Cell failure situation prediction method, apparatus, and related device

Abstract

The present disclosure provides a cell failure situation prediction method and device and related equipment, relating to the technical field of wireless communication, which comprises: acquiring a plurality of performance index data of a target cell; performing principal component analysis processing on the plurality of performance index data of the target cell, transforming the plurality of performance index data of the target cell to a low-dimensional space to obtain a plurality of target performance index data; inputting the plurality of target performance index data of the target cell into a pre-trained light gradient boosting machine model to output a failure probability of each performance index data; inputting the failure probability of each performance index data into a pre-trained random forest model to output a failure probability of the target cell; and determining a prediction result of the target cell according to the failure probability of the target cell. The present disclosure can predict whether the cell will fail in advance.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of wireless communication, and particularly relates to a cell failure situation prediction method and device and related equipment. BACKGROUND

[0002] The wireless network failure detection and processing technology in the prior art has the following problems:

[0003] 1) There is a lack of accurate and effective failure prediction technology, and only troubleshooting can be performed after the failure occurs, and the network has already been affected before the failure occurs, which will lead to network quality deterioration and increased user complaints. If the failure can be predicted before it occurs and the root cause of the failure is eliminated in advance, the stability and reliability of the network will be improved, and user complaints will be effectively reduced.

[0004] 2) The failure processing method in the prior art requires manual troubleshooting of various network elements and indicators, and the number of indicators in the existing network is large, so troubleshooting requires a large amount of manpower and time costs, and it is difficult to locate the root cause of the failure and perform corresponding repair in a short period of time, which will further amplify the impact of the failure.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The present disclosure provides a cell failure situation prediction method, device and related equipment, which at least partially overcomes the problem of lack of early prediction of cell failure in the prior art.

[0007] Other characteristics and advantages of the present disclosure will become apparent from the following detailed description, or will be learned by practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a cell failure situation prediction method is provided, comprising: obtaining a plurality of performance indicator data of a target cell; performing principal component analysis processing on the plurality of performance indicator data of the target cell, transforming the plurality of performance indicator data of the target cell to a low-dimensional space to obtain a plurality of target performance indicator data; inputting the plurality of target performance indicator data of the target cell into a pre-trained light gradient boosting machine model, outputting a failure probability of each performance indicator data; inputting the failure probability of each performance indicator data into a pre-trained random forest model, outputting a failure probability of the target cell; and determining a prediction result of the target cell according to the failure probability of the target cell.

[0009] In some example embodiments of the present disclosure, based on the foregoing scheme, the prediction result comprises that the target cell will fail or will not fail, and determining the prediction result of the target cell according to the failure probability of the target cell comprises: comparing the failure probability of the target cell with a first preset threshold; in the case that the failure probability of the target cell is greater than or equal to the first preset threshold, determining that the target cell will fail; in the case that the failure probability of the target cell is less than the first preset threshold, determining that the target cell will not fail.

[0010] In some example embodiments of the present disclosure, based on the foregoing scheme, when outputting the failure probability of the target cell, the random forest model is further configured to output a contribution value of each performance indicator data in the plurality of target performance indicator data to the prediction result, and after determining that the target cell will fail, the method further comprises: performing failure localization on the target cell according to the contribution value of each performance indicator data to the prediction result.

[0011] In some example embodiments of the present disclosure, based on the foregoing scheme, the principal component analysis processing on the plurality of performance indicator data of the target cell is performed to transform the plurality of performance indicator data of the target cell to a low-dimensional space to obtain the plurality of target performance indicator data, comprising: calculating the cross-correlation between the plurality of performance indicator data of the target cell; screening out performance indicator data less than a second preset threshold according to the cross-correlation; performing principal component analysis processing on the performance indicator data screened out less than the second preset threshold to obtain the plurality of target performance indicator data of the target cell.

[0012] In some example embodiments of the present disclosure, based on the foregoing scheme, the cross-correlation between the performance indicator data of the target cell is calculated, comprising: inputting the plurality of performance indicator data of the target cell into a pre-trained cross-correlation calculation model to output the cross-correlation between the performance indicator data of the target cell.

[0013] In some example embodiments of the present disclosure, based on the foregoing scheme, before inputting the plurality of performance indicator data of the target cell into the pre-trained cross-correlation calculation module to output the cross-correlation between the performance indicator data of the target cell, the method further comprises: preprocessing the performance indicator data of the target cell, the preprocessing at least comprising one of the following: missing value filling processing, data smoothing processing and threshold processing.

[0014] According to another aspect of the present disclosure, a cell failure situation prediction apparatus is also provided, comprising: a performance index data acquisition module configured to acquire a plurality of performance index data of a target cell; a principal component analysis processing module configured to perform principal component analysis processing on the plurality of performance index data of the target cell, transform the plurality of performance index data of the target cell to a low-dimensional space, and obtain a plurality of target performance index data; a performance index data failure probability output module configured to input the plurality of target performance index data of the target cell into a pre-trained light gradient boosting machine model, and output a failure probability of each performance index data; a target cell failure probability output module configured to input the failure probability of each performance index data into a pre-trained random forest model, and output a failure probability of the target cell; and a prediction result determination module configured to determine a prediction result of the target cell according to the failure probability of the target cell.

