A method for evaluating the intelligence level of information systems
Through the deep Topsis model and ANFIS adaptive fuzzy inference network, the problem of differences in index system and insufficient interpretability in the level of intelligence evaluation of information system is solved, and a comprehensive objective evaluation and efficient evaluation of the level of intelligence of information system is achieved, with better versatility and explanatory.
Patent Information
- Application Number
- CN202111299708.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-04
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2041-11-04
AI Technical Summary
The existing method of intelligent level evaluation of information systems has differences in index systems, mutual influence between evaluation indicators and is difficult to explain, and it is not universal and interpretable. Especially, machine learning-based methods require a large amount of data and are black-box models. , lack of explainability.
Using the deep Topsis model combined with the ANFIS adaptive fuzzy inference network, the intelligent level indicators of the information system are evaluated step by step, the highest evaluation indicators are generated and the intelligent level scores are output. The ANFIS adaptive inference neural network analysis indicators are used to provide parameters in the form of rules. Mapping to achieve stronger interpretability and universality.
A comprehensive and objective evaluation of the level of intelligence of information systems is realized, with good generalization and interpretability, and can process different types of data and prior knowledge, simplify system integration, and improve the reliability and confidence of evaluation. .
Smart Images

Figure CN114048115B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent system evaluation, and in particular to a method for evaluating the intelligence level of an information system. Background Art
[0002] With the development of big data and artificial intelligence, the evaluation of the intelligence level of information systems has become a hot research issue. The difficulties of this issue are as follows:
[0003] (1) There are certain differences in the existing information system intelligence level evaluation indicator systems. Different indicator systems have different focuses, and it is very difficult to propose a universal evaluation method.
[0004] (2) Information systems are usually very complex, and evaluation indicators may affect each other and have different degrees of influence on the level of intelligence;
[0005] (3) Information systems are widely used, and people need to understand the relationship between the system and its level of intelligence. Therefore, the evaluation method needs to be interpretable and easy to understand.
[0006] In recent years, people have developed many algorithms for evaluating the intelligence level of information systems, but these algorithms all have some shortcomings. Some algorithms are based on structural relationships for evaluation, such as the hierarchical analysis method, the grey correlation analysis method, and the fuzzy comprehensive evaluation method. This type of method is simple to implement but has limited evaluation performance. There are problems such as difficulty in determining the weight of evaluation indicators and weak interpretability, and poor versatility. Some algorithms are based on machine learning algorithms for evaluation, including principal component analysis (PCA), Bayesian networks, generative adversarial networks (GANs), and deep neural networks (DNNs). This type of method has good evaluation performance but requires a large number of data samples, and most of them are black box models and are not interpretable. Summary of the invention
[0007] To solve the above problems, we proposed an information system intelligence level evaluation method based on the deep Topsis model and adaptive fuzzy inference network ANFIS. The data used comes from real information systems and related fields, and it has good versatility and interpretability.
[0008] In view of the shortcomings of the prior art, the present invention provides a method for evaluating the intelligence level of an information system, comprising the following steps:
[0009] Step 1: For the information system to be evaluated, obtain the score of its intelligence level evaluation index pair;
[0010] Step 2: Based on the information system intelligence level evaluation index system, use the deep Tposis model to conduct a comprehensive evaluation of the index pairs level by level to generate the evaluation of the highest level evaluation index;
[0011] Step 3, determining the parameters of the ANFIS adaptive inference neural network, the parameters including architecture and algorithm parameters;
[0012] Step 4: Use ANFIS adaptive reasoning neural network to analyze the highest level indicator evaluation and output the intelligence level score of the information system;
[0013] Step 5: According to the benchmark or the corresponding intelligence level grade model, the intelligence level score is mapped to the final information system intelligence level evaluation result.
[0014] An information system intelligence level evaluation system includes the following modules:
[0015] Topsis model building module: used to build a deep Topsis model, conduct a comprehensive evaluation of the evaluation indicators of the intelligence level of the information system, and generate the evaluation of the highest-level evaluation indicators;
[0016] ANFIS adaptive reasoning neural network module: used to determine the parameters of the ANFIS adaptive reasoning neural network;
[0017] Information system intelligence level scoring module: used to score the information system intelligence level evaluation index using the highest level evaluation index obtained by the Topsis model module, and use the ANFIS adaptive reasoning neural network module to output the information system intelligence level score.
[0018] Intelligent level grade output module: used to map the intelligent level score into the final information system intelligent level evaluation result.
