Algorithm for automatically analyzing and judging problems of intelligent passing system

Abstract

The invention discloses an algorithm for automatically analyzing and judging problems of an intelligent passage system, and the algorithm comprises the following steps: S1, collecting the log and performance index data of the intelligent passage system, monitoring the real-time operation states of software and hardware in the intelligent passage system in real time, carrying out the cleaning, denoising and normalization processing of the collected data, and obtaining the real-time operation state of the software and hardware in the intelligent passage system; the data quality and consistency are ensured; and S2, selecting a support vector machine as an algorithm model, training the selected support vector machine algorithm model by using the historical data processed in the S1, and improving the accuracy and solving capability of the support vector machine algorithm model. According to the method, the software and hardware running states of the intelligent traffic system are monitored in real time through the support vector machine algorithm model, the states of the intelligent traffic system are distinguished, personnel can be reminded in time and a solution can be given in the abnormal state, the personnel can find and process the abnormal state in time conveniently, and the use requirement is met.

Description

Technical Field

[0001] The present invention relates to the technical field of smart passage systems, and in particular to an algorithm for automatically analyzing and determining problems in smart passage systems. Background Art

[0002] Smart access systems have always been difficult to maintain, especially when it comes to toll collection. The system requirements are relatively high, including software and hardware system stability and reliability, network environment stability, service response timeliness, etc. When problems arise during system operation, personnel are unable to discover and handle them in a timely manner. The prior art also lacks methods for analyzing and judging smart access systems. For this reason, this application proposes an algorithm for automatically analyzing and judging problems in smart access systems. Summary of the invention

[0003] Based on the technical problems existing in the background technology, the present invention proposes an algorithm for automatically analyzing and judging problems of the smart passage system.

[0004] The present invention proposes an algorithm for automatically analyzing and determining problems of a smart traffic system, comprising the following steps:

[0005] S1: Collect the logs and performance indicator data of the smart access system, monitor the real-time operation status of the software and hardware in the smart access system, and clean, denoise and normalize the collected data to ensure data quality and consistency;

[0006] S2: Select support vector machine as the algorithm model, and use the historical data processed in S1 to train the selected support vector machine algorithm model to improve the accuracy and solution ability of the support vector machine algorithm model;

[0007] S3: Input the real-time operation data of the smart access system into the trained support vector machine algorithm model. The support vector machine algorithm model classifies the operation status of the software and hardware in the smart access system and distinguishes between stable and abnormal states.

[0008] S4: Use big data analysis technology to identify and detect abnormal conditions identified in S3, analyze routine problems and professional problems, classify and organize different problems, and automatically generate solutions for different problems, quickly and effectively handle routine problems, and conduct model training for professional problems to improve problem-solving capabilities;

[0009] S5: Record the operation process of the analysis and solution in S4, form an operation optimization record for reference, and continuously monitor the operation of the support vector machine algorithm model. Adjust the parameters of the support vector machine algorithm model according to the operation optimization record to optimize the maintenance of the smart traffic system;

[0010] S6: Generate prediction capabilities based on the use of the support vector machine algorithm model to prevent processing and early warning notifications for human intervention.

[0011] Preferably, in S1, the logs include error logs, warning logs, and information logs; the performance indicator data include hardware performance indicator data and software performance indicator data; the hardware performance indicator data include CPU utilization, memory occupancy, disk read and write speed, and network bandwidth utilization; the software performance indicator data include response time, average load, and number of concurrent connections.

[0012] Preferably, in S2, the specific steps of training the support vector machine algorithm model are as follows:

[0013] S201: extracting features from the processed data to describe the operating status of various aspects of the smart access system, including CPU utilization, memory usage, network latency, etc.;

[0014] S202: Divide the extracted feature data into a training set and a test set for training and verifying the model;

[0015] S203: Using the training set data to train the support vector machine algorithm model, determine the best hyperplane, and maximize the intervals between different categories;

[0016] S204: Use the test set to evaluate the trained model, calculate the accuracy, recall, and precision indicators to determine the effect of the model, and use the cross-validation method to evaluate the generalization ability of the model to ensure the performance of the model on unknown data;

[0017] The formula used to calculate accuracy is:

[0018] The calculation formula for recall is:

[0019] The calculation formula of precision index is:

[0020] Where TP is the number of samples correctly predicted as positive, TN is the number of samples correctly predicted as negative, FP is the number of samples incorrectly predicted as positive, and FN is the number of samples incorrectly predicted as negative.

