Grade evaluation method for dynamic multi-dimensional power data and related equipment

Abstract

The invention relates to the field of power data management, and discloses a dynamic multi-dimensional power data grade evaluation method and related equipment, and the method comprises the steps: constructing a power data multi-dimensional evaluation index system, and obtaining a plurality of pieces of core dimension data; according to the determined index weight of each piece of core dimension data, a quantitative scoring standard is formulated, and the index weight is dynamically adjusted according to the data life cycle and the service priority in combination with the scoring standard; performing preliminary scoring on the core dimension data by using a mixed deep learning model to obtain a preliminary scoring result; performing dynamic optimization on the preliminary scoring result, adjusting parameters of the hybrid deep learning model through a real-time feedback mechanism, and generating a final comprehensive score; scoring grades are divided, and grade evaluation of the dynamic multi-dimensional power data is carried out according to the final comprehensive scores and the divided scoring grades. According to the method, the power data multi-dimensional evaluation index system is constructed, the power data are comprehensively considered from multiple key angles, and the comprehensiveness and accuracy of power data evaluation are ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric power data management, in particular to a dynamic multi-dimensional electric power data level evaluation method and related equipment. BACKGROUND

[0002] In the current booming electric power industry, data volume is growing explosively, which has become a key factor for enterprise operation, decision-making and industry development. With the construction of smart grid, frequent electric power market transactions and accelerated digital transformation of enterprises, massive electric power data covers power generation, power transmission, power transformation, power distribution, power consumption and other links, including equipment operation parameters, electric power load data, user power consumption behavior data, market transaction data and the like. These data are widely sourced, various in type and complex in structure, including structured database data, semi-structured log files and unstructured image and text data and the like.

[0003] The existing electric power data evaluation method has many deficiencies, such as single dimension, lack of industry adaptability, non-standard process and inability to quantitatively evaluate. Some enterprises only measure data quality according to a few basic indicators such as data accuracy and timeliness, ignoring the performance of data in sharing, application and social value. For example, when the traditional electric power data management system evaluates data, it only focuses on the data value range and update, and lacks effective evaluation means for the sharing degree of data in different business fields, business process optimization effect and social contribution such as energy saving and emission reduction, which is difficult to fully reflect the real value of data and provide sufficient support for enterprise data management, decision-making and social responsibility fulfillment. In determining the importance of data, it is often based on experience or simple data statistical indicators such as data access frequency or associated business size, which lacks systematicness and scientificity and is easily disturbed by subjective factors and local data characteristics, resulting in inaccurate importance judgment and affecting the rational allocation and efficient use of data resources. For example, emerging electric power data analysis projects may be underestimated due to low initial access frequency, and traditional high-frequency access data may have limited innovation effect but occupy resource advantage. These problems make it difficult for the existing method to fully reflect the real value of data and provide sufficient support for enterprise data management, decision-making and social responsibility fulfillment. SUMMARY

[0004] The purpose of the present application is to provide a dynamic multi-dimensional electric power data level evaluation method and related equipment to solve the technical problems of single dimension, lack of industry adaptability, non-standard process and inability to quantitatively evaluate in the process of electric power industry data level evaluation.

[0005] The present application is realized by the following technical solutions:

[0006] In a first aspect, the present application provides a dynamic multi-dimensional electric power data level evaluation method, comprising:

[0007] A multi-dimensional evaluation index system for power data was constructed, and several core dimension data were obtained based on the multi-dimensional evaluation index system for power data.

[0008] Based on the determined indicator weights for each core dimension of data, quantitative scoring standards are formulated, and the indicator weights are dynamically adjusted according to the data lifecycle and business priorities in conjunction with the scoring standards.

[0009] A preliminary scoring result is obtained by using a hybrid deep learning model to perform preliminary scoring on the core dimension data after adjusting the indicator weights.

[0010] The initial scoring results are dynamically optimized based on reinforcement learning algorithms, and the parameters of the hybrid deep learning model are adjusted through a real-time feedback mechanism to generate the final comprehensive score.

