System performance evaluation method and device based on big data

Abstract

The invention discloses a system performance evaluation method and device based on big data. The method comprises the steps that an equipment index information set and an equipment system historical data set are acquired; the equipment system historical data set comprises an actual test data set; the actual test data set comprises a technical index test data sequence of each type of equipment technical indexes; preprocessing the historical data set of the equipment system to obtain a preprocessed data set; constructing to obtain an equipment system efficiency evaluation model; assigning a value to an equipment system efficiency evaluation model by using the preprocessed data set; and processing the acquired performance index data set of the equipment system by using the assigned performance evaluation model of the equipment system to obtain an evaluation result value of the performance of the equipment system. The equipment system performance evaluation model constructed by the method is a complex reasoning network, and the accuracy of performance evaluation is improved.

Description

Technical Field

[0001] The present invention relates to the fields of industrial data processing and equipment system evaluation, and particularly relates to a system effectiveness evaluation method and device based on big data. Background Art

[0002] In the field of modern equipment system management and evaluation, with the continuous increase in the complexity of equipment systems and the improvement of informatization level, traditional equipment performance evaluation methods are facing many challenges. Most of the existing evaluation methods rely on single test data or expert experience, lacking the ability to deeply mine and comprehensively analyze a large amount of historical data, resulting in insufficient accuracy and reliability of evaluation results. In addition, when dealing with complex equipment systems, traditional methods often have difficulty in effectively integrating multi-dimensional equipment technical index information and cannot comprehensively reflect the overall performance of the equipment system. Due to the lack of quantitative analysis of the credibility of evaluation processes and methods, existing evaluation results often rely on expert experience and lack an objective verification mechanism, resulting in low credibility and repeatability of evaluation results.

[0003] There are many and complex types of equipment system evaluation data. How to accurately evaluate the effectiveness of the equipment system based on big data of equipment system evaluation is an urgent problem to be solved currently. Summary of the Invention

[0004] The present invention mainly solves the problem of how to accurately evaluate the effectiveness of the equipment system based on big data of equipment system evaluation, and discloses a system effectiveness evaluation method and device based on big data.

[0005] In the first aspect of the embodiments of the present invention, a system effectiveness evaluation method based on big data is disclosed, including:

[0006] S1, collecting an equipment index information set and an equipment system historical data set; the equipment system historical data set includes an actual test data set; the equipment index information set includes equipment technical index information of each type; the actual test data set includes a technical index test data sequence of the equipment technical index of each type;

[0007] S2, preprocessing the equipment system historical data set to obtain a preprocessed data set;

[0008] S3, constructing an equipment system effectiveness evaluation model;

[0009] S4, using the preprocessed data set to assign values to the equipment system effectiveness evaluation model;

[0010] S5, using the equipment system effectiveness evaluation model after value assignment to process the collected equipment system performance index data set to obtain an evaluation result value of the equipment system performance.

[0011] Preprocessing the historical dataset of the equipment system to obtain a preprocessed dataset, including:

[0012] S21. Performing format conversion processing on the historical dataset of the equipment system to obtain a first dataset;

[0013] S22. Performing data cleaning processing on the first dataset to obtain a second dataset;

[0014] S23. Performing data pattern checking processing on the second dataset to obtain a preprocessed dataset.

[0015] Performing data pattern checking processing on the second dataset to obtain a preprocessed dataset, including:

[0016] For the data of each type of data attribute in the second dataset, using the data acquisition information of the data as the known independent variable and the data value of the data as the known dependent variable, constructing an approximation curve to be approximated;

[0017] Using the polynomial function fitting method to perform curve fitting on the approximation curve to be approximated to obtain the best uniform approximation polynomial of the type of data attribute;

[0018] Extracting the coefficient vectors of the best uniform approximation polynomials of all types of data attributes, performing cross-correlation calculation processing on all the coefficient vectors to obtain a cross-correlation matrix;

[0019] Performing eigenvalue calculation processing on the cross-correlation matrix to obtain an eigenvalue vector; the eigenvalue vector is a vector constructed by all eigenvalues;

[0020] Constructing a fusion polynomial with the eigenvalue vector as the polynomial order; the total order of the fusion polynomial is the number of elements in the eigenvalue vector minus 1;

[0021] Using the fusion polynomial to perform calculation processing on the known independent variable to obtain an approximate dependent variable;

[0022] Determining whether the absolute value of the difference between the approximate dependent variable and the corresponding known dependent variable is greater than a set first regression discrimination threshold; if it is greater than the first regression discrimination threshold, deleting the data from the second dataset, and if it is less than or equal to the first regression discrimination threshold, not processing the data;

[0023] Performing fusion processing on all the data after performing data pattern checking processing on the second dataset to obtain a preprocessed dataset.

