Smart campus management method and system based on big data

Abstract

The invention discloses a smart campus management method and system based on big data, and particularly relates to the technical field of data management. The method comprises the following steps: deploying various types and numbers of sensors, collecting delay and throughput data in a peak period in real time in different monitoring areas in a campus, carrying out preprocessing and feature extraction on the data, and respectively extracting a delay abnormal fluctuation trend and a sudden reduction feature of a throughput decline rate; a machine learning model is constructed to predict a real-time response capability value of the smart campus management system, and the real-time response capability value is compared with a preset standard threshold value, so that the real-time response capability of the system is divided into high, medium and low levels, the system can take corresponding measures for different levels of response capabilities, and the real-time response capability of the smart campus management system is improved. The problems of data processing delay and unstable system response capability in a smart campus in the big data era can be effectively solved, the real-time performance, reliability and flexibility of the system are improved, and efficient operation of the campus management system in peak periods or large-scale activities is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a smart campus management method and system based on big data. Background Art

[0002] Smart campus management based on big data refers to optimizing the school's management and services based on intelligent technology by collecting, analyzing and utilizing a large amount of data inside and outside the campus. Big data technology on campus mainly involves real-time collection, analysis and processing of various information such as students, teachers, teaching resources, and facility use. These data include students' learning behavior, grades, health status, attendance, etc., as well as data on the school's energy consumption, course schedule, equipment maintenance, etc. Through big data analysis, schools can more accurately understand student needs, predict teaching effects, and improve teaching quality, thereby improving the efficiency and scientificity of school management.

[0003] Specifically, smart campus management based on big data can help schools achieve intelligent management in many aspects. For example, schools can provide personalized learning recommendations and tutoring services through the analysis of student learning data, thereby improving students' learning outcomes; through the analysis of campus facility usage data, schools can rationally plan resources, perform efficient equipment maintenance and save energy; at the same time, schools can also improve campus security through real-time analysis of monitoring data in terms of security management. In short, the application of big data technology makes school management more sophisticated and intelligent, and promotes the rational allocation and optimization of educational resources.

[0004] The prior art has the following deficiencies: Smart campus management based on big data usually requires real-time or near real-time data processing and feedback. However, the devices, sensors and data sources on campus are often widely distributed, and the amount of data is huge. How to ensure the real-time collection, transmission, processing and analysis of this data is still a technical problem. Especially during peak hours (such as during exams) or large-scale activities, the data traffic surges, and data processing delays or crashes may occur, affecting the real-time response capabilities of the smart campus management system. In addition, this delay not only affects the timeliness of personalized learning recommendations for students, but also affects the scheduling and energy efficiency management of equipment resources. Summary of the invention

[0005] The purpose of the present invention is to provide a smart campus management method and system based on big data to solve the shortcomings of the background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a smart campus management method based on big data, comprising the following steps: S1: Divide the campus into several monitoring areas and deploy different types and numbers of sensors in each monitoring area to collect real-time latency data and throughput data during peak hours; S2: preprocessing the collected delay data and throughput data during the peak time period, and extracting features from the preprocessed delay data and throughput data, respectively extracting the abnormal delay fluctuation trend features in the delay data and the sudden decrease feature of the throughput drop rate in the throughput data; S3: Based on the extracted abnormal delay fluctuation trend characteristics and throughput drop rate sudden reduction characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results; S4: Compare and analyze the obtained real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divide the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the result of the comparative analysis; S5: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan, and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low; S6: For the medium real-time response capability, further predict the real-time response capability of the smart campus management system in the subsequent fixed time period, and optimize the distribution of data traffic based on the prediction results.

[0007] Preferably, in S2, after analyzing the delay abnormal fluctuation trend characteristics in the extracted delay data, a delay abnormal fluctuation trend value is generated, wherein the delay abnormal fluctuation trend value is obtained by: The delay value Ds collected at time point s is obtained in real time, and the expected delay value T is set. At consecutive time points, the delay value continues to deviate from the target delay, and the deviation is accumulated until it exceeds the set threshold. : Used to detect whether the delay value exceeds the upper limit of the target delay, the lower bias : Used to detect whether the delay value is lower than the lower limit of the target delay; If the delay value exceeds the target delay, the deviation will accumulate until it exceeds the preset threshold, expressed as: ; In the formula, is the upper deviation value of the previous moment, Ds is the delay value of the current moment, c is the preset threshold, and the lower deviation is used to detect a sudden decrease in delay. If the delay value is lower than the target delay, the deviation will accumulate until it exceeds the set threshold. The expression is: ; In the formula, is the lower deviation value of the previous moment; calculate the ratio of the number of upper or lower deviations exceeding the preset control limit to the total number of delay values, that is, calculate the abnormal fluctuation trend value of the delay.