[0011] The scoring levels are divided, and the dynamic multidimensional power data is evaluated based on the final comprehensive score and the divided scoring levels.

[0012] Preferably, several core data dimensions include data on the scope of sharing, usage popularity, quality factors, security protection, application effectiveness, social value, data correlation analysis, and forward-looking data assessment.

[0013] Preferably, the specific process for determining the indicator weights based on the data of each core dimension is as follows:

[0014] Determine the most important and least important indicators among the indicators;

[0015] The importance of the most important and least important indicators relative to other evaluation indicators was determined using a 1-9 scale.

[0016] The indicator weights are determined based on the relative importance of the most important and least important indicators.

[0017] Preferably, in the step of dynamically adjusting indicator weights based on data lifecycle and business priority in conjunction with the scoring criteria, the dynamic adjustment of indicator weights specifically includes:

[0018] Real-time monitoring of data usage frequency, policy changes, and business scenario changes based on Bayesian optimization algorithms;

[0019] The optimal and worst-case methods are combined with indicator weights, and the indicator weights are periodically and dynamically adjusted using a sliding window algorithm.

[0020] Preferably, in the step of using a hybrid deep learning model to perform preliminary scoring on the core dimension data after adjusting the indicator weights to obtain preliminary scoring results, the fusion layer of the hybrid deep learning model adopts an adaptive weighting strategy, as expressed below:

[0021] G1=α*GCNN +(1-α)*G MLP +ΔG RL

[0022] Where α is the dynamic fusion coefficient, ΔG RL To reinforce learning correction terms; G CNN and G MLP These are the outputs of CNN and MLP, respectively; G1 is the initial score.

[0023] Preferably, in the step of dynamically optimizing the preliminary scoring results based on reinforcement learning algorithms and adjusting the parameters of the hybrid deep learning model through a real-time feedback mechanism to generate the final comprehensive score, the specific process is as follows:

[0024] The optimized score is obtained by designing a multi-dimensional reward function using a reinforcement learning algorithm, as shown in the following expression:

[0025]

[0026] Where Q(s) t ,a t ) represents state s t Take action a t Q value, For learning rate, r is the discount factor. t+1 For instant rewards;

[0027] By dynamically adjusting the parameters of the hybrid deep learning model through a real-time feedback mechanism that combines deep learning and reinforcement learning, a final comprehensive score is generated. The formula for the final comprehensive score is as follows:

[0028]

[0029] Wherein, G2 is the final comprehensive score; △G is the score correction item dynamically adjusted by reinforcement learning.

[0030] Preferably, in the step of dividing the rating levels, a fuzzy clustering algorithm is used to dynamically calibrate the rating interval threshold by combining the company's historical data and industry benchmarks to obtain the rating level. When the final comprehensive score falls within the divided rating level interval threshold, it is the corresponding level.

[0031] Secondly, the present invention also provides a dynamic multidimensional power data rating system, comprising:

[0032] The core dimension acquisition module is used to construct a multi-dimensional evaluation index system for power data and obtain several core dimension data based on the multi-dimensional evaluation index system for power data.

[0033] The indicator weight optimization module is used to determine the indicator weight based on the data of each core dimension, formulate quantitative scoring standards, and dynamically adjust the indicator weights according to the data lifecycle and business priorities in combination with the scoring standards.

[0034] The preliminary scoring model output module is used to perform preliminary scoring on the core dimension data after adjusting the indicator weights using a hybrid deep learning model to obtain preliminary scoring results.

[0035] The final comprehensive scoring model output module is used to dynamically optimize the preliminary scoring results based on the reinforcement learning algorithm, adjust the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generate the final comprehensive score.

[0036] The rating level determination module is used to divide the rating levels and to conduct dynamic multidimensional power data rating evaluation based on the final comprehensive score and the divided rating levels.

[0037] Thirdly, the present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the dynamic multidimensional power data rating method described above.

[0038] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the dynamic multidimensional power data grading method described above.

