Processing method and apparatus for wireless network, and electronic device

Abstract

The application discloses a processing method and device of a wireless network and electronic equipment, and the method comprises the following steps: acquiring a first characteristic index affecting the wireless network; determining a first root cause index set affecting the first characteristic index and a confidence degree corresponding to each first root cause index; calculating the value of the first characteristic index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characteristic index; obtaining root cause information and a score corresponding to the root cause information according to the first root cause index set, the confidence degree and the first dynamic threshold; and locating the root cause affecting the performance of the wireless network according to the root cause information and the score corresponding to the root cause information. Through the application, the problem that the index obtained by the clustering algorithm based on machine learning in the related art is easy to be affected and has poor interpretability, and the root cause affecting the performance of the wireless network cannot be accurately located is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless network, in particular to a wireless network processing method and device and electronic equipment. BACKGROUND

[0002] At present, wireless network technology has entered the era of high speed and large throughput. In order to meet the diversified development of wireless network service, the number of cells and the complexity of networking have been greatly improved. It is very challenging to quickly locate the root cause of poor quality cells in wireless network.

[0003] One of the prior art processing methods is to use KPI index to locate poor quality cells, but this method does not consider the specific perception of users to the performance of wireless network, and it relies too much on human experience to set KPI index combination and threshold, which is time-consuming and inefficient. In addition, there is another method that uses machine learning clustering algorithm to cluster and analyze multiple KPI indexes to obtain the threshold for locating poor quality cells, but the above method does not consider that the output result of the clustering algorithm is greatly affected by the clustering index data and the initial preset classification, and the output result has poor interpretability. At the same time, a large number of data samples are needed during calculation, which consumes a lot of computing resources and storage resources.

[0004] In view of the problem that the indexes obtained by the clustering algorithm based on machine learning in the related art are easily affected and have poor interpretability, resulting in the inability to accurately locate the root cause affecting the performance of wireless network, no effective solution has been proposed so far. SUMMARY

[0005] The main purpose of the present application is to provide a wireless network processing method and device and electronic equipment to solve the problem that the indexes obtained by the clustering algorithm based on machine learning in the related art are easily affected and have poor interpretability, resulting in the inability to accurately locate the root cause affecting the performance of wireless network.

[0006] In order to achieve the above object, according to one aspect of the present application, a processing method of a wireless network is provided, the method comprising: acquiring a first characteristic index affecting the wireless network, wherein the first characteristic index is a related index of user evaluation information of wireless network performance; determining a first root cause index set affecting the first characteristic index and a confidence degree corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indexes affecting the first characteristic index, and the first root cause index is an operating index of a base station or a network of a cell to which a user belongs; calculating a value of the first characteristic index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characteristic index; obtaining root cause information and a score corresponding to the root cause information according to the first root cause index set, the confidence degree and the first dynamic threshold; and locating a root cause affecting wireless network performance according to the root cause information and the score corresponding to the root cause information.

[0007] Further, acquiring the first characteristic index affecting the wireless network comprises: determining a collection period and a collection dimension of indexes affecting the wireless network, and acquiring the indexes affecting the operation of the wireless network, wherein the collection dimension is a dimension of collecting the indexes in a cell or a user; judging whether the indexes are associated with user evaluation information of wireless network performance; if the indexes are associated with the user evaluation information of wireless network performance, dividing the indexes into second characteristic indexes; if the indexes are not associated with the user evaluation information of wireless network performance, dividing the indexes into second root cause indexes; performing aggregation calculation on the second characteristic indexes and the second root cause indexes to obtain the first characteristic index and the first root cause index; and storing the first characteristic index and the first root cause index in a database.

[0008] Further, determining the first root cause index set affecting the first characteristic index and the confidence degree corresponding to each first root cause index comprises: constructing a data sample set of an FP-grown algorithm model according to the collection dimension and the collection period; inputting the first characteristic index and the first root cause index in the database into the FP-grown algorithm model to obtain the first root cause index set affecting the first characteristic index and the confidence degree corresponding to each first root cause index.

[0009] Further, the calculating, based on a time series prediction algorithm, a value of the first characteristic index to obtain a first dynamic threshold of the first characteristic index comprises: constructing long-term trend, periodic trend and random fluctuation three time series according to the collection period; preprocessing the first characteristic index according to the time series to remove outliers of the first characteristic index in the time series to obtain the first characteristic index after processing; training the first characteristic index after processing based on the time series using a Prophet algorithm to obtain the first dynamic threshold of the first characteristic index changing dynamically with time.

[0010] Further, the obtaining, according to the first root cause index set, the confidence and the first dynamic threshold, root cause information and a score corresponding to the root cause information comprises: obtaining a third root cause index set from the first root cause index set with the highest confidence; taking an inflection point of a value of a third root cause index in the third root cause index set as a second threshold; obtaining a target root cause index according to the first dynamic threshold; building a root cause tree, traversing the root cause tree according to the target root cause index and the second threshold to obtain the root cause information and the score corresponding to the root cause information.

[0011] Further, the obtaining, according to the first dynamic threshold, a target root cause index comprises: obtaining a third characteristic index lower than the first dynamic threshold in the first characteristic index; obtaining a root cause index corresponding to the third characteristic index from the database as the target root cause index.

[0012] Further, the building a root cause tree, traversing the root cause tree according to the target root cause index and the second threshold to obtain the root cause information and the score corresponding to the root cause information comprises: classifying root cause indexes affecting wireless networks according to a preset rule to obtain root cause indexes of different categories, and building the root cause tree according to root cause indexes of different categories to obtain the root cause tree of different categories; continuing to traverse child nodes of different branches of the target root cause tree node according to a comparison result of whether a value of the target root cause index is greater than the second threshold corresponding to a root cause index represented by a tree node of the target root cause tree, then continuing to traverse the child nodes of the target root cause tree node until the leaf node of the target root cause tree is traversed, obtaining the root cause information corresponding to the leaf node and calculating the score corresponding to the root cause information, wherein the target root cause tree is the root cause tree of the category to which the target root cause index belongs; outputting the root cause information and the score corresponding to the root cause information.

[0013] Further, according to a preset rule, the root cause indexes affecting the wireless network are classified to obtain root cause indexes of different categories, and the root cause trees of different categories are built according to the root cause indexes of different categories, including: according to a preset rule, acquiring root cause indexes of the same category affecting the wireless network to obtain a fourth root cause index set; in the fourth root cause index set, taking a first main cause index as a parent node of a first root cause tree, taking a secondary cause index affected by the first main cause index as a first child node of the parent node, taking a second main cause index irrelevant to the first main cause index as a second child node of the parent node, to obtain the parent node, the first child node and the second child node corresponding to the parent node, wherein the first main cause index is a root cause index with the largest influence range in the fourth root cause index set, and the second main cause index is a root cause index with the largest influence range among other root cause indexes irrelevant to the first main cause index in the fourth root cause index set; according to the generation process of the parent node, the first child node and the second child node corresponding to the parent node, the remaining root cause indexes in the fourth root cause index set are continuously generated into other nodes of the first root cause tree until all root cause indexes in the fourth root cause index set appear in the first root cause tree, to obtain the first root cause tree; according to the method of building the first root cause tree, the root cause trees of different categories are obtained.

[0014] To achieve the above-mentioned purpose, according to another aspect of the present application, a processing device of a wireless network is provided, which comprises: a first acquisition unit configured to acquire a first representation index affecting a wireless network, wherein the first representation index is a relevant index of user evaluation information of wireless network performance; a determination unit configured to determine a first root cause index set affecting the first representation index and a confidence degree corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indexes affecting the first representation index, and the first root cause index is an operating index of a base station or a network of a cell to which a user belongs; a second acquisition unit configured to calculate a value of the first representation index based on a time series prediction algorithm to obtain a first dynamic threshold of the first representation index; a third acquisition unit configured to obtain root cause information and a score corresponding to the root cause information according to the first root cause index set, the confidence degree and the first dynamic threshold; and a positioning unit configured to locate a root cause affecting wireless network performance according to the root cause information and the score corresponding to the root cause information.

[0015] Further, the first acquisition unit comprises: a first determination subunit, configured to determine a collection period and a collection dimension of an index affecting a wireless network, and acquire the index affecting operation of the wireless network, wherein the collection dimension is a dimension of collecting the index in terms of a cell or a user; a judgment subunit, configured to judge whether the index is associated with evaluation information of a user on performance of the wireless network; a first division subunit, configured to divide the index into a second representation index if the index is associated with the evaluation information of the user on the performance of the wireless network; a second division subunit, configured to divide the index into a second root cause index if the index is not associated with the evaluation information of the user on the performance of the wireless network; a first acquisition subunit, configured to perform aggregation calculation on the second representation index and the second root cause index to obtain the first representation index and the first root cause index; and a storage subunit, configured to store the first representation index and the first root cause index into a database.

