Government affair cloud service evaluation prediction method and system based on combination of improved TOPSIS and GRNN

Abstract

The invention provides a government affair cloud service evaluation prediction method and system based on combination of improved TOPSIS and GRNN, and relates to the technical field of cloud service evaluation prediction. The method comprises the steps of collecting and preprocessing evaluation index data corresponding to government affair cloud services; calculating the weights of the preprocessed evaluation indexes by adopting an entropy weight method to obtain the weights of the evaluation indexes; according to the weight of each evaluation index, performing preliminary sorting on each evaluation index based on an improved TOPSIS method to obtain a preliminary evaluation result; according to each evaluation index corresponding to each government affair cloud service and the preliminary evaluation result thereof, training a prediction model constructed based on the GRNN; and inputting new evaluation index data of the government affair cloud service into the trained prediction model, and predicting to obtain an evaluation result of the government affair cloud service. According to the method, the accuracy and objectivity of government affair cloud service evaluation results are improved, and the calculation efficiency and real-time performance of the system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cloud service evaluation prediction, and in particular to a government cloud service evaluation prediction method and system based on the combination of improved TOPSIS and GRNN. Background Art

[0002] As an important part of the informatization construction of government departments, the quality evaluation and prediction of government cloud services are of great significance to improving service levels. Traditional evaluation methods have the following problems: Strong subjectivity: Traditional cloud service evaluation methods often rely on expert scoring or user feedback, which is greatly affected by the subjective factors of the evaluators, resulting in poor objectivity and reliability of the evaluation results. Unreasonable weight distribution: Many evaluation methods are arbitrary in weight distribution and lack scientific basis, resulting in the evaluation results being unable to accurately reflect the importance of each indicator. High computational complexity: Some multi-attribute decision-making methods have cumbersome steps and large amount of calculations in the calculation process, and are less efficient in practical applications. Lack of dynamics: Traditional evaluation methods are generally based on static data and cannot reflect the dynamic changes of cloud services in a timely manner, resulting in delayed evaluation results and difficulty in guiding real-time decision-making. Poor ability to handle nonlinear relationships: Traditional evaluation methods are mostly linear models, which are difficult to effectively handle the complex nonlinear relationships in cloud service evaluation, resulting in insufficient accuracy of evaluation results.

[0003] Therefore, it is necessary to design a new government cloud service evaluation and prediction method to improve the accuracy and objectivity of the evaluation results. Summary of the invention

[0004] The purpose of the present invention is to provide a government cloud service evaluation and prediction method and system based on the combination of improved TOPSIS and GRNN, so as to solve the technical problems that traditional evaluation methods have strong subjectivity, unreasonable weight distribution, high computational complexity, lack of dynamics, poor nonlinear relationship processing ability, etc., which lead to insufficient accuracy of evaluation results.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a government cloud service evaluation prediction method based on the combination of improved TOPSIS and GRNN, comprising:

[0006] Collect various evaluation indicator data corresponding to each government cloud service;

[0007] Preprocessing the evaluation index data;

[0008] The entropy weight method is used to calculate the weights of various evaluation indicators after preprocessing to obtain the weights of various evaluation indicators;

[0009] According to the weight of each evaluation index, the evaluation indexes are preliminarily ranked based on the improved TOPSIS method to obtain the preliminary evaluation results;

[0010] According to the evaluation indicators corresponding to each government cloud service and its preliminary evaluation results, the prediction model constructed based on GRNN is trained; the new evaluation indicator data of the government cloud service is input into the trained prediction model for prediction, and the final evaluation result of the government cloud service is obtained and output.

[0011] Furthermore, the evaluation indicator data include service security, service performance, service support, and service commerciality.

[0012] Furthermore, preprocessing of the evaluation index data includes:

[0013] Normalize the collected data, including:

[0014] Perform positive processing on the collected indicator data;

[0015] Normalize the data after forward processing to remove the dimension of the original data;

[0016] Establish the initial decision matrix D = (d ij ) m×n .