[0015] According to still another aspect of the present disclosure, an electronic device is also provided, comprising: a processor; and a memory configured to store executable instructions of the processor; wherein the processor is configured to perform any of the above-mentioned cell failure situation prediction methods via execution of the executable instructions.

[0016] According to yet another aspect of the present disclosure, a computer-readable storage medium is also provided, having stored thereon a computer program, which, when executed by a processor, implements any of the above-mentioned cell failure situation prediction methods.

[0017] According to another aspect of the present disclosure, a computer program product is also provided, comprising: a computer program or instructions, which, when executed by a processor, implements any of the above-mentioned cell failure situation prediction methods.

[0018] The cell failure situation prediction method, apparatus and related device provided in the embodiments of the present disclosure can predict whether the target cell will fail by processing a plurality of performance index data of the target cell to obtain a plurality of target performance index data, training a light gradient boosting machine model on the plurality of target performance index data, outputting a failure probability of each target performance index data, and finally inputting the failure probabilities of all target performance index data into a pre-trained random forest model. Further, the embodiments of the present disclosure combine the fine analysis of the light gradient boosting machine model and the decision advantage of the random forest model to predict the failure of the target cell, which can improve the prediction accuracy and reduce misjudgment.

[0019] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and are not limiting to the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings, which are incorporated herein and constitute part of the specification, illustrate embodiments consistent with the present disclosure and, together with the description, further serve to explain the principles of the present disclosure. It is to be understood that the drawings are only schematic, and that they do not necessarily represent a limiting case of the embodiments of the present disclosure. For a better understanding, the embodiments will now be described, by way of example only, with reference to the drawings, in which:

[0021] Figure 1 An exemplary application system architecture schematic diagram of a cell failure condition prediction method in an embodiment of the present disclosure is shown.

[0022] Figure 2 A schematic diagram of a cell failure condition prediction method in an embodiment of the present disclosure is shown.

[0023] Figure 3 A schematic diagram of a method for determining whether a target cell will fail in an embodiment of the present disclosure is shown, which combines a light gradient boosting machine model and a random forest model.

[0024] Figure 4 A schematic diagram of a cell failure condition prediction process in an embodiment of the present disclosure is shown.

[0025] Figure 5 A schematic diagram of a cell failure condition prediction device in an embodiment of the present disclosure is shown.

[0026] Figure 6 A schematic diagram of an electronic device applying a cell failure condition prediction method in an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0027] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the implementations set forth herein; rather, these implementations are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the example implementations to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more implementations.

[0028] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more implementations. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the present disclosure. One skilled in the relevant art will recognize, however, that the embodiments of the present disclosure can be practiced without one or more of the specific details, or with other methods, components, materials, and so forth. In other instances, well-known structures, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the present disclosure.

[0029] The flowcharts shown in the drawings are only illustrative, and do not necessarily include all contents and operations / steps, nor are they necessarily executed in the order described. For example, some operations / steps can be further decomposed, and some operations / steps can be combined or partially combined, so the actual execution order can be changed according to actual conditions.

[0030] As shown in the figure, the system architecture includes a terminal device 101, a network 102, and a network side device 103. Figure 1

[0031] The network 102 is a medium for providing a communication link between the terminal device 101 and the network side device 103, and can be a wired network or a wireless network.

[0032] Optionally, the wireless network or wired network described above uses standard communication technology and / or protocols. The network is usually the Internet, but can also be any network, including but not limited to a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or any combination of virtual private networks). In some embodiments, technologies and / or formats including Hyper Text Mark-up Language (HTML), Extensible Markup Language (XML), etc. are used to represent data exchanged through the network. In addition, all or some links can be encrypted using conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPSec), etc. In other embodiments, custom and / or dedicated data communication technologies can be used instead of or in addition to the above data communication technologies.

[0033] ​Optionally, the terminal device in the embodiments of the present disclosure can also be referred to as a UE (User Equipment), and in specific implementation, the terminal device can be a mobile phone, a tablet computer, a laptop computer, a PDA (Personal Digital Assistant), a MID (Mobile Internet Device), a wearable device, or a vehicle-mounted device, and it should be noted that the specific type of the terminal device is not limited in the embodiments of the present disclosure.

[0034] The network side device can be a base station, a relay, or an access point, etc. The base station can be a base station of 5G and later versions (for example: 5G NR NB), or a base station in other communication systems (for example: eNB base station), and it should be noted that the specific type of the network side device is not limited in the embodiments of the present disclosure.

[0035] Those skilled in the art can know that, Figure 1 The number of terminals, networks and network side devices in the embodiments of the present disclosure is only illustrative, and any number of terminals, networks and network side devices can be provided according to actual needs. The embodiments of the present disclosure do not limit this.

