Wireless communication operation and maintenance data analysis method and system

Abstract

The invention relates to the technical field of wireless communication, and discloses a wireless communication operation and maintenance data analysis method and system, and the method comprises the following steps: in all sub-regions of a wireless communication operation and maintenance region, carrying out the high-frequency similarity analysis of the business flow value data of each historical time period, and constructing the data analyzed in each time period to obtain a first set; performing distance analysis on the first set of each time period, and constructing a second set through a distance analysis result; and performing influence analysis on the sub-regions in the second set, and constructing a third set according to an influence analysis result. According to the method, the analysis sub-regions which influence each other and are key in the wireless communication network can be accurately identified, so that operation and maintenance personnel can focus on key regions, carry out key monitoring on the key analysis sub-regions, distribute network resources more reasonably, increase bandwidth in the key regions, optimize base station configuration and the like, and the resource utilization efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a method and system for analyzing wireless communication operation and maintenance data. Background Art

[0002] With the rapid development of wireless communication technology, wireless communication networks are increasingly used in people's daily lives, work and various industries. From the popularization of smart phones to the large-scale access of IoT devices, the amount of data carried by wireless communication networks has exploded, which has put forward extremely high requirements on the performance and stability of the network. However, in the long-term wireless communication operation and maintenance practice, the operation and maintenance personnel first noticed the instability of network performance. When allocating network resources, they found that allocating resources based on experience could not better identify the mutually influential and critical analysis sub-areas in the wireless communication network, so that the operation and maintenance personnel could focus on key areas, reasonably allocate network resources, and improve resource utilization efficiency. For example, due to the failure to identify key areas in advance and predict peaks, the network will be severely congested and users will not be able to use network services normally. Summary of the invention

[0003] The object of the present invention is to provide a wireless communication operation and maintenance data analysis method and system to solve at least one of the above-mentioned prior art problems.

[0004] In a first aspect, the present invention provides a method for analyzing wireless communication operation and maintenance data, comprising the following steps: In all sub-areas of the wireless communication operation and maintenance area, high-frequency similarity analysis is performed on the service flow value data of each historical period, and the data analyzed in each period is constructed to obtain a first set; Perform distance analysis on the first set of each time period, and construct a second set based on the distance analysis results; An impact analysis is performed on the sub-regions in the second set, a third set is constructed based on the impact analysis results, and an impact range analysis is performed on the sub-regions in the third set to obtain key sub-regions.

[0005] As a further solution of the present invention: the process of obtaining the first set is: Obtain historical business traffic value data for each analysis area and set the analysis cycle and analysis period; For any analysis period in the same analysis cycle of all analysis sub-areas, the fourth set is determined by using the Chebyshev distance method; The number of occurrences of the analysis sub-regions contained in all similar analysis sub-region sets is obtained respectively, and the ratios are calculated with the total number of analysis cycles respectively to obtain the occurrence ratios; The analysis sub-regions with a proportion greater than or equal to the occurrence ratio threshold are marked as high-frequency similar sub-regions, and the first set of the analysis period is constructed.

[0006] As a further solution of the present invention: the process of obtaining the second set is: The distance between any two analysis areas in any high-frequency similarity analysis sub-area set is calculated by the distance formula, and the Euclidean distance matrix is ​​constructed; Calculate the mean jz and standard deviation bz of the non-zero elements in the Euclidean distance matrix. , then mark the second set, where α is the proportional coefficient.

[0007] As a further solution of the present invention: the process of obtaining the third set is: Data analysis is performed on the analysis sub-region combinations in the second set to obtain relevant impact degree values, and the analysis sub-region combinations in the second set that are greater than the relevant impact degree threshold are marked as the third set.

[0008] As a further solution of the present invention: the process of obtaining the relevant impact degree value is: The number of analysis periods in which two adjacent analysis sub-regions in the adjacent analysis sub-region combination appear in the high-frequency similar analysis sub-region set is counted respectively, and the ratio is calculated with the total number of analysis periods to obtain the regional similarity ratio; The distance values ​​of the two adjacent analysis sub-regions are calculated by ratio with the judgment distance threshold to obtain the regional distance ratio; The mutual information correlation coefficient is used to calculate the correlation representation value; The regional similarity ratio, regional distance ratio and related characterization values ​​are processed to obtain the relevant impact degree value.