[0017] Furthermore, preprocessing of the evaluation index data also includes:

[0018] Make the cost type indicators in the evaluation indicators positive;

[0019] The improved dynamic efficiency coefficient method is used to incentivize or punish cloud service providers.

[0020] Furthermore, the improved dynamic efficiency coefficient method is used to incentivize or punish cloud service providers, including:

[0021] Introducing an adjustment factor in the efficacy coefficient calculation process, and dynamically adjusting the efficacy coefficient using the adjustment factor;

[0022] The adjustment factor λ is expressed as follows:

[0023]

[0024] Based on the dynamically adjusted efficiency coefficient, cloud service providers with service levels above the average are incentivized, while cloud service providers with service levels below the average are punished.

[0025] Furthermore, the entropy weight method is used to calculate the weights of the preprocessed evaluation indicators to obtain the weights of the evaluation indicators, including:

[0026] The weight of the objective indicator is calculated by the entropy method to obtain the first indicator;

[0027] The weight of subjective indicators is calculated by analytic hierarchy process to obtain the second indicator;

[0028] The first indicator and the second indicator are respectively assigned weighted values, and then weighted calculation is performed to obtain the weights of various evaluation indicators.

[0029] Furthermore, the evaluation indicators are preliminarily ranked according to their weights based on the improved TOPSIS method to obtain preliminary evaluation results, including:

[0030] The initial matrix D is dimensionless, and the standard matrix X is obtained as (x ij ) m×n ,

[0031] According to the dimensionless standard matrix X combined with the weight values ​​of the evaluation indicators, a decision matrix R = (r ij ) m×n ,

[0032]

[0033] In the formula, ω1 to ωn represent the weight values ​​of each evaluation index, and n represents the total number of evaluation indicators.

[0034] Furthermore, the method of performing a preliminary ranking of the evaluation indicators based on the improved TOPSIS method according to the weights of the evaluation indicators to obtain a preliminary evaluation result also includes:

[0035] According to the decision matrix R, construct an absolute positive ideal solution and an absolute negative ideal solution;

[0036] Calculate the Euclidean distance between each evaluation index of each government cloud service in the government cloud service evaluation set and the absolute positive ideal solution and the absolute negative ideal solution respectively in, represents the Euclidean distance between each evaluation index and the absolute positive ideal solution, represents the Euclidean distance between each evaluation index and the absolute negative ideal solution;

[0037] According to the Euclidean distance between each evaluation index of each government cloud service and the absolute positive ideal solution and the absolute negative ideal solution Calculate the relative closeness of each government cloud service

[0038] Based on the relative proximity of each government cloud service Conduct a preliminary sorting and obtain preliminary evaluation results of each government cloud service.

[0039] Furthermore, constructing an absolute positive ideal solution and an absolute negative ideal solution according to the decision matrix R includes:

[0040] After the initial data of each evaluation index is standardized and dimensionless, the value range of the evaluation index is standardized to the interval [0,1]. In this interval, the closer the value is to 1, the better the performance of the index is, and the closer the value is to 0, the worse the performance is.

[0041] The construction of the absolute positive ideal solution includes:

[0042] Assuming that the result of the forward processing of each indicator value is 1, then all indicators have reached their possible optimal state, and the ideal solution at this time is the absolute positive ideal solution AC + , the expression is as follows:

[0043]

[0044] The construction of the absolute negative ideal solution includes:

[0045] Assuming that the result of the forward processing of each indicator value is 0, then all indicators have reached their possible worst state, and the ideal solution at this time is the absolute negative ideal solution AC - , the expression is as follows:

[0046]

[0047] In a second aspect, the present invention provides a government cloud service evaluation prediction system based on the combination of improved TOPSIS and GRNN, comprising:

[0048] The data collection module is used to collect the evaluation index data corresponding to each government cloud service;

[0049] A data preprocessing module, used for preprocessing the evaluation index data;

[0050] The weight calculation module is used to calculate the weights of various evaluation indicators after preprocessing by using the entropy weight method to obtain the weights of various evaluation indicators;

[0051] The preliminary evaluation module is used to preliminarily sort the evaluation indicators according to the weights of the evaluation indicators based on the improved TOPSIS method to obtain the preliminary evaluation results;

