CN117221910A

CN117221910A - Data processing method, device, equipment and medium for wireless network optimization

Info

Publication number: CN117221910A
Application number: CN202310936671.8A
Authority: CN
Inventors: 王粟; 王希; 殷旭; 王琳; 张磊
Original assignee: China Mobile Communications Group Co Ltd; China Mobile Group Fujian Co Ltd
Current assignee: China Mobile Communications Group Co Ltd; China Mobile Group Fujian Co Ltd
Priority date: 2023-07-27
Filing date: 2023-07-27
Publication date: 2023-12-12

Abstract

The disclosure provides a data processing method, device, equipment and medium for wireless network optimization, which belong to the technical field of network optimization, and the method comprises the following steps: acquiring an initial network problem work order, and preprocessing at least part of initial texts in the initial network problem work order to obtain a target network problem work order, wherein the target network problem work order comprises the following steps: the method comprises the steps of obtaining user query behavior data and user analysis behavior data, and identifying and obtaining a network anomaly identification result according to the user query behavior data, wherein the network anomaly identification result at least comprises; the network quality degradation cell determines a network problem root cause of the network quality degradation cell according to the user analysis behavior data, and determines a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the network problem root cause, so that the processing effect on the network problem work order can be effectively improved, and the suitability of the obtained wireless network optimization scheme and the network problem root cause can be effectively improved.

Description

Data processing method, device, equipment and medium for wireless network optimization

Technical Field

The disclosure relates to the technical field of network optimization, and in particular relates to a data processing method, device, equipment and medium for wireless network optimization.

Background

For the optimization of the daily network problems of the wireless network, a set of similar wireless network optimization management platform is currently distributed to network optimization specialists in the form of work orders through the platform, the network optimization specialists combine the experience of network optimization for years according to the work orders on personal work tables, perform data acquisition, data analysis and re-identification, analysis and formulation of the network problems by using various professional auxiliary tools, record the scheme and continue to flow to network first-line operators or an automatic execution system in the form of work orders for executing the scheme until the problem is solved and the work orders are closed.

In the related art, the processing effect of the wireless network optimization related data is poor, so that the network optimization scheme which is matched with the root cause of the network problem is difficult to determine.

Disclosure of Invention

The present disclosure aims to solve, at least to some extent, one of the technical problems in the related art.

Therefore, an object of the present disclosure is to provide a data processing method, apparatus, device, and medium for wireless network optimization, which can effectively improve the processing effect on a network problem work order, and can effectively improve the suitability of the obtained wireless network optimization scheme and the root cause of the network problem.

To achieve the above object, a data processing method for wireless network optimization according to an embodiment of a first aspect of the present disclosure includes:

acquiring an initial network problem work order, wherein the initial network problem work order comprises: a plurality of initial texts for describing network problems;

preprocessing at least part of the initial text in the initial network problem work order to obtain a target network problem work order, wherein the target network problem work order comprises: the type of problem;

acquiring user query behavior data and user analysis behavior data, wherein the user query behavior data is data obtained by query processing of network performance indexes related to the problem types by a user, and the user analysis behavior data is used for representing behavior data affecting decisions of user analysis network abnormality;

according to the user inquiry behavior data, identifying and obtaining a network abnormality identification result, wherein the network abnormality identification result at least comprises; a network quality degradation cell;

determining a network problem root cause of the network quality degradation cell according to the user analysis behavior data;

and determining a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the network problem root cause.

To achieve the above object, a data processing apparatus for wireless network optimization according to an embodiment of a second aspect of the present disclosure includes:

the first acquisition module is used for acquiring an initial network problem work order, wherein the initial network problem work order comprises: a plurality of initial texts for describing network problems;

the processing module is configured to pre-process at least part of the initial text in the initial network problem work order to obtain a target network problem work order, where the target network problem work order includes: the type of problem;

the second acquisition module is used for acquiring user query behavior data and user analysis behavior data, wherein the user query behavior data is data obtained by query processing of network performance indexes related to the problem types by a user, and the user analysis behavior data is used for representing behavior data affecting decision of a user for analyzing network abnormality;

the identification module is used for identifying and obtaining a network abnormality identification result according to the user inquiry behavior data, wherein the network abnormality identification result at least comprises; a network quality degradation cell;

a first determining module, configured to determine a root cause of a network problem of the network quality degradation cell according to the user analysis behavior data;

And the second determining module is used for determining a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the root cause of the network problem.

Embodiments of the third aspect of the present disclosure provide a computer device, including: a memory, a processor and a computer program stored on the memory and executable on the processor, which processor implements a data processing method for wireless network optimization as set forth in the embodiments of the first aspect of the present disclosure when the program is executed.

An embodiment of a fourth aspect of the present disclosure proposes a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a data processing method for wireless network optimization as proposed by an embodiment of the first aspect of the present disclosure.