[0036] In order to more clearly describe the embodiments of the present disclosure, the following professional terms are explained:

[0037] Feature engineering is an important step in machine learning and data science, which involves converting raw data into features that better reflect the nature of the problem. Good feature engineering can significantly improve the performance of the model. The main tasks of feature engineering include feature selection, feature construction and feature transformation.

[0038] Machine learning is an important branch of AI, which enables computers to learn from data without explicit programming and make predictions or decisions. The core of machine learning is to build algorithm models to enable computers to automatically improve their performance and improve accuracy with experience.

[0039] PCA (Principal Component Analysis) is a commonly used data dimensionality reduction technique, widely used in data preprocessing, feature extraction and visualization, etc. The main goal of PCA is to convert high-dimensional data into low-dimensional data through linear transformation, while preserving as much variance information of the original data as possible.

[0040] Light Gradient Boosting Machine (LightGBM): A high-efficiency machine learning framework based on Gradient Boosting Decision Tree (GBDT) algorithm, developed by Microsoft Research and open-sourced in 2016.

[0041] Random Forest (RF): An ensemble learning method that improves prediction accuracy and prevents overfitting by constructing multiple decision trees and taking their average results. Random Forest has high accuracy in predicting high-dimensional data and can effectively eliminate the influence of noise and outliers.

[0042] Figure 2 A schematic diagram of a cell failure condition prediction method in an embodiment of the present disclosure is shown, which includes the following steps:

[0043] S202, obtaining a plurality of performance indicator data of a target cell.

[0044] It should be noted that the target cell in the embodiment of the present disclosure can be any geographical area covered by one base station or multiple cooperatively working base stations, for example, the geographical area covered by base station A includes cell 1, cell 2, and cell 3, and the target cell can be cell 1, cell 2, or cell 3. In addition, the plurality of performance indicator data of the cell is used to evaluate and optimize the running condition of the network, for example, the performance indicator data includes: connection success rate, call drop rate / call drop rate, handover success rate, uplink / downlink throughput, etc. More specifically, the connection success rate is a measure of the success rate of user equipment attempting to access the network, usually expressed in percentage, a higher success rate means better network coverage and service quality; the call drop rate / call drop rate refers to the proportion of unexpected interruption in the process of ongoing calls or data transmission. Low call drop rate is an important indicator of network stability; the handover success rate refers to the ratio of successful handover from one cell to another when the user moves between different cells, an efficient handover process is crucial for maintaining call quality and data service continuity; the uplink / downlink throughput reflects the speed of data transmission on the network, corresponding to the data transmission rate from the user to the network (uplink) and from the network to the user (downlink), respectively.

[0045] S204, performing principal component analysis on the plurality of performance indicator data of the target cell, transforming the plurality of performance indicator data of the target cell to a low-dimensional space, and obtaining a plurality of target performance indicator data.

[0046] It should be noted that the principal component analysis processing in the embodiments of the present disclosure is used to reduce the dimension number of performance index data. Specifically, the principal component analysis processing includes data standardization, calculation of a covariance matrix or a correlation coefficient matrix, calculation of eigenvalues and eigenvectors, selection of principal components, and conversion of original data. More specifically, since different performance index data can have different dimensions and scales, the original multiple performance index data first needs to be standardized to make each variable have a zero mean and a unit variance. The covariance matrix shows the degree of covariation between different performance index data, while the correlation coefficient matrix measures the strength and direction of the relationship and is not affected by the scale. By solving the eigenvalues and eigenvectors of the covariance matrix or the correlation coefficient matrix, it can be determined which directions (i.e., principal components) explain the largest data variation. According to the size of the eigenvalues, the eigenvectors that explain most of the data variation are selected as the principal components. Generally, when the cumulative contribution rate reaches 70%-95%, it can be considered that these principal components can better represent the information of the original data set. The original multiple performance index data are projected onto the selected principal components to obtain multiple target performance index data after dimension reduction.

[0047] In S206, the multiple target performance index data of the target cell are input into the pre-trained light gradient boosting machine model, and the failure probability of each performance index data is output.

[0048] It should be noted that the light gradient boosting machine model in the embodiments of the present disclosure can be a model obtained by pre-training various artificial intelligence algorithm models (for example, a neural network model) or a combination model thereof through machine learning. The model can calculate the failure probability of each performance index data from the multiple target performance index data of the target cell. The input data of the model is the multiple target performance index data of the target cell, and the output data is the failure probability of each performance index data.

[0049] Through the above embodiments, the light gradient boosting machine model trained through machine learning is used to calculate the failure probability of the multiple target performance index data of the target cell to obtain the failure probability of each performance index data corresponding thereto, so that the rapid calculation of the failure probability of the target performance index data can be realized.