[0016] Further, the determination unit comprises: a first construction subunit, configured to construct a data sample set of an FP-grown algorithm model according to the collection dimension and the collection period; and a second acquisition subunit, configured to input the first representation index and the first root cause index in the database into the FP-grown algorithm model to perform calculation, to obtain a first root cause index set affecting the first representation index and the confidence corresponding to each first root cause index.

[0017] Further, the second acquisition unit comprises: a second construction subunit, configured to construct three time series of long-term trend, periodic trend and random fluctuation according to the collection period; a third acquisition subunit, configured to preprocess the first representation index according to the time series to remove abnormal values of the first representation index in the time series, to obtain a processed first representation index; and a fourth acquisition subunit, configured to train the processed first representation index using a Prophet algorithm based on the time series, to obtain the first dynamic threshold of dynamic change of the first representation index over time.

[0018] Further, the third acquisition unit comprises: a fifth acquisition subunit, configured to acquire a third root cause index with the highest confidence from the first root cause index set to obtain a third root cause index set; a second determination subunit, configured to determine an inflection point of a value of a third root cause index in the third root cause index set as a second threshold; a sixth acquisition subunit, configured to acquire a target root cause index according to the first dynamic threshold; and a seventh acquisition subunit, configured to build a root cause tree, traverse the root cause tree according to the target root cause index and the second threshold, and obtain the root cause information and the score corresponding to the root cause information.

[0019] Further, the sixth obtaining subunit comprises: a first obtaining module, configured to obtain, in the first representation index, a third representation index lower than the first dynamic threshold; and a second obtaining module, configured to obtain, from the database, a root cause index corresponding to the third representation index as the target root cause index.

[0020] Further, the seventh obtaining subunit comprises: a third obtaining module, configured to classify root cause indexes affecting a wireless network according to a preset rule to obtain root cause indexes of different categories, and build the root cause tree according to the root cause indexes of different categories to obtain the root cause tree of different categories; a fourth obtaining module, configured to continue traversing child nodes of different branches of a node of the target root cause tree according to a comparison result of whether a value of the target root cause index is greater than a second threshold corresponding to a root cause index represented by the node of the target root cause tree, obtain root cause information corresponding to the leaf node, and calculate a score corresponding to the root cause information, wherein the target root cause tree is the root cause tree of a category to which the target root cause index belongs; and an output module, configured to output the root cause information and the score corresponding to the root cause information.

[0021] Further, the third obtaining module comprises: a first obtaining submodule, configured to obtain root cause indexes of the same category affecting a wireless network according to a preset rule to obtain a fourth root cause index set; a second obtaining submodule, configured to, in the fourth root cause index set, take a first main cause index as a parent node of a first root cause tree, take a secondary cause index affected by the first main cause index as a first child node of the first main cause index node, and take a second main cause index irrelevant to the first main cause index as a second child node of the first main cause index to obtain the parent node, the first child node corresponding to the parent node, and the second child node, wherein the first main cause index is a root cause index with the largest influence range in the fourth root cause index set, and the second main cause index is a root cause index with the largest influence range among other root cause indexes irrelevant to the first main cause index in the fourth root cause index set; a third obtaining submodule, configured to continue generating other nodes of the first root cause tree from the fourth root cause index set according to a generation process of the parent node, the first child node corresponding to the parent node, and the second child node until all root cause indexes in the fourth root cause index set appear in the first root cause tree to obtain the first root cause tree; and a fourth obtaining submodule, configured to obtain the root cause tree of different categories according to the method of building the first root cause tree.

[0022] In order to achieve the above object, according to an aspect of the present application, an electronic device is provided, comprising one or more processors and a memory, the memory being configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned wireless network processing methods.

[0023] By the present application, the following steps are adopted: acquiring a first characteristic index affecting a wireless network, wherein the first characteristic index is a related index of user evaluation information of wireless network performance; determining a first root cause index set affecting the first characteristic index and a confidence degree corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indexes affecting the first characteristic index, and the first root cause index is an operating index of a base station or a network of a cell to which a user belongs; calculating a value of the first characteristic index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characteristic index; obtaining root cause information and a score corresponding to the root cause information according to the first root cause index set, the confidence degree and the first dynamic threshold; and locating a root cause affecting wireless network performance according to the root cause information and the score corresponding to the root cause information. The problem that indexes obtained based on a machine learning clustering algorithm are easily affected and have poor interpretability, resulting in an inability to accurately locate the root cause affecting wireless network performance, is solved. By acquiring a characteristic index, a root cause index associated with the characteristic index and a confidence degree of the root cause index, a threshold for distinguishing between poor quality cells is obtained through calculation, and finally the root cause index affecting the poor quality cell is analyzed in detail, the characteristic index and the root cause index that are more consistent with user perception are obtained, the root cause index affecting the poor quality cell is more reasonably determined, the root cause affecting the wireless network performance is more quickly and accurately located according to the root cause index affecting the poor quality cell, and thus the effect that the staff can be helped to quickly and accurately locate the root cause affecting the wireless network performance is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which form a part of the present application, are intended to provide further understanding of the present application, and the illustrative embodiments of the present application and their description serve the purpose of explaining the present application. The accompanying drawings should not be construed in a limiting manner. In the drawings:

[0025] Figure 1 is a flowchart of a wireless network processing method according to an embodiment of the present application;

[0026] Figure 2 is a schematic diagram of an optional wireless network processing method according to an embodiment of the present application Figure 1 ;

[0027] Figure 3is a schematic diagram of an optional wireless network processing method provided by an embodiment of the application Figure 2 ;

[0028] Figure 4 is a schematic diagram of an optional wireless network processing method provided by an embodiment of the application Figure 3 ;

[0029] Figure 5 is a schematic diagram of an optional wireless network processing method provided by an embodiment of the application Figure 4 ;

[0030] Figure 6 is a schematic diagram of a wireless network processing device provided by an embodiment of the application. DETAILED DESCRIPTION

[0031] It should be noted that the embodiments and features in the application can be combined with each other without conflict. The application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0032] In order to enable those skilled in the art to better understand the application scheme, the technical solutions in the embodiments of the application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only a part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the application.

[0033] It should be noted that the terms "first", "second" and the like in the specification and claims of the application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0034] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for display, analyzed data, etc.) involved in the present disclosure are all information and data authorized by the user or authorized by all parties.

[0035] The application will be described below in conjunction with the preferred implementation steps, Figure 1A flowchart of a processing method of a wireless network is provided according to an embodiment of the present application, as shown in Figure 1 The method comprises the following steps:

[0036] In step S101, a first characteristic index affecting the wireless network is acquired, wherein the first characteristic index is a related index of evaluation information of the user on the performance of the wireless network.

[0037] In the embodiment, the characteristic index refers to the evaluation information of the user on the performance and running condition of the wireless network, for example, the webpage opening success rate, the wireless drop-off rate, the HTTP page opening delay, etc. can be used as the characteristic index. In order to improve the user experience of the cell wireless network, the related index of the evaluation information of the user on the performance and running condition of the wireless network (i.e. the first characteristic index) needs to be acquired first.

[0038] In step S102, a first root cause index set affecting the first characteristic index and a confidence degree corresponding to each first root cause index are determined, wherein the first root cause index set is a set of multiple first root cause indexes affecting the first characteristic index, and the first root cause index is a running index of the network of the base station or the cell to which the user belongs.

[0039] In the embodiment, the root cause index refers to the running index of the wireless network of the base station or the cell (i.e. the first root cause index), which will affect the evaluation information (i.e. the characteristic index) of the user on the performance of the wireless network, for example, the average RSRP (Reference Signal Receiving Power) index, which is a key parameter indicating the wireless signal strength in the wireless network. If the average RSRP index of the cell is relatively high, the webpage opening success rate index of the cell will be improved accordingly, and the wireless drop-off rate index will be reduced accordingly. In addition to the average RSRP index, the near-end weak coverage proportion and the average power transmission margin are also root cause indexes. In order to obtain the root cause indexes affecting the characteristic index, the root cause index set affecting the running of the cell wireless network and the probability (i.e. the confidence degree) of each root cause index possibly affecting the characteristic index need to be acquired.

