Evaluation method of LTE wireless network and server

Abstract

The invention relates to the technical field of performance evaluation, in particular to an evaluation method of an LTE wireless network and a server, comprising the following steps: based on historical performance data of the LTE network, performing normalization processing, optimizing data deviation, identifying change modes of load fluctuation and signal strength, and comparing data differences between different time nodes to obtain an evaluation result of the LTE wireless network; and generating a historical performance analysis result. According to the method, through normalization processing and data deviation optimization, the availability of historical performance data and the accuracy of analysis are enhanced, more accurate network load prediction and performance expectation are allowed, rationalization and seasonal decomposition of resource allocation are promoted, trend and periodic components in the data are revealed, more detailed data insight is provided, and the method is suitable for popularization and application. The decision-making process of network adjustment and optimization is assisted, the ARIMA model is adopted to predict resource requirements, and the support vector machine is adopted to carry out real-time performance monitoring, so that timely adjustment and optimization of network configuration are ensured, and the response flexibility of the network and the continuity of service are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of performance evaluation, and in particular to an evaluation method and a server for an LTE wireless network. Background Art

[0002] The field of performance evaluation technology is used to quantify and analyze the operational performance of a system or device. In wireless network technology, especially in the field of LTE, performance evaluation is crucial, including measuring network coverage, signal quality, data transmission rate, latency, and service stability and reliability. The evaluation process uses simulated or actual user data flows to determine the responsiveness and carrying capacity of the network under different conditions. The results of performance evaluation support network planning, optimization and quality assurance, ensuring the optimization of user experience and the effective management of network resources.

[0003] Among them, the evaluation method of LTE wireless network focuses on evaluating and analyzing the performance parameters of LTE network, which is used to identify potential problems in the network, optimize network configuration, and improve data transmission efficiency and user service quality. By applying the evaluation method, network operators can ensure that network facilities meet the growing data demand while minimizing delays and interruptions, and ensuring efficient and continuous service supply. This method aims to provide scientific basis and technical support for the maintenance and upgrade of LTE network, especially in environments with dense users or complex geographical locations.

[0004] Traditional methods do not use pattern and trend analysis in historical data, and are unable to match the network tuning needs of densely populated users and complex geographical environments, resulting in uneven resource allocation and service quality issues. The data seasonality decomposition and trend prediction functions not implemented in traditional methods limit the foresight and adaptability of network tuning, affecting the network's scalability and upgrade efficiency. The lack of real-time performance monitoring and dynamic tuning mechanisms is particularly evident under high load conditions, resulting in increased service delays and reduced user experience, which overall limits the optimization and effective management of network performance. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an evaluation method and server for an LTE wireless network.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a method for evaluating an LTE wireless network, comprising the following steps:

[0007] S1: Based on the historical performance data of the LTE network, the data deviation is optimized through normalization processing, the change pattern of load fluctuation and signal strength is identified, and the data differences between different time nodes are compared to generate historical performance analysis results;

[0008] S2: Based on the historical performance analysis results, seasonal decomposition of the data series is performed, trend and cyclical components are extracted, and the data structure is adjusted to match the model input requirements to obtain a seasonal adjustment model;

[0009] S3: Based on the seasonal adjustment model, forecast the network load and performance for the next week, analyze trend changes and potential performance fluctuations, and obtain load forecast information;

[0010] S4: Based on the load forecast information, the ARIMA model is used to analyze the resource demand during the load peak period, calculate the required optimal bandwidth and signal adjustment amount, adjust the resource configuration according to the demand, and generate a resource optimization plan;

[0011] S5: Based on the resource optimization solution, using support vector machines, real-time monitoring of network performance, automatic adjustment of network configuration, including dynamic update of channels and frequencies, matching changes in network load, and generating dynamic tuning information;

[0012] S6: Based on the dynamic tuning information, evaluate the stability and efficiency of the network, analyze the performance difference before and after the tuning by comparing the performance indicators, determine the network operation efficiency, and obtain the terminal performance evaluation log.

[0013] The present invention has improvements in that the historical performance analysis results include normalized indicators, fluctuation trend graphs, and signal change curves; the seasonal adjustment model includes a trend component graph, a periodic component graph, and a residual graph; the load forecasting information includes a daily load forecast table and a performance stability forecasting graph; the resource optimization plan includes a bandwidth adjustment plan, a signal strength adjustment plan, and a resource allocation schedule; the dynamic tuning information includes channel update records, frequency adjustment details, and a performance monitoring log; and the terminal performance evaluation log includes tuning effect comparison data, operating efficiency data, and stability analysis results.

[0014] The present invention has the following improvements: based on the historical performance data of the LTE network, the data deviation is optimized through normalization processing, the change pattern of load fluctuation and signal strength is identified, and the data difference between the difference time nodes is compared. The specific steps of generating the historical performance analysis results are:

[0015] S101: Based on the historical performance data of the LTE network, outliers are removed through normalization processing, the data format is unified, and the records of different time points are time-standardized to obtain a normalized data set;

[0016] S102: constructing a time series based on the normalized data set, extracting key time points using a sliding window method, analyzing load fluctuations and signal strength changes at the time points, and obtaining fluctuation records at the time points;

[0017] S103: Based on the fluctuation records at the time points, the time node data is compared, and a differential method is used to identify key change trends and abnormal fluctuations, including change points and fluctuation ranges, to generate historical performance analysis results.