[0019] Compared with the prior art, the present invention has the following significant advantages:
[0020] (1) The present invention adopts the deep Topsis model, which can combine the knowledge in related fields to conduct a comprehensive and objective analysis of the information system, and then make a more reasonable and reliable evaluation of the intelligence level;
[0021] (2) The present invention adopts the ANFIS adaptive fuzzy inference network, and the model parameters can be presented to the user in the form of rules through mapping, which has strong interpretability, thereby making a more confident intelligent level assessment;
[0022] (3) The present invention utilizes the relevant theories of fuzzy mathematics, so that the model can process different types of data and prior knowledge, and has strong versatility;
[0023] (4) The method of the present invention is simple to implement, can be easily integrated with other systems or software, and has strong scalability.
[0024] The intelligent level assessment method of the present invention is simple and intuitive, and has good versatility and interpretability.
[0025] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 The figure is a schematic diagram of the workflow of the method for evaluating the intelligence level of an information system of the present invention.
[0027] Figure 2 It is a schematic diagram of the workflow model of the information system intelligence level evaluation method of the present invention.
[0028] Figure 3 Schematic diagram of intelligent level evaluation index in an embodiment of the present invention.
[0029] Figure 4 It is the score of the intelligent level evaluation index pair obtained in the embodiment of the present invention.
[0030] Figure 5 to Figure 6 This is a result diagram of comprehensive evaluation of indicators at each level using the deep Tposis model in an embodiment of the present invention.
[0031] Figure 7 It is a radar chart output in an embodiment of the present invention.
[0032] Figure 8 It is a schematic diagram of the rule-structure parameter conversion mapping process in the present invention.
[0033] Fig. 9 Schematic diagram of different connection modes formed by the ANFIS network based on different examples in an embodiment of the present invention. DETAILED DESCRIPTION
[0034] Combination Figure 1 , a method for evaluating the intelligence level of an information system, comprising the following steps:
[0035] Step 1: For the information system to be evaluated, obtain the score of its intelligence level evaluation index pair, which specifically includes the following steps:
[0036] Step 1-1: Based on the intelligent level evaluation index system, A[i][j] is used to represent the j-th secondary index of the i-th primary index in the dimension A of the intelligent level evaluation index system. Indicators of different dimensions are combined using Cartesian products to form different indicator pairs, such as 〈A[1][2], B[2][1], ...>;
[0037] Step 1-2: Obtain the score of the lowest level indicator pair in the intelligent level evaluation indicator system of the information system to be evaluated.
[0038] Step 2: Based on the information system intelligence level evaluation index system, use the deep Tposis model to conduct a comprehensive evaluation of the indicators level by level to generate the evaluation of the highest level evaluation indicators, which specifically includes the following steps:
[0039] Step 2-1, the original matrix is positively transformed. The evaluation index types in the intelligent level evaluation index system include extremely large indexes, extremely small indexes, intermediate indexes and interval indexes. All types of indexes are uniformly transformed into extremely large indexes, and different forms of conversion functions are adopted according to different index types;
[0040] If the indicators of the information system intelligence level evaluation indicator system are all extremely large indicators, skip this step;
[0041] Step 2-2: Calculate the score of the Topsis model for the lowest-level indicator pair and normalize it. Assuming that the lowest-level intelligent evaluation indicator system contains m evaluation indicator pairs, the score vector x is:
[0042] x=[x1,x2,…,x m ]
[0043] For the evaluation score vector x, define the ideal solution x + , each component of which is the highest score of the corresponding indicator pair; define the negative ideal solution x - , each component of which is the minimum score of the corresponding indicator pair;
[0044] The object to be evaluated and x + The distance is the Euclidean distance between corresponding vectors, expressed as Its and x - The distance is expressed as The normalized score is The score is between 0 and 1, and the larger the score, the closer it is to the ideal solution;
[0045] Step 2-4, establish a Topsis model for the evaluation indicators level by level, repeat steps 2-1 to 2-3, and establish a deep Topsis model until a score evaluation for the highest-level indicator pair, i.e., the first-level indicator pair, is generated.
[0046] The radar chart is provided for reference based on the first-level indicators output by the Topsis model.