[0040] In step S103, the value of the first characteristic index is calculated based on a time series prediction algorithm to obtain a first dynamic threshold of the first characteristic index.

[0041] In the actual use process of the cell wireless network, the characteristic index of the wireless network may change in the index value with the change of time. In order to more accurately locate the root cause affecting the performance of the cell wireless network, when the threshold of the poor-quality cell is acquired, the influence of the time change on the characteristic index of the wireless network needs to be excluded to obtain the dynamic threshold (i.e. the first dynamic threshold) of the characteristic index changing with time.

[0042] For example, webpages typically load less successfully on holidays than on weekdays. However, whether it's a holiday or not is not the root cause affecting the performance of the community's wireless network; it may simply be an external factor. Therefore, to more accurately pinpoint the root causes affecting the performance of the community's wireless network, it's necessary to reduce the impact of time variations on performance metrics.

[0043] Step S104: Based on the first root cause index set, confidence level and first dynamic threshold, obtain root cause information and the corresponding score of root cause information.

[0044] In this embodiment, in order to more accurately locate the root cause affecting the performance of the cell's wireless network, it is necessary to obtain detailed root cause information (i.e., root cause information) affecting the characterization index, and the probability (i.e., score) of each root cause affecting the performance of the cell's wireless network, based on the first root cause index set, confidence level and first dynamic threshold calculated according to the above scheme.

[0045] Step S105: Based on the root cause information and the corresponding score, locate the root cause affecting the performance of the wireless network.

[0046] After obtaining the root cause information and its corresponding score, the root cause affecting the performance of the cell's wireless network can be located.

[0047] In summary, the wireless network processing method provided in this application embodiment obtains a first characterization index affecting the wireless network, wherein the first characterization index is a relevant index of user evaluation information of wireless network performance; determines a first root cause index set affecting the first characterization index and a confidence level corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indices affecting the first characterization index, and the first root cause index is an operational index representing the network of the cell to which the user belongs, based on a base station or wireless network; calculates the value of the first characterization index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characterization index; obtains root cause information and a score corresponding to the root cause information based on the first root cause index set, the confidence level, and the first dynamic threshold; and locates the root cause affecting the performance of the wireless network based on the root cause information and the score corresponding to the root cause information. This solves the problem that the indicators obtained by the clustering algorithm based on machine learning are easily affected and have poor interpretability, resulting in the inability to accurately locate the root cause affecting the performance of the wireless network. By acquiring representative indicators, root cause indicators associated with these indicators, and the confidence levels of these root cause indicators, a threshold for distinguishing poor-quality cells is calculated. Finally, a detailed analysis of the root cause indicators affecting poor-quality cells is conducted. This approach achieves the acquisition of representative and root cause indicators that are more aligned with user perception, enabling a more reasonable determination of the root cause indicators affecting poor-quality cells. Based on these root cause indicators, the root causes affecting wireless network performance can be located more quickly and accurately, thus helping staff to more quickly and accurately pinpoint the root causes affecting wireless network performance.

[0048] Optionally, in the wireless network processing method provided in this application embodiment, obtaining the first characterization index affecting the wireless network includes: determining the collection period and collection dimension of the index affecting the wireless network, obtaining the index affecting the operation of the wireless network, wherein the collection dimension is the dimension of collecting indicators by cell or user; determining whether the index is related to the user's evaluation information on the performance of the wireless network; if the index is related to the user's evaluation information on the performance of the wireless network, then the index is classified as a second characterization index; if the index is not related to the user's evaluation information on the performance of the wireless network, then the index is classified as a second root cause index; performing convergence calculation on the second characterization index and the second root cause index to obtain the first characterization index and the first root cause index; and storing the first characterization index and the first root cause index in a database.

[0049] In existing technologies, multiple KPI indicators for a cell are typically determined, different KPI indicators are combined, and thresholds for these KPI indicators are defined. Then, cells with poor quality are identified based on the combination of KPI indicators and the thresholds. This method of selecting cells with poor quality relies on human experience to combine different KPI indicators and define thresholds. It is a relatively abstract method for identifying cells with poor quality, which is seriously out of touch with the actual user experience. Moreover, it cannot be adjusted in a timely manner according to the actual fluctuations of the wireless network, making it difficult to accurately pinpoint the root causes affecting the performance of the wireless network.

[0050] In this embodiment, to accurately pinpoint the root causes affecting wireless network performance, it is necessary to obtain relatively specific and detailed evaluation information (i.e., characterization indicators) from users regarding wireless network performance and operation. First, the collection period and dimensions for collecting wireless network indicators are determined. Based on whether the collected indicators are directly related to users' evaluation information on wireless network performance, the indicators are divided into two categories: characterization indicators and root cause indicators. Indicators related to users' evaluation information on wireless network performance are called characterization indicators, while indicators not directly related are called root cause indicators. Next, the obtained characterization indicators and root cause indicators are aggregated and calculated to obtain a more rationally categorized first characterization indicator and first root cause indicator. Finally, the first characterization indicator and first root cause indicator are stored in a database for subsequent identification of the root causes affecting wireless network performance.

[0051] By determining whether the collected indicators are directly correlated with users' evaluation information of wireless network performance, the indicators are divided into characterization indicators and root cause indicators. The indicators used to locate the root causes affecting wireless network performance are more in line with users' actual perception of the community's wireless network. The resulting characterization indicators and root cause indicators are also more scientific and reasonable, thus improving the accuracy of locating the root causes affecting wireless network performance.

[0052] Optionally, in the wireless network processing method provided in this application embodiment, determining the set of first root cause indicators affecting the first characterization indicator and the confidence level corresponding to each first root cause indicator includes: constructing a data sample set for the FP-grown algorithm model based on the collection dimension and collection period; inputting the first characterization indicator and the first root cause indicator in the database into the FP-grown algorithm model for calculation to obtain the set of first root cause indicators affecting the first characterization indicator and the confidence level corresponding to each first root cause indicator.

[0053] In existing technologies, multiple sets of multi-dimensional data are typically obtained from the big data of VoLTE signaling call detail records (VoLTE, a high-speed wireless communication standard for mobile phones and data terminals) to locate the root cause affecting wireless network performance. Each time the root cause is located using this method, all relevant indicators and multiple sets of multi-dimensional data need to be learned. However, the results obtained are often only suitable for analysis of a single cell and are difficult to generalize to the entire network, resulting in a large consumption of learning resources and time and low work efficiency.

[0054] In this embodiment, to broaden the applicability of the method for locating the root causes affecting wireless network performance, the FP-grown algorithm is used to calculate the first characterization metric and the first root cause metric, mining the correlation between the characterization metric and the root cause metric to obtain the set of first root cause metrics affecting the first characterization metric and the confidence level corresponding to each first root cause metric. The calculation process using the FP-grown algorithm is as follows: Figure 2 As shown, firstly, a dataset for training the model needs to be collected. The first representation indicator and the first root cause indicator are associated based on cell number and time period, and data cleaning, format unification, and data standardization are completed. Next, a data sample set is constructed according to the requirements of the FP-grown algorithm. The first representation indicator and the first root cause indicator are reorganized according to cell, resulting in multiple datasets, with each cell number corresponding to one dataset. Then, the FP-grown algorithm is used to mine frequent itemsets, and rule learning is performed on the rules associated with the first representation indicator and the first root cause indicator to obtain a rule base. This generates a set of first root cause indicators that have a strong correlation with the first representation indicator, and the probability (i.e., confidence score) of the first root cause indicator set affecting the first representation indicator is calculated. Finally, it is determined whether it is necessary to continue mining the association rules between the first representation indicator and the first root cause indicator set. If it is necessary to continue mining association rules, the process returns to the first step to continue collecting data to mine frequent itemsets and learn rules. If it is not necessary to continue mining association rules, the set of first root cause indicators associated with the first representation indicator and the confidence score corresponding to each first root cause indicator are output.

[0055] The FP-grown algorithm is used to calculate the first characterization index and the first root cause index, thereby obtaining a set of first root cause indices that are strongly correlated with the first characterization index and the confidence level corresponding to each first root cause index. This allows for more accurate results in locating the root causes affecting wireless network performance, thus making the method for locating the root causes affecting wireless network performance more widely applicable and the results more accurate.