[0050] In some embodiments, compared with traditional gradient boosting methods, the lightweight gradient boosting machine model in the embodiments of the present disclosure has a significant improvement in speed, especially on large data sets. The lightweight gradient boosting machine model in the embodiments of the present disclosure can greatly reduce memory consumption by using a histogram algorithm instead of a traditional presorting algorithm, so that the model training process is more lightweight. Moreover, the lightweight gradient boosting machine model can effectively handle problems with high feature dimension and large sample size and provide accurate prediction results. Furthermore, the lightweight gradient boosting machine model supports multi-thread parallel computing and distributed training, which further speeds up the training speed and is particularly suitable for large-scale data processing tasks in industrial applications.

[0051] S208, input the failure probability of each performance indicator data into the pre-trained random forest model, and output the failure probability of the target cell.

[0052] It should be noted that the random forest model in the embodiments of the present disclosure can be a model obtained by pre-training various artificial intelligence algorithm models (for example, neural network models) or combined models through machine learning. The model can calculate the failure probability of each performance indicator data of the target cell to obtain the failure probability of the target cell. The input data of the model is the failure probability of each performance indicator data of the target cell, and the output data is the failure probability of the target cell.

[0053] In some embodiments, the random forest model in the embodiments of the present disclosure is a powerful ensemble learning method that improves the performance of the model by constructing multiple decision trees and aggregating their prediction results. More specifically, the random forest model reduces the variance of a single model by integrating multiple decision trees, thereby reducing the risk of overfitting and improving the prediction accuracy of unseen data. When facing a data set with a large number of features, the random forest model can work effectively without feature selection or dimensionality reduction. Because each decision tree can be trained independently, the random forest model can use multi-core processors for parallel computing to speed up the training.

[0054] S210, determining the prediction result of the target cell according to the failure probability of the target cell.

[0055] The cell failure situation prediction method provided in the embodiments of the present disclosure first acquires a plurality of performance index data of a target cell; secondly, performs principal component analysis processing on the plurality of performance index data of the target cell, transforms the plurality of performance index data of the target cell to a low-dimensional space, and obtains a plurality of target performance index data; then, inputs the plurality of target performance index data of the target cell into a pre-trained light gradient boosting machine model, and outputs a failure probability of each performance index data; thereafter, inputs the failure probability of each performance index data into a pre-trained random forest model, and outputs a failure probability of the target cell; finally, determines a prediction result of the target cell according to the failure probability of the target cell. Compared with the problem of lacking cell failure prediction in the related art, the embodiments of the present disclosure obtain a plurality of target performance index data by processing a plurality of performance index data of a target cell, train a light gradient boosting machine model for the plurality of target performance index data, output a failure probability of each target performance index data, and finally input the failure probability of all target performance index data into a pre-trained random forest model to predict whether the target cell will fail, further, the embodiments of the present disclosure combine the light gradient boosting machine model fine analysis and the random forest model decision advantage to predict the failure of the target cell, which can improve the prediction accuracy and reduce the misjudgment.

[0056] In some embodiments, the prediction result in the embodiments of the present disclosure includes that the target cell will fail or will not fail, and determining the prediction result of the target cell according to the failure probability of the target cell includes: comparing the failure probability of the target cell with a first preset threshold; in a case where the failure probability of the target cell is greater than or equal to the first preset threshold, determining that the target cell will fail; and in a case where the failure probability of the target cell is less than the first preset threshold, determining that the target cell will not fail. Specifically, after the random forest model outputs the failure probability of the target cell, if the failure probability is greater than or equal to the first preset threshold, it is determined that the target cell will fail; if the failure probability is less than the first preset threshold, it is determined that the target cell will not fail. More specifically, the first preset threshold in the embodiments of the present disclosure is a threshold obtained based on multiple experiments, for example, the first preset threshold can be set to 30% in the embodiments of the present disclosure. It should be noted that the first preset threshold is not specifically limited in the embodiments of the present disclosure, and the first preset threshold in the embodiments of the present disclosure can be flexibly set according to actual conditions.

[0057] In some embodiments, the random forest model in the embodiments of the present disclosure, when outputting the failure probability of the target cell, is further configured to output a contribution value of each performance indicator data in the plurality of target performance indicator data to the prediction result, and after determining that the target cell will fail, the cell failure situation prediction method in the embodiments of the present disclosure further comprises: performing failure localization on the target cell according to the contribution value of each performance indicator data to the prediction result. Specifically, the random forest model in the embodiments of the present disclosure can output the contribution value of each performance indicator data to the prediction result when predicting, and the performance indicator data with a large contribution value is the root cause of the failure. Based on the failure prediction, the root cause of the failure is determined, the root cause of the failure can be quickly identified and located, the operation and maintenance cost is reduced, and the network stability is improved.