[0018] The present invention is improved in that, based on the historical performance analysis results, seasonal decomposition of the data series is performed, trend and periodic components are extracted, and the data structure is adjusted to match the model input requirements. The steps of obtaining the seasonal adjustment model are specifically as follows:

[0019] S201: Based on the historical performance analysis results, perform data hierarchical decomposition, isolate seasonal and non-seasonal factors, highlight key seasonal patterns through data segmentation, and obtain seasonal pattern details;

[0020] S202: Based on the seasonal pattern details, the trend components and periodic fluctuations are refined, and the trend line is adjusted and balanced by the moving average method to optimize the seasonal adaptability of the data and obtain a trend smoothing result;

[0021] S203: Based on the trend smoothing result, the data frequency and resolution are optimized to optimize the data structure and match the input requirements of the model to obtain a seasonal adjustment model.

[0022] The present invention is improved in that, based on the seasonal adjustment model, the network load and performance for the next week are predicted, trend changes and potential performance fluctuations are analyzed, and the steps of obtaining load prediction information are specifically as follows:

[0023] S301: Based on the seasonal adjustment model, quantitatively evaluate the daily network load and performance indicators in the next week, draw a future load trend chart, identify key trends and change ranges, and obtain trend change prediction details;

[0024] S302: Based on the trend change prediction details, compare with the historical data of the same period, determine the potential high-load time period and performance fluctuation range, mark the key fluctuation time points, and generate a fluctuation period identification record;

[0025] S303: Based on the fluctuation period identification record, the key period data is centrally analyzed, a load and performance fluctuation prediction chart is created, key prediction information is sorted and summarized, and load prediction information is obtained.

[0026] The present invention is improved in that, based on the load forecast information, the ARIMA model is used to analyze the resource demand during the load peak period, calculate the required optimal bandwidth and signal adjustment amount, and adjust the resource configuration according to the demand. The steps of generating a resource optimization plan are specifically as follows:

[0027] S401: Based on the load forecast information, using the ARIMA model, analyzing the network resource usage in multiple time periods, calculating the bandwidth demand and signal adjustment demand during the peak period, and obtaining bandwidth and signal demand indicators according to the network demand;

[0028] S402: Based on the bandwidth and signal demand indicators, design resource configuration measures that match the differential network loads, perform network resource allocation, match resource allocation with peak loads, and generate resource configuration details;

[0029] S403: Based on the resource configuration details, perform a network configuration update operation, adjust signal distribution and device settings, match the network configuration to the predicted load, and generate a resource optimization plan.

[0030] The present invention improves that the ARIMA model is based on the formula:

[0031] Φ(B)(1-B) d y t =θ(B)(1+αL+βL 2 ) t

[0032] Calculate the network load prediction value, where Φ(B) is the polynomial of the autoregressive parameter, Θ(B) is the polynomial of the moving average parameter, B is the backshift operator, d is the difference order, and y t is the load data in the time series, e t is the error term, L is the seasonal adjustment factor, α and β are the first and second weight coefficients of seasonal adjustment respectively.

[0033] The present invention has the following improvements: based on the resource optimization scheme, using a support vector machine, real-time monitoring of network performance, automatic adjustment of network configuration, including dynamic update of channels and frequencies, matching of network load changes, and generating dynamic tuning information:

[0034] S501: Based on the resource optimization solution, using a support vector machine, continuously tracking network performance indicators, recording performance fluctuations and network loads, identifying immediate performance degradation or abnormal fluctuation points, and obtaining real-time performance monitoring records;

[0035] S502: Based on the real-time performance monitoring record, analyze and identify abnormal points, automatically adjust channel and frequency parameters, and use network management tools to perform configuration updates to generate automatic configuration adjustment records;

[0036] S503: Based on the automatic configuration adjustment record, the adjusted network performance is evaluated, the effectiveness of the adjustment is verified through data comparison and analysis, and whether the network load change meets the expected standard is verified, and dynamic tuning information is generated.

[0037] The present invention is improved in that, based on the dynamic tuning information, the stability and efficiency of the network are evaluated, the performance difference before and after the tuning is analyzed by comparing the performance indicators, the network operation efficiency is determined, and the steps of obtaining the terminal performance evaluation log are specifically as follows:

[0038] S601: Based on the dynamic tuning information, continuously monitor performance indicators related to network stability and efficiency, perform data collection and preliminary trend analysis, and obtain network stability evaluation indicators;

[0039] S602: Based on the network stability evaluation index, calculate the change rate of the key performance index, compare the network response speed and data processing capacity at different time points, and evaluate the value of performance improvement or reduction to generate a performance efficiency comparison result;

[0040] S603: Based on the performance efficiency comparison result, key indicators before and after performance optimization are sorted and summarized, a network operation evaluation file is formulated, including key performance data and optimization records, and a terminal performance evaluation log is obtained.