[0052] The GNN prediction module is used to train the prediction model constructed based on GRNN according to the evaluation indicators corresponding to each government cloud service and its preliminary evaluation results; input the new evaluation indicator data of the government cloud service into the trained GRNN prediction model for prediction to obtain the corresponding evaluation results;

[0053] The result output module is used to output the final evaluation results of the government cloud service.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The various steps of the method work together, because data preprocessing ensures the consistency of data, the weight calculation module optimizes the weight distribution, the improved TOPSIS evaluation module realizes the preliminary sorting, and the GRNN prediction module provides accurate prediction results. Therefore, the overall method and system not only improve the accuracy and objectivity of the evaluation results, but also the calculation efficiency and real-time performance of the system are significantly improved from a structural point of view. In this way, because each module cooperates with each other, the government cloud service evaluation and prediction method can not only reflect the dynamic changes of cloud services in a timely manner, but also the evaluation results can effectively guide the optimization and improvement of cloud services. In other words, the evaluation method that used to rely on subjective scoring and linear models has been improved to be realized by data-driven multi-attribute decision analysis and neural network prediction. Therefore, the present invention realizes the scientificity and efficiency of the evaluation system, and makes the management and improvement of government cloud services more accurate and intelligent. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A flowchart of a government cloud service evaluation and prediction method based on a combination of improved TOPSIS and GRNN provided in an embodiment of the present invention;

[0057] Figure 2 This is a module diagram of the government cloud service evaluation and prediction system based on the improved TOPSIS and GRNN combination of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] Figure 1 A flowchart of a government cloud service evaluation prediction method based on improved TOPSIS and GRNN provided in an embodiment of the present invention is shown in FIG. Figure 1 As shown, the government cloud service evaluation prediction method 100 based on the combination of improved TOPSIS and GRNN includes:

[0060] S110: Collect various evaluation indicator data corresponding to each government cloud service.

[0061] Optionally, in step S110, the present invention selects government cloud service quality evaluation indicators based on the specific field characteristics of cloud computing, combined with the characteristics of different service models of cloud services and relevant domestic and international cloud service evaluation standards, based on the principles of system scientificity, integrity, observability, etc., and collects data on 20 indicators including service security, service performance, service support, and service commerciality.

[0062] S120: Preprocessing various evaluation index data.

[0063] In step S120, the collected data is normalized to eliminate the dimensional differences between different indicators and ensure the consistency and comparability of the data. The normalization process includes performing positive processing on the collected indicator data, integrating the collected indicator data of the government cloud service, and establishing the initial decision matrix D = (d ij ) m×n .

[0064] Specifically, in some embodiments, S120: preprocessing various evaluation index data includes:

[0065] S121: normalizing the collected data, including:

[0066] S1211: Perform positive processing on the collected indicator data;

[0067] S1212: normalizing the data after the forward processing to remove the dimension of the original data;

[0068] S1213: Establishing the initial decision matrix D = (d ij ) m×n .

[0069] Furthermore, S120: preprocessing the evaluation index data further includes:

[0070] S12111: Make the cost type indicators in the evaluation indicators positive for better comparison and decision-making;

[0071] S12112: Use the improved dynamic efficiency coefficient method to incentivize or punish cloud service providers, including:

[0072] S12113: Introduce an adjustment factor in the calculation process of the efficacy coefficient, and use the adjustment factor to dynamically adjust the efficacy coefficient;

[0073] The adjustment factor λ is expressed as follows:

[0074]

[0075] S12114: Based on the dynamically adjusted efficiency coefficient, cloud service providers with service levels above the average are incentivized, and cloud service providers with service levels below the average are punished.

[0076] In this embodiment, due to the introduction of an improved dynamic efficiency coefficient method, an adjustment factor is introduced in the efficiency coefficient calculation process, so that the efficiency coefficient can be dynamically adjusted as the adjustment factor changes, cloud service providers with higher-than-average service levels are incentivized, and cloud service providers with lower-than-average service levels are punished, thereby widening the gap, encouraging cloud service providers to continuously optimize and improve the cloud services they provide, encouraging the formation of a good competition mechanism among cloud service providers, and utilizing the adjustment factor to stimulate and motivate each cloud service provider.