[0058] In some embodiments, the embodiments of the present disclosure perform principal component analysis processing on the plurality of performance indicator data of the target cell, transform the plurality of performance indicator data of the target cell to a low-dimensional space, and obtain the plurality of target performance indicator data, comprising: calculating the cross-correlation between the plurality of performance indicator data of the target cell; filtering out the performance indicator data less than the second preset threshold according to the cross-correlation; performing principal component analysis processing on the performance indicator data less than the second preset threshold to obtain the plurality of target performance indicator data of the target cell. Specifically, the embodiments of the present disclosure first calculate the correlation between the plurality of performance indicator data of the target cell, retain the performance indicator data with a cross-correlation less than a second preset threshold, and then perform principal component analysis on the retained performance indicator data, realize data dimension reduction, improve the training and inference speed of the subsequent model, and reduce the influence of noise; more specifically, the second preset threshold in the embodiments of the present disclosure is a threshold obtained based on multiple experiments, for example, the second preset threshold in the embodiments of the present disclosure can be set to 60%, and it should be noted that the second preset threshold in the embodiments of the present disclosure is not specifically limited, and the second preset threshold in the embodiments of the present disclosure can be flexibly set according to actual conditions.

[0059] In some embodiments, the disclosure calculates the cross-correlation between the performance indicator data of the target cell, including: inputting the plurality of performance indicator data of the target cell into the pre-trained cross-correlation calculation model, and outputting the cross-correlation between the performance indicator data of the target cell. It should be noted that the cross-correlation is a statistical method for measuring the degree of similarity between two performance indicator data. Specifically, the neural network model can also be used to calculate or estimate the cross-correlation in certain specific scenarios, especially in complex data pattern recognition and high-dimensional data analysis. The cross-correlation calculation model in the disclosure can be obtained after training the neural network model. For example, the convolutional neural network is good at capturing local features, and the disclosure can indirectly estimate the cross-correlation by learning the features of the input data. For data with time series, a recurrent neural network, a long short-term memory network, or a gated recurrent unit can be used to capture the time dependence relationship, and the cross-correlation can be estimated by the learned features.

[0060] In some embodiments, before inputting the plurality of performance indicator data of the target cell into the pre-trained cross-correlation calculation module and outputting the cross-correlation between the performance indicator data of the target cell, the cell failure condition prediction method in the disclosure further includes: preprocessing the performance indicator data of the target cell, which at least includes one of the following: missing value filling processing, data smoothing processing and threshold processing. Specifically, preprocessing the performance indicator data of the target cell includes: supplementing missing values; detecting discontinuous zero values in the data and using adjacent data smoothing; detecting abnormal values exceeding the normal range and replacing them with the upper and lower limits of the data. By preprocessing the performance indicator data of the target cell, abnormal values can be removed to ensure the authenticity and reliability of the data. Missing data can be processed by interpolation, mean filling, etc. to avoid analysis bias caused by missing values. Preprocessing the performance indicator data is an important step in data analysis and optimization process. Preprocessing can significantly improve the quality of data, so that subsequent analysis, modeling and decision making are more accurate and effective.

[0061] In some embodiments, as shown in Figure 3 The flow of the disclosure combining the light gradient boosting machine model and the random forest model to obtain the prediction result of whether the target cell will fail includes:

[0062] S302, obtaining a plurality of performance indicator data of a target cell in a past preset time period, including historical data of performance indicator 1, historical data of performance indicator 2, …, historical data of performance indicator n.

[0063] S304, input the obtained plurality of performance index historical data into the corresponding pre-trained light gradient boosting machine model, and output the failure probability of each performance index data, including the failure probability of performance index 1, the failure probability of performance index 2, …, and the failure probability of performance index n.

[0064] S306, input the failure probability of each performance index data into the pre-trained random forest model, and output the failure probability of the target cell.

[0065] S308, determine the prediction result of whether the target cell will fail according to the failure probability of the target cell.

[0066] In some embodiments, the performance index data is constructed to correspond to the light gradient boosting machine model, the failure probability is obtained by inputting the performance index data, and all failure probabilities are input into the random forest model to compare with the first preset threshold to determine the failure condition of the target cell.

[0067] In some embodiments, as shown in Figure 4 The cell failure condition prediction method of the embodiments of the present disclosure specifically includes:

[0068] S402, obtain a plurality of performance index data of the target cell at an hour granularity in a past preset time period, and arrange them in chronological order.

[0069] S404, pre-process the obtained historical data (i.e., the plurality of performance index data at an hour granularity in the past preset time period).

[0070] S406, calculate the cross-correlation between each performance index data, and select one to keep according to a preset retention condition for two performance index data whose cross-correlation exceeds a second preset threshold, until the cross-correlation between all performance index data is less than the second preset threshold.

[0071] S408, perform principal component analysis on the retained performance index data, discard the performance index data whose eigenvalue is less than a second preset threshold, transform the original performance index data to a low-dimensional space, and obtain a plurality of target performance index data.

[0072] S410, input each target performance index data into a pre-trained light gradient boosting machine model to obtain the failure probability of each target performance index data.