[0024] The equipment system effectiveness evaluation model includes: one root node, five intermediate nodes, and ten indicator nodes; the root node is connected to each intermediate node by a hierarchical edge; the intermediate node is connected to each indicator node by a weight edge; each weight edge and hierarchical edge has a corresponding weight value;

[0025] The root node is used to represent the evaluation result value of the equipment system performance;

[0026] The five intermediate nodes are respectively used to represent the communication degree index, reliability index, credibility of the evaluation process, credibility of the evaluation method, and data product quality of the equipment system;

[0027] The ten indicator nodes are respectively used to represent the target detection ability value, target recognition ability value, communication ability value, mobility ability value, support ability value, survivability ability value, command and control ability value, stealth ability value, damage ability value, and computing ability value; each ability value has equipment technical index information of the corresponding category.

[0028] Assigning values to the equipment system effectiveness evaluation model by using the preprocessed data set includes:

[0029] Assigning values to the hierarchical edges for assigning values to the equipment system effectiveness evaluation model;

[0030] Performing probability conversion processing on the actual test data set to obtain a probability data set; the probability data set includes probability data of each indicator node; the probability data of the indicator node is used to determine the occurrence probability of the output quantity according to the input quantity of the indicator node;

[0031] According to the equipment technical index information of the corresponding category of each indicator node's ability value, obtaining the technical index test data sequence corresponding to the equipment technical index information of the corresponding category from the actual test data set, and using all the obtained technical index test data sequences to construct the acquisition data matrix corresponding to the indicator node; the row vector of the acquisition data matrix is the technical index test data sequence;

[0032] Calculating the weight value for the acquisition data matrix of each indicator node to obtain the weight value of the weight edge connected to each indicator node.

[0033] The performing probability conversion processing on the actual test data set to obtain a probability data set includes:

[0034] For the beneficial equipment technical indicators, using the first conversion model to perform conversion processing on the actual test data set to obtain the probability data of the indicator nodes corresponding to the beneficial equipment technical indicators;

[0035] For cost-type equipment technical indicators, a second conversion model is used to convert the equipment technical indicators to obtain probability data of indicator nodes corresponding to the cost-type equipment technical indicators;

[0036] The probability data of all indicator nodes are used to construct a probability data set.

[0037] The expression of the first conversion model is:

[0038]

[0039] Among them, Ψ(C1,E k ) represents the equipment technical index E of the actual test data set k The upper value is the number of actual collected data of C1. Indicates the index node When the input is C1, the output is V k The probability of K ,E k ) represents the equipment technical index E of the actual test data set k The upper value is C K The actual number of collected data, Indicates the index node Input is C K Output V k The probability of the first conversion model is expressed as The value of Ψ(C1,E k )、Ψ(C2,E k ),…,Ψ(C K ,E k ) is determined by the ratio value;

[0040] The first conversion model is used to calculate the combination of all inputs and outputs of each indicator node of all benefit-type equipment technical indicators to obtain the probability data of each indicator node; K is the number of categories of benefit-type equipment technical indicators;

[0041] The expression of the second conversion model is:

[0042]

[0043] Among them, the expression of the second conversion model is expressed as The value of The ratio value is determined, R is the number of categories of cost-type equipment technical indicators;

[0044] For each index node of all the technical indexes of cost-based equipment, calculate using the second conversion model for all combinations of inputs and outputs to obtain the probability data of each index node; K is the number of categories of the technical indexes of cost-based equipment.

[0045] In the second aspect of the implementation of the present invention, an architecture effectiveness evaluation device based on big data is disclosed. The device includes:

[0046] A memory storing executable program code;

[0047] A processor coupled to the memory;

[0048] The processor calls the executable program code stored in the memory to execute the above-mentioned architecture effectiveness evaluation method based on big data.

[0049] In the third aspect of the implementation of the present invention, a computer-storable medium is disclosed. The computer-storable medium stores computer instructions, which are used to execute the above-mentioned architecture effectiveness evaluation method when called by a computer.