[0009] As a further solution of the present invention: the process of obtaining the relevant characterization value is: For each analysis period, the mutual information correlation coefficient is used to calculate the correlation value of the business flow value data changes between the two analysis sub-areas; The correlation values ​​corresponding to all analysis periods in the same analysis period are averaged to obtain the correlation mean of the analysis period, and then the correlation means of all analysis periods are averaged to obtain the correlation representation value.

[0010] As a further solution of the present invention: the acquisition process of the key analysis sub-region is: Count the analysis sub-regions in the third set whose analysis sub-region combinations are not repeated, and mark them as analysis sub-regions to be marked; Traverse all the analysis sub-region combinations in the third set and mark each analysis sub-region, count the number of times each analysis region is marked, and calculate the ratio with the analysis sub-region to be marked to obtain the marking frequency ratio, and mark the analysis sub-regions that are greater than or equal to the marking frequency ratio threshold.

[0011] As a further solution of the present invention: also include: A business traffic forecasting model is built through historical business traffic value data of key analysis sub-areas to predict business traffic value data and identify peak analysis periods.

[0012] As a further solution of the present invention: obtaining a key analysis sub-area and obtaining corresponding historical business flow value data; Construct a business traffic forecasting model for each key analysis sub-area and obtain a business traffic forecasting value data sequence; Extract the number of service flow prediction values ​​that are greater than the service flow limit, and perform ratio processing with the total number of service flow prediction values ​​to obtain the proportion of high service flow values; The difference between the predicted service flow values ​​greater than the service flow limit and the service flow limit is calculated, and the average is processed, and then the ratio is calculated with the service flow limit to obtain the high service flow degree value; The peak characterization value is obtained by calculating the proportion of the number of high business traffic values ​​and the high business traffic degree value, and the analysis period corresponding to the key analysis sub-area greater than the peak characterization threshold is marked as the peak analysis period.

[0013] In a second aspect, the present invention provides a wireless communication operation and maintenance data analysis system, the system comprising: Similarity identification module: in all sub-areas of the wireless communication operation and maintenance area, high-frequency similarity analysis is performed on the service flow value data of each historical period, and the data analyzed in each period is constructed to obtain a first set; Distance recognition module: performs distance analysis on the first set of each time period and constructs the second set based on the distance analysis results; Impact and key identification module: perform impact analysis on the sub-regions in the second set, construct a third set based on the impact analysis results, and perform impact range analysis on the sub-regions in the third set to obtain key sub-regions.

[0014] Beneficial effects of the present invention: 1. The present invention can help operation and maintenance personnel clearly understand the distribution of similar areas of business traffic in different time periods, provide a basis for accurate planning of network resources, improve resource utilization efficiency, ensure network stability, and dig out the law and pattern of business traffic changes, and provide a reference for long-term network optimization and development planning. At the same time, by discovering business traffic similar areas with spatial aggregation characteristics, the present invention can solve the problems of unclear business traffic distribution, lack of pertinence in network resource planning, and low troubleshooting efficiency in wireless communication networks to a certain extent, and help wireless communication operation and maintenance work to be carried out more efficiently; 2. The present invention can accurately identify the mutually influential and critical analysis sub-areas in the wireless communication network, so that the operation and maintenance personnel can focus on the key areas, conduct key monitoring on these key analysis sub-areas, allocate network resources more reasonably, increase bandwidth in key areas, optimize base station configuration, etc., and improve resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 It is a flow chart of a wireless communication operation and maintenance data analysis method of the present invention; Figure 2 It is a structural schematic diagram of a wireless communication operation and maintenance data analysis system of the present invention. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme 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 only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0018] Embodiment 1 Figure 1A flowchart of a wireless communication operation and maintenance data analysis method provided in Embodiment 1 of the present invention. The embodiment of the present invention can be applied to the situation of analyzing the distribution of service traffic in a wireless communication network. The wireless communication operation and maintenance data analysis method can be executed by a wireless communication operation and maintenance data analysis system. The wireless communication operation and maintenance data analysis system can be implemented by software and / or hardware. The wireless communication operation and maintenance data analysis system can be configured in a wireless communication operation and maintenance data analysis device. Optionally, a wireless communication operation and maintenance data analysis device can be an electronic device, which can be a notebook, a desktop computer, a smart tablet, etc., and the embodiment of the present invention does not limit this.