[0073] S412, input the failure probability of all target performance index data into a pre-trained random forest model, output the failure probability of the target cell, and compare it with the first preset threshold to determine whether the target cell will fail in the future; repeat S410 and S412 when training the model until the error meets the requirement or the iteration number reaches the upper limit.

[0074] S414, if it is judged that the target cell will fail in the future, the contribution of each performance indicator data to the prediction result is calculated and sorted, and the performance indicator data causing the failure is output to assist in failure root cause positioning.

[0075] In some embodiments, the embodiments of the present disclosure control the optimization model according to the error and the number of iterations by repeating training, calculate and output key performance indicator data for predicting target cell failure to assist in root cause positioning. The random forest model in the embodiments of the present disclosure can cope with complex environments, and output root cause indicators based on failure prediction, reduce operation and maintenance costs, and improve network stability.

[0076] Based on the same inventive concept, the present disclosure also provides a cell failure condition prediction device, as follows. Since the principle of solving the problem of the device embodiments is similar to that of the above-mentioned method embodiments, the implementation of the device embodiments can be referred to the implementation of the above-mentioned method embodiments, and the repeated parts will not be repeated.

[0077] Figure 5 A schematic diagram of a cell failure condition prediction device in an embodiment of the present disclosure is shown. The device includes:

[0078] The performance indicator data acquisition module 501 is configured to acquire a plurality of performance indicator data of a target cell.

[0079] The principal component analysis processing module 502 is configured to perform principal component analysis processing on the plurality of performance indicator data of the target cell, transform the plurality of performance indicator data of the target cell to a low-dimensional space, and obtain a plurality of target performance indicator data.

[0080] The performance indicator data failure probability output module 503 is configured to input the plurality of target performance indicator data of the target cell into a pre-trained light gradient boosting machine model, and output a failure probability of each performance indicator data.

[0081] The target cell failure probability output module 504 is configured to input the failure probability of each performance indicator data into a pre-trained random forest model, and output a failure probability of the target cell.

[0082] The prediction result determination module 505 is configured to determine a prediction result of the target cell according to the failure probability of the target cell.

[0083] The cell failure situation prediction device provided in the embodiments of the present disclosure includes: a performance index data acquisition module, configured to acquire a plurality of performance index data of a target cell; a principal component analysis processing module, configured to perform principal component analysis processing on the plurality of performance index data of the target cell, transform the plurality of performance index data of the target cell to a low-dimensional space, and obtain a plurality of target performance index data; a performance index data failure probability output module, configured to input the plurality of target performance index data of the target cell into a pre-trained light gradient boosting machine model, and output a failure probability of each performance index data; a target cell failure probability output module, configured to input the failure probability of each performance index data into a pre-trained random forest model, and output a failure probability of the target cell; and a prediction result determination module, configured to determine a prediction result of the target cell according to the failure probability of the target cell. Compared with the problem of lacking cell failure prediction in the related art, the embodiments of the present disclosure can obtain a plurality of target performance index data by processing a plurality of performance index data of a target cell, train a light gradient boosting machine model for the plurality of target performance index data, output a failure probability of each target performance index data, and finally input the failure probability of all target performance index data into a pre-trained random forest model to predict whether the target cell will fail, thereby improving the prediction accuracy and reducing misjudgment.

[0084] In some embodiments, the prediction result in the embodiments of the present disclosure includes that the target cell will fail or will not fail, and the prediction result determination module in the embodiments of the present disclosure is further configured to compare the failure probability of the target cell with a first preset threshold, determine that the target cell will fail when the failure probability of the target cell is greater than or equal to the first preset threshold, and determine that the target cell will not fail when the failure probability of the target cell is less than the first preset threshold.

[0085] In some embodiments, when outputting the failure probability of the target cell, the random forest model in the embodiments of the present disclosure is further configured to output a contribution value of each performance index data in the plurality of target performance index data to the prediction result, and the cell failure situation prediction device in the embodiments of the present disclosure further includes: a failure positioning module, configured to perform failure positioning on the target cell according to the contribution value of each performance index data to the prediction result after determining that the target cell will fail.

[0086] In some embodiments, the principal component analysis processing module in the embodiments of the present disclosure is further configured to calculate the cross-correlation between the plurality of performance index data of the target cell; filter out the performance index data less than the second preset threshold according to the cross-correlation; and perform principal component analysis processing on the performance index data less than the second preset threshold to obtain a plurality of target performance index data of the target cell.

[0087] In some embodiments, the principal component analysis processing module in the embodiments of the present disclosure is further configured to input the plurality of performance index data of the target cell into a pre-trained cross-correlation calculation model, and output the cross-correlation between the performance index data of the target cell.

[0088] In some embodiments, the cell failure condition prediction apparatus in the embodiments of the present disclosure further comprises a performance index data preprocessing module configured to preprocess the performance index data of the target cell before inputting the performance index data of the target cell into the pre-trained cross-correlation calculation model and outputting the cross-correlation between the performance index data of the target cell, the preprocessing at least including one of the following: missing value filling processing, data smoothing processing, and threshold processing.