[0050] In the fourth aspect of the implementation of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the above-mentioned architecture effectiveness evaluation method based on big data.

[0051] The beneficial effects of the present invention are as follows:

[0052] The present invention uses big data technology and complex network analysis methods to integrate node and edge rules to form a systematic learning and analysis framework, which can reason about evaluation indexes, evaluation methods, etc., and improve the evaluation ability of the architecture effectiveness of equipment. Through effective data processing and analysis means, comprehensively master the effectiveness of the equipment system, realize the calculation and reasoning of relevant evaluation content, covering the architecture effectiveness learning and analysis process based on big data and complex networks, including big data and complex network-based evaluation data input, big data and complex network-based evaluation content correlation analysis, and big data and complex network-based evaluation task result recommendation.

[0053] Through multi-step preprocessing such as format conversion, data cleaning, and data pattern checking, the present invention effectively solves problems such as inconsistent data formats, noise interference, and abnormal data, significantly improves the quality of input data, and provides a strong guarantee for the accuracy and reliability of subsequent evaluation models.

[0054] The present invention realizes the improvement of dynamic evaluation ability: by using big data technology and combining historical data and real-time test data of the equipment architecture, it can dynamically reflect the change trend of equipment performance, realize the real-time evaluation of the performance of the equipment architecture, and solve the limitations of static evaluation of traditional methods.

[0055] The equipment system effectiveness evaluation model constructed by the present invention includes multiple intermediate nodes and index nodes, which can comprehensively consider multi-dimensional performance indicators of the equipment system (such as communication degree, reliability, data quality, etc.), fully reflect the overall performance of the equipment system, and solve the problem of one-sidedness in traditional evaluation methods. Brief Description of the Drawings

[0056] Figure 1 It is a flowchart of the implementation of the method of the present invention. Detailed Embodiment

[0057] To better understand the content of the present invention, two embodiments are given here.

[0058] Figure 1 It is a flowchart of the implementation of the method of the present invention.

[0059] Embodiment 1:

[0060] In the first aspect of the embodiment of the present invention, a system effectiveness evaluation method based on big data is disclosed, including:

[0061] S1, collecting an equipment index information set and an equipment system historical data set; the equipment system historical data set includes an actual test data set; the equipment index information set includes equipment technical index information of each type; the actual test data set includes a technical index test data sequence of the technical index of each type of equipment; the technical index test data sequence corresponds to an equipment of the equipment system; the equipment system includes several equipments;

[0062] S2, preprocessing the equipment system historical data set to obtain a preprocessed data set;

[0063] S3, constructing an equipment system effectiveness evaluation model;

[0064] S4, using the preprocessed data set to assign values to the equipment system effectiveness evaluation model;

[0065] S5, using the equipment system effectiveness evaluation model after assignment to process the collected equipment system performance index data set to obtain an evaluation result value of the equipment system performance.

[0066] The equipment system performance index data set includes the collected data of ten categories of technical indicators;

[0067] The preprocessing of the equipment system historical data set to obtain a preprocessed data set includes:

[0068] S21, performing format conversion processing on the equipment system historical data set to obtain a first data set;

[0069] S22. Perform data cleaning on the first data set to obtain a second data set;

[0070] S23. Perform data pattern checking on the second data set to obtain a preprocessed data set;

[0071] The step of performing data pattern checking on the second data set to obtain a preprocessed data set includes:

[0072] For the data of each type of data attribute in the second data set, using the data acquisition information of the data as the known independent variable and the data value of the data as the known dependent variable, construct an approximation curve to be approximated;

[0073] Use the polynomial function fitting method to perform curve fitting on the approximation curve to be approximated to obtain the best uniform approximation polynomial of the type of data attribute;

[0074] Extract the coefficient vectors of the best uniform approximation polynomials of all types of data attributes, perform cross-correlation calculation on all the coefficient vectors to obtain a cross-correlation matrix;

[0075] Perform eigenvalue calculation on the cross-correlation matrix to obtain an eigenvalue vector; the eigenvalue vector is a vector composed of all eigenvalues;

[0076] The element in the i-th row and j-th column of the cross-correlation matrix is the cross-correlation value of the coefficient vectors of the best uniform approximation polynomials of the i-th type of data attribute and the j-th type of data attribute;

[0077] The coefficient vector is a vector formed by arranging the coefficients of the best uniform approximation polynomial in descending order according to the corresponding orders;

[0078] Construct a fusion polynomial with the eigenvalue vector as the polynomial order; the total order of the fusion polynomial is the number of elements in the eigenvalue vector minus 1;

[0079] Taking the eigenvalue vector as the polynomial order includes: if the number of elements in the eigenvalue vector is N, the first element of the eigenvalue vector is the coefficient of the (N - 1)-th order, and the last element is the constant term of the polynomial.