[0041] An evaluation server for an LTE wireless network, the server comprising:

[0042] The historical data analysis module collects historical performance data of the LTE network, performs normalization processing, optimizes data deviation, and identifies load fluctuations and changes in signal strength through time node analysis to obtain historical performance analysis results;

[0043] The data structure adjustment module performs seasonal decomposition based on the historical performance analysis results, extracts trend and cyclical components, adjusts the components and matches the input requirements of the model, and generates a seasonal adjustment model;

[0044] The performance forecasting module predicts the load and performance of the network in the next week based on the seasonal adjustment model, compares it with the historical data of the same period, analyzes the potential trend changes, identifies the key fluctuation time points, and obtains the load forecast information;

[0045] The resource configuration module analyzes resource requirements during the load peak period based on the load forecast information, calculates the optimal bandwidth and signal adjustment amount, performs network configuration update operations, adjusts signal distribution and device settings, and generates a resource optimization plan;

[0046] Based on the resource optimization solution, the dynamic tuning module monitors network performance in real time, performs automatic adjustment of network configuration, dynamically updates channels and frequencies, matches changes in network load, and evaluates network stability and efficiency. Through performance indicators, it analyzes the performance differences before and after tuning and obtains terminal performance evaluation logs.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are:

[0048] In the present invention, through normalization processing and data deviation optimization, the availability of historical performance data and the accuracy of analysis are enhanced, allowing more accurate network load prediction and performance expectation, promoting the rationalization of resource allocation, seasonal decomposition reveals trends and cyclical components in the data, providing more detailed data insights, assisting the decision-making process of network tuning, using the ARIMA model to predict resource requirements and support vector machines for real-time performance monitoring, ensuring timely adjustment and optimization of network configuration, and improving the network's response flexibility and service continuity. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A flowchart of a method for evaluating an LTE wireless network is proposed for the present invention;

[0050] Figure 2 A schematic diagram of a detailed flow chart of step S1 in a method for evaluating an LTE wireless network is proposed in the present invention;

[0051] Figure 3 The present invention proposes a detailed flow chart of step S2 in a method for evaluating an LTE wireless network;

[0052] Figure 4 A schematic diagram of a detailed flow chart of step S3 in a method for evaluating an LTE wireless network is proposed in the present invention;

[0053] Figure 5 The present invention proposes a detailed flow chart of step S4 in a method for evaluating an LTE wireless network;

[0054] Figure 6 The present invention proposes a detailed flow chart of step S5 in a method for evaluating an LTE wireless network;

[0055] Figure 7 The present invention proposes a detailed flow chart of step S6 in a method for evaluating an LTE wireless network;

[0056] Figure 8 A server module diagram of an LTE wireless network is proposed for the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0058] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0059] Example

[0060] See also Figure 1 The present invention provides a technical solution: a method for evaluating an LTE wireless network, comprising the following steps:

[0061] S1: Based on the historical performance data of the LTE network, the data deviation is optimized through normalization processing, the change pattern of load fluctuation and signal strength is identified, and the data differences between different time nodes are compared to generate historical performance analysis results;

[0062] S2: Based on the historical performance analysis results, seasonal decomposition of the data series is performed to extract trend and cyclical components, and the data structure is adjusted to match the model input requirements to obtain a seasonal adjustment model;

[0063] S3: Based on the seasonal adjustment model, the network load and performance for the next week are predicted, trend changes and potential performance fluctuations are analyzed, and load forecast information is obtained;

[0064] S4: Based on the load forecast information, the ARIMA model is used to analyze the resource demand during the load peak period, calculate the required optimal bandwidth and signal adjustment amount, adjust the resource configuration according to the demand, and generate a resource optimization plan;

[0065] S5: Based on the resource optimization solution, support vector machines are used to monitor network performance in real time and automatically adjust network configuration, including dynamic updates of channels and frequencies, matching changes in network load, and generating dynamic tuning information;

[0066] S6: Based on the dynamic tuning information, evaluate the stability and efficiency of the network, analyze the performance difference before and after the tuning by comparing the performance indicators, determine the network operation efficiency, and obtain the terminal performance evaluation log.

[0067] Historical performance analysis results include normalized indicators, fluctuation trend graphs, and signal change curves. Seasonal adjustment models include trend component graphs, periodic component graphs, and residual graphs. Load forecast information includes daily load forecast tables and performance stability forecast graphs. Resource optimization plans include bandwidth adjustment plans, signal strength adjustment plans, and resource allocation schedules. Dynamic tuning information includes channel update records, frequency adjustment details, and performance monitoring logs. Terminal performance evaluation logs include tuning effect comparison data, operating efficiency data, and stability analysis results.

[0068] See also Figure 2 Based on the historical performance data of the LTE network, the steps to generate historical performance analysis results are as follows:

[0069] S101: Based on the historical performance data of the LTE network, outliers are removed through normalization processing, the data format is unified, and the records of different time points are time-standardized to obtain a normalized data set;

[0070] S102: Based on the normalized data set, a time series is constructed, key time points are extracted using a sliding window method, load fluctuations and signal strength changes at the time points are analyzed, and fluctuation records at the time points are obtained;

[0071] S103: Based on the time point fluctuation records, the time node data is compared, and the differential method is used to identify key change trends and abnormal fluctuations, including change points and fluctuation ranges, to generate historical performance analysis results.