[0008] Preferably, in S2, a throughput decrease rate sudden decrease index is generated for the throughput decrease rate sudden decrease feature extracted from the throughput data, wherein the throughput decrease rate sudden decrease index is obtained by: Collect and organize the throughput data T(t), where t is each time point in the time series, and apply fast Fourier transform to convert the time series data T(t) into frequency domain representation to reveal the frequency components in the data: ;in, is the complex number representation of the kth frequency component in the frequency domain, T(n) is the throughput data at the nth moment in the time domain, N is the length of the data sequence, k represents the index of the frequency component, is the complex exponential function in Fourier transform; The throughput number corresponding to the high-frequency component is: ;in, is the amplitude of the kth frequency component, The plural The real and imaginary parts of Based on the spectrum data and the amplitude change of the frequency component, the throughput drop rate sudden reduction index is calculated, and the frequency component amplitude is calculated by time window sliding to obtain the high-frequency component amplitude change rate at a certain moment. ;in, is the amplitude change rate of frequency component k at time t, is the kth frequency component in The amplitude at the moment is used to generate a throughput drop rate reduction index by statistically analyzing the rate of change of high-frequency amplitude: ; Among them, k∈high frequencies represents the index of the high-frequency part in the frequency component, and KMN is the throughput drop rate exponent.

[0009] Preferably, in S3, a machine learning model is constructed to perform feature training based on the extracted abnormal delay fluctuation trend characteristics and the throughput drop rate sudden decrease characteristics, and the real-time response capability value of the smart campus management system is determined based on the model output results, specifically: The delay abnormal fluctuation trend value and the throughput decrease rate sudden reduction index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the real-time response capability value label of the smart campus management system with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the real-time response capability value labels of all smart campus management systems as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The real-time response capability value of the smart campus management system is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0010] Preferably, in S4, the obtained real-time response capability value of the smart campus management system is compared and analyzed with the gradient standard threshold preset according to historical data, and the real-time response capability of the smart campus management system is divided into high real-time response capability, medium real-time response capability and low real-time response capability according to the result of the comparative analysis, specifically: The obtained real-time response capability value of the smart campus management system is compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the real-time response capability value of the smart campus management system is compared with the first standard threshold and the second standard threshold respectively; If the real-time response capability value of the smart campus management system is greater than the second standard threshold, it means that the real-time response capability of the smart campus management system is high, and a high real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as high real-time response capability; If the real-time response capability value of the smart campus management system is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the real-time response capability of the smart campus management system is medium, and a medium real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as medium real-time response capability; If the real-time response capability value of the smart campus management system is less than the first standard threshold, it means that the real-time response capability of the smart campus management system is low. At this time, a low real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as low real-time response capability.

[0011] Preferably, in S6, for the real-time response capability, the real-time response capability of the smart campus management system in a subsequent fixed time period is further predicted, and the allocation of data traffic is optimized according to the prediction results, specifically: For medium real-time response capability, that is, the real-time response capability value of the smart campus management system generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the real-time response capability values ​​of the smart campus management system generated within a subsequent fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold are collected, and a corresponding data set is established, and the standard deviation of the data set is calculated as the real-time response capability prediction index. According to the calculated real-time response capability prediction index, the real-time response capability of the smart campus management system within the subsequent fixed time period is further predicted; The obtained real-time response capability prediction index is compared with the real-time response capability prediction index reference threshold preset according to historical data. If the real-time response capability prediction index is greater than or equal to the real-time response capability prediction index reference threshold preset according to historical data, it means that there is no abnormality in the real-time response capability of the smart campus management system in the subsequent fixed time period. At this time, no warning signal is generated and no adjustment is required; if the real-time response capability prediction index is less than the real-time response capability prediction index reference threshold preset according to historical data, it means that there is an abnormality in the real-time response capability of the smart campus management system in the subsequent fixed time period. At this time, a warning signal is generated to optimize the distribution of data traffic.

[0012] Preferably, m monitoring areas are set, and the flow rate of each monitoring area is (i=1,2,…,m), where g is the current time. Based on the predicted real-time response capability prediction index R(g), the traffic distribution of each area is optimized. The optimization formula for traffic distribution is: ;in, is the predicted response capability value of region i at time g, is the total predicted response capability value of all monitoring areas, is the total system flow demand, that is, the sum of the flows in all regions; the total flow is divided into Assigned to each monitoring area.

[0013] The present invention also provides a smart campus management system based on big data, including a data acquisition module, a feature extraction module, a real-time response evaluation module, a real-time response capability division module, a resource scheduling module and a data flow optimization module; Data collection module: The campus is divided into several monitoring areas, and different types and numbers of sensors are deployed in each monitoring area to collect delay data and throughput data in real time during peak hours; Feature extraction module: pre-processes the collected delay data and throughput data during the peak time period, and extracts features from the pre-processed delay data and throughput data, respectively extracting the abnormal delay fluctuation trend features in the delay data and the sudden decrease in the throughput drop rate features in the throughput data; Real-time response evaluation module: Based on the extracted abnormal delay fluctuation trend characteristics and throughput drop rate characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results; Real-time response capability classification module: compares and analyzes the acquired real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divides the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the results of the comparative analysis; Resource scheduling module: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan, and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low; Data traffic optimization module: For the real-time response capability in the medium term, further prediction is made on the real-time response capability of the smart campus management system in the subsequent fixed time period, and the allocation of data traffic is optimized based on the prediction results.

[0014] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention can effectively solve the delay and system crash problems faced by the prior art when data traffic surges by accurately monitoring and analyzing the delay data and throughput data during peak hours in real time. By deploying sensors of different types and numbers, dividing multiple monitoring areas on campus, collecting key data in real time, and using machine learning models to analyze and predict the abnormal fluctuation trend characteristics of delays and the sudden decrease in throughput rate, real-time response capability values ​​are generated. This method can automatically adjust system configuration, resource allocation, and traffic optimization according to different response capability levels, ensuring that the smart campus management system can maintain stable and efficient operation during peak periods.