A fifth aspect embodiment of the present disclosure proposes a computer program product which, when executed by a processor, performs a data processing method for wireless network optimization as proposed by the first aspect embodiment of the present disclosure.

The data processing method, the data processing device, the computer equipment and the storage medium for wireless network optimization are provided, and at least part of initial texts in an initial network problem work order are preprocessed to obtain a target network problem work order, wherein the target network problem work order comprises the following steps: the method comprises the steps of obtaining user query behavior data and user analysis behavior data, and identifying and obtaining a network anomaly identification result according to the user query behavior data, wherein the network anomaly identification result at least comprises; the network quality degradation cell determines a network problem root cause of the network quality degradation cell according to the user analysis behavior data, and determines a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the network problem root cause, so that the processing effect on the network problem work order can be effectively improved, and the suitability of the obtained wireless network optimization scheme and the network problem root cause can be effectively improved.

Additional aspects and advantages of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure.

Drawings

The foregoing and/or additional aspects and advantages of the present disclosure will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:

FIG. 1 is a flow chart of a data processing method for wireless network optimization according to an embodiment of the present disclosure;

FIG. 2a is a flow chart of a data processing method for wireless network optimization according to another embodiment of the present disclosure;

FIG. 2b is a key phrase extraction flow chart presented in accordance with the present disclosure;

FIG. 3a is a flow chart of a data processing method for wireless network optimization according to another embodiment of the present disclosure;

fig. 3b is a flow chart for intelligently identifying bad cells of a wireless network according to the present disclosure;

FIG. 3c is a flow chart of bad cell network problem root cause positioning according to the present disclosure;

FIG. 4a is a flow chart of a data processing method for wireless network optimization according to another embodiment of the present disclosure;

fig. 4b is a quality difference cell optimization scheme rationality evaluation flow chart presented in accordance with the present disclosure;

FIG. 4c is a fully automatic intelligent closed-loop management flow chart for a wireless network optimization policy according to the present disclosure;

FIG. 4d is a flow chart of the operation of the network optimized full-service operation module according to the present disclosure;

FIG. 5 is a schematic diagram of a data processing apparatus for wireless network optimization according to an embodiment of the present disclosure;

FIG. 6 illustrates a block diagram of an exemplary computer device suitable for use in implementing embodiments of the present disclosure.

Detailed Description

Embodiments of the present disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are exemplary only for explaining the present disclosure and are not to be construed as limiting the present disclosure. On the contrary, the embodiments of the disclosure include all alternatives, modifications, and equivalents as may be included within the spirit and scope of the appended claims.

Fig. 1 is a flow chart illustrating a data processing method for wireless network optimization according to an embodiment of the present disclosure.

It should be noted that, the execution body of the data processing method for wireless network optimization in this embodiment is a data processing apparatus for wireless network optimization, where the apparatus may be implemented in software and/or hardware, and the apparatus may be configured in a computer device, where the computer device may include, but is not limited to, a terminal, a server, etc., and the terminal may be a mobile phone, a palm computer, etc.

As shown in fig. 1, the data processing method for wireless network optimization includes:

s101: acquiring an initial network problem work order, wherein the initial network problem work order comprises: a plurality of initial texts, the initial texts are used for describing network problems.

The initial problem worksheet may refer to an unprocessed worksheet acquired in a network optimization process. The initial problem worksheet may contain relevant information of the network problem to be optimized, such as scene attribute information and fault feature information of the network problem.

In the embodiment of the present disclosure, when the initial network problem work order is acquired, a communication link between the execution body and the big data server in the embodiment of the present disclosure may be pre-established, and then the initial network problem work order is acquired from the big data server, or any other possible method may be adopted to acquire the initial network problem work order, which is not limited.

S102: preprocessing at least part of the initial text in the initial network problem work order to obtain a target network problem work order, wherein the target network problem work order comprises: question type.

The target network problem work order refers to a network problem work order obtained by preprocessing at least part of initial texts in the initial network problem work order.

It can be appreciated that the initial network problem worksheet acquired by the embodiment of the present disclosure may have a situation that the description of the problem is unclear or effective information is lacking, so that in the embodiment of the present disclosure, at least a portion of initial text in the initial network problem worksheet may be preprocessed, so as to improve the practicability of text description contents in the worksheet.

S103: and acquiring user query behavior data and user analysis behavior data, wherein the user query behavior data is data obtained by performing query processing on network performance indexes related to the problem types by a user, and the user analysis behavior data is used for representing behavior data influencing decisions of the user on analysis of network anomalies.

The user queries behavior data, which refers to data obtained by querying the network problems described by the initial network problem work order in the network optimization process of the user. For example, may include network metric data.

The user analysis behavior data refers to related data of the user in an analysis process aiming at the initial network problem work order. For example, may include abnormal network metrics determined via a preliminary analysis.