[0019] A wireless communication operation and maintenance data analysis method provided by an embodiment of the present invention specifically includes the following steps: Step 1: In all sub-areas of the wireless communication operation and maintenance area, high-frequency similarity analysis is performed on the service flow value data of each historical period, and the data analyzed in each period is constructed to obtain a high-frequency similarity analysis sub-area set (first set); In some embodiments, the wireless communication operation and maintenance analysis area is divided into a plurality of analysis sub-areas, and historical service flow value data of each analysis area is obtained; Among them, the service flow value is a quantitative indicator to measure the service transmission volume in the wireless communication network, reflecting the amount of various service data passing through the network in a specific time period, and the acquisition method includes but is not limited to: network equipment statistics, service platform data interface; Set the analysis cycle and divide the analysis cycle into several analysis time periods. The analysis cycle and analysis time period are set by the implementers in this field. Generally, the analysis cycle is set to 24 hours (i.e., every day), and the analysis time period is set to 2 hours or 4 hours. Based on all analysis sub-areas, in any analysis period in the same analysis cycle, the Chebyshev distance method is used to determine similar analysis sub-areas. The specific process is as follows: Assuming that there are N analysis sub-areas in total, for the i-th analysis sub-area, its service flow value sequence during the analysis period is: , where m is the number of historical service flow values ​​in the analysis period; For example, suppose that wireless communication operation and maintenance data analysis is performed on a central business district of a city, and the area is divided into 8 analysis sub-areas, namely, sub-area 1 (first office building), sub-area 2 (first office building), sub-area 3 (shopping mall), sub-area 4 (restaurant concentration area), sub-area 5 (hotel), sub-area 6 (bank branch area), sub-area 7 (café block), and sub-area 8 (park green space), then N=8; The analysis cycle is set to one day (24 hours), and the analysis period is set to 2 hours, so there are 12 analysis periods (0-2 hours, 2-4 hours, ..., 22-24 hours). The business flow value data of each analysis sub-area in each analysis period for the past 15 days (15 analysis periods) are collected. Assume that the number of historical business flow values ​​in each analysis period is m=2 (that is, the business flow value is recorded once every hour); For any two analysis sub-regions i and j, calculate the Chebyshev distance , where the Chebyshev distance calculation formula is: , by looping through all combinations of analysis sub-regions, a Chebyshev distance matrix D of N is constructed, where represents the Chebyshev distance between analysis sub-regions i and j during the analysis period, , They represent the service flow value of the kth sampling point in the i-th analysis sub-area during the analysis period, and the service flow value of the kth sampling point in the j-th analysis sub-area during the analysis period; The Chebyshev distance focuses on the maximum difference between the service flow value sequences of two analysis sub-regions at corresponding positions. This ensures that when judging similar analysis sub-regions, the local significant differences will not be masked by the overall average difference. It can accurately reflect the degree of difference in service flow change patterns between sub-regions under extreme conditions, thereby accurately screening out sub-regions with similar service flow change characteristics. Set the distance threshold and traverse the Chebyshev distance matrix D. If , then the analysis sub-regions i and j are marked as having similar changes in the analysis period, and the analysis sub-regions with similar changes can be combined and stored in a set; Exemplarily, for the analysis period t, the set storing similar analysis sub-region combinations is ; After all analysis periods are processed according to the above steps, a set of similar analysis sub-regions corresponding to each analysis period is obtained (the fourth set); Based on any analysis period, obtain a set of similar analysis sub-regions of all analysis periods. For any similar analysis sub-region set, obtain the number of occurrences of the analysis sub-regions contained in the similar analysis sub-region set, and calculate the ratio with the total number of analysis periods to obtain the occurrence ratio. Set the occurrence ratio threshold, and mark the analysis sub-regions that are greater than or equal to the occurrence ratio threshold as high-frequency similar sub-regions, otherwise, mark them as non-high-frequency similar sub-regions; Obtain a high-frequency similarity analysis sub-region set, marked as the high-frequency similarity analysis sub-region set for the analysis period; Get all high-frequency similarity analysis sub-region sets for each analysis period; For example, taking the analysis period "10-12 o'clock" as an example, the similar analysis sub-area combination obtained through calculation is: , and store it in the corresponding collection ; For the analysis period "10-12 o'clock", the similar analysis sub-area sets of the past 15 analysis periods are counted; Assume that in these 15 cycles, the similar analysis sub-region set containing sub-region 1 appears 12 times, the similar analysis sub-region set containing sub-region 2 appears 10 times, the similar analysis sub-region set containing sub-region 3 appears 8 times, and the similar analysis sub-region set containing sub-region 5 appears 6 times; Then the occurrence frequency of sub-region 1 accounts for 80%; the occurrence frequency of sub-region 2 accounts for 67%; the occurrence frequency of sub-region 3 accounts for 53%; the occurrence frequency of sub-region 5 accounts for 40%; Based on the occurrence ratio threshold, sub-region 1 and sub-region 2 are marked as high-frequency similar sub-regions, and sub-region 3 and sub-region 5 are marked as non-high-frequency similar sub-regions; After the above steps, the high-frequency similarity analysis sub-region set for the analysis period "10-12 o'clock" is obtained as {1,2}; The remaining 11 analysis periods are processed according to the above process to obtain all high-frequency similarity analysis sub-region sets for each analysis period. For example, the high-frequency similarity analysis sub-region set for the analysis period "18:00-20:00" may be {3, 4, 7}, and the high-frequency similarity analysis sub-region set for the analysis period "2:00-4:00" may be an empty set. Based on these collections, operation and maintenance personnel can understand which areas have similar wireless communication service traffic changes at different times, so as to better plan network resources and perform operation and maintenance operations; By calculating the Chebyshev distance, we can screen out a set of high-frequency similar analysis sub-regions and then determine the peak hours. Operation and maintenance personnel can clearly understand the regional distribution of similar changes in business traffic at different times, and thus carry out targeted network resource planning during the time periods when similar areas are concentrated. For example, during peak hours, we can allocate more bandwidth to relevant areas in advance, increase base station capacity, etc., to improve resource utilization efficiency and ensure stable network operation. By analyzing a large amount of historical data based on Chebyshev distance, we can dig out the potential rules and patterns of business traffic changes in different analysis sub-areas at different times. These rules help operation and maintenance personnel to deeply understand the usage characteristics of wireless communication networks, such as the traffic change rules of different functional areas (office buildings, shopping malls, etc.) at different times, and provide reference for long-term network optimization and development planning; Step 2: Perform distance analysis on the high-frequency similarity analysis sub-region set in each time period, and construct a proximity analysis sub-region combination (the second set) based on the distance analysis results; Among them, "adjacent" means close to or near, which highlights that although there is a close distance relationship between regions in space, they are not necessarily directly adjacent; In some embodiments, a high-frequency similarity analysis sub-region set is obtained for each time period; It should be noted that there are sets of two or more analysis sub-regions in the high-frequency similarity analysis sub-region set; Obtaining the longitude and latitude coordinates of the center point of each analysis sub-area, wherein the longitude and latitude coordinates may be obtained by, but not limited to: extracting from map data, base station positioning data, or regional planning data through a geographic information system (GIS); For example, for sub-areas such as office buildings and shopping malls, the longitude and latitude of the center point can be calculated based on their boundary ranges on the map; Based on any set of high-frequency similarity analysis sub-regions, the distance between any two analysis regions is calculated using the distance formula; Among them, the distance value between any two analysis areas can be calculated by Euclidean distance. The specific formula is: ,in, is the center coordinate of the analysis sub-region m, is the center coordinate of the analysis sub-region n; For example, if the coordinates of the center point of the analysis sub-region m are (116.40, 39.90), and the coordinates of the center point of the analysis sub-region n are (116.42, 39.91), the Euclidean distance between them can be calculated by substituting the coordinates into the formula; If there are r regions in the high-frequency similarity analysis sub-region set, each pair of analysis sub-regions is traversed through a nested loop, and the calculated distance is stored in an r×r Euclidean distance matrix JL, where Represents the distance between region m and region n. When m=n, ; Calculate the mean jz and standard deviation bz of the non-zero elements in the Euclidean distance matrix, and set the distance threshold , where α is the proportionality coefficient, which is set by practitioners in this field based on data characteristics and experience; like , then the corresponding two analysis sub-regions are marked as adjacent analysis sub-regions; like , then the corresponding two analysis sub-regions are marked as non-adjacent analysis sub-regions; Respectively obtain adjacent analysis sub-region combinations of all markers in each high-frequency similarity analysis sub-region set; The effect of judging the adjacent sub-regions is that: performing distance analysis on high-frequency similar analysis sub-regions to obtain adjacent analysis sub-regions is helpful to discover similar business traffic regions with spatial aggregation characteristics; When troubleshooting wireless communication faults, since adjacent analysis sub-areas may be affected by the same factors, if a fault occurs in one area, you can focus on the adjacent areas to troubleshoot in advance to prevent the fault from spreading and improve the reliability and stability of the network. The technical solution of this embodiment is: divide the wireless communication operation and maintenance analysis area into several sub-areas, set the analysis cycle and time period, collect the historical business traffic value data of each sub-area, use the Chebyshev distance method to construct a distance matrix, set the distance threshold to screen out similar analysis sub-area combinations, and then determine the high-frequency similar analysis sub-area set by calculating the proportion of the number of occurrences, so as to clarify the business traffic similar change areas in different time periods, and then extract a set containing two or more sub-areas based on the high-frequency similar analysis sub-area set in step one, obtain the longitude and latitude coordinates of the center point of each sub-area, calculate the distance using the Euclidean distance formula, and mark the adjacent analysis sub-area combination by setting the judgment distance threshold; Based on the above process, operation and maintenance personnel can clearly understand the distribution of similar service traffic areas in different periods, provide a basis for accurate planning of network resources, improve resource utilization efficiency, ensure network stability, and dig out the rules and patterns of service traffic changes, providing a reference for long-term network optimization and development planning; At the same time, by discovering similar areas of business traffic with spatial clustering characteristics, it can solve the problems of unclear business traffic distribution, lack of targeted network resource planning, and low troubleshooting efficiency in wireless communication networks to a certain extent, thereby helping to carry out wireless communication operation and maintenance work more efficiently.