[0015] 2. The present invention further improves the intelligence and adaptability of the system through real-time prediction and dynamic adjustment of optimization strategies. For medium response capabilities, the present invention can adjust the distribution of data traffic according to the real-time response capability prediction index to ensure that traffic resources are reasonably allocated and avoid waste of resources. When the prediction shows that the system response capability is abnormal, it can quickly trigger an early warning signal and make corresponding adjustments. This method effectively enhances the stability, flexibility and real-time processing capabilities of the system, significantly improves the response efficiency of the smart campus management system during large-scale activities and peak hours, reduces the risk of delays and crashes, and optimizes the overall operation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0017] Figure 1 The present invention is a flow chart of the method.

[0018] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0020] Example 1, please refer to Figure 1 As shown, the smart campus management method based on big data described in this embodiment includes the following steps: S1: Divide the campus into several monitoring areas and deploy different types and numbers of sensors in each monitoring area to collect real-time latency data and throughput data during peak hours; S2: preprocessing the collected delay data and throughput data during the peak time period, and extracting features from the preprocessed delay data and throughput data, respectively extracting the abnormal delay fluctuation trend features in the delay data and the sudden decrease feature of the throughput drop rate in the throughput data; S3: Based on the extracted abnormal delay fluctuation trend characteristics and throughput drop rate sudden reduction characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results; S4: Compare and analyze the obtained real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divide the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the result of the comparative analysis; S5: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low; S6: For the medium real-time response capability, further predict the real-time response capability of the smart campus management system in the subsequent fixed time period, and optimize the distribution of data traffic based on the prediction results.

[0021] In S1, the campus is divided into several monitoring areas, and different types and numbers of sensors are deployed in each monitoring area to collect real-time delay data and throughput data during peak hours. Specifically: Divide the campus into multiple monitoring areas to ensure that the equipment, sensors and data sources in each area can be accurately covered and independently monitored. Basis for regional division: Building structure: Divide the area according to physical spaces such as teaching buildings, dormitories, canteens, libraries, laboratories, and sports fields. Functional area division: Different areas may involve monitoring needs of different functions, such as teaching areas, scientific research areas, living areas, activity areas, etc., and are divided according to functional differences. Traffic prediction: Considering that some areas may be subject to large data traffic and access loads during specific periods (such as exams, sports games, holidays, etc.), it is necessary to divide the area according to the expected traffic. High-traffic areas (such as large classrooms, dormitories, canteens, etc.) should have more monitoring points.

[0022] According to the characteristics and needs of each monitoring area, select the appropriate type and number of sensors for data collection to ensure comprehensive and accurate collection of latency and throughput data. Sensor types: Latency monitoring sensors: Network latency sensors: used to monitor network latency between different devices or areas, especially during peak data transmission periods, to help track the transmission latency of data packets. Response time sensors: can be used to monitor the response time of terminal devices, such as student devices, smart terminals in classrooms, access control systems, security monitoring systems, etc. Throughput monitoring sensors: Traffic monitoring sensors: such as bandwidth monitoring instruments, monitor the throughput of different network nodes to ensure that the load of data transmission channels is reasonably distributed. Packet counters: can accurately record the number of requests processed per second to help calculate the throughput of the system. Environmental monitoring sensors: used to monitor the operating environment of the equipment, such as temperature, humidity, etc., to ensure that the sensor or equipment works in a suitable environment and avoid system failures caused by environmental problems.

[0023] Sensor deployment: Including high-traffic areas, such as large classrooms, examination areas, canteens, etc., more sensors need to be deployed to ensure real-time monitoring of device status, data traffic and network conditions. Key equipment monitoring: For key areas such as computer rooms and server rooms in teaching buildings, sensors are deployed to monitor device load, response time, throughput, etc. Campus core network nodes: The campus' core routers, switches, access points and other network nodes should be equipped with traffic monitoring sensors to ensure real-time monitoring and feedback of data flows.

[0024] Collect latency and throughput data in real time through sensors. Record the time interval from data initiation to data reception, and track it through network latency monitoring tools (such as ping monitoring, response time measurement, etc.). Throughput data collection: the number of requests processed per second, data transmission rate, data packet size, etc., all of which can provide system load information. Transmit the collected data to the central data processing platform through the network to ensure that the data can flow quickly to the storage and processing center after collection to avoid delays. Especially during peak hours (such as exams, activities, etc.), based on traffic forecasts, centrally monitor the latency and throughput data in these areas to ensure that the system's real-time monitoring data is accurate.

[0025] S2: Preprocess the collected delay data and throughput data during the peak time period, and perform feature extraction on the preprocessed delay data and throughput data, respectively extracting the delay abnormal fluctuation trend features in the delay data and the throughput decrease rate sudden reduction features in the throughput data.

[0026] The goal of data preprocessing is to clean and optimize the original collected data into a form that can be effectively analyzed and modeled. Sensors or acquisition systems may cause noise in the data due to environmental changes, hardware failures, or interference. Therefore, it is necessary to remove outliers (such as invalid negative delay data or throughput values), which will affect the accuracy of subsequent analysis. Remove outliers, for example, by setting a reasonable threshold range to detect and remove sudden invalid data. For example, the delay value should not be less than 0ms, and the throughput value should not be lower than a reasonable lower limit. For possible short-term fluctuations (such as instantaneous delay peaks), sliding average, weighted average and other methods can be used for smoothing to reduce the interference of fluctuations on the model. The range of delay data and throughput data may vary greatly, so standardization is required to ensure the comparability between different features.

