RBM is achieved through the energy function. The specific formula of the energy function is as follows:

[0092]

[0093] Where v is the visible unit; h is the hidden unit; a and b are biases; and W is the weight.

[0094] According to the energy function, the joint probability distribution of the visible layer and the hidden layer is as follows:

[0095] P(v,h)=e -E(v’h) / Z

[0096] Where Z is the sum of all energy functions of visible units and hidden units.

[0097] The activation functions of the visible and hidden layers are as follows:

[0098]

[0099] Where sigm is the sigmoid activation function, DBN is composed of multiple RBMs stacked together, and the unknown parameters (W, a, and b) can be unsupervised layer-by-layer pre-training and reverse BP fine-tuning.

[0100] The entire network is fine-tuned using the back-propagation algorithm, and the parameter training is as follows:

[0101]

[0102]

[0103] In the formula, E P(h|v) <> is the expected value of input data; E recon <>Reconstruct the expected value of the distribution.

[0104] The deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning for multiple training cycles until the model reaches the optimal state, which is then determined as the final prediction model.

[0105] S50: The prediction model was evaluated using root mean square error and mean absolute percentage error.

[0106] Before prediction, the data is normalized, and the prediction data is denormalized to obtain the power supply reliability prediction results, as shown in Table 2. It is not difficult to see from Table 2 that the DBN power supply reliability prediction value is basically close to the true value, with the maximum difference between the predicted value and the actual value being 0.0389% and the minimum difference being 0.0045%.

[0107] time Power supply reliability rate / % Predicted value / % Forecast difference / % 2021-07 99.8054 99.8443 0.0389 2021-08 99.9485 99.9440 -0.0045 2021-09 99.9229 99.9366 0.0137 2021-10 99.9286 99.9146 -0.0140 2021-11 99.8362 99.8477 0.0115 2021-12 99.8623 99.8828 -0.0205

[0108] Table 2

[0109] In order to verify the performance of the model, the proposed model will be compared with the traditional artificial neural network (ANN), support vector regression (SVR), and differential integrated moving average autoregressive model (ARIMA) for prediction performance. The prediction results are shown in Table 3. It is not difficult to find that the prediction accuracy of ARIMA error is lower than that of other prediction methods. This is due to the nonlinear phenomenon in the input data, and ARIMA cannot capture the nonlinear relationship well. Among them, the proposed Pearson-DBN has the highest accuracy, with a maximum improvement of 0.0419% in MAPE and 0.0508% in RMSE. Artificial neural network ANN has limited ability to represent complex functions and cannot learn the essential characteristics of data sets from a small number of sample sets. Deep belief network DBN can overcome this limitation to a certain extent. Compared with ANN, DBN improves MAPE by 0.0092% and RMSE by 0.0139%.

[0110] Pearson can analyze the correlation between different indicators and select them. This method can reduce the model calculation complexity and improve the model prediction accuracy. Compared with DBN, Pearson-DBN improves MAPE by 0.0084% and RMSE by 0.0087%. The prediction results of the five models are as follows Figure 8 As shown in Figure 3, the prediction of the proposed model Pearson-DBN is closer to the true value and has better prediction effect.

[0111] algorithm MAPE / % RMSE / % Pearson-DBN 0.0172 0.0203 DBN 0.0256 0.0290 ANN 0.0348 0.0429 SVR 0.0423 0.0451 ARIMA 0.0591 0.0711

[0112] Table 3

[0113] This embodiment uses a deep belief network DBN to predict the power supply reliability rate, and overcomes the limitations of traditional artificial neural networks through a deep nonlinear network structure. First, the Person coefficient is introduced to analyze the relevant features of the power supply reliability rate, and the number of power outages, fault repair time and maximum load are selected as the input features of the prediction model. The correlation analysis reduces the computational complexity of the prediction model to a certain extent. Then the selected features are input into the prediction model, and the model parameters are optimized using a layer-by-layer unsupervised training method and a back propagation training method. Finally, the model is evaluated using MAPE and RMSE. The simulation results show that the deep belief network DBN prediction model is better than other models, with a maximum increase of 0.0419% in MAPE and a maximum increase of 0.0508% in RMSE, verifying the effectiveness of the model.