[0020] Embodiment 2 Based on the above embodiments, Figure 1 As shown, a wireless communication operation and maintenance data analysis method provided by an embodiment of the present invention specifically includes the following steps: Step 3: Based on the adjacent analysis sub-region combination, perform impact analysis on the analysis sub-regions in the adjacent analysis sub-region combination, and construct a mutual impact analysis sub-region combination (third set) based on the impact analysis results; In some embodiments, all combinations of analysis sub-regions marked as being adjacent in each set of analysis sub-regions to be determined are obtained; Based on any combination of adjacent analysis sub-regions, count the number of analysis periods in which the two adjacent analysis sub-regions appear in the set of high-frequency similar analysis sub-regions, and calculate the ratio with the total number of analysis periods to obtain the regional similarity ratio; The purpose of calculating the regional similarity ratio is to measure the similarity of the service traffic change patterns of adjacent analysis sub-regions, thereby providing a basis for evaluating the degree of mutual influence between them. In the wireless communication operation and maintenance scenario, regions with similar service traffic change patterns are more likely to influence each other in terms of network performance, service usage, etc. The distance values ​​of the two adjacent analysis sub-regions are calculated by ratio with the judgment distance threshold to obtain the regional distance ratio; The purpose of calculating the regional distance ratio is to introduce the spatial distance factor for quantitative consideration when evaluating the mutual influence degree of adjacent analysis sub-regions. In wireless communication operation and maintenance, spatial distance is one of the important factors affecting the mutual influence between regions. The smaller the regional distance ratio is, the closer the two adjacent analysis sub-regions are in space. Under the same other conditions, the possibility and degree of mutual influence between them are usually greater. When calculating the relevant impact value, the regional distance ratio is used as one of the key parameters and is calculated together with other data values ​​to help comprehensively evaluate the mutual impact of adjacent analysis sub-regions, thereby more accurately determining the combination of mutually impact analysis sub-regions, providing an important quantitative basis for subsequent network optimization and other operation and maintenance work; Count the analysis time periods in which the two adjacent analysis sub-areas appear in the high-frequency similar analysis sub-area set, and obtain corresponding business traffic value data; For each analysis period, the mutual information correlation coefficient is used to calculate the correlation value of the business traffic value data changes between the two analysis sub-areas. The calculation process is: The service flow values ​​of two analysis sub-areas A and B are as follows: , the service flow value of sub-area B is , where k represents the number of service flow values; The calculation formula of mutual information correlation coefficient is: ; The calculated mutual information correlation coefficient is the correlation value, and the correlation values ​​corresponding to all analysis periods of the same analysis period are averaged to obtain the correlation mean of the analysis period; The correlation degree means of all analysis periods are averaged to obtain the correlation representation value; The regional similarity ratio xs, regional distance ratio jz and related representation value xg are processed to obtain the related influence degree value YX; The specific data processing formula is: ,in, , , It is a preset proportionality coefficient, which is set by practitioners in this field based on experience; The purpose of calculating the relevant impact degree value is to enable operation and maintenance personnel to quickly and accurately identify sub-region combinations with a high degree of mutual influence between regions and exclude combinations with a small influence; for regional combinations with a high degree of mutual influence, they can be considered as a whole for resource allocation during network planning; they can effectively formulate fault troubleshooting and prevention strategies. When troubleshooting a certain area, they can simultaneously check and prevent the areas that affect it, discover potential faults in advance, prevent the spread of faults, and improve the reliability and stability of the network; A threshold of the relevant impact degree is set, and the adjacent analysis sub-region combinations that are greater than the relevant impact degree threshold are marked as mutual impact analysis sub-region combinations, otherwise, they are marked as non-relevant impact analysis sub-region combinations; Get all relevant impact analysis sub-area combinations; Step 4: Analyze the sub-region combinations based on mutual impact, and perform impact range analysis to obtain key analysis sub-regions; In some embodiments, all combinations of mutually-influenced analysis sub-regions are obtained, and all non-repetitive analysis sub-regions therein are counted and marked as analysis sub-regions to be marked; Traverse all the mutual influence analysis sub-region combinations and mark each analysis sub-region, count the number of times each analysis region is marked, and calculate the ratio of the number of times the analysis region is marked to the number of analysis sub-regions to be marked to obtain the marking number ratio; For example, if the mutual influence analysis sub-region combination is {(A,B), (A,C), (B,D)}, then during the traversal process, A appears 2 times, B appears 2 times, C appears 1 time, and D appears 1 time; A marking times ratio threshold is set, and the analysis sub-regions with a marking times ratio greater than or equal to the marking times ratio threshold are marked as key analysis sub-regions, and the analysis sub-regions with a marking times ratio less than the marking times ratio threshold are marked as non-key analysis sub-regions; The technical solution of this embodiment is as follows: first, obtain the adjacent analysis sub-region combinations in each set of analysis sub-regions to be judged, calculate the regional similarity ratio and regional distance ratio for these combinations, obtain the relevant characterization value by combining the mutual information correlation coefficient, calculate the relevant influence degree value by a preset formula, screen out the mutually influential analysis sub-region combinations according to the set threshold, and then based on these combinations, count the non-repeated analysis sub-regions, traverse the combinations to count the number of times each analysis sub-region is marked and calculate the ratio of the number of marks, and mark the key analysis sub-regions according to the set threshold after sorting; It can accurately identify the mutually influential and critical analysis sub-areas in the wireless communication network, so that operation and maintenance personnel can focus on key areas, conduct key monitoring of these key analysis sub-areas, allocate network resources more reasonably, increase bandwidth in key areas, optimize base station configuration, etc., improve resource utilization efficiency, and give priority to key areas when troubleshooting, quickly locate and solve problems, prevent faults from spreading, and ensure stable network operation. It can also provide a scientific basis for long-term network planning and optimization based on the business traffic characteristics and mutual influence relationships in key areas, and promote the sustainable development of wireless communication networks.