[0080] Use the fusion polynomial to perform calculation on the known independent variable to obtain an approximate dependent variable;

[0081] Determine whether the absolute value of the difference between the approximate dependent variable and the corresponding known dependent variable is greater than a set first regression discrimination threshold; if it is greater than the first regression discrimination threshold, delete the data from the second data set, and if it is less than or equal to the first regression discrimination threshold, do not process the data;

[0082] Perform a fusion process on all the data after the execution data pattern check process for the second data set to obtain a preprocessed data set;

[0083] The equipment system effectiveness evaluation model includes: a root node, five intermediate nodes, and ten index nodes; the root node is connected to each intermediate node through a hierarchical edge; the intermediate node is connected to each index node through a weight edge; each weight edge and hierarchical edge has a corresponding weight value;

[0084] The root node is used to represent the evaluation result value of the equipment system performance;

[0085] The five intermediate nodes are respectively used to represent the communication degree index, reliability index, credibility of the evaluation process, credibility of the evaluation method, and data product quality of the equipment system;

[0086] The ten index nodes are respectively used to represent the target detection ability value, target recognition ability value, communication ability value, mobility ability value, support ability value, survival ability value, command and control ability value, stealth ability value, damage ability value, and computing ability value;

[0087] The cost-type indicators include the stealth ability value, support ability value, and survival ability value, and the others are benefit-type indicators.

[0088] Each ability value has equipment technical index information of the corresponding category;

[0089] The construction of the equipment system effectiveness evaluation model can be realized by using a directed graph model construction method.

[0090] In the equipment system effectiveness evaluation model, according to the actually collected test data set, determine the values of the ten index nodes, and then according to the values of the ten index nodes, use the weight values of the weight edges to perform weighted summation to calculate the values of the five intermediate nodes. Finally, according to the values of the five intermediate nodes, use the weight values of the hierarchical edges to perform weighted summation to calculate the value of a root node.

[0091] Using the preprocessed data set to assign values to the equipment system effectiveness evaluation model includes:

[0092] Assign values to the hierarchical edges for assigning values to the equipment system effectiveness evaluation model;

[0093] Perform a probability conversion process on the actually collected test data set to obtain a probability data set; the probability data set includes the probability data of each index node; the probability data of the index node is used to determine the occurrence probability of the output quantity according to the input quantity of the index node;

[0094] According to the equipment technical index information of the corresponding category corresponding to the ability value of each index node, obtain the technical index test data sequence corresponding to the equipment technical index information of the corresponding category from the actual test data set, and use all the obtained technical index test data sequences to construct a collection data matrix corresponding to the index node; the row vector of the collection data matrix is the technical index test data sequence;

[0095] For the collection data matrix of each index node, calculate the weight value to obtain the weight value of the weighted edge connected by each index node.

[0096] The probability conversion process for the actual test data set to obtain a probability data set includes:

[0097] For the beneficial equipment technical index, use the first conversion model to perform conversion processing on the actual test data set to obtain the corresponding probability data;

[0098] For the cost-type equipment technical index, use the second conversion model to perform conversion processing on the equipment technical index to obtain the probability data of the index node;

[0099] Use the probability data corresponding to all equipment technical index information to construct a probability data set;

[0100] The expression of the first conversion model is:

[0101]

[0102] where, Ψ(C1,E k ) represents the number of actual acquisition data of the actual test data set with a value of C1 on the equipment technical index E k , represents the probability that the index node outputs V k when the input is C1, and so on, Ψ(C K ,E k ) represents the number of actual acquisition data of the actual test data set with a value of C k on the equipment technical index E K , represents the probability that the index node outputs V K when the input is C k , and the expression of the first conversion model represents that the value of is determined according to the proportional values of Ψ(C1,E k ), Ψ(C2,E k ), …, Ψ(C K ,E k );