In the embodiment of the disclosure, when the user query behavior data and the user analysis behavior data are acquired, reliable data support can be provided for the subsequent analysis user network optimization processing process.

S104: according to the user inquiry behavior data, identifying to obtain a network abnormality identification result, wherein the network abnormality identification result at least comprises; network quality deteriorates cells.

The network anomaly identification result is a result obtained by carrying out network anomaly identification based on user query behavior data.

The network quality degradation cell may refer to a cell with one or more abnormal network indexes.

In the embodiment of the disclosure, when the network anomaly identification result is identified according to the user query behavior data, the user query behavior data may be input into a pre-trained network anomaly identification model to obtain the network anomaly identification result, or may be a method based on digital combination, and the network anomaly identification result is identified according to the user query behavior data, which is not limited.

S105: and determining the network problem root cause of the network quality degradation cell according to the user analysis behavior data.

The network problem root causes that a normal cell becomes a network quality degradation cell.

It can be appreciated that, based on the user query behavior data, a plurality of abnormal indexes of the network quality degradation cell may be acquired, and one or more abnormal indexes may not have a correlation with the network problem root cause, so in the embodiment of the present disclosure, the network problem root cause of the network quality degradation cell may be determined according to the user analysis behavior data, so as to provide reliable reference information for subsequent determination of the wireless network optimization scheme.

S106: and determining a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the root cause of the network problem.

The wireless network optimization scheme may be an optimization scheme suitable for solving the root cause of the network problem.

Wherein, the evaluation information can be used for evaluating the optimizing effect of the wireless network optimizing scheme on the network quality degradation cell.

That is, in the embodiment of the present disclosure, after determining the root cause of the network problem of the network quality degradation cell, the wireless network optimization scheme and the evaluation information corresponding to the wireless network optimization scheme may be determined according to the root cause of the network problem, so as to ensure the applicability of the obtained wireless network optimization scheme to the network quality degradation cell, and realize the self-evaluation of the wireless network optimization scheme.

In this embodiment, an initial network problem work order is obtained, and at least a part of initial text in the initial network problem work order is preprocessed, so as to obtain a target network problem work order, where the target network problem work order includes: the method comprises the steps of obtaining user query behavior data and user analysis behavior data, and identifying and obtaining a network anomaly identification result according to the user query behavior data, wherein the network anomaly identification result at least comprises; the network quality degradation cell determines a network problem root cause of the network quality degradation cell according to the user analysis behavior data, and determines a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the network problem root cause, so that the processing effect on the network problem work order can be effectively improved, and the suitability of the obtained wireless network optimization scheme and the network problem root cause can be effectively improved.

Fig. 2a is a flow chart illustrating a data processing method for wireless network optimization according to another embodiment of the present disclosure.

As shown in fig. 2a, the data processing method for wireless network optimization includes:

s201: acquiring an initial network problem work order, wherein the initial network problem work order comprises: a plurality of initial texts, the initial texts are used for describing network problems.

The description of S201 may be specifically referred to the above embodiments, and will not be repeated here.

S202: the stop words and/or synonyms and/or paraphrasing words and/or invalidation information contained in each initial text are identified.

The stop words may be, for example, special '_symbol,' etc. in the initial text.

Wherein, synonyms can refer to characters in the initial text that have the same descriptive meaning.

Wherein, the near meaning word can refer to characters with similar descriptive meanings in the initial text.

Wherein, the invalid information refers to characters with lower availability in the initial text.

In the embodiment of the disclosure, when identifying the stop words and/or synonyms and/or near-meaning words and/or invalid information contained in each initial text, accurate processing objects can be provided for the subsequent preprocessing of the initial text.

S203: and determining whether each initial text contains valid information to obtain a determination result.

The effective information may refer to information having a reference value to the network optimization process. For example, may refer to the type of problem described above.

The determination result may be used to indicate whether valid information is included in the initial text.

It can be appreciated that, for an initial text that includes valid information and an initial text that does not include valid information, there may be a difference in processing procedures between the two, and thus, in the embodiment of the present disclosure, when determining whether each initial text includes valid information to obtain a determination result, reliable reference information may be provided for a processing procedure of a subsequent initial text.

S204: a text similarity value between different initial texts is determined, wherein the different initial texts are initial texts containing effective information and initial texts not containing effective information.

Wherein the text similarity value may be used to describe the similarity between different initial texts.

That is, there may be similar content between the plurality of initial texts acquired by the embodiments of the present disclosure, and thus, when determining text similarity values between different initial texts, reliable reference information may be provided for subsequent processing between different initial texts.

S205: and preprocessing at least part of the initial text in the initial network problem work order according to the stop words and/or the synonyms and/or the paraphrasing and/or the invalid information, the determination result and the text similarity value to obtain the target network problem work order.