[0056] Optionally, in the wireless network processing method provided in this application embodiment, calculating the value of the first characterization index based on the time series prediction algorithm to obtain the first dynamic threshold of the first characterization index includes: constructing three time series based on the collection period: long-term trend, periodic trend, and random fluctuation; preprocessing the first characterization index based on the time series to remove outliers in the time series to obtain the processed first characterization index; and training the processed first characterization index using the Prophet algorithm based on the time series to obtain the first dynamic threshold of the first characterization index that changes dynamically over time.

[0057] In this embodiment, to reduce the impact of time variations on the performance of the cell's wireless network, a first dynamic threshold for the dynamic change of the first characterization index over time is obtained through a time-series-based prediction algorithm. First, training samples for the Prophet algorithm are prepared, and the characterization index and its corresponding time series are combined to form a dataset. Next, the dataset is preprocessed, removing or replacing missing or abnormal values ​​to obtain a processed dataset (i.e., the processed first characterization index). Then, the processed first characterization index is input into the Prophet algorithm model for calculation. The Prophet algorithm decomposes the time series into three parts: long-term trend, periodic trend, and random fluctuation. The decomposed three parts are shown below. Figure 3 As shown, the three parts of the sequence are calculated separately, and then the values ​​of the three parts are summed to obtain the predicted value of the time series (i.e., the first dynamic threshold). The decomposition method of the three parts of the sequence is shown in the following formula:

[0058] y(t) = g(t) + s(t) + ∈(t)

[0059] In this equation, g(t) represents the trend term, which indicates the trend of the value of the first indicator in a non-periodic time series; s(t) represents the periodic term (or seasonal term), which indicates the trend of the value of the first indicator in a periodic time series, generally in units of weeks or years; ∈(t) represents the error term (or residual term), which indicates the numerical fluctuations not predicted by the model, and the error term follows a Gaussian distribution; y(t) represents the predicted value of the time series after the sum of the three parts; finally, the first dynamic threshold corresponding to the first indicator is output.

[0060] By decomposing the time series into three parts using the Prophet algorithm, the first characterization index is calculated to obtain the first dynamic threshold that changes over time. This reduces the impact of time variations on the first characterization index, thereby improving the accuracy of identifying the root cause of the location's impact on wireless network performance.

[0061] Optionally, in the wireless network processing method provided in this application embodiment, obtaining root cause information and the corresponding score of root cause information based on the first root cause index set, confidence level, and first dynamic threshold includes: obtaining the third root cause index with the highest confidence level from the first root cause index set to obtain the third root cause index set; taking the inflection point of the value of the third root cause index in the third root cause index set as the second threshold; obtaining the target root cause index based on the first dynamic threshold; constructing a root cause tree; traversing the root cause tree based on the target root cause index and the second threshold to obtain the root cause information and the corresponding score of root cause information.

[0062] In this embodiment, to pinpoint the root cause affecting wireless network performance, it is necessary to obtain the root cause indicators (i.e., target root cause indicators) affecting poor-quality wireless networks. Then, based on the constructed root cause tree, the root cause affecting the performance of the cell's wireless network (i.e., root cause information) and the probability (i.e., score) of the root cause information affecting the wireless network are obtained. First, the root cause indicator with the highest confidence (i.e., the third root cause indicator set) is selected from the first root cause indicator set. Next, the inflection point of the value of the third root cause indicator is selected from the third root cause indicator set as the threshold (i.e., the second threshold) for judging whether the root cause indicator is an abnormal indicator (or a poor-quality indicator). Then, based on the first dynamic threshold, the abnormal indicator (i.e., the target root cause indicator) is obtained from the first root cause indicator set. Finally, based on the target root cause indicator and the second threshold, the constructed root cause tree is traversed to obtain the root cause information affecting the performance of the cell's wireless network and the corresponding score of the root cause information.

[0063] By obtaining the abnormal threshold of the root cause index, abnormal indicators (i.e., target root cause indicators) can be distinguished from the first root cause index set. Then, the root cause tree is traversed according to the target root cause index to locate the root cause affecting the performance of the cell's wireless network. This enables a detailed analysis of the root cause affecting the performance of the cell's wireless network based on the target root cause index and the root cause tree. It also considers the impact of multiple factors on the performance of the wireless network, thereby improving the accuracy of locating the root cause affecting the performance of the cell's wireless network.

[0064] Optionally, in the wireless network processing method provided in this application embodiment, obtaining the target root cause index based on the first dynamic threshold includes: obtaining a third characterization index lower than the first dynamic threshold from the first characterization index; and obtaining the root cause index corresponding to the third characterization index from the database as the target root cause index.

[0065] In existing technologies, multiple KPI indicators are generally clustered and analyzed using machine learning-based clustering algorithms to obtain quality deviation thresholds for multiple KPI indicators. Although this method is simple in principle and easy to implement, the clustering indicator dataset and the initial preset classification have a significant impact on the output quality deviation threshold. That is, even a small fluctuation in the clustering indicator dataset and the initial preset classification may lead to a large change in the output quality deviation threshold. Another method is to add the mean of the indicator data for each dimension to three times the standard deviation of the indicator data, and use the resulting value as the quality deviation threshold for the corresponding dimension of indicator data. However, this method requires calculating the indicator data of all cells in the network, which requires acquiring and calculating a large amount of sample data, consuming a lot of computing and storage resources. At the same time, this approach requires the indicator data to conform to a normal distribution, which has a certain deviation from the actual data.

[0066] In this embodiment, in order to obtain the quality poor index that affects the performance of the cell wireless network, it is first necessary to obtain a characterization index (i.e., a third characterization index) that is lower than the first dynamic threshold from the first characterization index, so as to reduce the impact of time changes on the characterization index; then, based on the third characterization index, the root cause index (i.e., the target root cause index) associated with the third characterization index is obtained from the database.

[0067] The first dynamic threshold is used to filter the first characterization index to obtain the third characterization index. Then, the target root cause index associated with the third characterization index is obtained by the FP-grown algorithm. This realizes the selection of the target root cause index affecting the performance of the cell wireless network from the first root cause index set. In this way, the root cause index is located based on the target root cause index, thereby improving the accuracy of locating the root cause affecting the performance of the cell wireless network.

[0068] Optionally, in the wireless network processing method provided in this application embodiment, constructing a root cause tree and traversing the root cause tree according to the target root cause index and the second threshold to obtain root cause information and the corresponding score of the root cause information includes: classifying the root cause indexes affecting the wireless network according to preset rules to obtain different categories of root cause indexes, and constructing the root cause tree according to the different categories of root cause indexes to obtain different categories of root cause trees; based on the comparison result of whether the value of the target root cause index is greater than the second threshold corresponding to the root cause index represented by the target root cause tree node, continuing to traverse the child nodes of different branches of the target root cause tree node until the leaf node of the target root cause tree is reached, obtaining the root cause information corresponding to the leaf node, and calculating the score corresponding to the root cause information, wherein the target root cause tree is the root cause tree of the category to which the target root cause index belongs; and outputting the root cause information and the score corresponding to the root cause information.

[0069] In existing technologies, a set of multi-dimensional index data is generally input into a pre-trained bounding algorithm model. The random forest algorithm is then used to locate the root cause affecting the performance of the cell wireless network as any one of the following problems: weak coverage, interference, equipment failure, capacity, burstiness, and non-wireless issues. This method does not consider that the root cause affecting the performance of the cell wireless network may be composed of multiple factors, and that the probability of these factors affecting the performance of the cell wireless network also varies. Furthermore, it does not conduct further analysis and mining of the multiple factors affecting the performance of the cell wireless network.

[0070] In this embodiment, in order to conduct a detailed analysis of the root causes affecting the performance of the cell wireless network, it is necessary to analyze the impact of the target root cause indicators on the performance of the cell wireless network through a root cause tree analysis.

[0071] First, root cause trees are constructed for different categories of root cause indicators. Next, it is determined whether the value of the target root cause indicator is greater than the second threshold corresponding to the root cause indicator represented by the target root cause tree node. If the value of the target root cause indicator is greater than the second threshold, a child node under the branch satisfying the threshold is selected as the next tree node to be judged. If the value of the target root cause indicator is less than or equal to the second threshold, a child node under the branch not satisfying the threshold is selected as the next tree node to be judged. Then, the target root cause indicator is judged continuously according to the root cause tree until a leaf node of the target root cause tree is reached. When a leaf node is reached, the root cause information corresponding to the leaf node is obtained, and the probability (i.e., score) of the root cause information affecting the cell's wireless network is calculated. Finally, the root cause information and its corresponding score are output.