[0114] Embodiment 2

[0115] The embodiment of the present invention further provides a power supply reliability prediction system based on correlation analysis and deep belief network, the system comprising:

[0116] A data acquisition module, used to acquire power supply service data;

[0117] A feature extraction module is used to select features related to power supply reliability in power supply service data as influencing indicators;

[0118] A correlation analysis module is used to perform correlation analysis between power supply reliability and influencing indicators using the Pearson correlation coefficient, and select positively correlated influencing indicators or negatively correlated influencing indicators as input features;

[0119] The model structure determination module is used to establish a deep belief network, input the power supply service data of the input features into the deep belief network, and the deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning to optimize the network parameters to determine the prediction model structure;

[0120] The model evaluation module is used to evaluate the prediction model using root mean square error and mean absolute percentage error.

[0121] The beneficial effects of the system of the embodiment of the present invention are equivalent to the beneficial effects of the above-mentioned power supply reliability prediction method based on correlation analysis and deep belief network, which will not be elaborated here.

[0122] Furthermore, the system also includes a data processing module for preprocessing the power supply service data corresponding to the influencing index, that is, performing normalization and denormalization processing on the power supply service data corresponding to the influencing index.

[0123] Furthermore, the correlation analysis module includes:

[0124] The characteristic index selection module is used to select multiple influencing indicators according to actual needs. The influencing indicators include the average price coefficient of user electricity charges, the number of power outages, the fault repair time and the maximum load;

[0125] A calculation module is used to calculate the analysis value of the Pearson correlation coefficient between the power supply reliability rate and the influencing index;

[0126] The input feature selection module is used to select positively correlated influencing indicators or negatively correlated influencing indicators as input features according to the analysis values.

[0127] Furthermore, the calculation formula of the Pearson correlation coefficient is:

[0128]

[0129] Where: ρ is the analysis value corresponding to the Pearson correlation coefficient, X is the influencing indicator, S is the data set composed of the influencing indicators, Y is the power supply reliability index, cov is the covariance of the two indicators, and σ is the standard deviation of the corresponding indicator.

[0130] Furthermore, the model structure determination module includes:

[0131] Establish a model module, where the deep belief network includes several stacked restricted Boltzmann machines and a logistic regression layer stacked on top; obtain power supply service data corresponding to the input features and input them into the deep belief network;

[0132] The training module is used to train the network parameters layer by layer using the unsupervised learning method of the restricted Boltzmann machine;

[0133] The reverse fine-tuning module, which is used for the logistic regression layer including the BP neural network, receives the output of the last restricted Boltzmann machine using the supervised learning method of the BP neural network, and fine-tunes the entire network using the back-propagation algorithm of the BP neural network to determine the prediction model structure.

[0134] Furthermore, the specific steps of the training module include:

[0135] The restricted Boltzmann machine includes visible layers and hidden layers;

[0136] Train the visible layer of the first restricted Boltzmann machine;

[0137] After the visible layer of the first restricted Boltzmann machine is trained, its hidden layer is used as the visible layer of the next restricted Boltzmann machine to train the parameters of this layer, and the training is carried out layer by layer until the end.

[0138] Furthermore, the calculation formulas for the root mean square error (RMSE) and the mean absolute percentage error (MAPE) are:

[0139]

[0140] In the formula, N is the total number of observed data, Y i is the actual load value, is the load forecast value.

[0141] Embodiment 3

[0142] An embodiment of the present invention further provides a computer device, comprising:

[0143] one or more processors;

[0144] A storage device for storing one or more programs;

[0145] When one or more programs are executed by one or more processors, the one or more processors implement the above-mentioned power supply reliability prediction method based on correlation analysis and deep belief network.

[0146] The beneficial effects of the computer device of the embodiment of the present invention are equivalent to the beneficial effects of the above-mentioned power supply reliability prediction method based on correlation analysis and deep belief network, which will not be elaborated here.

[0147] An embodiment of the present invention further provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements a power supply reliability prediction method based on correlation analysis and a deep belief network.

[0148] The beneficial effects of the storage medium of the present invention are equivalent to the beneficial effects of the above-mentioned power supply reliability prediction method based on correlation analysis and deep belief network, and will not be elaborated here.

[0149] Fig. 9 A schematic block diagram of an example computer device 600 that can be used to implement an embodiment of the present invention is shown. The computer device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The computer device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0150] like Fig. 9As shown, the computer device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0151] A number of components in the computer device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the computer device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0152] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as a power supply reliability prediction method based on correlation analysis and a deep belief network. For example, in some embodiments, the power supply reliability prediction method based on correlation analysis and a deep belief network may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the computer device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the power supply reliability prediction method based on correlation analysis and a deep belief network described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured in any other appropriate manner (for example, by means of firmware) to execute the power supply reliability prediction method based on correlation analysis and a deep belief network.

[0153] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. Implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0154] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0155] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0156] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0157] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0158] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0159] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0160] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the instructions of the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise clearly stated in this document, the execution of these steps is not strictly limited in order and can be performed in other orders.