[0039] Compared with the prior art, the present invention has the following beneficial technical effects:

[0040] This invention provides a dynamic multidimensional power data grading method. By constructing a multidimensional evaluation index system for power data, it comprehensively considers power data from multiple key perspectives, avoiding the one-sidedness of single-dimensional evaluation and ensuring the comprehensiveness and accuracy of power data evaluation. The method determines the index weights based on each core dimension of data and formulates quantitative scoring standards. The index weights are dynamically adjusted in conjunction with the data lifecycle and business priorities, enabling the evaluation method to adapt to power data evaluation at different stages and under different business needs. A hybrid deep learning model is used to initially score the core dimension data after adjusting the index weights. Deep learning models have powerful feature extraction and learning capabilities, enabling them to uncover complex patterns and relationships in the data, improving the accuracy of the initial scoring. Scoring levels are divided, and the final comprehensive score is combined with the divided scoring levels to conduct dynamic multidimensional power data grading, making the evaluation results more intuitive, easy to understand, and convenient for practical application. The division of different levels can provide a clear basis for the classification management, resource allocation, and decision-making of power data, improving the efficiency and effectiveness of power data management.

[0041] Furthermore, based on the six dimensions of sharing scope, usage popularity, quality factors, security protection, application effectiveness, and social value, two innovative dimensions are introduced: data correlation analysis and data forward-looking assessment. Data correlation analysis uses graph neural network technology to assess the complex relationships between data and uncover potential business value; data forward-looking assessment uses machine learning models to predict the potential impact of data on future business, providing a basis for power companies to plan ahead.

[0042] Furthermore, a weight adjustment mechanism based on dynamic changes in data lifecycle and business priorities is introduced. By monitoring factors such as data usage frequency, update cycle, and changes in industry policies in real time, the weights of each indicator are dynamically adjusted to ensure that the evaluation methodology remains consistent with the latest developments in the power industry.

[0043] Furthermore, this invention introduces a deep learning model to automatically optimize data evaluation metrics. Through training on massive amounts of historical data, the model can automatically learn the complex mapping relationship between data features and scores, thereby more accurately assessing the overall value of the data. In addition, the model possesses self-learning capabilities, continuously optimizing the scoring logic based on new data samples, improving the accuracy and reliability of the evaluation.

[0044] Furthermore, this invention proposes an adaptive data level adjustment mechanism based on corporate strategic goals and business needs. By analyzing the company's core business processes, strategic planning, and market competition, the data level classification standards are dynamically adjusted to make the data level evaluation results more aligned with the company's actual needs, providing more targeted support for corporate decision-making. Attached Figure Description

[0045] Figure 1 This is a flowchart of the dynamic multidimensional power data level evaluation method in an embodiment of the present invention;

[0046] Figure 2 This is a flowchart illustrating the method for evaluating the level of dynamic multidimensional power data in an embodiment of the present invention.

[0047] Figure 3 This is a schematic diagram illustrating the construction of a multi-dimensional evaluation index system for power data in an embodiment of the present invention;

[0048] Figure 4 This is a schematic diagram of a dynamic multidimensional power data rating system in an embodiment of the present invention;

[0049] In the diagram: 1. Core dimension acquisition module; 2. Indicator weight optimization module; 3. Preliminary scoring model output module; 4. Final comprehensive scoring model output module; 5. Scoring level determination module. Detailed Implementation

[0050] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0051] The purpose of this invention is to provide a method and related equipment for evaluating the level of dynamic multidimensional power data, so as to solve the technical problems of single dimension, lack of industry adaptability, non-standard process and inability to quantitatively evaluate the data level evaluation process in the power industry.

[0052] The present invention will now be described in further detail with reference to the accompanying drawings:

[0053] Example 1

[0054] See Figure 1 and Figure 2 In one embodiment of the present invention, a method for evaluating the level of dynamic multidimensional power data is provided, comprising:

[0055] Step 1: Construct a multi-dimensional evaluation index system for power data, and obtain several core dimension data based on the multi-dimensional evaluation index system for power data;

[0056] Specifically, the core data dimensions include data on the scope of sharing, usage popularity, quality factors, security protection, application effectiveness, social value, data correlation analysis, and forward-looking data assessment.