[0072] S101: Based on the historical performance data of the LTE network, each measurement indicator is normalized. By analyzing the mean and standard deviation of each indicator, the Z-score method is used to standardize the data, thereby reducing the measurement deviation between devices. In addition, the outliers that exceed three times the interquartile range are identified and removed using box plots to ensure the purity of the data set. After that, for the data collected at different time points, the linear interpolation method is used to fill in the missing timestamps, so that the entire data set is unified in time, and a normalized and time-aligned data set is obtained.

[0073] S102: Based on the normalized data set, first define the time window, which is usually set to 30 minutes in length and moves for 5 minutes each time. Calculate basic statistics for the data in each window, such as average load and average signal strength. Then, apply the sliding window method to gradually extract windows in the entire time range, and record the changes in load and signal strength at each time point. The data is converted into corresponding fluctuation records through the time series method, and the fluctuation characteristics of each window are recorded with the timestamp as the index, recording the performance fluctuation of the network at different time points.

[0074] S103: Based on the fluctuation records of time points, the differential method is used, that is, the data points of the continuous time window are processed by first-order differential processing, so as to highlight the changing trend of the data, and the differential results are analyzed. By setting the threshold, it is determined which differential results represent normal trend changes and which represent abnormal fluctuations. The threshold is set based on the fluctuation distribution of historical data, and generally the 95% confidence interval of the differential value is taken. In addition, for the differential results exceeding the threshold, the change points and fluctuation ranges are further analyzed, such as by calculating the change rate and duration before and after each abnormal point to describe the nature of the fluctuation in detail. A historical performance analysis result is generated, which records the performance change trend of each key time node and the identified abnormal fluctuations.

[0075] See also Figure 3 Based on the historical performance analysis results, seasonal decomposition of the data series is performed, trend and cyclical components are extracted, and the data structure is adjusted to match the model input requirements. The specific steps to obtain the seasonal adjustment model are as follows:

[0076] S201: Based on the historical performance analysis results, perform data hierarchical decomposition, isolate seasonal and non-seasonal factors, highlight key seasonal patterns through data segmentation, and obtain seasonal pattern details;

[0077] S202: Based on the details of the seasonal pattern, the trend components and periodic fluctuations are refined, and the trend line is adjusted and balanced by the moving average method to optimize the seasonal adaptability of the data and obtain the trend smoothing result;

[0078] S203: Based on the trend smoothing results, the data structure is optimized by optimizing the data frequency and resolution to match the input requirements of the model and obtain a seasonal adjustment model.

[0079] S201: Based on the historical performance analysis results, we first use the time series decomposition method to divide the data hierarchy into seasonal and non-seasonal components, and use the polynomial fitting method to model the trend and seasonal components in the historical data set. The non-seasonal factors are isolated by polynomial fitting of the long-term trend, and the seasonal factors are identified by the contemporaneous averaging method. In this process, the data at the same time point each year are averaged to extract obvious seasonal patterns, such as the communication load peak in a specific month or week. Subsequently, the seasonal and non-seasonal data are processed separately through data segmentation technology to highlight the key seasonal patterns and establish a detailed pattern description for each season, so that the network operation and maintenance team can better understand the seasonal fluctuations in the data and their impact on network performance.

[0080] S202: Based on the details of the seasonal pattern, further refine the trend components and periodic fluctuations in the data, apply the moving average method to smooth the trend line of the seasonal data, and adjust the trend line of the data by calculating the average value of several consecutive time points to reduce the impact of short-term fluctuations on trend analysis. In this step, choosing an appropriate time window length is the key, which is usually set according to the periodic characteristics of the data. The application of moving average not only optimizes the seasonal adaptability of the data, but also helps reveal trend changes in a longer period. Through this process, a trend smoothing result is generated.

[0081] S203: Based on the trend smoothing results, the frequency and resolution of the data are optimized and adjusted to match the input requirements of the prediction model, including data resampling, converting the data from the original high-frequency sampling rate to the lower frequency required by the model, and ensuring the continuity of the data through interpolation methods. At the same time, the structure of the data is optimized, such as adjusting the data storage format and indexing method to improve the efficiency of data processing and the speed of model operation. Finally, through optimization measures, an adjustment model is constructed that meets the model input requirements and can accurately reflect seasonal changes.

[0082] See also Figure 4 , based on the seasonal adjustment model, forecast the network load and performance for the next week, analyze trend changes and potential performance fluctuations, and the steps to obtain load forecast information are as follows:

[0083] S301: Based on the seasonal adjustment model, quantitatively evaluate the daily network load and performance indicators in the next week, draw a future load trend chart, identify key trends and change ranges, and obtain trend change prediction details;

[0084] S302: Based on the trend change prediction details, compare with the historical data of the same period, determine the potential high-load time period and performance fluctuation range, mark the key fluctuation time points, and generate fluctuation period identification records;

[0085] S303: Based on the fluctuation period identification records, the key period data is centrally analyzed, a load and performance fluctuation prediction chart is created, key prediction information is sorted and summarized, and load prediction information is obtained.