[0089] Those skilled in the art can understand that each aspect of the present disclosure can be implemented as a system, a method or a program product. Therefore, each aspect of the present disclosure can be specifically implemented as a complete hardware embodiment, a complete software embodiment (including firmware, microcode, etc.), or an embodiment combining software and hardware aspects, which can be collectively referred to as "circuitry", "module" or "system".

[0090] Based on the same inventive concept, the embodiments of the present disclosure further provide an electronic device, which comprises a processor and a memory for storing executable instructions of the processor, wherein the processor is configured to execute any one of the above-mentioned cell failure condition prediction methods by executing the executable instructions. Since the principle of solving problems of the electronic device embodiments is similar to that of the above-mentioned method embodiments, the implementation of the electronic device embodiments can be referred to the implementation of the above-mentioned method embodiments, and the repeated parts will not be described here.

[0091] The electronic device 600 according to this embodiment of the present disclosure will be described below with reference to Figure 6 Figure 6 The displayed electronic device 600 is only an example and should not impose any limitation on the functions and use range of the embodiments of the present disclosure.

[0092] As Figure 6 ​As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 can include, but are not limited to, the at least one processing unit 601 described above, the at least one storage unit 602 described above, and a bus 603 that connects the different system components including the storage unit 602 and the processing unit 601.

[0093] The storage unit stores program codes which can be executed by the processing unit 601, so that the processing unit 601 performs the steps described in the above “Exemplary Methods” section according to various exemplary embodiments of the present disclosure.

[0094] In some embodiments, when the electronic device is used to control, for example, the cell failure situation prediction method described above, the processing unit 601 can perform the following steps of the above method embodiments:

[0095] Obtaining a plurality of performance indicator data of a target cell; performing principal component analysis processing on the plurality of performance indicator data of the target cell, transforming the plurality of performance indicator data of the target cell to a low-dimensional space to obtain a plurality of target performance indicator data; inputting the plurality of target performance indicator data of the target cell into a pre-trained light gradient boosting machine model, and outputting a failure probability of each performance indicator data; inputting the failure probability of each performance indicator data into a pre-trained random forest model, and outputting a failure probability of the target cell; determining a prediction result of the target cell according to the failure probability of the target cell.

[0096] The storage unit 602 can include a readable medium in the form of a volatile storage unit, such as a random access memory (RAM) 6021 and / or a cache memory 6022, and can further include a read-only memory (ROM) 6023.

[0097] The storage unit 602 can further include program / utility 6024 having a set of program modules 6025, including but not limited to, an operating system, one or more application programs, other program modules, and program data, and each of these examples or some combination thereof, can include implementation of a network environment.

[0098] The bus 603 can represent one or more of several types of bus structures, including a storage unit bus or storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit bus, or a local bus using any of a variety of bus architectures.

[0099] The electronic device 600 can also communicate with one or more external devices 604 such as a keyboard or pointing device, a Bluetooth device, or a database, etc.; and / or one or more devices that enable a user to interact with the electronic device 600; and / or any devices (e.g., routers, modems, etc.) that enable the electronic device 600 to communicate with one or more other computing devices. Such communication can occur via Input / Output (I / O) interface 605. Still yet, the electronic device 600 can communicate with one or more networks, such as a local area network (LAN), a wide area network (WAN), and / or the Internet, through a network adapter 606. As depicted, the network adapter 606 communicates with the other components of the electronic device 600 via bus 603. It should be appreciated that although not shown, other hardware and / or software modules could be used in connection with the electronic device 600. Such modules can include, but are not limited to, microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems, etc.

[0100] Those skilled in the art will readily understand that the example embodiments described herein can be implemented by software and / or by hardware coupled with software, as described above. Thus, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, a USB flash disk, a mobile hard disk, etc.) or a network, and includes a number of instructions to enable a computing device (such as a personal computer, a server, a terminal device, or a network device, etc.) to execute the methods according to the embodiments of the present disclosure.

[0101] Based on the same inventive concept, the present disclosure also provides a computer readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the cell failure situation prediction method of any one of the above. Since the computer readable storage medium embodiment solves problems in the same principle as the above method embodiments, the implementation of the computer readable storage medium embodiment can be referred to the implementation of the above method embodiments, and the repeated parts will not be described here.

[0102] More specific examples of the computer readable storage medium in the present disclosure can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0103] In this disclosure, a computer readable storage medium can include a data signal transporting or broadcasting computer readable program code embodied in the data signal. The data signal can also be transmitted over a network including the Internet. Additionally, the computer readable storage medium can also be any tangible storage medium which can be used to store and / or carry the program codes.

[0104] Optionally, program code embodied on a computer readable storage medium can also be transmitted using any appropriate medium, including but not limited to wireless, wired, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0105] In an implementation, the program code for carrying out operations of the present disclosure can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computing device, partly on the user's computing device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, such as through the Internet using an Internet Service Provider.