[0021] Embodiment 3 Based on the above embodiments, Figure 1 As shown, a wireless communication operation and maintenance data analysis method provided by an embodiment of the present invention specifically includes the following steps: Step 5: Build a business traffic forecasting model through the historical business traffic value data of the key analysis sub-area, forecast the business traffic value data, and identify the peak analysis period; In some embodiments, a key analysis sub-area is obtained, and corresponding historical service flow value data is obtained; Based on any key analysis sub-area, select an appropriate prediction model to build a business traffic prediction model according to the characteristics of the data and analysis requirements; Among them, the selection of appropriate prediction models includes but is not limited to: long short-term memory network (LSTM), moving average model (SARIMA); Taking the LSTM model as an example, the preprocessed data is divided into a training set and a test set according to a certain ratio, such as 70% as a training set and 30% as a test set; Determine the structure and parameters of the LSTM model, including but not limited to: number of hidden layers, number of neurons, learning rate, and number of iterations; The training set data is input into the LSTM model for training. The model continuously adjusts the weights and biases through the back-propagation algorithm to minimize the error between the predicted value and the actual value. During training, you can use the early stopping method to prevent overfitting, that is, stop training when the performance of the model on the validation set no longer improves; Use the test set data to evaluate the trained model and calculate the evaluation indicators, which include but are not limited to: root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE); If the model evaluation results are not ideal, you can try to adjust the model parameters, which include but are not limited to: increasing the number of hidden layer neurons and adjusting the learning rate; Until a business traffic prediction model that meets the performance indicators is obtained; Based on the business traffic prediction model, predict the business traffic value data of a future analysis period to obtain a business traffic prediction value data sequence; Extract the number of service flow prediction values ​​that are greater than the service flow limit, and perform ratio processing with the total number of service flow prediction values ​​to obtain the proportion of high service flow values; Among them, the business flow limit is set by the implementers in this field based on the historical business flow value and operation and maintenance work experience; The difference between the predicted service flow values ​​greater than the service flow limit and the service flow limit is calculated respectively, and the average is processed, and the obtained value is calculated by ratio with the service flow limit to obtain the high service flow degree value; The peak characterization value is obtained by multiplying the proportion of high business traffic values ​​by the high business traffic degree value; A peak characterization threshold is set, and a key analysis sub-region with a value greater than the peak characterization threshold is marked as a peak analysis sub-region, and its analysis period is a peak analysis period, otherwise, it is a non-peak analysis period; Based on the peak analysis sub-area, other analysis sub-areas in the combination of the mutually-influencing analysis sub-areas are obtained, and these analysis sub-areas and the peak analysis sub-area are monitored in real time; Real-time monitoring of other analysis sub-areas in the combination of peak analysis sub-areas can detect network failures that may be caused by business traffic peaks in advance. Since these areas have a mutual influence relationship, high traffic in the peak analysis sub-area may be transmitted to other areas, triggering a chain reaction. Real-time monitoring helps operation and maintenance personnel to detect abnormalities in time and take measures to prevent the spread of faults, thereby improving network reliability and stability and reducing the impact of network failures on users. At the same time, practitioners in this field can also set the monitoring frequency of the analysis area that interacts with the peak analysis sub-area according to the relevant influence degree value between the peak analysis sub-area and the analysis area that interacts with it. For example, if the relevant influence degree value is small, a larger monitoring frequency is set; if the relevant influence degree value is large, a smaller monitoring frequency is set; The technical solution of this embodiment is: obtain key analysis sub-areas and their historical business traffic value data, select a suitable model according to data characteristics and analysis requirements, build a business traffic prediction model, use the model to predict the traffic data of a future analysis period, calculate the proportion of high business traffic values ​​and the high business traffic degree value, and then obtain the peak characterization value, compare it with the peak characterization threshold to identify the peak analysis sub-area and its period, and finally perform real-time monitoring on the peak analysis sub-area and other analysis sub-areas in the combination of the mutually influential analysis sub-areas; This can predict and identify peak analysis sub-areas and cycles, provide a basis for forward-looking allocation of network resources, and monitor relevant areas in real time to promptly detect and prevent the spread of network failures caused by business traffic peaks, reduce the impact of network failures on users, improve network reliability and stability, and help wireless communication network operation and maintenance work to be carried out more efficiently and intelligently to meet users' growing network service needs.