[0077] S130: using the entropy weight method to calculate the weights of the preprocessed evaluation indicators to obtain the weights of the evaluation indicators.

[0078] Preferably, in some embodiments, step S130 includes:

[0079] S131: Calculate the weight of the objective indicator by using an entropy method to obtain a first indicator;

[0080] S132: Calculate the subjective index weight by analytic hierarchy process to obtain the second index;

[0081] S133: Assign weighted values ​​to the first indicator and the second indicator respectively, and then perform weighted calculation to obtain the weights of various evaluation indicators.

[0082] Traditional TOPSIS usually adopts the method of introducing expert experience to assign weights to indicators, but such an operation is too dependent on the professional knowledge and personal experience of the expert, and is very subjective. In this embodiment, the entropy weight method is used to calculate the weights of each evaluation indicator to ensure the scientificity and rationality of the weight distribution. The present invention introduces a combined weighting method to ensure that the evaluation indicators are both subjective and objective in the process of calculating the weights. The calculation of the objective weights is realized by the entropy method, and the calculation of the subjective indicator weights is realized by hierarchical analysis. Finally, the two are weighted and combined to obtain the weights of each evaluation indicator.

[0083] S140: According to the weights of various evaluation indicators, various evaluation indicators are preliminarily ranked based on the improved TOPSIS method to obtain preliminary evaluation results;

[0084] Further, step S140 includes:

[0085] S141: Perform dimensionless processing on the initial matrix D to obtain a standard matrix X=(x ij ) m×n .

[0086] S142: Construct a decision matrix R=(r ij ) m×n ,

[0087]

[0088] In the formula, ω1 to ωn represent the weight values ​​of each evaluation index, and n represents the total number of evaluation indicators.

[0089] S143: According to the decision matrix R, construct an absolute positive ideal solution and an absolute negative ideal solution.

[0090] In the traditional TOPSIS method, as the service solution set changes, the positive and negative ideal solutions also change. Therefore, when the government cloud service evaluation solution set changes, the evaluation results are uncertain. To address this problem, the present invention proposes absolute positive and negative ideal solutions to replace relative positive and negative ideal solutions.

[0091] Specifically, step S143 includes:

[0092] After the initial data of each evaluation index is standardized and dimensionless, the value range of the evaluation index is standardized to the interval [0,1]. In this interval, the closer the value is to 1, the better the performance of the index is, and the closer the value is to 0, the worse the performance is.

[0093] Constructing an absolute positive ideal solution includes: assuming that the result of the forward processing of each indicator value is 1, then all indicators have reached their possible optimal state, and the ideal solution at this time is the absolute positive ideal solution AC + , the expression is as follows:

[0094]

[0095] Constructing an absolute negative ideal solution includes: assuming that the result of the forward processing of each indicator value is 0, then all indicators have reached their possible worst state, and the ideal solution at this time is the absolute negative ideal solution AC - , the expression is as follows:

[0096]

[0097] S144: Calculating the Euclidean distance between each evaluation index of each government cloud service in the government cloud service evaluation set and the absolute positive ideal solution and the absolute negative ideal solution in, represents the Euclidean distance between each evaluation index and the absolute positive ideal solution, It represents the Euclidean distance between each evaluation index and the absolute negative ideal solution.

[0098] S145: calculating the Euclidean distance between each evaluation index of each government cloud service and the absolute positive ideal solution and the absolute negative ideal solution Calculate the relative closeness of each government cloud service

[0099] S146: Based on the relative proximity of each government cloud service Conduct a preliminary sorting and obtain preliminary evaluation results of each government cloud service.

[0100] S150: Train the GRNN model according to the evaluation indicators corresponding to each government cloud service and its preliminary evaluation results; input the new evaluation indicator data of the government cloud service into the trained GRNN model for prediction to obtain the final evaluation result of the government cloud service.