[0057] Among them, the multi-dimensional evaluation index system for power data, constructed from shared scope data, usage popularity data, quality factor data, safety protection data, application effectiveness data, and social value data, is as follows: Figure 3 As shown in Tables 1-6, data evaluation indicators are quantitatively designed based on the characteristics, uses, and circulation methods (open data, shared data, and transactional data) of data assets.

[0058] Table 1 Data Sharing Scope Scoring Criteria

[0059]

[0060]

[0061]

[0062] Table 2 uses popularity rating criteria for data.

[0063]

[0064]

[0065]

[0066]

[0067] Table 3 Data Quality Factors Scoring Criteria

[0068]

[0069]

[0070]

[0071] Table 4 Data Security Protection Scoring Criteria

[0072]

[0073]

[0074]

[0075]

[0076] Table 5. Scoring Criteria for Data Application Effectiveness

[0077]

[0078]

[0079]

[0080]

[0081] Table 6. Social Value Scoring Criteria for Data

[0082]

[0083]

[0084] Step 2: Based on the determined indicator weights for each core dimension of data, formulate quantitative scoring standards, and dynamically adjust the indicator weights according to the data lifecycle and business priorities in conjunction with the scoring standards.

[0085] Specifically, the process for determining the weight of indicators based on the data of each core dimension is as follows:

[0086] The process by which the i-th expert determines the indicator weights is as follows: The i-th expert first determines the most important indicator among the m indicators mentioned above. and the least important indicators Then, the most important indicator was determined using a 1-9 scale. The higher the value, the more important the most important indicator is relative to the other indicators; similarly, a 1-9 scale is used to determine the importance of other indicators relative to the least important indicator. The higher the value, the more important the other evaluation indicators are relative to the least important indicator.

[0087] Specifically, dynamically adjusting indicator weights includes:

[0088] Real-time monitoring of data usage frequency, policy changes, and business scenario changes based on Bayesian optimization algorithms;

[0089] The optimal and worst-case methods are combined with indicator weights, and the indicator weights are periodically and dynamically adjusted using a sliding window algorithm.

[0090] In order to determine the optimal weight of the j-th indicator based on the i-th expert, and We should minimize the weight of each indicator and construct the following model using the best-worst method to determine the optimal weight of each indicator. The specific model is as follows:

[0091] minσ

[0092]

[0093] j = 1, 2, m, i = 1, 2, n

[0094] Based on the weight a of the i-th expert i The expression for calculating the optimal weight of the j-th indicator is as follows:

[0095]

[0096] Step 3: Use a hybrid deep learning model to perform preliminary scoring on the core dimension data after adjusting the indicator weights to obtain preliminary scoring results;

[0097] Specifically, based on determining the optimal weight of each indicator, each indicator is scored according to the specific scoring criteria in each evaluation indicator, and the overall score obtained by adding up the scores of each specific indicator is used as the preliminary comprehensive score result.

[0098] The specific formula for calculating the scoring results is as follows:

[0099] G=(Ia1b1xGa1b1+......+Ia1bnxGa1bn)xA1+(Ia2b1xGa2b1+......+Ia2bnxGa2b1)xA2+(Ia2b1xGa3b1+......+Ia2bnxGa3bn)xA 3+(Ia4b1xGa4b1+......+Ia4bnxGa4bn)xA4+(Ia5b1xGa5b1+......+Ia5bnxGa5b1)xA5+(Ia6b1xGa6b1+......+Ia6bnxGa6bn)xA6

[0100] Where: G represents the final score result;

[0101] A1, A2, A3, A4, A5, and A6 represent the weights corresponding to the six primary indicators: scope of sharing, usage popularity, quality factors, security protection, application effectiveness, and social value, respectively.