[0072] For example, interference-type root cause trees such as Figure 4As shown, when the target root cause indicator "Nearby cell fault" is analyzed based on the interference-type root cause tree, the tree node "Uplink interference" in the interference-type root cause tree is first compared. If the value of the target root cause indicator "Nearby cell fault" is less than or equal to the second threshold corresponding to the root cause indicator "Uplink interference", then the child node "Downlink interference" under the "No" branch that does not meet the root cause indicator "Uplink interference" is selected for further comparison. Next, if the value of the target root cause indicator "Nearby cell fault" is determined to be greater than the second threshold corresponding to the tree node "Nearby cell fault", then the leaf node "Due to the fault in the nearby cell, the coverage has changed, causing interference" under the "Fault exists" branch that meets the tree node "Nearby cell fault" is selected for further traversal. The root cause information "Due to the fault in the nearby cell, the coverage has changed, causing interference" and the optimization suggestion for the root cause information "Perform engineering processing on the faulty cell" are obtained from the leaf node. Then, the score of the root cause information is calculated, and finally the root cause information "Due to the fault in the nearby cell, the coverage has changed, causing interference" and the score of the root cause information are output.

[0073] By conducting detailed analysis of target root cause indicators based on different types of root cause trees, the system fully considers multiple factors affecting the performance of the cell's wireless network when locating the root cause, and assesses the probability of each factor affecting the performance of the cell's wireless network. This improves the accuracy of locating the root cause affecting the performance of the cell's wireless network.

[0074] Optionally, in the wireless network processing method provided in this application embodiment, the root cause indicators affecting the wireless network are classified according to preset rules to obtain different categories of root cause indicators, and root cause trees are built based on the different categories of root cause indicators. Obtaining different categories of root cause trees includes: obtaining root cause indicators of the same category affecting the wireless network according to preset rules to obtain a fourth set of root cause indicators; in the fourth set of root cause indicators, the first primary indicator is used as the parent node of the first root cause tree, the secondary indicators under the influence of the first primary indicator are used as the first child nodes of the first primary indicator node, and the second primary indicators unrelated to the first primary indicator are used as the second child nodes of the first primary indicator. This process involves obtaining the parent node, its corresponding first child node, and its second child node. The first primary causal indicator is the root cause indicator with the largest influence range in the fourth root cause indicator set, and the second primary causal indicator is the root cause indicator with the largest influence range among other root cause indicators unrelated to the first primary causal indicator in the fourth root cause indicator set. Based on the generation process of the parent node and its corresponding first and second child nodes, the remaining root cause indicators in the fourth root cause indicator set are used to generate other nodes of the first root cause tree until all root cause indicators in the fourth root cause indicator set appear in the first root cause tree, thus obtaining the first root cause tree. Based on the method for constructing the first root cause tree, different categories of root cause trees are obtained.

[0075] In this embodiment, in order to locate the root cause affecting the performance of the cell's wireless network based on the root cause tree, it is necessary to build the root cause tree to analyze in detail the impact of the target root cause index on the cell's wireless network.

[0076] First, based on preset rules, the root causes affecting wireless networks are divided into six categories, and six root cause trees are constructed. These six categories are parameter-related, fault-related, interference-related, coverage-related, capacity-related, and other categories.

[0077] Next, a first root cause tree is constructed based on the fourth root cause indicator set of the same category affecting the wireless network: In the fourth root cause indicator set, the first primary indicator with the largest influence range under this category is obtained, that is, the first primary indicator contains the widest range of root cause indicators under this category. The first primary indicator is used as the parent node of the first root cause tree, the secondary indicators under the influence of the first primary indicator are used as the first child nodes of the first primary indicator node, and the second primary indicators unrelated to the first primary indicator are used as the second child nodes of the first primary indicator. Among them, the secondary indicator is the root cause indicator with the largest influence range among the root cause indicators under the influence of the first primary indicator in the fourth root cause indicator set, and the second primary indicator is the root cause indicator with the largest influence range among other root cause indicators unrelated to the first primary indicator in the fourth root cause indicator set. This results in a parent node, the first child node and the second child node corresponding to the parent node.

[0078] Then, based on the generation process of the parent node, the first child node corresponding to the parent node, and the second child node corresponding to the parent node, the remaining root cause indicators in the fourth root cause indicator set are used to construct other nodes of the first root cause tree until all root cause indicators in the fourth root cause indicator set appear in the first root cause tree, thus obtaining the first root cause tree.

[0079] Finally, based on the method used to generate the first root tree, five other root trees are constructed, resulting in six root trees of different categories.

[0080] For example, interference-type root cause trees such as Figure 4 As shown, the tree nodes of the interference root cause tree are root cause indicators, the leaf nodes of the interference root cause tree are detailed root causes under the main root cause, and the boxes connected by dashed lines near the leaf nodes of detailed root causes record optimization suggestions for the detailed root causes. Figure 4 In the interference root cause tree shown, "upstream interference," "downstream interference," and "external interference" are all tree nodes. "Upstream interference" is the primary cause indicator in the interference root cause tree, and it also has the largest impact among the interference root cause indicators. "Upstream interference" has two child nodes, one of which is the child node resulting from the judgment of the tree node "upstream interference," i.e. Figure 4The tree node "Uplink Interference" in the tree structure has a branch called "External Interference," which is a secondary cause indicator. Another child node of "Uplink Interference" is "Downlink Interference," which, after assessing "Uplink Interference," is considered the most influential among root cause indicators unrelated to "Uplink Interference"—it is the second primary cause indicator. To the left of the "External Interference" tree node are two boxes. The leaf node "Caused by External Interference Sources" is the detailed root cause of the "External Interference" root cause indicator. The box "External Interference Investigation," connected to the leaf node "Caused by External Interference Sources" by a dotted line, provides optimization suggestions for the detailed root cause "Caused by External Interference Sources."

[0081] By constructing a root cause tree based on the influence range of root cause indicators, a detailed analysis of the root causes affecting the performance of the cell's wireless network can be achieved, thereby improving the accuracy of locating the root causes affecting the performance of the cell's wireless network.

[0082] Optionally, in this embodiment, the process of locating the root cause affecting the performance of the cell's wireless network can be as follows: Figure 5 As shown, in step S1, the indicator system construction stage, the indicators are divided into representative indicators and root cause indicators. The storage granularity and dimension of each indicator are determined, and aggregation calculations are performed and stored in the database. In step S2, the indicator correlation analysis stage, the dataset required for calculation is prepared, an algorithm model is established based on the FP-growth algorithm, and the correlation between representative indicators and root cause indicators is explored. In step S3, the indicator dynamic threshold determination stage, the threshold of the representative indicators is dynamically adjusted, and the threshold of the root cause indicators is determined. In step S4, the root cause location stage, the root cause tree is built, poor-quality cells are identified, the root cause is located, and the root cause and related detailed information (i.e., detailed root cause information) are output.

[0083] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0084] This application also provides a wireless network processing apparatus. It should be noted that the wireless network processing apparatus of this application can be used to execute the wireless network processing method provided in this application. The wireless network processing apparatus provided in this application will be described below.

[0085] Figure 6 This is a schematic diagram of a wireless network processing apparatus according to an embodiment of this application. Figure 6As shown, the device includes: a first acquisition unit 601, a determination unit 602, a second acquisition unit 603, a third acquisition unit 604, and a positioning unit 605.

[0086] Specifically, the first acquisition unit 601 is used to acquire a first characterization index that affects the wireless network, wherein the first characterization index is a relevant index of the user's evaluation information on the performance of the wireless network.

[0087] The determining unit 602 is used to determine the set of first root cause indicators that affect the first characterization indicator and the confidence level corresponding to each first root cause indicator. The set of first root cause indicators is a set of multiple first root cause indicators that affect the first characterization indicator. The first root cause indicator is an operational indicator of the network of the cell to which the user of the base station or wireless network belongs.

[0088] The second acquisition unit 603 is used to calculate the value of the first characterization index based on the time series prediction algorithm to obtain the first dynamic threshold of the first characterization index.

[0089] The third acquisition unit 604 is used to obtain root cause information and the corresponding score of root cause information based on the first root cause index set, confidence level and first dynamic threshold.

[0090] The positioning unit 605 is used to locate the root cause affecting the performance of the wireless network based on the root cause information and the corresponding score of the root cause information.