[0047] Step 3: Determine the parameters of the ANFIS adaptive inference neural network, which include architecture and algorithm parameters, and specifically include the following steps:
[0048] Step 3-1: Analyze relevant evaluation cases and abstract them into “if-then” rules.<A[1][2],B[2][1]> The closer it is to its corresponding ideal solution<A[1][2],B[2][1]> + Or (negative ideal solution), the higher (lower) the intelligence level y of the information system is, the conversion rule is as follows:
[0049] if D(<A[1][2],B[2][1]> ,<A[1][2],B[2][1]> + )→0,then y→y max ,
[0050] if D(<A[1][2],B[2][1]> ,<A[1][2],B[2][1]> - )→0,then y→y min
[0051] Among them, the “if-part” can be expanded through the “and” operation. This rule can be used for any level of indicators and reflects that some evaluation indicators are more important, that is, they have a greater impact on the intelligence level of the information system.
[0052] Step 3-2, determine the Gaussian membership function:
[0053] First, determine the fuzzy partition, and divide each primary indicator into two partitions, namely 1 and 0, through the Gaussian membership function, expressed as:
[0054]
[0055] Where d = 1 or 0 is used to distinguish the partition, the value of σ is 1, and {σ, d} is the initial value of the antecedent parameter in the algorithm parameters of the ANFIS adaptive reasoning network;
[0056] Step 3-3: Based on the transformed rule set R = {r1, ..., r m}, first r i The evaluation index pairs involved in the "if-part" in the formula are converted into scores of the corresponding levels, where the scores of the indicators involved are replaced by positive ideal values or negative ideal values, and the scores of the indicator pairs not involved are replaced by average values or default values. Then, the deep Topsis model is used to convert them into the corresponding first-level indicator pair scores. Finally, the Gaussian membership function is used to determine whether it belongs to partition 0 or partition 1, so that the membership calculation node with the strongest response is connected to the rule node to complete the mapping of a rule. Finally, all rule mappings are converted into network architectures to obtain the architecture parameters of the ANFIS adaptive fuzzy reasoning network.
[0057] Step 3-4: Based on the results of the above rule transformation, r i The "then-part" in the formula is converted into the corresponding ideal or negative ideal information system intelligent score, and the least squares estimation is used for training to obtain the algorithm parameters of the ANFIS adaptive fuzzy inference network.
[0058] Step 4: Use ANFIS adaptive reasoning neural network to analyze the highest level indicator evaluation and output the intelligence level score of the information system, specifically:
[0059] By using the ANFIS adaptive fuzzy inference network and the parameters of the ANFIS adaptive fuzzy inference network obtained in step 3, the scores of the highest-level evaluation indicators obtained in step 2 are analyzed, and the intelligence level score of the information system to be evaluated is inferred and output.
[0060] Step 5: According to the benchmark or the corresponding intelligence level grade model, the intelligence level score is mapped to the final information system intelligence level evaluation result.
[0061] An information system intelligence level evaluation system includes the following modules:
[0062] Topsis model building module: used to build a deep Topsis model, conduct a comprehensive evaluation of the evaluation indicators of the intelligence level of the information system, and generate the evaluation of the highest-level evaluation indicators;
[0063] ANFIS adaptive reasoning neural network module: used to determine the parameters of the ANFIS adaptive reasoning neural network;
[0064] Information system intelligence level scoring module: used to score the information system intelligence level evaluation index using the highest level evaluation index obtained by the Topsis model module, and use the ANFIS adaptive reasoning neural network module to output the information system intelligence level score.
[0065] Intelligent level grade output module: used to map the intelligent level score into the final information system intelligent level evaluation result.
[0066] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:
[0067] Step 1: For the information system to be evaluated, obtain the score of its intelligence level evaluation index pair;
[0068] Step 2: Based on the information system intelligence level evaluation index system, use the deep Tposis model to conduct a comprehensive evaluation of the index pairs level by level to generate the evaluation of the highest level evaluation index;
[0069] Step 3, determining the parameters of the ANFIS adaptive inference neural network, the parameters including architecture and algorithm parameters;
[0070] Step 4: Use ANFIS adaptive reasoning neural network to analyze the highest level indicator evaluation and output the intelligence level score of the information system;
[0071] Step 5: According to the benchmark or the corresponding intelligence level grade model, the intelligence level score is mapped to the final information system intelligence level evaluation result.