[0103] For all combinations of inputs and outputs of an index node, use the first conversion model to calculate and obtain corresponding probability data; K is the number of categories of equipment technical indicators;

[0104] For each index node under all beneficial equipment technical indicator conditions, use the first conversion model to obtain the probability data of each index node;

[0105] The expression of the second conversion model is:

[0106]

[0107] Among them, the expression of the second conversion model indicates that The value of is determined according to the proportional value of

[0108] For all combinations of inputs and outputs of an index node, use the second conversion model to calculate and obtain corresponding probability data;

[0109] For each index node under all cost-type equipment technical indicator conditions, use the second conversion model to obtain the probability data of each index node;

[0110] The calculation of the weight value of the weighted edge connected to each index node for the collected data matrix of each index node includes:

[0111] Perform principal component analysis on the collected data matrix of each index node to obtain a principal component matrix;

[0112] Perform difference calculation on the principal component matrix to obtain a difference vector;

[0113] Perform fusion calculation on the difference vector to obtain the weight value of the weighted edge connected to the index node;

[0114] The expression of the fusion calculation is:

[0115]

[0116] Among them, β i is the i-th item of the difference vector, and λ is the weight value of the weighted edge connected to the index node. The expression of the difference calculation is:

[0117] β = (P T V + aI) -1 P T y,

[0118] V = QR -1 ,

[0119] Among them, β is the calculated difference vector, Q and R are the Q matrix and R matrix obtained by performing QR decomposition on the principal component matrix Y respectively, V is the intermediate matrix, a is the largest eigenvalue of the principal component matrix P, y is the eigenvector corresponding to the largest eigenvalue of the principal component matrix P, and I represents the identity matrix.

[0120] The expression for the principal component analysis process is:

[0121] Y = CX,

[0122] where Y is the principal component matrix, C is the coefficient matrix, X is the collected data matrix, the principal component analysis process can be implemented by the PCA algorithm, and both the principal component matrix and the coefficient matrix are determined through the principal component analysis process;

[0123] The equipment technical indicators are divided into two types: benefit type and cost type. For benefit type indicators, the larger the value, the better the equipment; for cost type indicators, the smaller the value, the better the equipment.

[0124] The present invention provides an objective verification mechanism for the evaluation results by quantifying the credibility of the evaluation process and the evaluation method, and combining the analysis of data product quality, significantly improving the credibility and repeatability of the evaluation results.

[0125] The evaluation model of the present invention can be flexibly adjusted according to different equipment types and technical indicators through the dynamic assignment mechanism of weight edges and hierarchical edges, has strong adaptability and versatility, and can meet the diverse evaluation needs of complex equipment systems.

[0126] The processed equipment system performance index data set collected is processed by using the equipped system effectiveness evaluation model after assignment to obtain the evaluation result value of the equipment system performance, including:

[0127] According to the equipment technical indicator information corresponding to the category of the ability value of each index node, each index node is used to receive the technical indicator data corresponding to the category in the equipment system performance index data set, and according to the probability data of the index node, the probability value of each output data is determined under the condition of the input technical indicator data, and the output data with the maximum probability value is determined as the output value of the index node;

[0128] For each intermediate node, according to the output values of each index node connected to the intermediate node, a weighted sum is performed with the weight values of the weight edges connecting the index nodes and the intermediate node to obtain the value of the intermediate node;

[0129] For the value of each intermediate node, a weighted sum is performed with the weight value of the connected hierarchical edge to obtain the evaluation result value of the equipment system performance.

[0130] Assign values to the hierarchical edges for the effectiveness evaluation model of the equipment system. The weights of the hierarchical edges of the intermediate nodes can be determined by randomly generating weight values or by calculating the variance of the historical data of each intermediate node, and using the determined variance value as the weight value of the hierarchical edge of the intermediate node.

[0131] The data cleaning process includes: filling missing values, smoothing noisy data, and smoothing or deleting outliers. For smoothing noisy data, first identify the noisy data, and then smooth the noisy data based on the data before and after the noisy data. The noisy data is the value that is less than the detection sensitivity of the sensor of the observed data or greater than the measurement upper limit of the sensor of the observed data. The outliers can be identified using the Kalman filter method. To determine the filling value for missing values, the average of the measured values within a certain sampling interval before and after the missing value can be taken.

