CN-121996998-A - Method and device for evaluating using efficiency of equipment and electronic equipment

CN121996998ACN 121996998 ACN121996998 ACN 121996998ACN-121996998-A

Abstract

The application provides a method, a device and electronic equipment for evaluating using efficiency of equipment, and the method comprises the steps of collecting an evaluation data set corresponding to the equipment to be evaluated, preprocessing the evaluation data set to obtain a target evaluation data set, inputting the target evaluation data set into an efficiency evaluation model comprising an input layer, a feature extraction layer and a regression prediction layer three-level architecture, and outputting an evaluation index, wherein the efficiency evaluation model is a model obtained by training an initial evaluation model by adopting training samples in advance, and determining the using efficiency evaluation result of the equipment to be evaluated according to the evaluation index. By adopting the method for evaluating the using efficiency of the equipment, the complex nonlinear problem can be well processed, and the method can be suitable for complex data distribution.

Inventors

DONG WENTAO
CHEN TONG
Luan Xinrui
LIU YING
YIN FEI
ZHANG YIMENG
DONG YUCAI
ZHANG XIAOWEI
Kong Zining
XIAO LONGBIN
WANG SHENGXU
WANG QIANG
CUI WEI
LIN YUANYUAN
ZHANG SHITAI

Assignees

中国电子科技集团公司第十五研究所

Dates

Publication Date: 20260508
Application Date: 20260112

Claims (10)

1. A method for evaluating the performance of a device, comprising: collecting an evaluation data set corresponding to equipment to be evaluated, and preprocessing the evaluation data set to obtain a target evaluation data set; Inputting the target evaluation data set into a performance evaluation model comprising an input layer, a feature extraction layer and a regression prediction layer three-level architecture, and outputting evaluation indexes, wherein the performance evaluation model is a model obtained by training an initial evaluation model by adopting a training sample in advance; and determining a use efficiency evaluation result of the equipment to be evaluated according to the evaluation index.
2. The method of claim 1, wherein inputting the target evaluation dataset into a performance evaluation model comprising an input layer, a feature extraction layer, and a regression prediction layer three-level architecture, outputting evaluation data, comprises: The input layer is used for receiving the target evaluation data set, splicing at least one piece of structured data contained in the target evaluation data set to obtain structured spliced data, and splicing at least one piece of unstructured data contained in the target evaluation data set to obtain unstructured spliced data; the feature extraction layer respectively adopts different feature extraction sub-models to perform feature extraction on the structured spliced data and the unstructured spliced data to obtain feature extraction data, and the feature extraction data is used as input of a regression prediction layer; and the regression prediction layer carries out regression prediction according to the feature extraction data and outputs an evaluation index.
3. The method of claim 2, wherein the feature extraction layer performs feature extraction on the structured stitching data and the unstructured stitching data using different feature extraction sub-models, respectively, to obtain feature extraction data, including: The first sub-model in the feature extraction layer performs feature extraction on the structured spliced data to obtain first feature data; the second sub-model in the feature extraction layer performs feature extraction on unstructured spliced data to obtain second feature data; And the fusion sub-model in the feature extraction layer performs feature fusion on the first feature data and the second feature data to obtain the feature extraction data.
4. A method according to claim 3, wherein the feature extraction of the structured stitching data by the first sub-model in the feature extraction layer results in first feature data, comprising: Performing feature analysis on the structured spliced data to obtain an analysis data set, wherein the analysis data set comprises at least one data of semantic data, structural data and mode data; Selectively activating a plurality of feature extraction layers according to the analysis data set to respectively perform feature extraction on each part of data in the structured spliced data to obtain a plurality of first reference feature data, wherein each feature extraction layer correspondingly extracts the features of at least one part of data in the structured spliced data; The feature extraction layers respectively adopt a learnable activation function to carry out dimension mapping on the corresponding first reference feature data to obtain a plurality of second reference feature data; And carrying out feature fusion on the extracted multiple sub-feature data according to the weight corresponding to each feature extraction layer to obtain the first feature data.
5. A method according to claim 3, wherein the feature extraction of the unstructured spliced data by the second sub-model in the feature extraction layer to obtain second feature data comprises: performing first dimension reduction operation on the unstructured data by adopting a preset dimension reduction mode to obtain a dimension reduction characteristic data set; performing feature fusion on the dimension reduction feature data set to obtain first dimension reduction feature data; performing feature extraction operation on the first dimension reduction feature data to obtain third reference feature data; And adopting a hidden layer in the second sub-model to perform second dimension reduction operation according to the evaluation requirement to obtain the second characteristic data.
6. The method of claim 1, wherein preprocessing the evaluation data set to obtain a target evaluation data set comprises: preprocessing the evaluation data set to obtain a first reference evaluation data set, wherein the preprocessing comprises missing value processing, outlier processing and data normalization; performing first classification processing on the data in the first reference evaluation data set to obtain a plurality of second reference evaluation data sets; Respectively carrying out preset processing on the data in the plurality of second reference evaluation data sets according to a preset data processing mode corresponding to each second reference evaluation data set to obtain a plurality of third reference evaluation data sets, wherein the preset data processing mode corresponds to a data category corresponding to the second reference evaluation data set; Performing a second classification process on a plurality of the third reference evaluation data sets to obtain the target evaluation data set comprising a structured data subset and an unstructured data subset.
7. The method of claim 2, wherein training the initial assessment model with training samples comprises: Initializing model weight, activation function parameters and super parameters corresponding to an initial regression prediction layer in the initial evaluation model; And performing iterative training on the initial evaluation model by adopting an automatic differential framework according to the training sample, wherein the iterative training comprises the steps of calculating a global kernel matrix and a noise variance corresponding to the training sample through a forward propagation mode, calculating NLML loss according to the global kernel matrix and the noise variance, and performing gradient calculation and parameter updating through a backward propagation mode, and determining the evaluation model meeting the iteration stop condition as the efficiency evaluation model, wherein the parameter updating comprises updating the model weight, the activation function parameter and the super parameter.
8. The method of claim 1, wherein the evaluation metrics include a numerical metric and a visual metric, the visual metrics including a 3D thermodynamic diagram for indicating the impact of the set of evaluation data on the numerical metric and a dynamic simulation diagram for simulating the motion trajectories and the destructive effects of the device under evaluation.
9. A device for evaluating the performance of use of an apparatus, comprising: the acquisition processing unit is used for acquiring an evaluation data set corresponding to the equipment to be evaluated, and preprocessing the evaluation data set to obtain a target evaluation data set; The input unit is used for inputting the target evaluation data set into a performance evaluation model comprising an input layer, a feature extraction layer and a regression prediction layer three-level architecture and outputting evaluation indexes, wherein the performance evaluation model is a model obtained by training an initial evaluation model by adopting a training sample in advance; And the determining unit is used for determining a use efficiency evaluation result of the equipment to be evaluated according to the evaluation index.
10. An electronic device comprising at least one processor and a memory communicatively coupled to the at least one processor, wherein the memory stores a computer program executable by the at least one processor to cause the at least one processor to perform the method of evaluating usage performance of the device of any of claims 1-8.