[0072] A computer storable medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:
[0073] Step 1: For the information system to be evaluated, obtain the score of its intelligence level evaluation index pair;
[0074] Step 2: Based on the information system intelligence level evaluation index system, use the deep Tposis model to conduct a comprehensive evaluation of the index pairs level by level to generate the evaluation of the highest level evaluation index;
[0075] Step 3, determining the parameters of the ANFIS adaptive inference neural network, the parameters including architecture and algorithm parameters;
[0076] Step 4: Use ANFIS adaptive reasoning neural network to analyze the highest level indicator evaluation and output the intelligence level score of the information system;
[0077] Step 5: According to the benchmark or the corresponding intelligence level grade model, the intelligence level score is mapped to the final information system intelligence level evaluation result.
[0078] The present invention will be further described below in conjunction with the embodiments.
[0079] Example
[0080] exist The experiment was conducted in CoreTMi7-8700 CPU@3.20GHz 3.19GHz environment, and the running software environment was Python3.7.
[0081] Combination Figure 1 and Figure 2 , a method for evaluating the intelligence level of an information system, comprising the following steps:
[0082] Step 1: For the information system to be evaluated, obtain the score of its intelligence level evaluation index pair, which specifically includes the following steps:
[0083] Step 1-1: Based on the intelligent level evaluation index system, use A[i][j] to represent the jth secondary index of the i-th primary index in dimension A of the intelligent level evaluation index system. Indicators of different dimensions are combined using Cartesian products to form different indicator pairs, such as<A[1][2],B[2][1],...> ,like Figure 3 As shown;
[0084] Step 1-2: Obtain the score of the lowest level indicator pair in the intelligent level evaluation indicator system of the information system to be evaluated, such as Figure 4 shown.
[0085] Step 2: Based on the information system intelligence level evaluation index system, use the deep Tposis model to conduct a comprehensive evaluation of the indicators level by level to generate the evaluation of the highest level evaluation indicators, which specifically includes the following steps:
[0086] Step 2-1, the original matrix is positively transformed. The evaluation index types in the intelligent level evaluation index system include extremely large indexes, extremely small indexes, intermediate indexes and interval indexes. All types of indexes are uniformly transformed into extremely large indexes, and different forms of conversion functions are adopted according to different index types;
[0087] If the indicators of the information system intelligence level evaluation indicator system are all extremely large indicators, skip this step;
[0088] Step 2-2: Calculate the score of the Topsis model for the lowest-level indicator pair and normalize it. Assuming that the lowest-level intelligent evaluation indicator system contains m evaluation indicator pairs, the score vector x is:
[0089] x=[x1,x2,…,x m ]
[0090] For the evaluation score vector x, define the ideal solution x + , each component of which is the highest score of the corresponding indicator pair; define the negative ideal solution x - , each component of which is the minimum score of the corresponding indicator pair;
[0091] The object to be evaluated and x + The distance is the Euclidean distance between corresponding vectors, expressed as Its and x - The distance is expressed as The normalized score is The score is between 0 and 1, and the larger the score, the closer it is to the ideal solution;
[0092] Step 2-4, build a Topsis model for the evaluation indicators level by level, repeat steps 2-1 to 2-3, build a deep Topsis model, until the highest level indicator pair, i.e., the score evaluation of the first level indicator pair, is generated, such as Figure 5 to Figure 6 shown.
[0093] According to the Topsis model, the first-level indicators are output for reference on the radar chart, such as Figure 7 shown.
[0094] Step 3: Determine the parameters of the ANFIS adaptive inference neural network, which include architecture and algorithm parameters, and specifically include the following steps:
[0095] Step 3-1: Analyze relevant evaluation cases and abstract them into "if-then" rules to give the following rules:
[0096] if D(<A[1][2],B[2][1]> ,<A[1][2],B[2][1]> + )→0
[0097] and D(<A[3][1],B[2][2]> ,<A[3][1],B[2][2]> + )→0
[0098] and D(<A[2][2],B[1][4]> ,<A[2][2],B[1][4]> + )→0
[0099] and D(<A[1][3],B[3][1]> ,<A[1][3],B[3][1]> + )→0
[0100] then y→y max ,
[0101] if D(<A[3][2],B[1][1]> ,<A[3][2],B[1][1]> - )→0
[0102] and D(<A[2][3],B[2][2]> ,<A[2][3],B[2][2]> - )→0
[0103] then y→y min
[0104] Rules-Sample 1 (including two items)
[0105] if D(<A[1][2],B[2][1]> ,<A[1][2],B[2][1]>+ )→0
[0106] then y→y max ,
[0107] if D(<A[2][2],B[1][4]> ,<A[2][2],B[1][4]> + )→0
[0108] then y→y max ,
[0109] if D(<A[3][2],B[1][1]> ,<A[3][2],B[1][1]> )→0
[0110] then y→y min
[0111] Rules-Sample 2 (including three rules)
[0112] Step 3-2, determine the Gaussian membership function:
[0113] First, determine the fuzzy partition, and divide each primary indicator into two partitions, namely 1 and 0, through the Gaussian membership function, expressed as:
[0114]
[0115] Where d = 1 or 0 is used to distinguish the partition, the value of σ is 1, and {σ, d} is the initial value of the antecedent parameter in the algorithm parameters of the ANFIS adaptive reasoning network;
[0116] Step 3-3: Based on the transformed rule set R = {r1, ..., r m}, first r i The evaluation index pairs involved in the "if-part" in the formula are converted into scores of the corresponding levels, where the scores of the indicators involved are replaced by positive ideal values or negative ideal values, and the scores of the indicator pairs not involved are replaced by average values or default values. Then, the deep Topsis model is used to convert them into the corresponding first-level indicator pair scores. Finally, the Gaussian membership function is used to determine whether it belongs to partition 0 or partition 1, so that the membership calculation node with the strongest response is connected to the rule node to complete the mapping of a rule. Finally, all rule mappings are converted into network architectures to obtain the architecture parameters of the ANFIS adaptive fuzzy reasoning network.