[0132] When using the polynomial function fitting method to fit the curve to be approximated, the best uniform linear approximation method can be used. The best uniform approximation polynomial f(Ix) has the following expression:

[0133] f(Ix) = α P1 (Ix) P1 + α P1-1 (Ix) P1-1 + … + α2(Ix) 2 + α1(Ix) + α0,

[0134] where P1 is the order of the best uniform approximation polynomial f(Ix), and α0, α1, α2, …, α P1 are the coefficients of the best uniform approximation polynomial f(Ix);

[0135] The data collection information is the collection time information.

[0136] The format conversion process can be implemented using an ETL tool.

[0137] In the second aspect of the implementation of the present invention, a system effectiveness evaluation device based on big data is disclosed. The device includes:

[0138] A memory storing executable program code;

[0139] A processor coupled to the memory;

[0140] The processor calls the executable program code stored in the memory and executes the system effectiveness evaluation method based on big data as described above.

[0141] In the third aspect of the implementation of the present invention, a computer-readable storage medium is disclosed. The computer-readable storage medium stores computer instructions, which are used to execute the method for evaluating the system effectiveness based on big data when called by a computer.

[0142] In the fourth aspect of the implementation of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the method for evaluating the system effectiveness based on big data.

[0143] Example Two:

[0144] The embodiment of the present invention discloses a method for evaluating the system effectiveness based on big data, including:

[0145] (1) Data collection and integration

[0146] Data collection is the basis of the entire evaluation process, which involves obtaining relevant data from various sources. The data sources for evaluating the effectiveness of the equipment system are extensive, including sensor data, operation records, maintenance logs, experimental test results, historical performance data, etc. These data sources usually have different formats and structures, so it is necessary to converge them into a unified platform through data integration technology.

[0147] Integrate various data sources through big data technology to form an evaluation data set. Let the input data set be D input :

[0148] D input = Integrate(D source1 , D source2 ,..., D sourceN )

[0149] During the integration process, it is necessary to handle data format conversion, structure matching, and data redundancy problems. Identify multi-source data to determine various data sources and understand their data formats and contents. Sensor data may include real-time operation parameters, while historical performance data records the long-term device performance. Use data collection tools and interfaces (such as APIs, data scraping tools, etc.) to extract data from different sources to ensure the accuracy and integrity of the data. Data integration: Convert the formats of the collected various data and merge them, and use ETL (Extract, Transform, Load) tools to store the data uniformly in the data platform for further processing and analysis.

[0150] (2) Data preprocessing and storage

[0151] Data preprocessing ensures data quality and prepares for complex network analysis. This process includes data cleaning, data transformation, and data storage. Data cleaning denoises the data, handles missing values, and corrects errors. This step involves algorithm application and manual inspection to ensure data consistency and accuracy. Data transformation: The data is transformed into a format suitable for complex network analysis. For example, structured data is converted into the format of nodes and edges, or unstructured data is extracted and transformed. Data storage selects appropriate data storage technologies for data management. In a big data environment, distributed storage systems (such as Hadoop HDFS, NoSQL databases, etc.) can handle massive amounts of data, while graph databases (such as Neo4j) are suitable for storing and analyzing complex network data.

[0152] (3) Complex network construction

[0153] Complex network construction maps the effectiveness of the equipment system to a network model for structural and relationship analysis. This step includes the definition of nodes and edges, network construction, and feature extraction.

[0154] Definition of nodes and edges Define nodes (such as equipment, system components, tasks, etc.) and edges (such as dependency relationships, interaction relationships) in the network model. Each node and edge needs to have clear attribute and relationship definitions. Network construction Use graph theory techniques and network analysis tools to construct a complex network model. By defining the connection relationships and weights between nodes, an effectiveness network diagram of the equipment system is formed. Feature extraction Extract key features from the network, such as node centrality (indicating the importance of a node in the network), network density (indicating the tightness of network connections), and path length (indicating the efficiency of information propagation). These features help to understand the structure of the system and key nodes.