[0091] The wireless network processing apparatus provided in this application embodiment includes a first acquisition unit 601 for acquiring a first characterization index affecting the wireless network, wherein the first characterization index is a relevant index of user evaluation information of wireless network performance; a determination unit 602 for determining a first root cause index set affecting the first characterization index and a confidence level corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indices affecting the first characterization index, and the first root cause index is an operational index representing the network of the cell to which the user belongs to the base station or wireless network; a second acquisition unit 603 for calculating the value of the first characterization index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characterization index; a third acquisition unit 604 for obtaining root cause information and the corresponding score of the root cause information based on the first root cause index set, confidence level, and first dynamic threshold; and a location unit 605 for locating the root cause affecting the performance of the wireless network based on the root cause information and the corresponding score of the root cause information. This solves the problem that the indicators obtained by the clustering algorithm based on machine learning are easily affected and have poor interpretability, resulting in the inability to accurately locate the root cause affecting the performance of the wireless network. By acquiring representative indicators, root cause indicators associated with these indicators, and the confidence levels of these root cause indicators, a threshold for distinguishing poor-quality cells is calculated. Finally, a detailed analysis of the root cause indicators affecting poor-quality cells is conducted. This approach achieves the acquisition of representative and root cause indicators that are more aligned with user perception, enabling a more reasonable determination of the root cause indicators affecting poor-quality cells. Based on these root cause indicators, the root causes affecting wireless network performance can be located more quickly and accurately, thus helping staff to more quickly and accurately pinpoint the root causes affecting wireless network performance.

[0092] Optionally, in the wireless network processing apparatus provided in this application embodiment, the first acquisition unit 601 includes: a first determining subunit, used to determine the collection period and collection dimension of the indicators affecting the wireless network, and acquire the indicators affecting the operation of the wireless network, wherein the collection dimension is the dimension of the indicators collected by the cell or user; a judging subunit, used to judge whether the indicator is related to the user's evaluation information on the performance of the wireless network; a first dividing subunit, used to divide the indicator into a second characterization indicator if the indicator is related to the user's evaluation information on the performance of the wireless network; a second dividing subunit, used to divide the indicator into a second root cause indicator if the indicator is not related to the user's evaluation information on the performance of the wireless network; a first acquisition subunit, used to perform convergence calculation on the second characterization indicator and the second root cause indicator to obtain a first characterization indicator and a first root cause indicator; and a storage subunit, used to store the first characterization indicator and the first root cause indicator in a database.

[0093] Optionally, in the wireless network processing apparatus provided in this application embodiment, the determining unit 602 includes: a first construction subunit, configured to construct a data sample set for the FP-grown algorithm model based on the acquisition dimension and acquisition period; and a second acquisition subunit, configured to input the first characterization index and the first root cause index from the database into the FP-grown algorithm model for calculation, thereby obtaining a set of first root cause indices affecting the first characterization index and the confidence level corresponding to each first root cause index.

[0094] Optionally, in the wireless network processing apparatus provided in this application embodiment, the second acquisition unit 603 includes: a second construction subunit, used to construct three time series—long-term trend, periodic trend, and random fluctuation—based on the acquisition period; a third acquisition subunit, used to preprocess the first characterization index based on the time series, removing outliers in the time series to obtain the processed first characterization index; and a fourth acquisition subunit, used to train the processed first characterization index using the Prophet algorithm based on the time series to obtain a first dynamic threshold for the dynamic change of the first characterization index over time.

[0095] Optionally, in the wireless network processing apparatus provided in this application embodiment, the third acquisition unit 604 includes: a fifth acquisition subunit, configured to acquire the third root cause indicator with the highest confidence from the first root cause indicator set, thereby obtaining a third root cause indicator set; a second determination subunit, configured to determine the inflection point of the value of the third root cause indicator in the third root cause indicator set as a second threshold; a sixth acquisition subunit, configured to acquire the target root cause indicator based on the first dynamic threshold; and a seventh acquisition subunit, configured to build a root cause tree, traverse the root cause tree based on the target root cause indicator and the second threshold, and obtain root cause information and the score corresponding to the root cause information.

[0096] Optionally, in the wireless network processing apparatus provided in this application embodiment, the sixth acquisition subunit includes: a first acquisition module, configured to acquire a third characterization index lower than a first dynamic threshold from the first characterization index; and a second acquisition module, configured to acquire the root cause index corresponding to the third characterization index from the database as a target root cause index.

[0097] Optionally, in the wireless network processing device provided in this application embodiment, the seventh acquisition subunit includes: a third acquisition module, configured to classify the root cause indicators affecting the wireless network according to preset rules to obtain different categories of root cause indicators, and to build root cause trees based on the different categories of root cause indicators to obtain different categories of root cause trees; a fourth acquisition module, configured to continue traversing the child nodes of different branches of the target root cause tree node according to the comparison result of whether the value of the target root cause indicator is greater than the second threshold corresponding to the root cause indicator represented by the target root cause tree node, until the leaf node of the target root cause tree is reached, to obtain the root cause information corresponding to the leaf node, and to calculate the score corresponding to the root cause information, wherein the target root cause tree is the root cause tree of the category to which the target root cause indicator belongs; and an output module, configured to output the root cause information and the score corresponding to the root cause information.

[0098] Optionally, in the wireless network processing apparatus provided in this application embodiment, the third acquisition module includes: a first acquisition submodule, configured to acquire root cause indicators of the same category affecting the wireless network according to preset rules, thereby obtaining a fourth root cause indicator set; and a second acquisition submodule, configured to, in the fourth root cause indicator set, take a first primary cause indicator as the parent node of a first root cause tree, take secondary cause indicators influenced by the first primary cause indicator as the first child node of the first primary cause indicator node, and take second primary cause indicators unrelated to the first primary cause indicator as the second child node of the first primary cause indicator, thereby obtaining a parent node, the first child node corresponding to the parent node, and the second child node. The first principal causative indicator is the root cause indicator with the largest influence range in the fourth root cause indicator set. The second principal causative indicator is the root cause indicator with the largest influence range among other root cause indicators unrelated to the first principal causative indicator in the fourth root cause indicator set. The third acquisition submodule is used to continue generating other nodes of the first root cause tree from the remaining root cause indicators in the fourth root cause indicator set based on the generation process of the parent node and the first and second child nodes corresponding to the parent node, until all root cause indicators in the fourth root cause indicator set appear in the first root cause tree, thus obtaining the first root cause tree. The fourth acquisition submodule is used to obtain root cause trees of different categories based on the device used to construct the first root cause tree.

[0099] The processing device of the wireless network includes a processor and a memory. The first acquisition unit 601, the determination unit 602, the second acquisition unit 603, the third acquisition unit 604, and the positioning unit 605 are all stored in the memory as program units. The processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0100] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters can improve the accuracy of pinpointing the root causes affecting the performance of a cell's wireless network.

[0101] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0102] This invention provides a computer-readable storage medium storing a program thereon, which, when executed by a processor, implements a method for processing wireless networks.

[0103] This invention provides a processor for running a program, wherein the program executes a wireless network processing method during runtime.

[0104] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps: obtaining a first characterization index affecting the wireless network, wherein the first characterization index is a relevant index of user evaluation information on the performance of the wireless network; determining a first root cause index set affecting the first characterization index and a confidence level corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indices affecting the first characterization index, and the first root cause index is an operational index representing the network of the cell to which the user belongs (e.g., a base station or wireless network); calculating the value of the first characterization index based on a time series prediction algorithm to obtain a first dynamic threshold for the first characterization index; obtaining root cause information and a score corresponding to the root cause information based on the first root cause index set, the confidence level, and the first dynamic threshold; and locating the root cause affecting the performance of the wireless network based on the root cause information and the score corresponding to the root cause information.

[0105] When the processor executes the program, it also performs the following steps: obtaining the first characterization index affecting the wireless network includes: determining the collection period and collection dimension of the index affecting the wireless network, and obtaining the index affecting the operation of the wireless network, wherein the collection dimension is the dimension of the index collected by cell or user; determining whether the index is related to the user's evaluation information on the performance of the wireless network; if the index is related to the user's evaluation information on the performance of the wireless network, then the index is classified as the second characterization index; if the index is not related to the user's evaluation information on the performance of the wireless network, then the index is classified as the second root cause index; performing convergence calculation on the second characterization index and the second root cause index to obtain the first characterization index and the first root cause index; and storing the first characterization index and the first root cause index in the database.

[0106] When the processor executes the program, it also performs the following steps: determining the set of first root cause indicators affecting the first representation indicator and the confidence level corresponding to each first root cause indicator, including: constructing a data sample set for the FP-grown algorithm model based on the collection dimension and collection period; inputting the first representation indicator and the first root cause indicator in the database into the FP-grown algorithm model for calculation to obtain the set of first root cause indicators affecting the first representation indicator and the confidence level corresponding to each first root cause indicator.