[0102] Ia1b1, Ia1b2, ..., Ia1bn represent the weights of the secondary indicators within the application scope, and Ga1b1, Ga1b2, ..., Ga1bn represent the scores of the secondary indicators within the shared scope.

[0103] Ia2b1, Ia2b2, ..., Ia2bn represent the weights of the secondary indicators within the call frequency, and Ga2b1, Ga2b2, ..., Ga2b1 represent the scores of the secondary indicators within the usage popularity.

[0104] Ia3b1, Ia3b2, ..., Ia3bn represent the weights of the secondary indicators within the application effectiveness, and Ga3b1, Ga3b2, ..., Ga3bn represent the scores of the secondary indicators within the quality factors.

[0105] Ia4b1, Ia4b2, ..., Ia4bn represent the weights of the secondary indicators within the application scope, and Ga4b1, Ga4b2, ..., Ga4bn represent the scores of the secondary indicators within the security protection scope.

[0106] Ia5b1, Ia5b2, ..., Ia5bn represent the weights of the secondary indicators within the call frequency, and Ga5b1, Ga5b2, ..., Ga5b1 represent the scores of the secondary indicators within the application effectiveness.

[0107] Ia6b1, Ia6b2, ..., Ia6bn represent the weights of the secondary indicators within the application effectiveness, while Ga6b1, Ga6b2, ..., Ga6bn represent the scores of the secondary indicators within the social value.

[0108] This embodiment employs a hybrid model combining a multilayer perceptron (MLP) and a convolutional neural network (CNN) to process the complex features and multidimensional information of power data. The MLP is used to process structured data (such as indicator scores), while the CNN is used to extract spatial correlations between data (such as data correlation analysis).

[0109] Before scoring, the data is standardized by normalizing all indicator values ​​to the range [1, 10] to eliminate the influence of different indicator units. The normalization formula is as follows:

[0110]

[0111] Where x is the original index value, and min(x) and max(x) are the minimum and maximum values ​​of the index, respectively.

[0112] The hybrid deep learning model architecture in this embodiment includes the following:

[0113] Input layer: The input layer receives normalized indicator data, including secondary indicators in dimensions such as sharing scope, usage popularity, quality factors, security protection, application effectiveness, social value, data correlation analysis, and data forward-looking assessment.

[0114] Convolutional layers: The CNN part is used to extract spatial relationships between data points. The kernel size is dynamically adjusted according to the data dimension. The convolution operation formula is as follows:

[0115]

[0116] in, For convolution output, W k,l For the convolution kernel weights, I i+k,j+l Here is the input data, and b is the bias term. For example, ReLU is an activation function.

[0117] The MLP part processes structured data, using a multi-layer fully connected network to perform weighted summation and non-linear transformations on the metrics. The output formula of the fully connected layer is as follows:

[0118]

[0119] Where y is the output, W is the weight matrix, x is the input vector, and b is the bias vector.

[0120] The outputs of CNN and MLP are fused to obtain a preliminary score. The fusion formula is as follows:

[0121]

[0122] Among them, G1 is the preliminary score result, G CMN and GMLP These are the outputs of CNN and MLP, respectively, where α is the fusion weight, dynamically adjusted based on data features; ΔG RL To reinforce learning corrections.

[0123] Step 4: Dynamically optimize the preliminary scoring results based on reinforcement learning algorithms, adjust the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generate the final comprehensive score;

[0124] Specifically, a reinforcement learning algorithm is introduced to dynamically adjust the parameters of the scoring model to adapt to the dynamic changes in electricity data. The goal of the reinforcement learning algorithm is to maximize the long-term cumulative reward, and the update formula is as follows:

[0125]

[0126] Where Q(s) t ,a t ) represents state s t Take action a t Q value, For learning rate, r is the discount factor. t+1 For instant rewards.

[0127] Wherein, state s t : The normalized values ​​of current data indicators and historical scoring results.

[0128] Action a t : For scoring model parameters (such as weights) Convolution kernel weights W k,l Adjustments (etc.).