That is, in the embodiment of the present disclosure, after determining the stop word and/or the synonym and/or the paraphrasing and/or the invalidation information, the determination result, and the text similarity value, the corresponding initial text may be preprocessed according to the stop word and/or the synonym and/or the paraphrasing and/or the invalidation information, the determination result, and the text similarity value, so that the reliability of the initial text processing process is effectively ensured.

Optionally, in some embodiments, when preprocessing at least part of the initial text in the initial network problem work order according to the stop word and/or the synonym and/or the near-meaning word and/or the invalid information, the determination result and the text similarity value to obtain the target network problem work order, deleting the stop word and/or the invalid information contained in the initial text in the initial network problem work order, merging the synonym and/or the near-meaning word in the initial text in the initial network problem work order, and performing target processing on the initial text which does not contain the valid information in the initial network problem work order according to the text similarity value and the initial text which contains the valid information, where the initial network problem work order obtained by processing is used as the target network problem work order, so that the initial text initial process can be adapted to a personalized application scene, and the processing effect on the initial text can be effectively improved.

Optionally, in some embodiments, when performing the target processing on the initial text that does not include the effective information in the initial network problem work order according to the text similarity value and the initial text that includes the effective information, if the text similarity value is greater than the similarity threshold, the effective information may be filled into the initial text that does not include the effective information, if the text similarity value is less than or equal to the similarity threshold, the processing experience data of the network problem is obtained, the effective information that matches the problem type of the initial network problem work order is obtained by parsing the processing experience data, and the matched effective information is filled into the initial text that does not include the effective information, thereby ensuring that the processed initial text includes the effective information, and effectively improving the practicability of the processed initial text.

The similarity threshold value refers to a threshold value configured for text similarity values in advance.

The processing experience data may refer to historical data for optimizing network problems.

That is, after the initial network problem worksheet is acquired, the stop words and/or synonyms and/or near-meaning words and/or invalid information included in each initial text may be identified, whether each initial text includes valid information or not is determined to obtain a determination result, a text similarity value between different initial texts is determined, wherein the different initial texts are initial texts including valid information and initial texts not including valid information, and at least part of the initial texts in the initial network problem worksheet are preprocessed according to the stop words and/or synonyms and/or near-meaning words and/or invalid information, the determination result and the text similarity value to obtain the target network problem worksheet, so that corresponding feature information can be accurately identified for the different initial texts, and preprocessing is performed on the initial texts based on the feature information, so that suitability of the processing procedure to the initial texts can be effectively improved.

For example, the initial network problem worksheet in the embodiments of the present disclosure may be as shown in table 1:

TABLE 1

This section is mainly to inform network optimization engineers of network bad cell objects, network problem descriptions and network environment related basic information. For the cleaning work of the part of data, the scheme mainly extracts key information (and the effective information) of short text data of 'problem description', the extraction process is shown in fig. 2b, and fig. 2b is a key phrase extraction flow chart according to the disclosure.

For example, the scheme can pre-process short text data of a history work order (including eliminating stop words (such as ' special symbol ', ' and other auxiliary words) and synonym/near-meaning word combination (such as ' low call completion rate ' and ' wireless low call completion rate '), then extract text key phrases by using an NLP technology, after extracting the key phrases (such as ' 5G low call completion rate '), a network optimization expert eliminates invalid phrases (such as ' city B county C ', ' coal mine C-515 ', and the like) in the key phrases based on service experience, after eliminating the wireless key phrases, part of work order ' problem description ' no valid key phrases exist, and for the work order problem description of the type, the scheme utilizes cosine similarity to calculate text similarity with the history containing valid key phrases, sets a similarity threshold s, matches the work order problem description with similarity greater than s and highest similarity to obtain the valid phrases, and performs matching if the work order ' problem description ' with no valid key phrases ' is recorded, and the work order problem description ' is filled with the key phrases by the monthly service experience optimization expert periodically. Finally, filling the most recently added field of the extracted key phrase, wherein the field name is 'problem type', and finally, the basic information data of the work order is obtained as shown in table 2:

TABLE 2

The newly added problem type field (and the effective information) can provide an analysis direction for the user to locate the root cause of the network problem, and the user can quickly inquire relevant performance indexes for analysis based on own experience.

S206: and obtaining a user query log table, wherein the user query log table is used for describing information related to query processing, and the information related to query processing comprises at least one of work order identification, query time, query field name and query condition of a target network problem work order.

The work order identifier can be used for identifying the initial network problem work order. The query time may refer to the time when the user performs the query processing. The query field name may refer to a field name corresponding to the query network indicator. The query condition may be used to indicate a query range, for example, a time range, a space range, or the like, which is not limited.

For example, the user query log table may be as shown in table 3:

TABLE 3 Table 3

In the embodiment of the disclosure, when the user query log table is acquired, reliable data support can be provided for subsequent acquisition of user query behavior data.

S207: and analyzing the first content information corresponding to at least one of the work order identification, the query time, the query field name and the query condition from the user query log table, and taking the first content information as user query behavior data.