[0107] The processor also performs the following steps when executing the program: calculating the value of the first characterization index based on the time series prediction algorithm to obtain the first dynamic threshold of the first characterization index, including: constructing three time series based on the collection period: long-term trend, periodic trend, and random fluctuation; preprocessing the first characterization index based on the time series to remove outliers in the time series to obtain the processed first characterization index; and training the processed first characterization index using the Prophet algorithm based on the time series to obtain the first dynamic threshold of the first characterization index that changes dynamically over time.

[0108] When the processor executes the program, it also performs the following steps: obtaining root cause information and the corresponding score based on the first root cause indicator set, confidence level, and first dynamic threshold, including: obtaining the third root cause indicator with the highest confidence level from the first root cause indicator set to obtain the third root cause indicator set; taking the inflection point of the value of the third root cause indicator in the third root cause indicator set as the second threshold; obtaining the target root cause indicator based on the first dynamic threshold; building a root cause tree, traversing the root cause tree based on the target root cause indicator and the second threshold to obtain the root cause information and the corresponding score.

[0109] When the processor executes the program, it also performs the following steps: Based on the first dynamic threshold, the target root cause index is obtained by: obtaining a third characterization index that is lower than the first dynamic threshold from the first characterization index; and obtaining the root cause index corresponding to the third characterization index from the database as the target root cause index.

[0110] The processor also performs the following steps when executing the program: building a root cause tree, traversing the root cause tree according to the target root cause index and the second threshold to obtain root cause information and the corresponding score of the root cause information, including: classifying the root cause indexes affecting the wireless network according to preset rules to obtain root cause indexes of different categories, and building root cause trees according to the root cause indexes of different categories to obtain root cause trees of different categories; based on the comparison result of whether the value of the target root cause index is greater than the second threshold corresponding to the root cause index represented by the root cause index node of the target root cause tree, continuing to traverse the child nodes of different branches of the target root cause index node until the leaf node of the target root cause tree is reached, obtaining the root cause information corresponding to the leaf node, and calculating the score corresponding to the root cause information, wherein the target root cause tree is the root cause tree of the category to which the target root cause index belongs; outputting the root cause information and the score corresponding to the root cause information.

[0111] When the processor executes the program, it also performs the following steps: According to preset rules, it classifies the root cause indicators affecting the wireless network into different categories of root cause indicators, and constructs root cause trees based on these categories. Obtaining root cause trees of different categories includes: according to preset rules, acquiring root cause indicators of the same category affecting the wireless network to obtain a fourth set of root cause indicators; in the fourth set of root cause indicators, using the first primary root cause indicator as the parent node of the first root cause tree, using secondary root cause indicators influenced by the first primary root cause indicator as the first child node of the first primary root cause indicator node, and using second primary root cause indicators unrelated to the first primary root cause indicator as the second child node of the first primary root cause indicator, thus obtaining the parent root cause indicator. The first and second child nodes corresponding to the parent node are defined as follows: the first primary causal indicator is the root cause indicator with the largest influence range in the fourth root cause indicator set; the second primary causal indicator is the root cause indicator with the largest influence range among other root cause indicators unrelated to the first primary causal indicator in the fourth root cause indicator set. Based on the generation process of the parent node and its corresponding first and second child nodes, the remaining root cause indicators in the fourth root cause indicator set are used to generate other nodes of the first root cause tree until all root cause indicators in the fourth root cause indicator set appear in the first root cause tree, thus obtaining the first root cause tree. Based on the method of constructing the first root cause tree, different categories of root cause trees are obtained.

[0112] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0113] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having the following method steps: obtaining a first characterization index affecting the wireless network, wherein the first characterization index is a relevant index of user evaluation information of the wireless network performance; determining a first root cause index set affecting the first characterization index and a confidence level corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indices affecting the first characterization index, and the first root cause index is an operational index representing the network of the cell to which the user belongs in the base station or wireless network; calculating the value of the first characterization index based on a time series prediction algorithm to obtain a first dynamic threshold for the first characterization index; obtaining root cause information and a score corresponding to the root cause information based on the first root cause index set, the confidence level, and the first dynamic threshold; and locating the root cause affecting the performance of the wireless network based on the root cause information and the score corresponding to the root cause information.

[0114] When executed on a data processing device, it is also suitable to execute an initialization program with the following steps: obtaining a first characterization indicator affecting the wireless network, including: determining the collection period and collection dimension of the indicator affecting the wireless network; obtaining the indicator affecting the operation of the wireless network, wherein the collection dimension is the dimension of collecting indicators by cell or user; determining whether the indicator is related to the user's evaluation information on the performance of the wireless network; if the indicator is related to the user's evaluation information on the performance of the wireless network, then the indicator is classified as a second characterization indicator; if the indicator is not related to the user's evaluation information on the performance of the wireless network, then the indicator is classified as a second root cause indicator; performing convergence calculation on the second characterization indicator and the second root cause indicator to obtain the first characterization indicator and the first root cause indicator; and storing the first characterization indicator and the first root cause indicator in the database.

[0115] When executed on a data processing device, it is also suitable to execute an initialization program with the following method steps: determining the set of first root cause indicators affecting the first characterization indicator and the confidence level corresponding to each first root cause indicator, including: constructing a data sample set for the FP-grown algorithm model based on the collection dimension and collection period; inputting the first characterization indicator and the first root cause indicator in the database into the FP-grown algorithm model for calculation to obtain the set of first root cause indicators affecting the first characterization indicator and the confidence level corresponding to each first root cause indicator.

[0116] When executed on a data processing device, it is also suitable to execute an initialization program with the following steps: calculating the value of the first characterization index based on a time series prediction algorithm to obtain the first dynamic threshold of the first characterization index, including: constructing three time series—long-term trend, periodic trend, and random fluctuation—based on the collection period; preprocessing the first characterization index based on the time series to remove outliers in the time series and obtain the processed first characterization index; and training the processed first characterization index using the Prophet algorithm based on the time series to obtain the first dynamic threshold of the first characterization index that changes dynamically over time.

[0117] When executed on a data processing device, it is also suitable to execute an initialization program with the following steps: obtaining root cause information and the corresponding score based on a first root cause indicator set, confidence level, and a first dynamic threshold, including: obtaining the third root cause indicator with the highest confidence level from the first root cause indicator set to obtain a third root cause indicator set; using the inflection point of the value of the third root cause indicator in the third root cause indicator set as a second threshold; obtaining the target root cause indicator based on the first dynamic threshold; constructing a root cause tree; traversing the root cause tree based on the target root cause indicator and the second threshold to obtain root cause information and the corresponding score.

[0118] When executed on a data processing device, it is also suitable to execute an initialization program with the following method steps: obtaining a target root cause index based on a first dynamic threshold, including: obtaining a third characterization index that is lower than the first dynamic threshold from the first characterization index; and obtaining the root cause index corresponding to the third characterization index from the database as the target root cause index.

[0119] When executed on a data processing device, it is also suitable to execute an initialization program with the following steps: building a root cause tree, traversing the root cause tree based on the target root cause index and a second threshold to obtain root cause information and the corresponding score of the root cause information, including: classifying the root cause indexes affecting the wireless network according to preset rules to obtain different categories of root cause indexes, and building root cause trees based on the different categories of root cause indexes to obtain different categories of root cause trees; based on the comparison result of whether the value of the target root cause index is greater than the second threshold corresponding to the root cause index represented by the target root cause tree node, continuing to traverse the child nodes of different branches of the target root cause tree node until the leaf node of the target root cause tree is reached, obtaining the root cause information corresponding to the leaf node, and calculating the score corresponding to the root cause information, wherein the target root cause tree is the root cause tree of the category to which the target root cause index belongs; outputting the root cause information and the score corresponding to the root cause information.