[0027] Transform the data into a distribution with a mean of 0 and a variance of 1. Common methods include Z-score standardization. Scale the data to a specified range (such as 0 to 1) to eliminate dimensional differences and prevent certain features from dominating the analysis results. The collected data may be missing. Common processing methods include filling in missing values ​​(for example, using the mean, median, or interpolation method to fill in missing values) or directly discarding missing data points.

[0028] Latency refers to the time from when a data request is initiated to when a response is completed. During peak hours, latency may fluctuate significantly, especially in situations such as network congestion and device overload. By extracting the characteristics of abnormal latency fluctuation trends, the system can identify abnormal latency and make corresponding adjustments.

[0029] Variance of Delay: Calculate the variance of the delay data. The larger the variance, the more obvious the delay fluctuation of the system. A larger variance may mean that the network or device is overloaded, and there is a risk of slow response.

[0030] Delay Trend: Extract the rising or falling trend of delay through time series analysis. Capture the long-term fluctuation trend of delay through sliding window or linear regression.

[0031] Burst Delay: Monitors burst delays that occur in a short period of time. A threshold can be defined, and when the delay exceeds a certain critical value, it is marked as a burst delay.

[0032] Latency outlier detection: Use statistical methods (such as z-score or IQR) to detect outliers in latency data and capture abnormal fluctuations in a timely manner.

[0033] Throughput refers to the amount of requests or data that a system can process or transmit per unit time. During peak hours, throughput may drop sharply due to system overload, resource bottlenecks, or network congestion. By extracting the characteristics of the sudden drop in throughput rate, the system can detect the risk of throughput drop in a timely manner and take preventive measures.

[0034] Throughput Decline Rate: Calculates the rate at which throughput changes over a certain period of time. If throughput decreases rapidly (for example, the amount of data decreases by more than 20% per second), it indicates that the system has a performance bottleneck or is overloaded.

[0035] Sudden Throughput Drop: By analyzing the throughput change curve, we can extract the time point and magnitude of the sudden drop in throughput. Usually, when the throughput drops suddenly, it means that the system performance is seriously degraded, and a crash or response timeout may occur.

[0036] Throughput Drop Magnitude: The drop in throughput from peak to trough can be used to quantify the severity of the system bottleneck. By calculating the maximum drop in throughput, the bottleneck of the system under high load can be evaluated.

[0037] After analyzing the delay abnormal fluctuation trend characteristics in the extracted delay data, a delay abnormal fluctuation trend value is generated, wherein the delay abnormal fluctuation trend value is obtained by: Obtain the delay value Ds collected at time point s in real time, and set the expected delay value T, which is usually a fixed value. For example, it can be the average delay or target delay of the system under normal load. If the delay value continues to deviate from the target delay at consecutive time points, the deviation is accumulated until it exceeds the set threshold. : Used to detect whether the delay value exceeds the upper limit of the target delay. : Used to detect whether the delay value is lower than the lower limit of the target delay.

[0038] The upper deviation is used to detect a sudden increase in latency. If the latency value significantly exceeds the target latency, it will accumulate the deviation until it exceeds the preset threshold, expressed as: ; In the formula, is the upper deviation value of the previous moment, Ds is the delay value of the current moment, and k is the preset threshold, which is usually a multiple of the standard deviation σ of the delay data. This parameter is used to control the sensitivity. If a larger k is set, the delay value needs to deviate more from the target to be considered abnormal. If a smaller k is set, it will be more sensitive and easier to detect smaller fluctuations.

[0039] Downward deviation is used to detect a sudden decrease in latency. If the latency value is lower than the target latency, it accumulates these deviations until it exceeds the set threshold, expressed as: ; In the formula, is the lower deviation value of the previous moment.

[0040] When the upper or lower deviation exceeds the preset control limit, it indicates that the delay has fluctuated abnormally and corresponding response measures need to be taken. The control limit is a set threshold value H, which represents the maximum tolerance of the cumulative deviation and can be a multiple of a certain standard deviation (for example, H=5σ). Calculate the ratio of the number of upper or lower deviations exceeding the preset control limit to the total number of delay values, that is, calculate the abnormal fluctuation trend value of the delay.

[0041] In the smart campus management system, the change in the delay abnormal fluctuation trend value reflects the real-time response capability of the system. When the delay abnormal fluctuation trend value is large, it means that there is a large delay fluctuation in the system, which may be caused by factors such as traffic surge during peak hours, insufficient hardware resources, and network problems. This fluctuation usually leads to data processing delays, affects the system's ability to respond to emergencies, and reduces the system's real-time responsiveness. Therefore, the larger the delay abnormal fluctuation trend value, the worse the system's real-time response capability is, and there is a higher risk of failure.

[0042] On the contrary, when the delay abnormal fluctuation trend value is small, it means that the system's delay fluctuation is relatively stable, and the data processing capacity and resource scheduling can remain stable during peak hours or large-scale activities. This shows that the system has a strong real-time response capability and can quickly process real-time data and make corresponding adjustments, thereby ensuring the timeliness and efficiency of services such as student learning recommendations and resource scheduling. Therefore, the smaller the delay abnormal fluctuation trend value, the stronger the real-time response capability of the smart campus management system, and the better it can cope with challenges under different load conditions.