[0021] S205: Adjust and optimize the support vector machine algorithm model according to the evaluation results.

[0022] Preferably, in S3, when distinguishing between a stable state and an abnormal state, the support vector machine algorithm finds the best segmentation hyperplane in the input data to maximize the interval between different categories, and classifies the data points according to the relationship between the hyperplane, identifies the stable state and the abnormal state, and marks whether the smart traffic system is currently in an abnormal state;

[0023] The mathematical expression used to identify stable and abnormal states is:

[0024] f(x)=sign(Σα_i*y_i*K(x_i,x)+b);

[0025] Where f(x) is the decision function, x is the feature vector of the input sample, α_i is the Lagrange multiplier of the support vector, y_i is the category label of the corresponding support vector, K(x_i,x) is the calculation result of the function, which represents the similarity between the support vector x_i and the input sample x, and b is the bias term;

[0026] By inputting the new sample feature vector x into the decision function, the positive and negative signs of the function output are used to determine which category the sample belongs to. If the function value is greater than zero, it is predicted to be a positive class (i.e., a stable state); if the function value is less than zero, it is predicted to be a negative class (i.e., an abnormal state).

[0027] Preferably, in S4, the specific logical steps for analyzing routine problems and professional problems are as follows:

[0028] S401: Extract features from abnormal status data to extract key features that can distinguish between routine problems and professional problems, including various performance indicators, operation status, and processing time;

[0029] S402: Perform cluster analysis on the abnormal state data using a clustering algorithm, and cluster data points with high similarity together to form different groups;

[0030] S403: Based on the result of data analysis, the abnormal status data is divided into two groups: routine problems and professional problems.

[0031] Preferably, in S4, conventional problems can be automatically processed according to the pre-trained support vector machine algorithm model, and the conventional problems can be organized into fixed templates with standard solutions attached. For conventional problems, solutions are directly given. In addition, the optimal solution can be searched online based on the conventional problems, and the optimal solution is input into the fixed template for optimization and update;

[0032] For professional problems, we can collect processing records and solutions for new problems, and input their data into the support vector machine algorithm model to train the model and optimize the solution templates in the model. When professional problems are input, the optimized solution templates are extracted based on the characteristics of the input problems, and highly adaptable solutions are automatically generated.

[0033] Preferably, the specific logical steps of S6 are as follows:

[0034] S601: Continuously monitor the software and hardware operating status of the smart access system in real time and collect real-time data;

[0035] S602: extracting feature data from real-time data, and inputting the feature data into a trained support vector machine algorithm model, and performing prediction through the decision function of the support vector machine algorithm model. During the prediction, the data point is predicted to belong to a stable state or an abnormal state according to its position in the feature space, and a prediction result is generated;

[0036] S603: Based on the prediction results, the support vector machine algorithm model finds that the smart access system is in an abnormal state, triggers the early warning system, and sends a notification to relevant personnel. At the same time, manual intervention is recommended.

[0037] Preferably, the specific steps of S402 are as follows:

[0038] S4021: Randomly select K center points (c_1, c_2, ..., c_K) as the initial cluster centers;

[0039] S4022: For each data point (x_i), calculate its distance from the center point of each cluster and assign it to the cluster with the closest distance;

[0040] S4023: For each cluster (k), calculate the mean of all its member data points as the new cluster center (c_k), the formula is:

[0041] Where C_K represents the center of the kth cluster, |S K | represents the number of sample points belonging to the Kth cluster, Represents the sum of all data points belonging to the Kth cluster;

[0042] S4024: Repeat steps S4022 and S4023 until the cluster center no longer changes or a convergence condition is reached.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention uses a support vector machine algorithm model to monitor the software and hardware operating status of the smart access system in real time, distinguish the status of the smart access system, and can promptly remind personnel and provide solutions in abnormal states, making it convenient for personnel to discover and handle them in a timely manner to meet usage requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of an algorithm proposed by the present invention for automatically analyzing and determining problems of a smart traffic system. DETAILED DESCRIPTION

[0046] The present invention will be further explained below in conjunction with specific embodiments.