Description

Method and device for evaluating using efficiency of equipment and electronic equipment Technical Field The present application relates to the field of device evaluation technologies, and in particular, to a method and an apparatus for evaluating use performance of a device, and an electronic device. Background The use efficiency of the equipment refers to the comprehensive representation of the capacity and effect of the equipment for completing specified operation tasks and achieving expected operation targets under specific environmental conditions, and covers the whole process from the operation of the equipment to the completion of the operation. The use efficiency is the combination of the operation capability and the operation effect, the capability refers to the physical characteristics and the operation performance of the equipment, and the characteristics determine the possible action range and degree of the equipment in operation, and the efficiency is the actual effect of the equipment in operation. Thus, the development and optimization of equipment usage performance assessment has a profound impact. The device has a relatively complex structure and numerous influencing factors for determining the using efficiency, and when the device is used for evaluating the efficiency, the traditional evaluation methods such as an ADC evaluation method, an analytic hierarchy process and a support vector product are adopted, but the ADC evaluation method can reduce the expression constraint on each index, the analytic hierarchy process can be difficult to calculate due to large judgment matrix order, and the support vector product is sensitive to real data, so that a method capable of processing complex nonlinear problems is needed, the efficiency simulation analysis method based on the neural network better solves the problems of the traditional evaluation methods, but the neural network has the defects of long optimizing time, easiness in sinking into local extremum, dependence on initial weight and threshold value, incapability of adapting to complex data distribution and the like, and the proper initial weight and threshold value are needed to be set so as to generate better evaluation effect. Therefore, in order to improve the use efficiency of the equipment, ensuring the scientificity of use and application and promoting the infiltration research of the use theory, further optimizing the evaluation method is particularly important. At least one of the problems of the existing neural network-based performance simulation analysis method is solved, and no effective solution is proposed at present. Disclosure of Invention The embodiment of the application provides a method and a device for evaluating using efficiency of equipment and electronic equipment, which are used for at least solving at least one of the problems existing in the traditional neural network-based efficiency simulation analysis method. According to one aspect of the embodiment of the application, a method for evaluating the use efficiency of equipment is provided, which comprises the steps of collecting an evaluation data set corresponding to equipment to be evaluated, preprocessing the evaluation data set to obtain a target evaluation data set, inputting the target evaluation data set into an efficiency evaluation model comprising an input layer, a feature extraction layer and a regression prediction layer three-level architecture, outputting evaluation indexes, wherein the efficiency evaluation model is a model obtained by training an initial evaluation model by adopting training samples in advance, and determining the use efficiency evaluation result of the equipment to be evaluated according to the evaluation indexes. According to another aspect of the embodiment of the application, the device for evaluating the use efficiency of the equipment comprises an acquisition processing unit, an input unit and a determining unit, wherein the acquisition processing unit is used for acquiring an evaluation data set corresponding to the equipment to be evaluated and preprocessing the evaluation data set to obtain a target evaluation data set, the input unit is used for inputting the target evaluation data set into an efficiency evaluation model comprising an input layer, a feature extraction layer and a regression prediction layer three-level framework, outputting evaluation indexes, the efficiency evaluation model is a model obtained by training an initial evaluation model by adopting training samples in advance, and the determining unit is used for determining the use efficiency evaluation result of the equipment to be evaluated according to the evaluation indexes. According to still another aspect of the embodiments of the present application, there is provided a computer-readable storage medium storing computer instructions for causing a computer to execute the usage effectiveness evaluation method of the above apparatus. According to yet