[0155] (4) Correlation analysis and pattern recognition

[0156] Correlation analysis mines potential patterns and relationships in the effectiveness of the equipment system through network analysis methods. This step includes two main aspects: relationship mining and pattern recognition. Relationship mining applies network analysis algorithms (such as community detection, clustering analysis, path analysis, etc.) to identify key nodes and important relationships in the network. Community detection can discover functional modules in the system, and path analysis can reveal the paths of effectiveness transmission. Pattern recognition uses data mining techniques to identify effectiveness patterns. For example, by comparing the performance patterns of different equipment in similar tasks, key factors affecting effectiveness are found, and the performance of the system in different scenarios is predicted.

[0157] Based on the complex network analysis method, correlation analysis is performed on the input data to identify the relationships between evaluation contents. Let the obtained correlation matrix be R, then the correlation analysis of evaluation contents can be expressed as:

[0158] R = Analyze_Relevance(D input , G),

[0159] where G is a complex network structure representing the effectiveness evaluation of the equipment system.

[0160] (5) Evaluation Result Generation and Recommendation

[0161] The evaluation result generation integrates the analysis results into a comprehensive report and provides optimization suggestions and action guidelines. The result integration integrates the results of the relevance analysis into an effectiveness evaluation report. The report includes the analysis of effectiveness indicators, the evaluation of the performance of system components, the analysis of the suitability of task scenarios, etc. The intelligent recommendation develops an intelligent recommendation system based on the analysis results to provide optimization solutions for users. For example, it recommends the best equipment configuration or operation strategy according to task requirements to help optimize system performance. The result visualization uses visualization tools (such as charts, dashboards, etc.) to display complex analysis results. Visualization can help users intuitively understand data patterns and analysis conclusions, thus making more accurate decisions.

[0162] Based on the results of the relevance analysis, recommend relevant evaluation tasks and results. Let the set of recommended task results be R recommended :

[0163] R recommended = Recommend(R),

[0164] (6) Feedback and Optimization

[0165] Feedback and optimization is a continuous improvement process that continuously adjusts and optimizes the evaluation model and analysis process according to the evaluation results and user feedback. Feedback on the evaluation results and recommendation system to understand the actual needs and improvement suggestions of users. Model adjustment adjusts and optimizes the evaluation model and analysis algorithm according to feedback and new data. For example, update the definition of nodes and edges of the network model and improve the parameter settings of the analysis algorithm. Process improvement continuously improves the processes of data collection, preprocessing, network construction, and result generation to improve the efficiency and effectiveness of evaluation. Introduce new technologies and methods to optimize the work process and enhance the overall analysis ability.

[0166] The above are only embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the scope of the claims of the present invention.

Claims

1. A system effectiveness evaluation method based on big data, characterized in that Including: S1. Collect an equipment index information set and an equipment system historical data set; the equipment system historical data set includes an actual test data set; the equipment index information set includes equipment technical index information of each type; the actual test data set includes a technical index test data sequence of the equipment technical index of each type. S2. Preprocess the equipment system historical data set to obtain a preprocessed data set. S3. Construct an equipment system effectiveness evaluation model. S4. Use the preprocessed data set to assign values to the equipment system effectiveness evaluation model. S5. Use the equipment system effectiveness evaluation model after value assignment to process the collected equipment system performance index data set to obtain an evaluation result value of the equipment system performance.

2. The method for evaluating system effectiveness based on big data according to claim 1, wherein, The preprocessing of the equipment system historical data set to obtain a preprocessed data set includes: S21. Perform format conversion processing on the equipment system historical data set to obtain a first data set. S22. Perform data cleaning processing on the first data set to obtain a second data set. S23. Perform data pattern check processing on the second data set to obtain a preprocessed data set.

3. The system effectiveness evaluation method based on big data according to claim 2, wherein, The data pattern check processing of the second data set to obtain a preprocessed data set includes: For the data of each type of data attribute in the second data set, using the data acquisition information of the data as the known independent variable and the data value of the data as the known dependent variable, construct a curve to be approximated using the known independent variable and the known dependent variable. Use the polynomial function fitting method to perform curve fitting on the curve to be approximated to obtain the best uniform approximation polynomial of the type of data attribute. Extract the coefficient vectors of the best uniform approximation polynomials of all types of data attributes, perform cross-correlation calculation processing on all the coefficient vectors to obtain a cross-correlation matrix. Perform eigenvalue calculation processing on the cross-correlation matrix to obtain an eigenvalue vector; the eigenvalue vector is a vector constructed by all eigenvalues. Construct a fusion polynomial with the eigenvalue vector as the polynomial order; the total order of the fusion polynomial is the number of elements in the eigenvalue vector minus 1. Use the fusion polynomial to perform calculation processing on the known independent variable to obtain an approximate dependent variable. Determine whether the absolute value of the difference between the approximate dependent variable and the corresponding known dependent variable is greater than a set first regression discrimination threshold; if it is greater than the first regression discrimination threshold, delete the data from the second data set, and if it is less than or equal to the first regression discrimination threshold, do not process the data. Perform fusion processing on all the data after the data pattern check processing of the second data set to obtain a preprocessed data set.