[0047] Example

[0048] Reference Figure 1 This embodiment proposes an algorithm for automatically analyzing and determining problems of a smart traffic system, including the following steps:

[0049] S1: Collect the logs and performance indicator data of the smart access system, monitor the real-time operating status of the software and hardware in the smart access system, and clean, denoise and normalize the collected data to ensure data quality and consistency. The smart access system usually records various events and operation logs during system operation. The operating status of the system software and hardware can be understood in real time by monitoring system log records;

[0050] The logs include error logs, warning logs, and information logs. The performance indicator data include hardware performance indicator data and software performance indicator data. The hardware performance indicator data include CPU utilization, memory usage, disk read and write speed, and network bandwidth utilization. The software performance indicator data include response time, average load, and number of concurrent connections.

[0051] S2: Select support vector machine as the algorithm model, and use the historical data processed in S1 to train the selected support vector machine algorithm model to improve the accuracy and solution ability of the support vector machine algorithm model;

[0052] The specific steps are as follows:

[0053] S201: extracting features from the processed data to describe the operating status of various aspects of the smart access system, including CPU utilization, memory usage, network latency, etc.;

[0054] S202: Divide the extracted feature data into a training set and a test set for training and verifying the model;

[0055] S203: Using the training set data to train the support vector machine algorithm model, determine the best hyperplane, and maximize the intervals between different categories;

[0056] S204: Use the test set to evaluate the trained model, calculate the accuracy, recall, and precision indicators to determine the effect of the model, and use the cross-validation method to evaluate the generalization ability of the model to ensure the performance of the model on unknown data;

[0057] The formula used to calculate accuracy is:

[0058] The calculation formula for recall is:

[0059] The calculation formula of precision index is:

[0060] Where TP is the number of samples correctly predicted as positive, TN is the number of samples correctly predicted as negative, FP is the number of samples incorrectly predicted as positive, and FN is the number of samples incorrectly predicted as negative.

[0061] S205: adjusting and optimizing the support vector machine algorithm model according to the evaluation results;

[0062] S3: Input the real-time operation data of the smart access system into the trained support vector machine algorithm model. The support vector machine algorithm model classifies the operation status of the software and hardware in the smart access system and distinguishes between stable and abnormal states.

[0063] When distinguishing between stable and abnormal states, the support vector machine algorithm finds the best segmentation hyperplane in the input data to maximize the interval between different categories, and classifies the data points according to their relationship with the hyperplane, identifying stable and abnormal states, and marking whether the smart traffic system is currently in an abnormal state.

[0064] The mathematical expression used to identify stable and abnormal states is:

[0065] f(x)=sign(Σα_i*y_i*K(x_i,x)+b);

[0066] Where f(x) is the decision function, x is the feature vector of the input sample, α_i is the Lagrange multiplier of the support vector, y_i is the category label of the corresponding support vector, K(x_i,x) is the calculation result of the function, which represents the similarity between the support vector x_i and the input sample x, and b is the bias term;

[0067] By inputting the new sample feature vector x into the decision function, the positive and negative signs of the function output are used to determine which category the sample belongs to. If the function value is greater than zero, it is predicted to be a positive class (i.e., a stable state); if the function value is less than zero, it is predicted to be a negative class (i.e., an abnormal state);

[0068] S4: Use big data analysis technology to identify and detect abnormal conditions identified in S3, analyze routine problems and professional problems, classify and organize different problems, and automatically generate solutions for different problems, quickly and effectively handle routine problems, and conduct model training for professional problems to improve problem-solving capabilities;

[0069] The specific logical steps for analyzing routine and professional issues are as follows:

[0070] S401: Extract features from abnormal status data to extract key features that can distinguish between routine problems and professional problems, including various performance indicators, operation status, and processing time;

[0071] S402: Perform cluster analysis on the abnormal state data using a clustering algorithm, and cluster data points with high similarity together to form different groups;

[0072] The specific steps are as follows:

[0073] S4021: Randomly select K center points (c_1, c_2, ..., c_K) as the initial cluster centers;

[0074] S4022: For each data point (x_i), calculate its distance from the center point of each cluster and assign it to the cluster with the closest distance;

[0075] S4023: For each cluster (k), calculate the mean of all its member data points as the new cluster center (c_k), the formula is:

[0076] Where C_K represents the center of the kth cluster, |S K | represents the number of sample points belonging to the Kth cluster, Represents the sum of all data points belonging to the Kth cluster;