[0120] When executed on a data processing device, it is also suitable to execute an initialization program with the following steps: According to preset rules, root cause indicators affecting the wireless network are classified to obtain different categories of root cause indicators, and root cause trees are constructed based on these different categories. The different categories of root cause trees include: according to preset rules, obtaining root cause indicators of the same category affecting the wireless network to obtain a fourth set of root cause indicators; in the fourth set of root cause indicators, the first primary root cause indicator is used as the parent node of the first root cause tree, secondary root cause indicators influenced by the first primary root cause indicator are used as the first child nodes of the first primary root cause indicator node, and second primary root cause indicators unrelated to the first primary root cause indicator are used as the first child nodes of the first primary root cause indicator. Two child nodes are generated to obtain the parent node, the first child node and the second child node corresponding to the parent node. The first primary cause indicator is the root cause indicator with the largest influence range in the fourth root cause indicator set, and the second primary cause indicator is the root cause indicator with the largest influence range among the other root cause indicators in the fourth root cause indicator set that are unrelated to the first primary cause indicator. Based on the generation process of the parent node, the first child node and the second child node corresponding to the parent node, the remaining root cause indicators in the fourth root cause indicator set are used to generate other nodes of the first root cause tree until all root cause indicators in the fourth root cause indicator set appear in the first root cause tree, thus obtaining the first root cause tree. Based on the method of constructing the first root cause tree, different categories of root cause trees are obtained.

[0121] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0122] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0123] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0124] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0125] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0126] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0127] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0128] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0129] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0130] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A processing method of a wireless network, characterized by, The method comprises the following steps: obtaining a first characteristic index affecting a wireless network, wherein the first characteristic index is a relevant index of user evaluation information of wireless network performance; determining a first root cause index set affecting the first characteristic index and a confidence degree corresponding to each first root cause index, wherein the first root cause index set is a set of multiple first root cause indexes affecting the first characteristic index, and the first root cause index is an operating index of a base station or a network of a cell to which a user belongs; calculating a value of the first characteristic index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characteristic index; obtaining root cause information and a score corresponding to the root cause information according to the first root cause index set, the confidence degree and the first dynamic threshold; locating a root cause affecting wireless network performance according to the root cause information and the score corresponding to the root cause information; wherein obtaining the first characteristic index affecting the wireless network comprises determining a collection period of indexes affecting the wireless network; calculating a value of the first characteristic index based on a time series prediction algorithm to obtain a first dynamic threshold of the first characteristic index comprises: constructing long-term trend, periodic trend and random fluctuation three time series according to the collection period; preprocessing the first characteristic index according to the time series to remove abnormal values of the first characteristic index in the time series to obtain the first characteristic index after processing; training the first characteristic index after processing based on the time series using a Prophet algorithm to obtain the first dynamic threshold of the first characteristic index changing dynamically with time; obtaining root cause information and a score corresponding to the root cause information according to the first root cause index set, the confidence degree and the first dynamic threshold comprises: obtaining a third root cause index set from the first root cause index set with the highest confidence degree; taking an inflection point of a value of a third root cause index in the third root cause index set as a second threshold; obtaining a target root cause index according to the first dynamic threshold; building a root cause tree, traversing the root cause tree according to the target root cause index and the second threshold to obtain the root cause information and the score corresponding to the root cause information.

2. The method of claim 1, wherein, Obtaining the first characteristic index affecting the wireless network comprises: determining a collection dimension of indexes affecting the wireless network to obtain the indexes affecting the operation of the wireless network, wherein the collection dimension is a dimension of collecting the indexes in a cell or a user; judging whether the indexes are associated with user evaluation information of wireless network performance; if the indexes are associated with user evaluation information of wireless network performance, dividing the indexes into second characteristic indexes; if the indexes are not associated with user evaluation information of wireless network performance, dividing the indexes into second root cause indexes; performing aggregation calculation on the second characteristic indexes and the second root cause indexes to obtain the first characteristic indexes and the first root cause indexes; storing the first characteristic indexes and the first root cause indexes in a database.

3. The method of claim 2, wherein, The determining of the first root cause indicator set affecting the first characterization indicator and the confidence corresponding to each first root cause indicator comprises: According to the collection dimension and the collection cycle, a data sample set of the FP-grown algorithm model is constructed; The first characterization indicator and the first root cause indicator in the database are input into the FP-grown algorithm model for calculation, so as to obtain the first root cause indicator set affecting the first characterization indicator and the confidence corresponding to each first root cause indicator.

4. The method of claim 2, wherein, According to the first dynamic threshold, the target root cause indicator comprises: In the first characterization indicator, a third characterization indicator lower than the first dynamic threshold is obtained; The root cause indicator corresponding to the third characterization indicator is obtained from the database as the target root cause indicator.

5. The method of claim 1, wherein, The root cause tree is built, and the root cause information and the score corresponding to the root cause information are obtained by traversing the root cause tree according to the target root cause indicator and the second threshold, comprising: According to a preset rule, the root cause indicators affecting the wireless network are classified to obtain root cause indicators of different categories, and the root cause tree is built according to the root cause indicators of different categories to obtain the root cause tree of different categories; According to the comparison result of whether the numerical value of the target root cause indicator is greater than the second threshold corresponding to the root cause indicator represented by the target root cause tree node, the child nodes of different branches of the target root cause tree node are continuously traversed until the leaf node of the target root cause tree is traversed, the root cause information corresponding to the leaf node is obtained, and the score corresponding to the root cause information is calculated, wherein the target root cause tree is the root cause tree of the category to which the target root cause indicator belongs; The root cause information and the score corresponding to the root cause information are output.

6. The method of claim 5, wherein, The root cause tree is built, and the root cause indicators affecting the wireless network are classified according to a preset rule to obtain root cause indicators of different categories, and the root cause tree is built according to the root cause indicators of different categories to obtain the root cause tree of different categories, comprising: According to a preset rule, the root cause indicators of the same category affecting the wireless network are obtained to obtain a fourth root cause indicator set; In the fourth root cause indicator set, a first main cause indicator is taken as a parent node of a first root cause tree, a secondary cause indicator under the influence of the first main cause indicator is taken as a first child node of the first main cause indicator node, and a second main cause indicator irrelevant to the first main cause indicator is taken as a second child node of the first main cause indicator, so as to obtain the parent node, the first child node and the second child node corresponding to the parent node, wherein the first main cause indicator is a root cause indicator with the largest influence range in the fourth root cause indicator set, and the second main cause indicator is a root cause indicator with the largest influence range among other root cause indicators irrelevant to the first main cause indicator in the fourth root cause indicator set; According to the generation process of the parent node, the first child node and the second child node corresponding to the parent node, the remaining root cause indicators in the fourth root cause indicator set are continuously generated to generate other nodes of the first root cause tree until all root cause indicators in the fourth root cause indicator set appear in the first root cause tree, and the first root cause tree is obtained; According to the method for constructing the first root cause tree, the root cause trees of different categories are obtained.

7. A processing device of a wireless network, characterized by The method comprises the following steps: A first acquisition unit is configured to acquire a first representation indicator affecting a wireless network, wherein the first representation indicator is a relevant indicator of user evaluation information of wireless network performance; A determination unit is configured to determine a first root cause indicator set affecting the first representation indicator and a confidence degree corresponding to each first root cause indicator, wherein the first root cause indicator set is a set of multiple first root cause indicators affecting the first representation indicator, and the first root cause indicator is an operation indicator of a network representing a base station or a cell to which a user belongs; A second acquisition unit is configured to calculate a value of the first representation indicator based on a time series prediction algorithm to obtain a first dynamic threshold of the first representation indicator; A third acquisition unit is configured to obtain root cause information and a score corresponding to the root cause information according to the first root cause indicator set, the confidence degree and the first dynamic threshold; A positioning unit is configured to locate a root cause affecting wireless network performance according to the root cause information and the score corresponding to the root cause information. The first acquisition unit comprises a determination unit configured to determine a collection period affecting a wireless network indicator. The second acquisition unit comprises a second construction subunit configured to construct three time series of long-term trend, periodic trend and random fluctuation according to the collection period; a third acquisition subunit configured to pre-process the first representation indicator to remove abnormal values of the first representation indicator in the time series according to the time series; and a fourth acquisition subunit configured to train the processed first representation indicator based on the time series using a Prophet algorithm to obtain a first dynamic threshold of the first representation indicator changing dynamically over time. The third acquisition unit comprises a fifth acquisition subunit configured to obtain a third root cause indicator set by obtaining a third root cause indicator with the highest confidence degree from the first root cause indicator set; a second determination subunit configured to determine a second threshold value by determining an inflection point of a value of the third root cause indicator in the third root cause indicator set; a sixth acquisition subunit configured to obtain a target root cause indicator according to the first dynamic threshold; and a seventh acquisition subunit configured to build a root cause tree, traverse the root cause tree according to the target root cause indicator and the second threshold value, and obtain the root cause information and the score corresponding to the root cause information.

8. An electronic device, comprising: The apparatus comprises one or more processors and a memory configured to store one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the processing method of the wireless network according to any one of claims 1 to 6.

Images