[0022] Embodiment 4 Based on the above embodiments, Figure 2 As shown, a wireless communication operation and maintenance data analysis system provided by an embodiment of the present invention specifically includes: Similarity identification module: obtains the service flow value data of each analysis sub-area in the wireless communication operation and maintenance analysis area from historical data, and performs time period analysis to obtain a high-frequency similar analysis sub-area set in each time period; Proximity recognition module: Based on the high-frequency similarity analysis sub-region set in each time period, the high-frequency similarity analysis sub-region performs distance analysis to obtain the proximity analysis sub-region combination; Impact identification module: Based on the combination of adjacent analysis sub-areas, the mutual influence relationship analysis is performed to evaluate the mutual influence degree of adjacent analysis sub-areas and obtain the combination of mutual influence analysis sub-areas; Key identification module: analyzes the sub-region combination based on mutual influence and obtains key analysis sub-regions; Prediction and Identification Module: Based on the historical business traffic value data of the key analysis sub-areas, a business traffic prediction model is built to predict business traffic value data and identify peak analysis periods. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0023] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0024] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A wireless communication operation and maintenance data analysis method, characterized in that: The following steps are involved: In all sub-areas of the wireless communication operation and maintenance area, high-frequency similarity analysis is performed on the service flow value data of each historical period, and the data analyzed in each period is constructed to obtain a first set; Perform distance analysis on the first set of each time period, and construct a second set based on the distance analysis results; An impact analysis is performed on the sub-regions in the second set, a third set is constructed based on the impact analysis results, and an impact range analysis is performed on the sub-regions in the third set to obtain key sub-regions.