That is, in the embodiment of the present disclosure, after the user query log table is obtained, first content information corresponding to at least one of the work order identifier, the query time, the query field name, and the query condition may be parsed from the user query log table, and the first content information is used as user query behavior data, so that flexibility and reliability of the obtained user query behavior data may be effectively improved.

S208: and obtaining a user analysis behavior log table, wherein the user analysis behavior log table is used for describing analysis behavior related information, the analysis behavior related information comprises at least one of a work order identifier, operation time, an analysis field name and a target query field name, and the target query field name is a query field name corresponding to the analysis field name.

The operation time may refer to a time for performing an analysis operation. The analysis field name refers to a field name corresponding to the network index that the user analyzes. The target query field name refers to the query field name corresponding to the analysis field name.

For example, the user analysis behavior log table may be as shown in table 4:

TABLE 4 Table 4

In the embodiment of the disclosure, when the user analysis behavior log table is obtained, reliable data support can be provided for subsequent determination of the user analysis behavior data.

S209: and analyzing second content information corresponding to at least one of the work order identification, the operation time, the analysis field name and the target query field name from the user analysis behavior log table, and taking the second content information as user analysis behavior data.

In the embodiment of the disclosure, after the user analysis behavior log table is obtained, the second content information corresponding to at least one of the work order identifier, the operation time, the analysis field name and the target query field name in the user analysis behavior log table can be flexibly selected, and the second content information is used as the user analysis behavior data.

That is, in the embodiment of the present disclosure, a user query log table may be obtained, where the user query log table is used to describe information related to query processing, where the information related to query processing includes at least one of a work order identifier, a query time, a query field name, and a query condition of a target network problem work order, first content information corresponding to at least one of the work order identifier, the query time, the query field name, and the query condition is parsed from the user query log table, and the first content information is used as user query behavior data, and a user analysis behavior log table is obtained, where the user analysis behavior log table is used to describe information related to analysis behavior, where the information related to analysis behavior includes at least one of the work order identifier, the operation time, the analysis field name, and the target query field name is a query field name corresponding to the analysis field name, and second content information corresponding to at least one of the work order identifier, the operation time, the analysis field name, and the second content information is used as user analysis behavior data, thereby effectively improving user query behavior data and user analysis behavior indicating effect data.

S210: according to the user inquiry behavior data, identifying to obtain a network abnormality identification result, wherein the network abnormality identification result at least comprises; network quality deteriorates cells.

S211: and determining the network problem root cause of the network quality degradation cell according to the user analysis behavior data.

S212: and determining a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the root cause of the network problem.

The descriptions of S210-S212 may be specifically referred to the above embodiments, and are not repeated here.

In this embodiment, by identifying the stop words and/or synonyms and/or near-meaning words and/or invalid information included in each initial text, determining whether each initial text includes valid information to obtain a determination result, determining a text similarity value between different initial texts, where the different initial texts are initial texts including valid information and initial texts not including valid information, preprocessing at least part of the initial texts in the initial network problem worksheet according to the stop words and/or synonyms and/or near-meaning words and/or invalid information, the determination result and the text similarity value to obtain a target network problem worksheet, thereby accurately identifying corresponding feature information for different initial texts, preprocessing the initial texts based on the feature information, and effectively improving suitability of the processing procedure to the initial texts. Combining the synonyms and/or the paraphrasing in the initial texts in the initial network problem worksheet by deleting the stop words and/or the invalid information contained in the initial texts in the initial network problem worksheet, and performing target processing on the initial texts which do not contain the valid information in the initial network problem worksheet according to the text similarity value and the initial texts which contain the valid information, wherein the initial network problem worksheet obtained by processing is used as the target network problem worksheet, so that the initial text initial process can be adapted to the personalized application scene, and the processing effect on the initial texts can be effectively improved. If the text similarity value is larger than the similarity threshold value, the effective information is filled into the initial text which does not contain the effective information, if the text similarity value is smaller than or equal to the similarity threshold value, processing experience data of the network problem is obtained, the effective information matched with the problem type of the work order of the initial network problem is obtained through analysis from the processing experience data, and the matched effective information is filled into the initial text which does not contain the effective information, so that the processed initial text can be guaranteed to contain the effective information, and the practicability of the processed initial text can be effectively improved. The method comprises the steps of obtaining a user query log table, wherein the user query log table is used for describing information related to query processing, the information related to query processing comprises at least one of a work order identifier, query time, query field names and query conditions of a target network problem work order, analyzing first content information corresponding to at least one of the work order identifier, the query time, the query field names and the query conditions from the user query log table, taking the first content information as user query behavior data, and obtaining a user analysis behavior log table, wherein the user analysis behavior log table is used for describing information related to analysis behavior, the information related to analysis behavior comprises at least one of the work order identifier, the operation time, the analysis field names and the target query field names, the target query field names are query field names corresponding to the analysis field names, analyzing second content information corresponding to at least one of the work order identifier, the operation time, the analysis field names and the target query field names from the user analysis behavior log table, and taking the second content information as user analysis behavior data.