[0117] like Figure 8As shown in the figure, the depth of the color indicates the score (0-1), which reflects the importance of the rule to the indicator; the circular node represents the membership calculation, and each component contains two fuzzy partitions (upper: 0 partition, lower: 1 partition). The higher the membership, the greater the probability of belonging to this partition; finally, the membership calculation node with the strongest response is connected to the rule node to complete the mapping of a rule.
[0118] Specific results such as Fig. 9 As shown, Fig. 9 The left picture is sample one, and the right picture is sample two.
[0119] Step 3-4: Based on the results of the above rule transformation, r i The "then-part" in the figure is converted into the corresponding ideal or negative ideal information system intelligent score, where the ideal score is 100 and the negative ideal score is 1. The least squares estimation is used for training to obtain the algorithm parameters of the ANFIS adaptive fuzzy inference network.
[0120] Step 4: Use ANFIS adaptive reasoning neural network to analyze the highest level indicator evaluation and output the intelligence level score of the information system, specifically:
[0121] Using the ANFIS adaptive fuzzy inference network and the parameters of the ANFIS adaptive fuzzy inference network obtained in step 3, the scores of the highest-level evaluation indicators obtained in step 2 are analyzed, and the intelligence level scores of the information system to be evaluated are inferred and output;
[0122] It can be obtained that the intelligent score obtained based on the rules in sample one is 17.44, and the intelligent score obtained based on the rules in sample two is 29.32.
[0123] Step 5: According to the benchmark or the corresponding intelligent level model, the intelligent level score is mapped to the final information system intelligent level evaluation result. The intelligent level score can be converted into an intelligent level according to the needs, for example:
[0124] Intelligent level Corresponding score Level 1 1-10 Level 2 11-20 Level 3 21-30 Level 4 31-40 Level 5 41-50 Level 6 51-60 Level 7 61-70 Level 8 71-80 Level 9 80 and above
[0125] The present invention adopts the deep Topsis model and ANFIS adaptive fuzzy reasoning network, which can conduct a comprehensive and objective analysis of the information system, and then make a more reasonable and reliable evaluation of the intelligence level, with good versatility and interpretability.
Claims
1. A method for evaluating the intelligence level of information systems. It is characterized in that The following steps are involved: Step 1: For the information system to be evaluated, obtain the score of its intelligence level evaluation index pair; Step 2: Based on the information system intelligence level evaluation index system, use the deep Tposis model to conduct a comprehensive evaluation of the index pairs level by level to generate the evaluation of the highest level evaluation index; Step 3, determining the parameters of the ANFIS adaptive inference neural network, the parameters including architecture and algorithm parameters; Step 4: Use ANFIS adaptive reasoning neural network to analyze the highest level indicator evaluation and output the intelligence level score of the information system; Step 5: According to the benchmark or the corresponding intelligence level grade model, the intelligence level score is mapped to the final information system intelligence level evaluation result.
2. The method for evaluating the intelligence level of an information system according to claim 1, It is characterized in that The step 1 of scoring the intelligent level evaluation index of the information system to be evaluated specifically includes the following steps: Step 1-1: Based on the intelligent level evaluation index system, A[i][j] is used to represent the j-th secondary index of the i-th primary index in the dimension A of the intelligent level evaluation index system. Indicators of different dimensions are combined using Cartesian products to form different indicator pairs. Step 1-2: Obtain the score of the lowest level indicator pair in the intelligent level evaluation indicator system of the information system to be evaluated.