[0077] S4024: repeat steps S4022 and S4023 until the cluster center no longer changes or convergence conditions are reached;

[0078] S403: Based on the result of data analysis, the abnormal status data is divided into two groups: routine problems and professional problems;

[0079] For routine problems, it can automatically process them according to the pre-trained support vector machine algorithm model, organize the routine problems into fixed templates, attach standard solutions, and directly give solutions to routine problems. In addition, it can also search for the best solution online based on routine problems, and input the best solution into the fixed template for optimization and update;

[0080] For professional problems, we can collect the processing records and solutions of new problems, and input their data into the support vector machine algorithm model to train the model and optimize the solution template in the model. When professional problems are input, we can extract the optimized solution template according to the characteristics of the input problems and automatically generate highly adaptable solutions.

[0081] S5: Record the operation process of the analysis and solution in S4, form an operation optimization record for reference, and continuously monitor the operation of the support vector machine algorithm model. Adjust the parameters of the support vector machine algorithm model according to the operation optimization record to optimize the maintenance of the smart traffic system;

[0082] S6: Generate prediction capabilities based on the use of the support vector machine algorithm model, and prevent processing and early warning notification of manual intervention in advance. The specific logical steps are as follows:

[0083] S601: Continuously monitor the software and hardware operating status of the smart access system in real time and collect real-time data;

[0084] S602: extracting feature data from real-time data, and inputting the feature data into a trained support vector machine algorithm model, and performing prediction through the decision function of the support vector machine algorithm model. During the prediction, the data point is predicted to belong to a stable state or an abnormal state according to its position in the feature space, and a prediction result is generated;

[0085] S603: Based on the prediction results, the support vector machine algorithm model finds that the smart access system is in an abnormal state, triggers the early warning system, and sends a notification to relevant personnel. At the same time, manual intervention is recommended;

[0086] This embodiment uses a support vector machine algorithm model to monitor the software and hardware operating status of the smart access system in real time, distinguish the status of the smart access system, and can promptly remind personnel and provide solutions in abnormal states, making it convenient for personnel to discover and handle them in a timely manner to meet usage requirements.

[0087] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An algorithm for automatically analyzing and determining problems in a smart traffic system, characterized in that: The following steps are involved: S1: Collect the logs and performance indicator data of the smart access system, monitor the real-time operation status of the software and hardware in the smart access system, and clean, denoise and normalize the collected data to ensure data quality and consistency; S2: Select support vector machine as the algorithm model, and use the historical data processed in S1 to train the selected support vector machine algorithm model to improve the accuracy and solution ability of the support vector machine algorithm model; S3: Input the real-time operation data of the smart access system into the trained support vector machine algorithm model. The support vector machine algorithm model classifies the operation status of the software and hardware in the smart access system and distinguishes between stable and abnormal states. S4: Use big data analysis technology to identify and detect abnormal conditions identified in S3, analyze routine problems and professional problems, classify and organize different problems, and automatically generate solutions for different problems, quickly and effectively handle routine problems, and conduct model training for professional problems to improve problem-solving capabilities; S5: Record the operation process of the analysis and solution in S4, form an operation optimization record for reference, and continuously monitor the operation of the support vector machine algorithm model. Adjust the parameters of the support vector machine algorithm model according to the operation optimization record to optimize the maintenance of the smart traffic system; S6: Generate prediction capabilities based on the use of the support vector machine algorithm model to prevent processing and early warning notifications for human intervention.

2. The algorithm for automatically analyzing and judging problems of a smart traffic system according to claim 1, characterized in that: In S1, the logs include error logs, warning logs, and information logs. The performance indicator data include hardware performance indicator data and software performance indicator data. The hardware performance indicator data include CPU utilization, memory occupancy, disk read and write speed, and network bandwidth utilization. The software performance indicator data include response time, average load, and number of concurrent connections.