4. The system effectiveness evaluation method based on big data according to claim 1, characterized in that The equipment system effectiveness evaluation model includes: a root node, five intermediate nodes, and ten index nodes; the root node is connected to each intermediate node through a hierarchical edge; the intermediate node is connected to each index node through a weight edge; each weight edge and hierarchical edge has a corresponding weight value. The root node is used to represent the evaluation result value of the equipment system performance. The five intermediate nodes are respectively used to characterize the communication degree index, reliability index, credibility of evaluation process, credibility of evaluation method, and data product quality of the equipment system; The ten index nodes are respectively used to characterize the target detection ability value, target recognition ability value, communication ability value, mobility ability value, support ability value, survivability ability value, command and control ability value, stealth ability value, damage ability value, and computing ability value; each ability value has equipment technical index information of the corresponding category.

5. The system effectiveness evaluation method based on big data according to claim 1, wherein, Using the preprocessed data set to assign values to the equipment system effectiveness evaluation model includes: Assigning values to the hierarchical edges for assigning values to the equipment system effectiveness evaluation model; Performing probability conversion processing on the actual test data set to obtain a probability data set; the probability data set includes probability data of each index node; the probability data of the index node is used to determine the occurrence probability of the output quantity according to the input quantity of the index node; According to the equipment technical index information of the corresponding category of each ability value of the index node, obtaining a technical index test data sequence corresponding to the equipment technical index information of the corresponding category from the actual test data set, and using all the obtained technical index test data sequences to construct an acquisition data matrix corresponding to the index node; the row vector of the acquisition data matrix is the technical index test data sequence; Calculating the weight value for the acquisition data matrix of each index node to obtain the weight value of the weight edge connected to each index node.

6. The system effectiveness evaluation method based on big data according to claim 5, wherein The performing probability conversion processing on the actual test data set to obtain a probability data set includes: For the benefit-type equipment technical index, using the first conversion model to perform conversion processing on the actual test data set to obtain the probability data of the index node corresponding to the benefit-type equipment technical index; For the cost-type equipment technical index, using the second conversion model to perform conversion processing on the equipment technical index to obtain the probability data of the index node corresponding to the cost-type equipment technical index; Using the probability data of all index nodes to construct a probability data set.

7. The method for evaluating the system effectiveness based on big data according to claim 6, characterized in that The expression of the first conversion model is: Among them, Ψ(C1, E k ) represents the number of actual collected data with the value of C1 on the equipment technical index E k . represents the index node The probability of outputting V when the input is C1, and so on. Ψ(C k , E K ) represents the number of actual collected data with the value of C k on the equipment technical index E k . K The probability of outputting V when the input of the index node is C . K The value of the first conversion model's expression represents k is determined according to the proportional values of Ψ(C1, E ), Ψ(C2, E k ), …, Ψ(C k , E K ), E k );​ Calculating for each index node of all benefit-type equipment technical indexes in all combinations of inputs and outputs using the first conversion model to obtain the probability data of each index node; K is the number of categories of benefit-type equipment technical indexes; The expression of the second conversion model is: Among them, the expression of the second conversion model indicates that takes values according to The proportional value is determined, and R is the number of categories of cost-based equipment technical indicators; Calculating for each index node of all cost-type equipment technical indexes in all combinations of inputs and outputs using the second conversion model to obtain the probability data of each index node; K is the number of categories of cost-type equipment technical indexes.

8. An apparatus for evaluating the effectiveness of a system based on big data, characterized in that, The device includes: A memory storing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory and executes the big data-based system effectiveness evaluation method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, which are used to execute the big data-based system effectiveness evaluation method according to any one of claims 1 to 7 when called by a computer.

10. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the big data-based system effectiveness evaluation method described in any one of claims 1 to 7.

Images