3. The method for evaluating the intelligence level of an information system according to claim 1, It is characterized in that The step 2 uses the deep Tposis model to comprehensively evaluate the indicators step by step, specifically including the following steps: Step 2-1, the original matrix is positively transformed. The evaluation index types in the intelligent level evaluation index system include extremely large indexes, extremely small indexes, intermediate indexes and interval indexes. All types of indexes are uniformly transformed into extremely large indexes, and different forms of conversion functions are adopted according to different index types; If the indicators of the information system intelligence level evaluation indicator system are all extremely large indicators, skip this step; Step 2-2: Calculate the score of the Topsis model for the lowest-level indicator pair and normalize it. Assuming that the lowest-level intelligent evaluation indicator system contains m evaluation indicator pairs, the score vector x is: x=[x 1 ,x 2 ,…,x m ] For the evaluation score vector x, define the ideal solution x + , each component of which is the highest score of the corresponding indicator pair; define the negative ideal solution x - , each component of which is the minimum score of the corresponding indicator pair; The object to be evaluated and x + The distance is the Euclidean distance between corresponding vectors, expressed as Its and x - The distance is expressed as The normalized score is The score is between 0 and 1, and the larger the score, the closer it is to the ideal solution; Step 2-4, establish a Topsis model for the evaluation indicators level by level, repeat steps 2-1 to 2-3, and establish a deep Topsis model until a score evaluation for the highest-level indicator pair, i.e., the first-level indicator pair, is generated.
4. The method for evaluating the intelligence level of an information system according to claim 1, It is characterized in that Determining the parameters of the ANFIS adaptive reasoning network in step 3 specifically includes the following steps: Step 3-1: Analyze relevant evaluation cases and abstract them into "if-then" rules; Step 3-2, determine the Gaussian membership function: First, determine the fuzzy partition, and divide each primary indicator into two partitions, namely 1 and 0, through the Gaussian membership function, expressed as: Where d = 1 or 0 is used to distinguish the partition, the value of σ is 1, and {σ, d} is the initial value of the antecedent parameter in the algorithm parameters of the ANFIS adaptive reasoning network; Step 3-3, based on the transformed rule set R = {r 1 ,...,r m }, first r i The evaluation index pairs involved in the "if-part" in the formula are converted into scores of the corresponding levels, where the scores of the indicators involved are replaced by positive ideal values or negative ideal values, and the scores of the indicator pairs not involved are replaced by average values or default values. Then, the deep Topsis model is used to convert them into the corresponding first-level indicator pair scores. Finally, the Gaussian membership function is used to determine whether it belongs to partition 0 or partition 1, so that the membership calculation node with the strongest response is connected to the rule node to complete the mapping of a rule. Finally, all rule mappings are converted into network architectures to obtain the architecture parameters of the ANFIS adaptive fuzzy reasoning network. Step 3-4: Based on the results of the above rule transformation, r i The "then-part" in the formula is converted into the corresponding ideal or negative ideal information system intelligent score, and the least squares estimation is used for training to obtain the algorithm parameters of the ANFIS adaptive fuzzy inference network.
5. The method for evaluating the intelligence level of an information system according to claim 1, It is characterized in that The output information system intelligence level score in step 4 is specifically: By using the ANFIS adaptive fuzzy inference network and the parameters of the ANFIS adaptive fuzzy inference network obtained in step 3, the scores of the highest-level evaluation indicators obtained in step 2 are analyzed, and the intelligence level score of the information system to be evaluated is inferred and output.
6. An information system intelligence level evaluation system, It is characterized in that Includes the following modules: Topsis model building module: used to build a deep Topsis model, conduct a comprehensive evaluation of the evaluation indicators of the intelligence level of the information system, and generate the evaluation of the highest-level evaluation indicators; ANFIS adaptive reasoning neural network module: used to determine the parameters of the ANFIS adaptive reasoning neural network; Information system intelligence level scoring module: used to use the highest level evaluation index obtained by the Topsis model module to score the information system intelligence level evaluation index, and use the ANFIS adaptive reasoning neural network module to output the information system intelligence level score; Intelligent level grade output module: used to map the intelligent level score into the final information system intelligent level evaluation result.
7. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.
8. A computer storable medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.