[0043] A throughput reduction rate sudden reduction index is generated for the throughput reduction rate sudden reduction feature in the extracted throughput data, wherein the throughput reduction rate sudden reduction index is obtained by: Collect and organize throughput data T(t), where t is each time point in the time series. Normalize or smooth the throughput data T(t) to remove outliers or noise. Apply FFT to convert the time series data T(t) into a frequency domain representation to reveal the frequency components in the data: ;in, is the complex number representation of the kth frequency component in the frequency domain, T(n) is the throughput data at the nth time in the time domain, N is the length of the data sequence, k represents the index of the frequency component (from 0 to N−1), is the complex exponential function in Fourier transform.

[0044] By calculating the amplitude of the frequency domain data, we can understand the change trend of the throughput, especially the high-frequency components. High-frequency components usually correspond to rapid changes in the throughput data, and the expression is: ;in, is the amplitude of the kth frequency component, The plural The real and imaginary parts of .

[0045] In the frequency domain, a sharp drop in throughput data usually corresponds to the low-frequency part (slow change) and the high-frequency part (fast change) in the spectrum. In order to analyze the sudden drop in the throughput drop rate, the key is to detect the sudden drop in the high-frequency component. Based on the spectrum data and the amplitude change of the frequency component, the throughput drop rate sudden drop index is calculated. The frequency component amplitude is calculated by sliding the time window to obtain the change rate of the high-frequency component amplitude at a certain moment. ;in, is the amplitude change rate of frequency component k at time t, is the kth frequency component in The amplitude at the moment is used to generate a throughput drop rate reduction index by statistically analyzing the rate of change of high-frequency amplitude: ; Among them, k∈high frequencies represents the index of the high-frequency part in the frequency component, and KMN is the throughput drop rate exponent.

[0046] The larger the throughput drop rate exponent, the faster and more drastic the drop in throughput occurs, indicating that the processing capacity of the smart campus management system under high load or emergencies is seriously affected. This situation usually occurs during peak hours or large-scale activities, when the system throughput drops sharply, resulting in a significant decrease in the speed of data transmission. The increase in the throughput drop rate exponent often reflects that the system cannot effectively cope with burst traffic, which may cause system crashes, data loss or response delays, seriously affecting the stability and real-time response capabilities of the smart campus management system.

[0047] On the contrary, the smaller the throughput drop rate exponent, the more stable the throughput remains in the face of load changes, the system can effectively adapt to various load fluctuations, and has strong real-time response capabilities. In this case, the smart campus management system can ensure that the throughput does not fluctuate much during peak hours or large-scale activities by optimizing resource allocation and traffic scheduling, avoiding excessive delays or crashes. The smaller value of the throughput drop rate exponent indicates that the system has strong fault tolerance and dynamic adjustment capabilities, which can ensure the stable operation of various functions and provide reliable real-time services.

[0048] S3: Based on the extracted delay abnormal fluctuation trend characteristics and throughput drop rate sudden reduction characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results.

[0049] The delay abnormal fluctuation trend value and the throughput decrease rate sudden reduction index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the real-time response capability value label of the smart campus management system with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the real-time response capability value labels of all smart campus management systems as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The real-time response capability value of the smart campus management system is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0050] The method for obtaining the real-time response capability value of the smart campus management system is to obtain the corresponding function expression from the comprehensive feature vector training data of the trained machine learning model: ; In the formula, is the output function of the model, SDF is the abnormal delay fluctuation trend value, KMN is the throughput drop rate sudden decrease index, It is the real-time response capability value of the smart campus management system.

[0051] S4: Compare and analyze the acquired real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divide the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the results of the comparative analysis.

[0052] The obtained real-time response capability value of the smart campus management system is compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the real-time response capability value of the smart campus management system is compared with the first standard threshold and the second standard threshold respectively; If the real-time response capability value of the smart campus management system is greater than the second standard threshold, it means that the real-time response capability of the smart campus management system is high, and a high real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as high real-time response capability; If the real-time response capability value of the smart campus management system is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the real-time response capability of the smart campus management system is medium, and a medium real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as medium real-time response capability; If the real-time response capability value of the smart campus management system is less than the first standard threshold, it means that the real-time response capability of the smart campus management system is low. At this time, a low real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as low real-time response capability.

[0053] S5: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan, and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low.

[0054] When the system's real-time response capability is evaluated as high, it means that the system can maintain stable operation under high load, and its processing capacity and response time meet current needs. At this time, there is no need to make major adjustments to the system. Specific treatment measures include: Based on the evaluation results, continue to use the current smart campus management system architecture and maintain the existing resource allocation and system configuration. At this point, the system can effectively handle large data traffic and high-load scenarios, the response time is in line with expectations, and the operation is stable. Although the system performs well, regular performance monitoring and resource usage monitoring are still required. Regular real-time data collection ensures that the system can still maintain efficient operation under changing load conditions. At this stage, the system can be fine-tuned to further improve performance. For example, optimize algorithms, upgrade hardware equipment, etc., but these measures are long-term optimizations and are not urgent operations. Although the system performs well, it should remain flexible to potential emergencies (such as special exams, cultural activities, etc.). Dynamic load balancing or automatic elastic expansion technology can be used to ensure that the system can automatically adapt to real-time loads.