[0101] Generalized regression neural network (GRNN) is based on mathematical statistics and establishes implicit relationship between input and output through training samples. GRNN method has fast convergence speed and small amount of calculation, which can solve the problem of small samples well, and is better than ordinary RBF and BP neural networks in learning speed and generalization.

[0102] In step S150, the present invention utilizes a generalized regression neural network (GRNN) to construct a prediction model, uses the historical 30-day indicator collection data of government cloud services as the training model input, and uses the various evaluation indicators corresponding to each government cloud service and the corresponding preliminary evaluation results based on TOPSIS as the model output, constructs a training sample, and trains the GRNN prediction model to obtain a trained prediction model; finally, the quality evaluation indicators and indicator collection data of the government cloud service are selected as input factors for establishing the prediction model, and the generalized regression neural network algorithm (GRNN) is optimized and modeled to achieve quality evaluation prediction of each government cloud service.

[0103] S160: Output the final evaluation results of the government cloud service.

[0104] This step S160 is used to output the final evaluation result and generate a corresponding evaluation report for user reference.

[0105] According to the above embodiment, the prediction method ensures the consistency of data, optimizes the weight distribution, and realizes preliminary sorting based on the improved TOPSIS according to the collaborative work of the above steps. The GRNN prediction model provides accurate prediction results. Therefore, the overall system not only improves the accuracy and objectivity of the evaluation results, but also realizes the scientificity and efficiency of the evaluation system through data-driven multi-attribute decision analysis and neural network prediction, and makes the management and improvement of government cloud services more accurate and intelligent.

[0106] The above is an introduction to the method embodiment. The following is a further explanation of the solution of the present invention through a system embodiment.

[0107] Figure 2 Schematic diagram of the module of the government cloud service evaluation and prediction system based on the improved TOPSIS and GRNN combination of the present invention, as shown in Figure 2 As shown, the government cloud service evaluation prediction system 200 based on the combination of improved TOPSIS and GRNN includes:

[0108] The data collection module 210 is used to collect various evaluation index data corresponding to each government cloud service;

[0109] The data preprocessing module 220 is used to preprocess the evaluation index data;

[0110] The weight calculation module 230 is used to calculate the weights of the preprocessed evaluation indicators by using the entropy weight method to obtain the weights of the evaluation indicators;

[0111] A preliminary evaluation module 240 is used to perform preliminary sorting of the evaluation indicators based on the weights of the evaluation indicators and the improved TOPSIS method to obtain preliminary evaluation results;

[0112] The GNN prediction module 250 is used to train the prediction model constructed based on the GRNN according to the evaluation indicators corresponding to each government cloud service and the preliminary evaluation results thereof; input the new evaluation indicator data of the government cloud service into the trained prediction model for prediction to obtain the corresponding evaluation results;

[0113] The result output module 260 is used to output the final evaluation result of the government cloud service.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0115] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The government cloud service evaluation prediction method based on the combination of improved TOPSIS and GRNN is characterized by: include: Collect various evaluation indicator data corresponding to each government cloud service; Preprocessing the evaluation index data; The entropy weight method is used to calculate the weights of various evaluation indicators after preprocessing to obtain the weights of various evaluation indicators; According to the weight of each evaluation index, the evaluation indexes are preliminarily ranked based on the improved TOPSIS method to obtain the preliminary evaluation results; According to the evaluation indicators corresponding to each government cloud service and its preliminary evaluation results, the prediction model constructed based on GRNN is trained; the new evaluation indicator data of the government cloud service is input into the trained GRNN prediction model for prediction, and the final evaluation result of the government cloud service is obtained and output.

2. The method according to claim 1, characterized in that The evaluation indicator data include service security, service performance, service support, and service commerciality.

3. The method according to claim 1, characterized in that The preprocessing of various evaluation index data includes: Normalize the collected data, including: Perform positive processing on the collected indicator data; Normalize the data after forward processing to remove the dimension of the original data; Establish the initial decision matrix D = (d ij ) m×n .

4. The method according to claim 3, characterized in that The preprocessing of the evaluation index data also includes: Make the cost type indicators in the evaluation indicators positive; The improved dynamic efficiency coefficient method is used to incentivize or punish cloud service providers.