[0086] S301: Based on the seasonal adjustment model, the model is first used to predict the daily network load and performance indicators in the next week. In the specific process, the model will input information including date, time, historical load data, etc., and use time series prediction technology, such as ARIMA or seasonal ARIMA, to calculate future load values ​​based on seasonal patterns and trend analysis. Each predicted value will be converted into a percentage increase or decrease relative to historical data, thereby drawing a network load trend chart for the next week. This chart not only shows the predicted load for each day, but also marks the expected highest and lowest load points. Through data, key trends and changes can be identified to provide decision support for network management.

[0087] S302: Based on the trend change prediction details, the predicted data is compared and analyzed with the historical data of the same period. The usual load level and performance fluctuations are determined through historical data, and compared with the predicted results to identify the high-load time periods and performance fluctuation ranges that occur compared to the normal state. This comparative analysis helps to determine the high-risk intervals that require extra attention during specific dates and time periods, mark the key fluctuation time points, and record them in the fluctuation period identification record, which describes the specific date, time and predicted load or performance indicators of each identified time point.

[0088] S303: Based on the records of fluctuation period identification, the data of key periods are analyzed centrally. Through data visualization tools, a fluctuation forecast chart containing load and performance is created. The chart shows the load trend and performance changes in key periods, providing an intuitive visual representation to facilitate the understanding and communication of complex data information. Next, the chart and its analysis results are sorted and summarized to form a detailed report. The report not only contains the time point and predicted value of each key prediction, but also outlines the impact of the prediction on future network operations.

[0089] See also Figure 5 Based on the load forecast information, the ARIMA model is used to analyze the resource demand during the load peak period, calculate the required optimal bandwidth and signal adjustment amount, and adjust the resource configuration according to the demand. The specific steps to generate the resource optimization plan are as follows:

[0090] S401: Based on the load forecast information, the ARIMA model is used to analyze the network resource usage in multiple time periods, calculate the bandwidth demand and signal adjustment demand during the peak period, and obtain bandwidth and signal demand indicators according to the network demand;

[0091] S402: Based on the bandwidth and signal demand indicators, design resource configuration measures that match the differential network loads, perform network resource allocation, match resource allocation with peak loads, and generate resource configuration details;

[0092] S403: Based on the resource configuration details, perform a network configuration update operation, adjust signal distribution and device settings, match the network configuration to the predicted load, and generate a resource optimization plan.

[0093] S401: Based on the load forecast information, the ARIMA model is used to conduct a detailed analysis of the predicted network load within the next week, especially focusing on the predicted high-load period. The model analysis will be refined into each time period in units of hours, evaluate the specific usage of network resources during high-load periods, and calculate the required bandwidth and signal strength to adjust the demand. To this end, a prediction model is constructed based on known user behavior and network equipment performance parameters, and historical load data and time variables are input to output the predicted bandwidth demand and signal strength for each time period. In this way, the peak and valley values ​​of the demand at different time points can be accurately calculated, and the bandwidth and signal demand indicators can be obtained based on this, providing a quantitative basis for resource optimization.

[0094] ARIMA model, according to the formula:

[0095] Φ(B)(1-B) d y t =θ(B)(1+αL+βL 2 ) t

[0096] Calculate the network load prediction value, where Φ(B) is the polynomial of the autoregressive parameter, Θ(B) is the polynomial of the moving average parameter, B is the backshift operator, d is the difference order, and y t is the load data in the time series, e t is the error term, L is the seasonal adjustment factor, α and β are the first and second weight coefficients of seasonal adjustment respectively.

[0097] Execution process:

[0098] First, the time series data y t The d-order difference of the data is taken to achieve data stationarity, and then the backshift operator B is applied to adjust the time series. The autoregressive polynomial Φ(B) and the moving average polynomial Θ(B) are used to estimate the future load. The seasonal adjustment factor L and the first and second weight coefficients α and β are added to adjust the impact of seasonal changes. This method increases the responsiveness of the model to seasonal fluctuations. The weight coefficients α and β are determined by fitting historical data to minimize the forecast error. t , thereby calculating the bandwidth requirements and signal strength requirements for each time period.

[0099] S402: Based on the bandwidth and signal demand indicators, formulate detailed resource allocation measures. First, design a set of flexible resource allocation strategies that can dynamically adjust bandwidth and signal allocation according to real-time changes in network load, including automatically increasing bandwidth supply and optimizing signal allocation during high-load periods, and reducing resource input accordingly during low-load periods to save costs. Then, through software-defined networking (SDN) technology, realize real-time allocation and adjustment of network resources, and record the time, scope and corresponding load forecast value of each adjustment operation in detail to ensure that resource allocation accurately matches actual demand, thereby generating resource configuration details.

[0100] S403: Based on the resource configuration details, perform the network configuration update operation. In this step, the network administrator will update the configuration settings of the network equipment through the management software, adjust the bandwidth allocation parameters of the routers and switches and the output strength of the signal amplifier, and pay special attention to the priority settings of key devices and links to ensure that sufficient resources can be provided to support communication needs during the predicted high-load period. The update operation will be performed according to the resource configuration details to ensure that each configuration adjustment can accurately correspond to the predicted network load, forming a complete resource optimization plan. The plan records every step from resource demand prediction to resource configuration adjustment to ensure the optimal performance of the network under different load conditions.