2. A wireless communication operation and maintenance data analysis method according to claim 1, characterized in that: The process of obtaining the first set is as follows: Set an analysis period consisting of multiple time periods. For any time period in the same analysis period; The similarity of each time period is determined by the Chebyshev distance method, and the fourth set is constructed based on the similarity results; Get all the different sub-regions contained in the fourth set, and calculate the occurrence ratio of each sub-region; All sub-areas with a percentage greater than or equal to the occurrence ratio threshold are obtained to construct the first set of the time period.

3. A wireless communication operation and maintenance data analysis method according to claim 1, characterized in that: The process of obtaining the second set is as follows: For each first set, the distance between any two analysis areas in the first set is calculated, and a Euclidean distance matrix is ​​constructed; Calculate the mean jz and standard deviation bz of the non-zero elements in the Euclidean distance matrix. , it is marked as a combination of adjacent analysis sub-regions, where α is the proportional coefficient.

4. A wireless communication operation and maintenance data analysis method according to claim 1, characterized in that: The acquisition process of the third set is: Data analysis is performed on the sub-regions in the second set to obtain relevant influence degree values, and a combination of sub-regions in the second set that is greater than a relevant influence degree threshold is obtained and marked as a third set.

5. A wireless communication operation and maintenance data analysis method according to claim 4, characterized in that: The process of obtaining the relevant impact value is as follows: Calculate the proportion of the number of time periods in which the sub-region combinations in the second set appear in the first set, and obtain the regional similarity proportion; Calculate the ratio of the distance value of the sub-region combination in the second set to the judgment distance threshold to obtain the region distance ratio; Calculating the relevant characterization value of the combination of sub-regions in the second set; The relevant impact degree value is calculated by calculating the regional similarity ratio, regional distance ratio and relevant characterization value.

6. A wireless communication operation and maintenance data analysis method according to claim 5, characterized in that: The process of obtaining the relevant characterization value is as follows: For each time period, the mutual information correlation coefficient is used to calculate the correlation value of the business flow value data changes between the two sub-areas; The correlation values ​​corresponding to all analysis periods in the same period are averaged, and then the correlation means of all analysis periods are averaged to obtain the correlation representation value.

7. A wireless communication operation and maintenance data analysis method according to claim 1, characterized in that: The acquisition process of the key sub-area is as follows: Combining the non-repeated sub-regions in the third set and marking them as sub-regions to be marked for analysis; Traverse all the third set of sub-region combinations and mark each sub-region, and count the number of times each sub-region is marked relative to the number of times each sub-region is marked; The analysis sub-regions with a marking frequency ratio greater than or equal to a threshold are marked as key sub-regions.

8. A wireless communication operation and maintenance data analysis method according to claim 1, characterized in that: Also includes: A business traffic prediction model is built through historical business traffic value data of key sub-areas to predict business traffic value data and identify peak analysis periods.

9. A wireless communication operation and maintenance data analysis method according to claim 8, characterized in that: The acquisition process of identifying the peak analysis period is as follows: Construct a business traffic forecasting model for each key analysis sub-area and obtain a business traffic forecasting value data sequence; Calculate the proportion of the number of service flow forecast values ​​that are greater than the service flow limit; Calculate the difference between all service flow prediction values ​​greater than the service flow limit and the service flow limit respectively, perform average processing, and then calculate the ratio with the service flow limit to obtain the high service flow degree value; The peak characterization value is obtained by calculating the proportion of the number of high business traffic values ​​and the high business traffic degree value, and the analysis period corresponding to the key analysis sub-area greater than the peak characterization threshold is marked as the peak analysis period.

10. A wireless communication operation and maintenance data analysis system, characterized in that: The system is used to execute the method described in any one of claims 1 to 9, and the system comprises: Similarity identification module: in all sub-areas of the wireless communication operation and maintenance area, high-frequency similarity analysis is performed on the service flow value data of each historical period, and the data analyzed in each period is constructed to obtain a first set; Distance recognition module: performs distance analysis on the first set of each time period and constructs the second set based on the distance analysis results; Impact and key identification module: perform impact analysis on the sub-regions in the second set, construct a third set based on the impact analysis results, and perform impact range analysis on the sub-regions in the third set to obtain key sub-regions.

Images