[0129] Reward r t+1 Based on the degree of matching between the scoring results and the actual data value, the reward function is designed as follows:

[0130] r t+1 =-|G predicterd -G actual |

[0131] Among them, G predicterd G is the score predicted by the model. actual For actual data value.

[0132] Combining deep learning models and reinforcement learning dynamic adjustment mechanisms, the final scoring formula is as follows:

[0133]

[0134] Wherein, △G is a score correction term dynamically adjusted by reinforcement learning, which is dynamically updated according to the Q-learning algorithm.

[0135] Step 5: Divide the scoring levels and conduct a dynamic multidimensional power data level evaluation based on the final comprehensive score and the divided scoring levels.

[0136] Specifically, in the process of dividing the rating levels, a fuzzy clustering algorithm is used to dynamically calibrate the rating interval threshold by combining the company's historical data and industry benchmarks to obtain the rating level. When the final comprehensive score falls within the threshold of the divided rating level interval, it is the corresponding level.

[0137] Among them, data is divided into four levels based on dimensions such as influence, stability, commercial value, and guiding weight for decision-making: star data, important data, potential data, and marginal data, as shown in Table 7.

[0138] Table 7 shows the data levels.

[0139] Data Tier Score Interval Star Data (85,100] Important Data (60-85] Potential Data (10,60] Edge Data (0,10]

[0140] In summary, this invention provides a dynamic multidimensional power data grading method. By constructing a multidimensional data evaluation index system, dynamically adjusting index weights, introducing artificial intelligence to optimize the scoring mechanism, and adaptively adjusting data grades, this method accurately selects high-value data, enhances the power company's data-driven decision-making capabilities and data asset value creation capabilities, and ensures the safe and stable operation of the power system and the sustainable development of the enterprise.

[0141] Example 2

[0142] according to Figure 4 As shown, the present invention also provides a dynamic multidimensional power data rating system, comprising:

[0143] Core Dimension Acquisition Module 1 is used to construct a multi-dimensional evaluation index system for power data and obtain several core dimension data based on the multi-dimensional evaluation index system for power data.

[0144] The indicator weight optimization module 2 is used to determine the indicator weight based on the data of each core dimension, formulate quantitative scoring standards, and dynamically adjust the indicator weights according to the data lifecycle and business priority in combination with the scoring standards.

[0145] Preliminary scoring model output module 3 is used to perform preliminary scoring on the core dimension data after adjusting the indicator weights using a hybrid deep learning model to obtain preliminary scoring results.

[0146] The final comprehensive scoring model output module 4 is used to dynamically optimize the preliminary scoring results based on the reinforcement learning algorithm, adjust the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generate the final comprehensive score.

[0147] The rating level determination module 5 is used to divide the rating levels and to conduct dynamic multidimensional power data rating evaluation based on the final comprehensive score and the divided rating levels.

[0148] Example 3

[0149] The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, such as a dynamic multidimensional power data rating program.

[0150] When the processor executes the computer program, it implements the steps of the above-mentioned dynamic multidimensional power data rating method, for example:

[0151] A multi-dimensional evaluation index system for power data was constructed, and several core dimension data were obtained based on the multi-dimensional evaluation index system for power data.

[0152] Based on the determined indicator weights for each core dimension of data, quantitative scoring standards are formulated, and the indicator weights are dynamically adjusted according to the data lifecycle and business priorities in conjunction with the scoring standards.

[0153] A preliminary scoring result is obtained by using a hybrid deep learning model to perform preliminary scoring on the core dimension data after adjusting the indicator weights.

[0154] The initial scoring results are dynamically optimized based on reinforcement learning algorithms, and the parameters of the hybrid deep learning model are adjusted through a real-time feedback mechanism to generate the final comprehensive score.

[0155] The scoring levels are divided, and the dynamic multidimensional power data is evaluated based on the final comprehensive score and the divided scoring levels.