[0101] See also Figure 6 ,Based on the resource optimization scheme, the support vector machine is used to monitor the network performance in real time and automatically adjust the network configuration, including the dynamic update of channels and frequencies, matching the changes in network load, and generating dynamic tuning information. The specific steps are as follows:

[0102] S501: Based on the resource optimization solution, using support vector machines, continuously track network performance indicators, record performance fluctuations and network loads, identify immediate performance degradation or abnormal fluctuation points, and obtain real-time performance monitoring records;

[0103] S502: Based on the real-time performance monitoring record, analyze and identify abnormal points, automatically adjust channel and frequency parameters, and use network management tools to update configuration and generate automatic configuration adjustment records;

[0104] S503: Based on the automatic configuration adjustment record, the adjusted network performance is evaluated, the effectiveness of the adjustment is verified through data comparison and analysis, and whether the network load change meets the expected standard is verified, and dynamic tuning information is generated.

[0105] S501: Based on the resource optimization solution, the support vector machine (SVM) model is used to continuously monitor network performance indicators such as latency, packet loss rate and throughput. The SVM model is trained to identify performance degradation or abnormal fluctuation points that deviate from normal fluctuations. The SVM model uses its strong classification ability and good adaptability to perform real-time analysis of network performance. Whenever the network load changes, the model will analyze the data in real time and quickly identify any deviation from the predetermined performance threshold. Each abnormal point or performance degradation event identified will be recorded in the real-time performance monitoring record. This process not only improves the speed of problem discovery, but also enhances the network's self-healing ability.

[0106] S502: Based on the real-time performance monitoring records, the system will automatically analyze and identify anomalies and adjust channel and frequency parameters according to the analysis results. The adjustments are performed through the network management tool, which supports automated scripts and commands and can quickly respond to the output of the performance monitoring system. For example, if the monitoring records show that the load of a certain channel is continuously too high, the system will automatically reduce the load of the channel and increase the resource allocation of adjacent channels. Similarly, frequency adjustments are also automatically performed based on the signal quality monitoring results to optimize coverage and signal strength. After each automatic adjustment, the relevant configuration updates will be recorded in the automatic configuration adjustment record, which describes the time, reason and specific operations performed for each adjustment.

[0107] S503: Based on the automatic configuration adjustment records, a detailed evaluation of the adjusted network performance is performed by comparing the network performance data before and after the adjustment. Statistical analysis and data visualization techniques are used to demonstrate the effect of the adjustment to ensure that each adjustment achieves the expected improvement. In addition, the evaluation process will also verify whether the network load changes meet the pre-set standards and the output of the prediction model. Through comparative analysis, not only can the effectiveness of the adjustment be verified, but the prediction model can also be adjusted according to the actual results to improve the accuracy of future predictions. The dynamic tuning information generated after all information is aggregated provides a closed-loop feedback mechanism for continuous optimization and self-adjustment for network management.

[0108] See also Figure 7 Based on the dynamic tuning information, the stability and efficiency of the network are evaluated. The performance difference before and after the tuning is analyzed by comparing the performance indicators to determine the network operation efficiency. The specific steps to obtain the terminal performance evaluation log are as follows:

[0109] S601: Based on the dynamic tuning information, continuously monitor the performance indicators related to network stability and efficiency, collect data and conduct preliminary trend analysis, and obtain network stability evaluation indicators;

[0110] S602: Based on the network stability evaluation index, calculate the change rate of the key performance index, compare the network response speed and data processing capacity at different time points, and evaluate the performance improvement or reduction value to generate a performance efficiency comparison result;

[0111] S603: Based on the performance efficiency comparison results, key indicators before and after performance optimization are sorted and summarized, a network operation evaluation file is prepared, including key performance data and optimization records, and a terminal performance evaluation log is obtained.

[0112] S601: Based on dynamic tuning information, an integrated monitoring platform is used to continuously track and record performance indicators that are closely related to network stability and efficiency, such as network throughput, response time, and connection stability. This involves collecting data from multiple network nodes to ensure that all aspects of the network are covered. The collected data is subjected to preliminary trend analysis using advanced data analysis methods, and time series analysis is used to identify long-term trends and periodic fluctuations in performance. Through analysis, key factors that affect network stability are identified and converted into specific network stability evaluation indicators. The evaluation indicators provide network managers with real-time and dynamic network health status reports.

[0113] S602: Based on the network stability evaluation indicators, the change rate of key performance indicators is further calculated, including comparing the network response speed and data processing capacity at different time points, such as peak and trough periods, so as to evaluate the volatility and stability of network performance. The change rate is calculated by analyzing the performance data at consecutive time points, using the difference and percentage change formula to measure the performance improvement or decline. This method can accurately quantify the magnitude of performance changes and generate performance efficiency comparison results. The results show the performance changes of the network after optimization and adjustment, including which indicators have been improved and which require further attention, providing the network maintenance team with a clear performance evaluation and optimization direction.