[0055] When the system's real-time response capability value is evaluated as low real-time response capability, it means that the system performs poorly under high load and may face the risk of delays, crashes, or insufficient resource scheduling. Therefore, emergency measures must be taken immediately to make rapid adjustments to ensure the stability and real-time response capability of the system. The specific processing steps include: Once the system is classified as having low real-time response capability, emergency response signals should be triggered immediately. These signals will automatically trigger the predetermined emergency process, including notifying relevant personnel, starting system performance monitoring and alarm systems. If the existing hardware resources are insufficient, the system should be able to automatically expand, such as expanding computing and storage capabilities through cloud computing resources, quickly adding more servers or virtual machines, and sharing the load pressure.

[0056] Prioritize resources, such as high-priority tasks (for example, key teaching applications during exams) can get priority access to system resources, while low-priority tasks (for example, background statistics tasks) are delayed. If necessary, part of the data processing can be transferred from the central server to the edge of the network, and the edge computing nodes can be used to share the tasks of data preprocessing and preliminary analysis, thereby reducing the pressure on the main server. To avoid delays and congestion in data transmission, data compression and caching technologies can be used to reduce the amount of data transmitted and ensure rapid response of important data.

[0057] Strengthen real-time monitoring of various system indicators, especially key performance indicators such as network latency, throughput, processing capacity, and resource usage. The monitoring system needs to capture any anomalies in a timely manner so that the emergency plan can be launched more quickly. The first step of the emergency plan is to start the automatic early warning system, monitoring indicators such as latency fluctuations and throughput drops, which can trigger system alarms, quickly respond to potential risks, and notify administrators in time to handle them. After taking emergency measures in an emergency situation with low real-time response capabilities, the system should record all event details, including changes in data such as latency and throughput, for later analysis. These data will help the development team discover system bottlenecks and further optimize the architecture. For the problem of low response capabilities, retrospective analysis should be conducted regularly to summarize the effects of the emergency plan, evaluate the effects of system adjustments, and continuously improve the plan and technical architecture to reduce the probability of similar problems in the future.

[0058] S6: For the medium real-time response capability, further predict the real-time response capability of the smart campus management system in the subsequent fixed time period, and optimize the distribution of data traffic based on the prediction results.

[0059] For medium real-time response capability, that is, the real-time response capability value of the smart campus management system generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the real-time response capability values ​​of the smart campus management system generated within a subsequent fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold are collected, and a corresponding data set is established, and the standard deviation of the data set is calculated as the real-time response capability prediction index. According to the calculated real-time response capability prediction index, the real-time response capability of the smart campus management system in the subsequent fixed time period is further predicted.

[0060] The obtained real-time response capability prediction index is compared with the real-time response capability prediction index reference threshold preset according to historical data. If the real-time response capability prediction index is greater than or equal to the real-time response capability prediction index reference threshold preset according to historical data, it means that there is no abnormality in the real-time response capability of the smart campus management system in the subsequent fixed time period. At this time, no warning signal is generated and no adjustment is required; if the real-time response capability prediction index is less than the real-time response capability prediction index reference threshold preset according to historical data, it means that there is an abnormality in the real-time response capability of the smart campus management system in the subsequent fixed time period. At this time, a warning signal is generated to optimize the distribution of data traffic.

[0061] There are m monitoring areas, and the flow rate in each monitoring area is (i=1,2,…,m), where g is the current time. Based on the predicted real-time response capability prediction index R(g), the traffic distribution of each area is optimized. The optimization formula for traffic distribution is: ;in, is the predicted response capability value of region i at time g, is the total predicted response capability value of all monitoring areas, is the total system flow demand, which is the sum of flows in all zones.

[0062] According to the ratio of the predicted response capacity of each area to the total response capacity, the total flow Reasonably distribute it to each monitoring area. If the response capability of a certain area is stronger, more traffic will be allocated. If the response capability of a certain area is weaker, the data traffic in that area will be limited to avoid overload.

[0063] The optimization effect can be further evaluated by monitoring the deviation between the actual response capability of the real-time monitoring system and the predicted value. For example, the balance of flow distribution in each monitoring area can be evaluated. If the flow distribution in each area is balanced, it means that the optimization effect is good. The expression is: . When the balance is smaller, it means that the traffic distribution in each monitoring area is close to equal, and the system load is well balanced. Each area is reasonably allocated traffic according to its real-time response capability, which helps to avoid overload or waste of resources in a specific area due to excessive or insufficient traffic. By optimizing the distribution of traffic, the smart campus management system can maintain efficient resource utilization, reduce response delays, and improve the overall stability and responsiveness of the system. In this case, the system performs better during peak hours, can handle large-scale data traffic and multiple task requests, and ensure service quality and user experience.

[0064] When the balance is larger, it means that the traffic is unevenly distributed. Some monitoring areas may carry too much traffic, causing overload, resulting in delayed response time, system crash or waste of resources. This imbalance may make some areas insufficiently responsive or even unable to cope with high-load tasks, affecting the stability of the overall system. For example, during peak hours, too little traffic may be allocated to some key areas, which may lead to insufficient resources; while too much traffic may be allocated to other unimportant areas, which will waste computing resources and bandwidth. In general, systems with a large balance are prone to bottlenecks when demand fluctuates and cannot fully utilize the system's optimal performance.