Fig. 3a is a flow chart illustrating a data processing method for wireless network optimization according to another embodiment of the present disclosure.

As shown in fig. 3a, the data processing method for wireless network optimization includes:

s301: acquiring an initial network problem work order, wherein the initial network problem work order comprises: a plurality of initial texts, the initial texts are used for describing network problems.

S302: preprocessing at least part of the initial text in the initial network problem work order to obtain a target network problem work order, wherein the target network problem work order comprises: question type.

S303: and acquiring user query behavior data and user analysis behavior data, wherein the user query behavior data is data obtained by performing query processing on network performance indexes related to the problem types by a user, and the user analysis behavior data is used for representing behavior data influencing decisions of the user on analysis of network anomalies.

The descriptions of S301 to S303 may be specifically referred to the above embodiments, and are not repeated herein.

S304: and identifying and obtaining a plurality of abnormal indexes according to the user query behavior data.

The anomaly index may be a network index in which an anomaly exists.

In the embodiment of the disclosure, when a plurality of abnormal indexes are identified according to the user query behavior data, a reliable analysis object can be provided for a subsequent network abnormal analysis process.

S305: and acquiring an index dynamic early warning threshold value related to each abnormal index.

The index dynamic early warning threshold value can be used for determining whether the corresponding abnormal index really has an abnormality or not.

In the embodiment of the present disclosure, when the dynamic early warning threshold of the index related to each abnormal index is obtained, the communication link between the execution body and the big data server in the embodiment of the present disclosure may be pre-established, and then the dynamic early warning threshold of the index related to each abnormal index is obtained from the big data server, or the dynamic early warning threshold of the index related to each abnormal index may be obtained by the real device based on the third party threshold, which is not limited.

Optionally, in some embodiments, when the index dynamic early warning threshold value related to each abnormal index is acquired, network state data related to each abnormal index may be acquired, and the index dynamic early warning threshold value related to the corresponding abnormal index is determined according to the network state data, so that applicability of the obtained index dynamic early warning threshold value may be effectively improved based on the network state data.

The network state data may refer to state data that is composed of abnormal indexes with different time dimensions and is used for describing the change condition of the abnormal indexes.

Optionally, in some embodiments, when determining the indicator dynamic early warning threshold related to the corresponding abnormal indicator according to the network state data, the quarter-bit number and the three-quarter-bit number of the corresponding abnormal indicator in the preset duration may be obtained by parsing from the network state data, the bit distance of the abnormal indicator is calculated according to the quarter-bit number and the three-quarter-bit number, and the indicator dynamic early warning threshold corresponding to the abnormal indicator is determined according to the quarter-bit number and the bit distance, so that the indication effect of the obtained indicator dynamic early warning threshold may be effectively improved.

The preset duration may be, for example, one week or one month, which is not limited.

Wherein the quantile may be used to describe the difference information between the quarter and the three-quarters.

S306: and selecting at least part of abnormal indexes from the plurality of abnormal indexes according to the dynamic early warning thresholds of the plurality of indexes.

That is, in the embodiment of the present disclosure, after the index dynamic early warning threshold value related to each abnormal index is obtained, the second filtering of the multiple index dynamic early warning threshold values may be implemented according to the multiple index dynamic early warning threshold values, so as to select at least a part of abnormal indexes from the multiple abnormal indexes.

S307: and determining the network quality degradation cell according to the abnormal fluctuation days corresponding to at least part of the abnormal indexes.

In the embodiment of the disclosure, when determining the network quality degradation cell according to the abnormal fluctuation days corresponding to at least part of the abnormal indexes, a cell whose abnormal fluctuation days exceed a preset number of days threshold may be used as the network quality degradation cell, or a cell whose abnormal index number is greater than a preset threshold and whose abnormal fluctuation days correspond to each abnormal index exceeds a preset number of days threshold may be used as the network quality degradation cell.

S308: and taking the abnormal index, the index dynamic early warning threshold value and the network quality degradation cell as network abnormal recognition results.

That is, in the embodiment of the present disclosure, not only the data query log of the user is recorded, and the data queried by the user is stored, but also the abnormal value recognition is performed on the index queried by the user by using the abnormal detection algorithm (such as a box graph, an isolated forest, etc.), and the abnormal result is output.

For example: for the index "wireless connection rate", by calculating the quarter-digit q1=96.98 and the three-quarter-digit q3= 99.43 of the index for nearly 30 days, and further calculating the digit distance iqr=q3-q1=2.75, the lower limit value of the index is: q1-1.5 iqr= 93.305, then less than 93.305 is the outlier.