5. The method according to claim 4, characterized in that The improved dynamic efficacy coefficient method is used to incentivize or punish cloud service providers, including: Introducing an adjustment factor in the efficacy coefficient calculation process, and dynamically adjusting the efficacy coefficient using the adjustment factor; The adjustment factor λ is expressed as follows: Based on the dynamically adjusted efficiency coefficient, cloud service providers with service levels above the average are incentivized, while cloud service providers with service levels below the average are punished.

6. The method according to any one of claims 3 to 5, characterized in that: The entropy weight method is used to calculate the weights of the preprocessed evaluation indicators to obtain the weights of the evaluation indicators, including: The weight of the objective indicator is calculated by the entropy method to obtain the first indicator; The weight of subjective indicators is calculated by analytic hierarchy process to obtain the second indicator; The first indicator and the second indicator are respectively assigned weighted values, and then weighted calculation is performed to obtain the weights of various evaluation indicators.

7. The method according to claim 6, characterized in that According to the weight of each evaluation index, the evaluation indexes are preliminarily ranked based on the improved TOPSIS method to obtain preliminary evaluation results, including: The initial matrix D is dimensionless, and the standard matrix X is obtained as (x ij ) m×n , According to the dimensionless standard matrix X combined with the weight values ​​of the evaluation indicators, a decision matrix R = (r ij ) m×n , In the formula, ω1 to ωn represent the weight values ​​of each evaluation index, and n represents the total number of evaluation indicators.

8. The method according to claim 7, characterized in that The method of performing a preliminary ranking of the evaluation indicators based on the weights of the evaluation indicators and the improved TOPSIS method to obtain a preliminary evaluation result also includes: According to the decision matrix R, construct an absolute positive ideal solution and an absolute negative ideal solution; Calculate the Euclidean distance between each evaluation index of each government cloud service in the government cloud service evaluation set and the absolute positive ideal solution and the absolute negative ideal solution respectively in, represents the Euclidean distance between each evaluation index and the absolute positive ideal solution, represents the Euclidean distance between each evaluation index and the absolute negative ideal solution; According to the Euclidean distance between each evaluation index of each government cloud service and the absolute positive ideal solution and the absolute negative ideal solution Calculate the relative closeness of each government cloud service Based on the relative proximity of each government cloud service Conduct a preliminary sorting and obtain preliminary evaluation results of each government cloud service.

9. The method according to claim 8, characterized in that The step of constructing an absolute positive ideal solution and an absolute negative ideal solution according to the decision matrix R comprises: After the initial data of each evaluation index is standardized and dimensionless, the value range of the evaluation index is standardized to the interval [0,1]. In this interval, the closer the value is to 1, the better the performance of the index is, and the closer the value is to 0, the worse the performance is. The construction of the absolute positive ideal solution includes: Assuming that the result of the forward processing of each indicator value is 1, then all indicators have reached their possible optimal state, and the ideal solution at this time is the absolute positive ideal solution AC + , the expression is as follows: The construction of the absolute negative ideal solution includes: Assuming that the result of the forward processing of each indicator value is 0, then all indicators have reached their possible worst state, and the ideal solution at this time is the absolute negative ideal solution AC - , the expression is as follows:

10. The government cloud service evaluation and prediction system based on the combination of improved TOPSIS and GRNN is characterized by: include: The data collection module is used to collect the evaluation index data corresponding to each government cloud service; A data preprocessing module, used for preprocessing the evaluation index data; The weight calculation module is used to calculate the weights of various evaluation indicators after preprocessing by using the entropy weight method to obtain the weights of various evaluation indicators; The preliminary evaluation module is used to preliminarily sort the evaluation indicators according to the weights of the evaluation indicators based on the improved TOPSIS method to obtain the preliminary evaluation results; The GNN prediction module is used to train the prediction model constructed based on GRNN according to the evaluation indicators corresponding to each government cloud service and its preliminary evaluation results; input the new evaluation indicator data of the government cloud service into the trained GRNN prediction model for prediction to obtain the corresponding evaluation results; The result output module is used to output the final evaluation results of the government cloud service.

Images