[0065] Therefore, the smaller the balance degree, the better the optimization effect, which means that the traffic distribution is more even and the system can maintain good responsiveness and stability under different load conditions. On the contrary, the larger the balance degree, the greater the problem of traffic distribution and the system may not be able to effectively cope with large-scale loads.

[0066] In this embodiment, the campus is divided into multiple monitoring areas and sensors are deployed in each area to collect delay data and throughput data in peak time periods in real time. Then, the collected data is preprocessed and feature extracted to extract abnormal fluctuation trend features in the delay data and sudden drop in the rate of decline features in the throughput data. Using these features, a machine learning model is constructed for training to predict the real-time response capability value of the system, and the real-time response capability is classified according to the gradient standard threshold preset by historical data: high, medium, and low real-time response capability. For high real-time response capability, the system continues to maintain the existing architecture; for low real-time response capability, an emergency plan is generated and the system configuration is quickly adjusted; for medium real-time response capability, the response capability of subsequent time periods is further predicted, and the distribution of data traffic is optimized to ensure that the system can operate efficiently during peak load periods, thereby improving the response speed and stability of the overall system.

[0067] Example 2, please refer to Figure 2 As shown, the smart campus management system based on big data described in this embodiment includes a data acquisition module, a feature extraction module, a real-time response evaluation module, a real-time response capability division module, a resource scheduling module and a data flow optimization module; Data collection module: The campus is divided into several monitoring areas, and different types and numbers of sensors are deployed in each monitoring area to collect delay data and throughput data in real time during peak hours; Feature extraction module: pre-processes the collected delay data and throughput data during the peak time period, and extracts features from the pre-processed delay data and throughput data, respectively extracting the abnormal delay fluctuation trend features in the delay data and the sudden decrease in the throughput drop rate features in the throughput data; Real-time response evaluation module: Based on the extracted abnormal delay fluctuation trend characteristics and throughput drop rate characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results; Real-time response capability classification module: compares and analyzes the acquired real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divides the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the results of the comparative analysis; Resource scheduling module: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan, and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low; Data traffic optimization module: For the real-time response capability in the medium term, further prediction is made on the real-time response capability of the smart campus management system in the subsequent fixed time period, and the allocation of data traffic is optimized based on the prediction results.

[0068] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A smart campus management method based on big data, characterized by: The following steps are involved: S1: Divide the campus into several monitoring areas and deploy different types and numbers of sensors in each monitoring area to collect real-time latency data and throughput data during peak hours; S2: preprocessing the collected delay data and throughput data during the peak time period, and extracting features from the preprocessed delay data and throughput data, respectively extracting the abnormal delay fluctuation trend features in the delay data and the sudden decrease feature of the throughput drop rate in the throughput data; S3: Based on the extracted abnormal delay fluctuation trend characteristics and throughput drop rate sudden reduction characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results; S4: Compare and analyze the obtained real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divide the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the result of the comparative analysis; S5: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan, and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low; S6: For the medium real-time response capability, further predict the real-time response capability of the smart campus management system in the subsequent fixed time period, and optimize the distribution of data traffic based on the prediction results.

2. A smart campus management method based on big data according to claim 1, characterized in that: In S2, after analyzing the delay abnormal fluctuation trend characteristics in the extracted delay data, a delay abnormal fluctuation trend value is generated, wherein the delay abnormal fluctuation trend value is obtained by: The delay value Ds collected at time point s is obtained in real time, and the expected delay value T is set. At consecutive time points, the delay value continues to deviate from the target delay, and the deviation is accumulated until it exceeds the set threshold. : Used to detect whether the delay value exceeds the upper limit of the target delay, the lower bias : Used to detect whether the delay value is lower than the lower limit of the target delay; If the delay value exceeds the target delay, the deviation will accumulate until it exceeds the preset threshold, expressed as: ; In the formula, is the upper deviation value of the previous moment, Ds is the delay value of the current moment, c is the preset threshold, and the lower deviation is used to detect a sudden decrease in delay. If the delay value is lower than the target delay, the deviation will accumulate until it exceeds the set threshold. The expression is: ; In the formula, is the lower deviation value of the previous moment; calculate the ratio of the number of upper or lower deviations exceeding the preset control limit to the total number of delay values, that is, calculate the abnormal fluctuation trend value of the delay.

3. A smart campus management method based on big data according to claim 2, characterized in that: In S2, a throughput reduction rate sudden reduction index is generated based on the throughput reduction rate sudden reduction feature extracted from the throughput data, wherein the throughput reduction rate sudden reduction index is obtained by: Collect and organize the throughput data T(t), where t is each time point in the time series, and apply fast Fourier transform to convert the time series data T(t) into frequency domain representation to reveal the frequency components in the data: ;in, is the complex number representation of the kth frequency component in the frequency domain, T(n) is the throughput data at the nth moment in the time domain, N is the length of the data sequence, k represents the index of the frequency component, is the complex exponential function in Fourier transform; The high-frequency components correspond to the rapid changes in the throughput data and are expressed as: ;in, is the amplitude of the kth frequency component, The plural The real and imaginary parts of Based on the spectrum data and the amplitude change of the frequency component, the throughput drop rate sudden reduction index is calculated, and the frequency component amplitude is calculated by time window sliding to obtain the high-frequency component amplitude change rate at a certain moment. ;in, is the amplitude change rate of frequency component k at time t, is the kth frequency component in The amplitude at the moment is used to generate a throughput drop rate reduction index by statistically analyzing the rate of change of high-frequency amplitude: ; Among them, k∈high frequencies represents the index of the high-frequency part in the frequency component, and KMN is the throughput drop rate exponent.