3. The algorithm for automatically analyzing and judging problems of a smart traffic system according to claim 1, characterized in that: In S2, the specific steps of training the support vector machine algorithm model are as follows: S201: extracting features from the processed data to describe the operating status of various aspects of the smart access system, including CPU utilization, memory usage, and network latency; S202: Divide the extracted feature data into a training set and a test set for training and verifying the model; S203: Using the training set data to train the support vector machine algorithm model, determine the best hyperplane, and maximize the intervals between different categories; S204: Use the test set to evaluate the trained model, calculate the accuracy, recall, and precision indicators to determine the effect of the model, and use the cross-validation method to evaluate the generalization ability of the model to ensure the performance of the model on unknown data; The formula used to calculate accuracy is: The calculation formula for recall is: The calculation formula of precision index is: Where TP is the number of samples correctly predicted as positive, TN is the number of samples correctly predicted as negative, FP is the number of samples incorrectly predicted as positive, and FN is the number of samples incorrectly predicted as negative. S205: Adjust and optimize the support vector machine algorithm model according to the evaluation results.

4. The algorithm for automatically analyzing and judging problems of a smart traffic system according to claim 1, characterized in that: In S3, when distinguishing between a stable state and an abnormal state, the support vector machine algorithm finds the best segmentation hyperplane in the input data to maximize the interval between different categories, and classifies the data points according to the relationship between the hyperplane, identifies the stable state and the abnormal state, and marks whether the smart traffic system is currently in an abnormal state; The mathematical expression used to identify stable and abnormal states is: f(x)=sign(Σα_i*y_i*K(x_i,x)+b); Where f(x) is the decision function, x is the feature vector of the input sample, α_i is the Lagrange multiplier of the support vector, y_i is the category label of the corresponding support vector, K(x_i,x) is the calculation result of the function, which represents the similarity between the support vector x_i and the input sample x, and b is the bias term; By inputting the new sample feature vector x into the decision function, the positive and negative signs of the function output are used to determine which category the sample belongs to. If the function value is greater than zero, it is predicted to be a positive class; if the function value is less than zero, it is predicted to be a negative class.

5. The algorithm for automatically analyzing and determining problems of a smart traffic system according to claim 1 is characterized in that: In S4, the specific logical steps for analyzing routine and professional issues are as follows: S401: Extract features from abnormal status data to extract key features that can distinguish between routine problems and professional problems, including various performance indicators, operation status, and processing time; S402: Perform cluster analysis on the abnormal state data using a clustering algorithm, and cluster data points with high similarity together to form different groups; S403: Based on the result of data analysis, the abnormal status data is divided into two groups: routine problems and professional problems.

6. The algorithm for automatically analyzing and determining problems of a smart traffic system according to claim 1, characterized in that: In S4, routine problems can be automatically processed according to the pre-trained support vector machine algorithm model, and the routine problems can be organized into fixed templates with standard solutions attached. For routine problems, solutions are directly given. In addition, the optimal solution can be searched online based on the routine problems, and the optimal solution is input into the fixed template for optimization and update; For professional problems, we can collect processing records and solutions for new problems, and input their data into the support vector machine algorithm model to train the model and optimize the solution templates in the model. When professional problems are input, the optimized solution templates are extracted based on the characteristics of the input problems, and highly adaptable solutions are automatically generated.

7. The algorithm for automatically analyzing and determining problems of a smart traffic system according to claim 1, characterized in that: The specific logical steps of S6 are as follows: S601: Continuously monitor the software and hardware operating status of the smart access system in real time and collect real-time data; S602: extracting feature data from real-time data, and inputting the feature data into a trained support vector machine algorithm model, and performing prediction through the decision function of the support vector machine algorithm model. During the prediction, the data point is predicted to belong to a stable state or an abnormal state according to its position in the feature space, and a prediction result is generated; S603: Based on the prediction results, the support vector machine algorithm model finds that the smart access system is in an abnormal state, triggers the early warning system, and sends a notification to relevant personnel. At the same time, manual intervention is recommended.

8. The algorithm for automatically analyzing and determining problems of a smart traffic system according to claim 5, characterized in that: The specific steps of S402 are as follows: S4021: Randomly select K center points (c_1, c_2, ..., c_K) as the initial cluster centers; S4022: For each data point (x_i), calculate its distance from the center point of each cluster and assign it to the cluster with the closest distance; S4023: For each cluster (k), calculate the mean of all its member data points as the new cluster center (c_k), the formula is: Where C_K represents the center of the kth cluster, |S K | represents the number of sample points belonging to the Kth cluster, X i Represents the sum of all data points belonging to the Kth cluster; S4024: Repeat steps S4022 and S4023 until the cluster center no longer changes or a convergence condition is reached.

Images