[0156] Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system, for example:

[0157] Core Dimension Acquisition Module 1 is used to construct a multi-dimensional evaluation index system for power data and obtain several core dimension data based on the multi-dimensional evaluation index system for power data.

[0158] The indicator weight optimization module 2 is used to determine the indicator weight based on the data of each core dimension, formulate quantitative scoring standards, and dynamically adjust the indicator weights according to the data lifecycle and business priority in combination with the scoring standards.

[0159] Preliminary scoring model output module 3 is used to perform preliminary scoring on the core dimension data after adjusting the indicator weights using a hybrid deep learning model to obtain preliminary scoring results.

[0160] The final comprehensive scoring model output module 4 is used to dynamically optimize the preliminary scoring results based on the reinforcement learning algorithm, adjust the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generate the final comprehensive score.

[0161] The rating level determination module 5 is used to divide the rating levels and to conduct dynamic multidimensional power data rating evaluation based on the final comprehensive score and the divided rating levels.

[0162] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the mobile terminal.

[0163] For example, the computer program can be divided into a core dimension acquisition module 1, an indicator weight optimization module 2, a preliminary scoring model output module 3, a final comprehensive scoring model output module 4, and a scoring level determination module 5;

[0164] The specific functions of each module are as follows:

[0165] Core Dimension Acquisition Module 1 is used to construct a multi-dimensional evaluation index system for power data and obtain several core dimension data based on the multi-dimensional evaluation index system for power data.

[0166] The indicator weight optimization module 2 is used to determine the indicator weight based on the data of each core dimension, formulate quantitative scoring standards, and dynamically adjust the indicator weights according to the data lifecycle and business priority in combination with the scoring standards.

[0167] Preliminary scoring model output module 3 is used to perform preliminary scoring on the core dimension data after adjusting the indicator weights using a hybrid deep learning model to obtain preliminary scoring results.

[0168] The final comprehensive scoring model output module 4 is used to dynamically optimize the preliminary scoring results based on the reinforcement learning algorithm, adjust the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generate the final comprehensive score.

[0169] The rating level determination module 5 is used to divide the rating levels and to conduct dynamic multidimensional power data rating evaluation based on the final comprehensive score and the divided rating levels.

[0170] The mobile terminal can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The mobile terminal may include, but is not limited to, a processor and memory.

[0171] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the mobile terminal, connecting various parts of the mobile terminal via various interfaces and lines.

[0172] The memory can be used to store the computer program and / or module. The processor implements various functions of the mobile terminal by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0173] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function (such as sound playback or image playback). The data storage area may store data created based on the use of the phone (such as audio data or a phonebook). Furthermore, the memory may include high-speed random access memory (RAM) and non-volatile memory, such as hard disks, RAM, plug-in hard disks, SmartMediaCards (SMC), Secure Digital (SD) cards, flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0174] Example 4

[0175] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the dynamic multidimensional power data grading evaluation method.

[0176] If the modules / units integrated in the mobile terminal are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0177] Based on this understanding, all or part of the processes in the above method can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of the above-described aggregated reinforcement learning resource scheduling method. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form.

[0178] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0179] It should be noted that the content contained in the computer-readable medium may be appropriately added to or subtracted from the content as required by the legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium may not include electrical carrier signals and telecommunication signals.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for evaluating the grading of dynamic multidimensional power data, characterized in that, include: A multi-dimensional evaluation index system for power data was constructed, and several core dimension data were obtained based on the multi-dimensional evaluation index system for power data. Based on the determined indicator weights for each core dimension of data, quantitative scoring standards are formulated, and the indicator weights are dynamically adjusted according to the data lifecycle and business priorities in conjunction with the scoring standards. A preliminary scoring result is obtained by using a hybrid deep learning model to perform preliminary scoring on the core dimension data after adjusting the indicator weights. The initial scoring results are dynamically optimized based on reinforcement learning algorithms, and the parameters of the hybrid deep learning model are adjusted through a real-time feedback mechanism to generate the final comprehensive score. The scoring levels are divided, and the dynamic multidimensional power data is evaluated based on the final comprehensive score and the divided scoring levels.