4. The smart campus management method based on big data according to claim 3 is characterized by: In S3, based on the extracted abnormal delay fluctuation trend characteristics and the throughput drop rate sudden reduction characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results, specifically: The abnormal delay fluctuation trend value and the throughput decrease rate reduction index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the real-time response capability value label of the smart campus management system with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the real-time response capability value labels of all smart campus management systems as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The real-time response capability value of the smart campus management system is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

5. A smart campus management method based on big data according to claim 4, characterized in that: In S4, the obtained real-time response capability value of the smart campus management system is compared and analyzed with the gradient standard threshold preset according to historical data, and the real-time response capability of the smart campus management system is divided into high real-time response capability, medium real-time response capability and low real-time response capability according to the result of the comparative analysis, specifically: The obtained real-time response capability value of the smart campus management system is compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the real-time response capability value of the smart campus management system is compared with the first standard threshold and the second standard threshold respectively; If the real-time response capability value of the smart campus management system is greater than the second standard threshold, it means that the real-time response capability of the smart campus management system is high, and a high real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as high real-time response capability; If the real-time response capability value of the smart campus management system is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the real-time response capability of the smart campus management system is medium, and a medium real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as medium real-time response capability; If the real-time response capability value of the smart campus management system is less than the first standard threshold, it means that the real-time response capability of the smart campus management system is low. At this time, a low real-time response signal is generated, and the real-time response capability of the smart campus management system is classified as low real-time response capability.

6. A smart campus management method based on big data according to claim 5, characterized in that: In S6, for the real-time response capability, the real-time response capability of the smart campus management system in the subsequent fixed time period is further predicted, and the allocation of data traffic is optimized according to the prediction results, specifically: For medium real-time response capability, that is, the real-time response capability value of the smart campus management system generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the real-time response capability values ​​of the smart campus management system generated within a subsequent fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold are collected, and a corresponding data set is established, and the standard deviation of the data set is calculated as the real-time response capability prediction index. According to the calculated real-time response capability prediction index, the real-time response capability of the smart campus management system within the subsequent fixed time period is further predicted; The obtained real-time response capability prediction index is compared with the real-time response capability prediction index reference threshold preset according to historical data. If the real-time response capability prediction index is greater than or equal to the real-time response capability prediction index reference threshold preset according to historical data, it means that there is no abnormality in the real-time response capability of the smart campus management system in the subsequent fixed time period. At this time, no warning signal is generated and no adjustment is required; if the real-time response capability prediction index is less than the real-time response capability prediction index reference threshold preset according to historical data, it means that there is an abnormality in the real-time response capability of the smart campus management system in the subsequent fixed time period. At this time, a warning signal is generated to optimize the distribution of data traffic.

7. A smart campus management method based on big data according to claim 6, characterized in that: There are m monitoring areas, and the flow rate in each monitoring area is , , where g is the current time. Based on the predicted real-time response capability prediction index R(g), the traffic distribution of each area is optimized. The optimization formula for traffic distribution is: ;in, is the predicted response capability value of region i at time g, is the total predicted response capability value of all monitoring areas, is the total system flow demand, that is, the sum of the flows in all regions; the total flow is divided into Assigned to each monitoring area.

8. A smart campus management system based on big data, used to implement a smart campus management method based on big data as described in any one of claims 1 to 7, characterized in that: It includes data collection module, feature extraction module, real-time response evaluation module, real-time response capability division module, resource scheduling module and data flow optimization module; Data collection module: The campus is divided into several monitoring areas, and different types and numbers of sensors are deployed in each monitoring area to collect delay data and throughput data in real time during peak hours; Feature extraction module: pre-processes the collected delay data and throughput data during the peak time period, and extracts features from the pre-processed delay data and throughput data, respectively extracting the abnormal delay fluctuation trend features in the delay data and the sudden decrease in the throughput drop rate features in the throughput data; Real-time response evaluation module: Based on the extracted abnormal delay fluctuation trend characteristics and throughput drop rate characteristics, a machine learning model is constructed for feature training, and the real-time response capability value of the smart campus management system is determined based on the model output results; Real-time response capability classification module: compares and analyzes the acquired real-time response capability value of the smart campus management system with the gradient standard threshold preset according to historical data, and divides the real-time response capability of the smart campus management system into high real-time response capability, medium real-time response capability and low real-time response capability according to the results of the comparative analysis; Resource scheduling module: For high real-time response capability, continue to maintain the existing smart campus management system architecture without further adjustment; for low real-time response capability, immediately generate an emergency plan, and quickly adjust the system configuration and resource allocation for emergency situations that occur when the real-time response capability is low; Data traffic optimization module: For the real-time response capability in the medium term, further prediction is made on the real-time response capability of the smart campus management system in the subsequent fixed time period, and the allocation of data traffic is optimized based on the prediction results.

Images