[0114] S603: Based on the performance efficiency comparison results, key indicators before and after performance optimization are sorted and summarized. In this process, all relevant performance data and optimization measures are recorded together to form a comprehensive network operation evaluation file. The file not only includes key performance data, but also records the results of the optimization operations and adjustments, as well as the specific impact of the changes on network performance. Finally, the generated terminal performance evaluation log provides a detailed performance and optimization history record, which is convenient for network operation and maintenance personnel to track performance trends, evaluate the effectiveness of optimization measures, and make further operation and maintenance decisions accordingly.

[0115] See also Figure 8 , an evaluation server for an LTE wireless network, the server comprising:

[0116] The historical data analysis module collects historical performance data of the LTE network, performs normalization processing, optimizes data deviation, and identifies load fluctuations and changes in signal strength through time node analysis to obtain historical performance analysis results;

[0117] The data structure adjustment module performs seasonal decomposition based on historical performance analysis results, extracts trend and cyclical components, adjusts the components and matches the input requirements of the model to generate a seasonal adjustment model;

[0118] The performance forecasting module predicts the network load and performance in the next week based on the seasonal adjustment model, compares it with historical data for the same period, analyzes potential trend changes, identifies key fluctuation time points, and obtains load forecast information;

[0119] The resource configuration module analyzes resource requirements during load peaks based on load forecast information, calculates optimal bandwidth and signal adjustment, performs network configuration update operations, adjusts signal distribution and device settings, and generates resource optimization plans;

[0120] The dynamic tuning module monitors network performance in real time based on resource optimization solutions, performs automatic adjustment of network configuration, dynamically updates channels and frequencies, matches changes in network load, and evaluates network stability and efficiency. It analyzes performance differences before and after tuning through performance indicators and obtains terminal performance evaluation logs.

[0121] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for evaluating an LTE wireless network, characterized in that: The following steps are involved: Based on the historical performance data of the LTE network, the data deviation is optimized through normalization processing, the change pattern of load fluctuation and signal strength is identified, and the data differences between different time nodes are compared to generate historical performance analysis results; Based on the historical performance analysis results, seasonal decomposition of the data series is performed to extract trend and cyclical components, and the data structure is adjusted to match the model input requirements to obtain a seasonal adjustment model; Based on the seasonal adjustment model, forecast network load and performance for the next week, analyze trend changes and potential performance fluctuations, and obtain load forecast information; Based on the load forecast information, the ARIMA model is used to analyze the resource demand during the load peak period, calculate the required optimal bandwidth and signal adjustment amount, and adjust the resource configuration according to the demand to generate a resource optimization plan; Based on the resource optimization scheme, support vector machines are used to monitor network performance in real time and automatically adjust network configuration, including dynamic updates of channels and frequencies, to match changes in network load and generate dynamic tuning information; Based on the dynamic tuning information, the stability and efficiency of the network are evaluated, the performance difference before and after the tuning is analyzed by comparing the performance indicators, the network operation efficiency is determined, and the terminal performance evaluation log is obtained.

2. The method for evaluating an LTE wireless network according to claim 1, wherein: The historical performance analysis results include normalized indicators, fluctuation trend graphs, and signal change curves; the seasonal adjustment model includes trend component graphs, periodic component graphs, and residual graphs; the load forecasting information includes daily load forecast tables and performance stability forecast graphs; the resource optimization plans include bandwidth adjustment plans, signal strength adjustment plans, and resource allocation schedules; the dynamic tuning information includes channel update records, frequency adjustment details, and performance monitoring logs; and the terminal performance evaluation logs include tuning effect comparison data, operating efficiency data, and stability analysis results.

3. The method for evaluating an LTE wireless network according to claim 1, wherein: Based on the historical performance data of the LTE network, the steps to generate historical performance analysis results are as follows: Based on the historical performance data of the LTE network, we normalized the data, removed outliers, unified the data format, and performed time standardization on the records of different time points to obtain a normalized data set. Based on the normalized data set, a time series is constructed, key time points are extracted using a sliding window method, load fluctuations and signal strength changes at the time points are analyzed, and fluctuation records at the time points are obtained; Based on the fluctuation records at the time points, the time node data are compared and the differential method is used to identify key change trends and abnormal fluctuations, including change points and fluctuation ranges, to generate historical performance analysis results.

4. The method for evaluating an LTE wireless network according to claim 1, wherein: Based on the historical performance analysis results, seasonal decomposition of the data series is performed, trend and cyclical components are extracted, and the data structure is adjusted to match the model input requirements. The steps of obtaining the seasonal adjustment model are as follows: Based on the historical performance analysis results, perform data hierarchical decomposition, isolate seasonal and non-seasonal factors, highlight key seasonal patterns through data segmentation, and obtain seasonal pattern details; Based on the seasonal pattern details, the trend components and periodic fluctuations are refined, and the trend line is adjusted and balanced by the moving average method to optimize the seasonal adaptability of the data and obtain a trend smoothing result; Based on the trend smoothing results, the data structure is optimized by optimizing the data frequency and resolution to match the input requirements of the model and obtain the seasonal adjustment model.