2. The method for evaluating the level of dynamic multidimensional power data according to claim 1, characterized in that, The core data dimensions include data on the scope of sharing, usage popularity, quality factors, security protection, application effectiveness, social value, data correlation analysis, and forward-looking data assessment.

3. The method for evaluating the level of dynamic multidimensional power data according to claim 1, characterized in that, The specific process for determining the indicator weights based on the data of each core dimension is as follows: Determine the most important and least important indicators among the indicators; The importance of the most important and least important indicators relative to other evaluation indicators was determined using a 1-9 scale. The indicator weights are determined based on the relative importance of the most important and least important indicators.

4. The method for evaluating the level of dynamic multidimensional power data according to claim 1, characterized in that, The step of dynamically adjusting indicator weights based on data lifecycle and business priority according to the scoring criteria specifically includes: Real-time monitoring of data usage frequency, policy changes, and business scenario changes based on Bayesian optimization algorithms; The optimal and worst-case methods are combined with indicator weights, and the indicator weights are periodically and dynamically adjusted using a sliding window algorithm.

5. The method for evaluating the level of dynamic multidimensional power data according to claim 1, characterized in that, In the step of using a hybrid deep learning model to perform preliminary scoring on the core dimension data after adjusting the indicator weights to obtain preliminary scoring results, the fusion layer of the hybrid deep learning model adopts an adaptive weighting strategy, as expressed below: G1=α*G CNN +(1-a)*G MLP +ΔG RL Where α is the dynamic fusion coefficient, ΔG RL To reinforce learning correction terms; G CNN and G MLP These are the outputs of CNN and MLP, respectively; G1 is the initial score.

6. The method for evaluating the level of dynamic multidimensional power data according to claim 1, characterized in that, The specific process of dynamically optimizing the preliminary scoring results based on the reinforcement learning algorithm, adjusting the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generating the final comprehensive score is as follows: The optimized score is obtained by designing a multi-dimensional reward function using a reinforcement learning algorithm, as shown in the following expression: Where Q(s) t ,a t ) represents state s t Take action a t Q value, For learning rate, r is the discount factor. t+1 For instant rewards; By dynamically adjusting the parameters of the hybrid deep learning model through a real-time feedback mechanism that combines deep learning and reinforcement learning, a final comprehensive score is generated. The formula for the final comprehensive score is as follows: Wherein, G2 is the final comprehensive score; △G is the score correction item dynamically adjusted by reinforcement learning.

7. The method for evaluating the level of dynamic multidimensional power data according to claim 1, characterized in that, In the step of dividing the rating levels, a fuzzy clustering algorithm is used to dynamically calibrate the rating interval threshold by combining the company's historical data and industry benchmarks to obtain the rating level. When the final comprehensive score falls within the divided rating level interval threshold, it is the corresponding level.

8. A dynamic multidimensional power data rating system, characterized in that, include: The core dimension acquisition module is used to construct a multi-dimensional evaluation index system for power data and obtain several core dimension data based on the multi-dimensional evaluation index system for power data. The indicator weight optimization module is used to determine the indicator weight based on the data of each core dimension, formulate quantitative scoring standards, and dynamically adjust the indicator weights according to the data lifecycle and business priorities in combination with the scoring standards. The preliminary scoring model output module is used to perform preliminary scoring on the core dimension data after adjusting the indicator weights using a hybrid deep learning model to obtain preliminary scoring results. The final comprehensive scoring model output module is used to dynamically optimize the preliminary scoring results based on the reinforcement learning algorithm, adjust the parameters of the hybrid deep learning model through a real-time feedback mechanism, and generate the final comprehensive score. The rating level determination module is used to divide the rating levels and to conduct dynamic multidimensional power data rating evaluation based on the final comprehensive score and the divided rating levels.

9. A mobile terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the dynamic multidimensional power data grading method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the dynamic multidimensional power data grading method as described in any one of claims 1-7.

Images