[0161] Moreover, at least part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The order of execution is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0162] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A power supply reliability prediction method based on correlation analysis and deep belief network, characterized in that: The method comprises: Obtain power supply service data; Selecting features related to the power supply reliability rate in the power supply service data as influencing indicators; The Pearson correlation coefficient is used to perform a correlation analysis between the power supply reliability rate and the influencing index, and a positively correlated influencing index or a negatively correlated influencing index is selected as an input feature; Establishing a deep belief network, inputting the power supply service data of the input features into the deep belief network, wherein the deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning to optimize network parameters to determine a prediction model structure; The root mean square error and mean absolute percentage error were used to evaluate the prediction model.

2. The power supply reliability prediction method based on correlation analysis and deep belief network according to claim 1 is characterized in that: The method further includes preprocessing the power supply service data corresponding to the influencing index, that is, performing normalization and denormalization processing on the power supply service data corresponding to the influencing index.

3. The power supply reliability prediction method based on correlation analysis and deep belief network according to claim 1 is characterized in that: The specific steps of using the Pearson correlation coefficient to perform correlation analysis between the power supply reliability rate and the influencing index and selecting a positively correlated influencing index or a negatively correlated influencing index as an input feature include: Select multiple influencing indicators according to actual needs, wherein the influencing indicators include the average price coefficient of user electricity charges, the number of power outages, the fault repair time and the maximum load; Calculating the analysis value of the Pearson correlation coefficient between the power supply reliability rate and the influencing index; A positively correlated influencing index or a negatively correlated influencing index is selected as an input feature according to the analysis value.

4. The power supply reliability prediction method based on correlation analysis and deep belief network according to claim 3 is characterized in that: The calculation formula of the Pearson correlation coefficient is: Where ρ is the analysis value corresponding to the Pearson correlation coefficient, X is the influencing indicator, S is the data set consisting of the influencing indicators, Y is the power supply reliability index, cov is the covariance of the two indicators, and σ is the standard deviation of the corresponding indicator.

5. The power supply reliability prediction method based on correlation analysis and deep belief network according to claim 1 is characterized in that: The specific steps of establishing a deep belief network, inputting the power supply service data of the input features into the deep belief network, and optimizing network parameters by layer-by-layer pre-training and BP reverse fine-tuning of the deep belief network to determine the prediction model structure include: The deep belief network includes a plurality of stacked restricted Boltzmann machines and a logistic regression layer stacked on top; Obtaining the power supply service data corresponding to the input feature, and inputting it into the deep belief network; The network parameters are trained layer by layer using the unsupervised learning method of restricted Boltzmann machine; The logistic regression layer includes a BP neural network, which uses a supervised learning method of the BP neural network to receive the output of the last restricted Boltzmann machine, and uses a back propagation algorithm of the BP neural network to fine-tune the entire network to determine the prediction model structure.

6. The power supply reliability prediction method based on correlation analysis and deep belief network according to claim 5 is characterized in that: The specific steps of using the unsupervised learning method of the restricted Boltzmann machine to train the network parameters layer by layer include: The restricted Boltzmann machine includes a visible layer and a hidden layer; Training the visible layer of the first restricted Boltzmann machine; After the visible layer of the first restricted Boltzmann machine is trained, its hidden layer is used as the visible layer of the next restricted Boltzmann machine to perform parameter training on the layer, and the training is performed layer by layer until completion.

7. The power supply reliability prediction method based on correlation analysis and deep belief network according to claim 1 is characterized in that: The calculation formulas of the root mean square error (RMSE) and the mean absolute percentage error (MAPE) are respectively: In the formula, N is the total number of observed data, Y i is the actual load value, is the load forecast value.

8. The power supply reliability prediction system based on correlation analysis and deep belief network is characterized by: The system comprises: A data acquisition module, used to acquire power supply service data; A feature extraction module, used for selecting features related to the power supply reliability rate in the power supply service data as influencing indicators; A correlation analysis module, used for performing a correlation analysis between the power supply reliability rate and the influencing index by using the Pearson correlation coefficient, and selecting the influencing index with a positive correlation or the influencing index with a negative correlation as an input feature; A model structure determination module is used to establish a deep belief network, input the power supply service data of the input features into the deep belief network, and the deep belief network uses layer-by-layer pre-training and BP reverse fine-tuning to optimize network parameters to determine the prediction model structure; The model evaluation module is used to evaluate the prediction model using root mean square error and mean absolute percentage error.

9. A computer device, characterized in that: include: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images