Reminding the user of the abnormal indexes and abnormal points, and supplementing the table as shown in table 5:

TABLE 5

According to the macro in the embodiment of the disclosure, based on the user query log data, the AI algorithm can be utilized to perform daily monitoring on the user query index, identify the index degradation cell and perform early warning, timely find the suspected problem cell, intelligently identify the wireless network poor quality cell, and realize network perception intellectualization without a work order trigger. The specific implementation process is shown in fig. 3b, and fig. 3b is a flowchart for intelligently identifying a poor quality cell of a wireless network according to the present disclosure.

According to the method, a daily monitoring index set is extracted through inquiring and learning user data, abnormal points are identified and removed through abnormal detection algorithms such as box line graphs and isolated forests, then index dynamic early warning thresholds are calculated through algorithms such as clustering and time sequence prediction, abnormal fluctuation index sets are identified, and network quality degradation cells are further identified according to abnormal fluctuation days of indexes of approximately 7 days.

For example: further calculating the quantile iqr=q3-q1=2.75 by calculating the quantile q1=96.98 and the quantile q3= 99.43 for 30 days of "wireless connectivity", the lower limit value of the index is: q1-1.5 iqr= 93.305, then less than 93.305 is the outlier. After the abnormal points are removed, a data set with index values higher than 93.305 is obtained, a density clustering algorithm is adopted for the partial data, and parameters (E=sigma, minPts=n/2) are set to describe the sample distribution compactness of the neighborhood. Wherein, e describes a neighborhood distance threshold of a certain sample, minPts describes a threshold of the number of samples in a neighborhood of a certain sample with a distance e, σ represents a standard deviation of data, n represents the number of samples, a minimum value of a maximum cluster of classification is used as a dynamic early warning threshold of an index (for example, the calculated dynamic threshold is 95.27), then the data of the index in seven days is compared with the value, a mark smaller than the value is abnormal, and abnormal days of the last 7 days are counted.

That is, in the embodiment of the disclosure, after the user query behavior data is obtained, a plurality of abnormal indexes can be identified according to the user query behavior data, an index dynamic early warning threshold related to each abnormal index is obtained, at least part of abnormal indexes are selected from the plurality of abnormal indexes according to the plurality of index dynamic early warning thresholds, a network quality degradation cell is determined according to the number of abnormal fluctuation days corresponding to at least part of abnormal indexes, the index dynamic early warning threshold and the network quality degradation cell are used as network abnormal recognition results, so that the reliability of the network abnormal recognition result obtaining process can be effectively improved, and the indicating effect of the obtained network abnormal recognition results can be ensured.

S309: initial index data corresponding to the target query field name is obtained.

The initial index data refers to unprocessed index data corresponding to the target query field name.

In the embodiment of the disclosure, when the initial index data corresponding to the target query field name is acquired, reliable data support can be provided for subsequent determination of the root cause of the network problem.

S310: and carrying out characteristic engineering treatment on the initial index data to obtain target index data.

In the embodiment of the disclosure, when the characteristic engineering processing is performed on the initial index data, the screening processing of the initial index data can be realized, so that the practicability of the obtained target index data is ensured.

For example, a network engineer may obtain the root cause of a network problem by uploading a queried and downloaded data set, preprocessing the data set (including data type conversion, missing value processing, summing, computation between columns, etc.), and analyzing the data set. For the module, the present disclosure mainly designs the content output finally for data analysis, and the user can perform row and column sorting, and sequentially sort out rows/columns or a specific cell which has an influence on the analysis decision. And extracting the field of the initial query table corresponding to the field of the final output data by recording the final output data of the user and sequentially checking the row/column data which has influence on the analysis decision of the final output data and analyzing the SQL of the background data. Meanwhile, for the index screening part, the method automatically selects important features through a feature engineering technology to perform preliminary screening on data for users, a specific flow is shown in fig. 3c, and fig. 3c is a flow chart of the root cause positioning of the bad cell network problem according to the method. The method comprises the steps of analyzing a behavior log based on collected user data, extracting indexes related to user data analysis, preprocessing the data by using a machine learning method, extracting important characteristic variables, then combining actual problem root causes of a quality difference cell, constructing quality difference cell network problem root cause label data, and finally constructing a quality difference root cause positioning model by using a classification algorithm, and intelligently judging the quality difference cell network problem root cause.

For example, for a user to query an index data set (such as RRC connection average number, downlink PRB average utilization, wireless connection rate, uplink and downlink traffic, coverage rate, etc.), an index with a deletion rate exceeding 50% is removed, a row with a small amount of missing data is removed, the data set is normalized, important feature variables are screened by calculating information gain, a network problem root cause determined by historical field investigation is used as a classification label variable, a bad cell root cause positioning model is constructed by adopting a lightGBM classification algorithm, and a bad cell network root cause is output.