[0106] Based on the same inventive concept, the present disclosure further provides a computer program product, comprising a computer program or instructions, which, when executed by a processor, implement the cell failure situation prediction method of any one of the above method embodiments. Since the principle of solving problems by the computer program product embodiment is similar to that of the above method embodiments, implementation of the computer program product embodiment can be referred to the implementation of the above method embodiments, and repeated descriptions are not repeated.

[0107] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, such division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into a plurality of modules or units.

[0108] Moreover, although individual steps of the methods in the present disclosure are described in a particular order in the drawings, this is not required or implied, nor is it necessary to perform all of the steps shown to achieve the desired result. Additionally or alternatively, certain steps can be omitted, combined into a single step, performed in a different order, broken down into multiple steps, and / or the like.

[0109] Those skilled in the art will readily understand that the example embodiments described herein can be implemented by software and / or by hardware coupled with software, as described above. Thus, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product. The software product can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash disk, a mobile hard disk, or the like) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to perform the methods according to the embodiments of the present disclosure.

[0110] Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure following the general principles thereof and including such departures from the present disclosure that come within known use or custom in the art to which the present disclosure pertains. The specification and examples are to be regarded as illustrative only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A method of predicting a cell failure situation, characterized by, The method comprises the following steps: obtaining a plurality of performance index data of a target cell; performing principal component analysis processing on the plurality of performance index data of the target cell, and transforming the plurality of performance index data of the target cell to a low-dimensional space to obtain a plurality of target performance index data; inputting the plurality of target performance index data of the target cell into a pre-trained light gradient boosting machine model, and outputting a failure probability of each performance index data; inputting the failure probability of each performance index data into a pre-trained random forest model, and outputting a failure probability of the target cell; determining a prediction result of the target cell according to the failure probability of the target cell.

2. The cell failure situation prediction method according to claim 1, characterized by, The prediction result includes that the target cell will fail or will not fail, and determining the prediction result of the target cell according to the failure probability of the target cell comprises the following steps: comparing the failure probability of the target cell with a first preset threshold; in the case that the failure probability of the target cell is greater than or equal to the first preset threshold, determining that the target cell will fail; in the case that the failure probability of the target cell is less than the first preset threshold, determining that the target cell will not fail.

3. The cell failure situation prediction method according to claim 2, characterized by, When the random forest model outputs the failure probability of the target cell, it is also used to output a contribution value of each performance index data in the plurality of target performance index data to the prediction result, and after determining that the target cell will fail, the method further comprises the following steps: locating a failure of the target cell according to the contribution value of each performance index data to the prediction result.

4. The cell failure situation prediction method according to claim 1, characterized by, The principal component analysis processing on the plurality of performance index data of the target cell, and transforming the plurality of performance index data of the target cell to a low-dimensional space to obtain a plurality of target performance index data comprises the following steps: calculating the cross-correlation between the plurality of performance index data of the target cell; screening performance index data less than a second preset threshold according to the cross-correlation; performing principal component analysis processing on the performance index data screened less than the second preset threshold to obtain the plurality of target performance index data of the target cell.

5. The cell failure situation prediction method according to claim 4, characterized by, The calculation of the cross-correlation between the performance index data of the target cell comprises the following steps: inputting the plurality of performance index data of the target cell into a pre-trained cross-correlation calculation model to output the cross-correlation between the performance index data of the target cell.

6. The cell failure condition prediction method according to claim 4, characterized by, Before inputting the plurality of performance index data of the target cell into the pre-trained cross-correlation calculation module to output the cross-correlation between the performance index data of the target cell, the method further comprises the following steps: performing preprocessing on the performance index data of the target cell, wherein the preprocessing at least comprises one of the following: missing value filling processing, data smoothing processing and threshold processing.

7. A cell failure situation prediction apparatus characterized by comprising: The method comprises the following steps: a performance index data acquisition module for acquiring a plurality of performance index data of a target cell; a principal component analysis processing module for performing principal component analysis processing on the plurality of performance index data of the target cell, and transforming the plurality of performance index data of the target cell to a low-dimensional space to obtain a plurality of target performance index data; The performance index data fault probability output module is configured to input a plurality of target performance index data of the target cell into a pre-trained light gradient boosting machine model, and output a fault probability of each performance index data. The target cell fault probability output module is configured to input the fault probability of each performance index data into a pre-trained random forest model, and output a fault probability of the target cell. The prediction result determination module is configured to determine a prediction result of the target cell according to the fault probability of the target cell.

8. An electronic device, comprising: The cell fault condition prediction method comprises: a processor; and a memory configured to store executable instructions of the processor; wherein the processor is configured to execute the executable instructions to perform the cell fault condition prediction method according to any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the cell fault condition prediction method according to any one of claims 1-6.

10. A computer program product, comprising: The computer program or instructions are executed by the processor to implement the cell fault condition prediction method according to any one of claims 1-6.

Images