5. The method for evaluating an LTE wireless network according to claim 1, wherein: Based on the seasonal adjustment model, the network load and performance for the next week are predicted, and trend changes and potential performance fluctuations are analyzed. The steps for obtaining load forecast information are as follows: Based on the seasonal adjustment model, quantitatively evaluate the network load and performance indicators of each day in the next week, draw a future load trend chart, identify key trends and change ranges, and obtain trend change prediction details; Based on the trend change prediction details, compare with the historical data of the same period, determine the potential high-load time period and performance fluctuation range, mark the key fluctuation time points, and generate fluctuation period identification records; Based on the fluctuation period identification records, the key period data is analyzed centrally, a load and performance fluctuation forecast chart is created, key forecast information is sorted and summarized, and load forecast information is obtained.

6. The method for evaluating an LTE wireless network according to claim 1, wherein: Based on the load forecast information, the ARIMA model is used to analyze the resource demand during the load peak period, calculate the required optimal bandwidth and signal adjustment amount, and adjust the resource configuration according to the demand. The steps of generating a resource optimization plan are as follows: Based on the load forecast information, the ARIMA model is used to analyze the network resource usage in multiple time periods, calculate the bandwidth demand and signal adjustment demand during the peak period, and obtain bandwidth and signal demand indicators based on network demand; Based on the bandwidth and signal demand indicators, design resource configuration measures that match the differential network loads, perform network resource allocation, match resource allocation with peak loads, and generate resource configuration details; Based on the resource configuration details, a network configuration update operation is performed to adjust signal distribution and device settings, match the network configuration to the predicted load, and generate a resource optimization plan.

7. The method for evaluating an LTE wireless network according to claim 6, wherein: The ARIMA model, according to the formula: Φ(B)(1-B) d y t =Θ(B)(1+αL+βL 2 )e t Calculate the network load prediction value, where Φ(B) is the polynomial of the autoregressive parameter, Θ(B) is the polynomial of the moving average parameter, B is the backshift operator, d is the difference order, and y t is the load data in the time series, e t is the error term, L is the seasonal adjustment factor, α and β are the first and second weight coefficients of seasonal adjustment respectively.

8. The method for evaluating an LTE wireless network according to claim 1, wherein: Based on the resource optimization scheme, support vector machines are used to monitor network performance in real time and automatically adjust network configuration, including dynamic update of channels and frequencies, to match changes in network loads, and to generate dynamic tuning information in the following specific steps: Based on the resource optimization solution, support vector machines are used to continuously track network performance indicators, record performance fluctuations and network loads, identify immediate performance degradation or abnormal fluctuation points, and obtain real-time performance monitoring records; Based on the real-time performance monitoring records, analyzing and identifying abnormal points, automatically adjusting channel and frequency parameters, and using network management tools to perform configuration updates to generate automatic configuration adjustment records; Based on the automatic configuration adjustment record, the adjusted network performance is evaluated, the effectiveness of the adjustment is verified through data comparison and analysis, and whether the network load change meets the expected standard is verified to generate dynamic tuning information.

9. The method for evaluating an LTE wireless network according to claim 1, wherein: Based on the dynamic tuning information, the stability and efficiency of the network are evaluated, the performance difference before and after the tuning is analyzed by comparing the performance indicators, the network operation efficiency is determined, and the steps of obtaining the terminal performance evaluation log are specifically as follows: Based on the dynamic tuning information, continuously monitor performance indicators related to network stability and efficiency, perform data collection and preliminary trend analysis, and obtain network stability evaluation indicators; Based on the network stability evaluation indicators, calculate the change rate of key performance indicators, compare the network response speed and data processing capacity at different time points, and evaluate the value of performance improvement or decline to generate performance efficiency comparison results; Based on the performance efficiency comparison results, key indicators before and after performance optimization are sorted and summarized, a network operation evaluation file is developed, including key performance data and optimization records, and a terminal performance evaluation log is obtained.

10. An evaluation server for an LTE wireless network, characterized in that: According to the LTE wireless network evaluation method according to any one of claims 1 to 9, the server comprises: The historical data analysis module collects historical performance data of the LTE network, performs normalization processing, optimizes data deviation, and identifies load fluctuations and changes in signal strength through time node analysis to obtain historical performance analysis results; The data structure adjustment module performs seasonal decomposition based on the historical performance analysis results, extracts trend and cyclical components, adjusts the components and matches the input requirements of the model, and generates a seasonal adjustment model; The performance forecasting module predicts the load and performance of the network in the next week based on the seasonal adjustment model, compares it with the historical data of the same period, analyzes the potential trend changes, identifies the key fluctuation time points, and obtains the load forecast information; The resource configuration module analyzes resource requirements during the load peak period based on the load forecast information, calculates the optimal bandwidth and signal adjustment amount, performs network configuration update operations, adjusts signal distribution and device settings, and generates a resource optimization plan; Based on the resource optimization solution, the dynamic tuning module monitors network performance in real time, performs automatic adjustment of network configuration, dynamically updates channels and frequencies, matches changes in network load, and evaluates network stability and efficiency. Through performance indicators, it analyzes the performance differences before and after tuning and obtains terminal performance evaluation logs.

Images