S311: and inputting target index data into a pre-trained quality difference cell root cause positioning model, and obtaining a network problem root cause of a network quality degradation cell output by the quality difference cell root cause positioning model.

That is, in the embodiment of the present disclosure, a bad cell root cause positioning model may be trained in advance, for processing target index data, to obtain a network problem root cause of a network quality degradation cell.

Optionally, in some embodiments, the quality difference cell root cause positioning model is trained based on the following manner: constructing an initial classification model, and acquiring a plurality of sample data, wherein the sample data comprises: sample index data and problem root cause label data corresponding to the sample index data, and training an initial classification model based on the sample index data and the problem root cause label data to obtain a quality difference cell root cause positioning model, so that the output accuracy of the obtained quality difference cell root cause positioning model can be effectively improved.

The initial classification model may be, for example, a lightGBM classification model, or may be any other type of model, which is not limited thereto.

The sample index data may be abnormal index data in the history network optimization process.

The problem root cause label data may refer to a problem root cause determined for sample index data in a history network optimization process.

That is, in the embodiment of the disclosure, after the user analysis behavior data is obtained, the initial index data corresponding to the target query field name may be obtained, the feature engineering processing may be performed on the initial index data, so as to obtain the target index data, the target index data may be input into the pre-trained quality difference cell root cause positioning model, and the network problem root cause of the network quality degradation cell output by the quality difference cell root cause positioning model may be obtained, so that the network problem root cause of the network quality degradation cell may be rapidly and accurately determined based on the quality difference cell root cause positioning model.

S312: and determining a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the root cause of the network problem.

The description of S312 may be specifically referred to the above embodiments, and will not be repeated here.

In this embodiment, by acquiring the network status data related to each abnormal indicator, and determining the indicator dynamic early warning threshold related to the corresponding abnormal indicator according to the network status data, the applicability of the obtained indicator dynamic early warning threshold can be effectively improved based on the network status data. The method comprises the steps of analyzing the network state data to obtain the quarter-bit number and the three-quarter-bit number of the corresponding abnormal index in the preset time period, calculating the bit distance of the abnormal index according to the quarter-bit number and the three-quarter-bit number, and determining the index dynamic early warning threshold corresponding to the abnormal index according to the quarter-bit number and the bit distance, so that the indication effect of the obtained index dynamic early warning threshold can be effectively improved. According to the method, a plurality of abnormal indexes are obtained through identification according to user query behavior data, an index dynamic early warning threshold value related to each abnormal index is obtained, at least part of abnormal indexes are selected from the plurality of abnormal indexes according to the plurality of index dynamic early warning thresholds, a network quality degradation cell is determined according to abnormal fluctuation days corresponding to at least part of abnormal indexes, the index dynamic early warning thresholds and the network quality degradation cell are used as network abnormal recognition results, and therefore reliability of the network abnormal recognition result obtaining process can be effectively improved, and indication effect of the obtained network abnormal recognition results can be guaranteed. Obtaining a plurality of sample data by constructing an initial classification model, wherein the sample data comprises: sample index data and problem root cause label data corresponding to the sample index data, and training an initial classification model based on the sample index data and the problem root cause label data to obtain a quality difference cell root cause positioning model, so that the output accuracy of the obtained quality difference cell root cause positioning model can be effectively improved. The initial index data corresponding to the target query field name is obtained, characteristic engineering processing is carried out on the initial index data, target index data are obtained, the target index data are input into a pre-trained quality difference cell root cause positioning model, and the network problem root cause of the network quality degradation cell output by the quality difference cell root cause positioning model is obtained, so that the network problem root cause of the network quality degradation cell can be rapidly and accurately determined based on the quality difference cell root cause positioning model.

Fig. 4a is a flow chart illustrating a data processing method for wireless network optimization according to another embodiment of the present disclosure.

As shown in fig. 4a, the data processing method for wireless network optimization includes:

s401: acquiring an initial network problem work order, wherein the initial network problem work order comprises: a plurality of initial texts, the initial texts are used for describing network problems.

S402: preprocessing at least part of the initial text in the initial network problem work order to obtain a target network problem work order, wherein the target network problem work order comprises: question type.

S403: and acquiring user query behavior data and user analysis behavior data, wherein the user query behavior data is data obtained by performing query processing on network performance indexes related to the problem types by a user, and the user analysis behavior data is used for representing behavior data influencing decisions of the user on analysis of network anomalies.

S404: according to the user inquiry behavior data, identifying to obtain a network abnormality identification result, wherein the network abnormality identification result at least comprises; network quality deteriorates cells.

S405: and determining the network problem root cause of the network quality degradation cell according to the user analysis behavior data.

The descriptions of S401 to S405 may be specifically referred to the above embodiments, and are not repeated herein.

S406: and determining a reference problem root cause matched with the network problem root cause from a preset relation table, and taking a reference network optimization scheme corresponding to the matched reference problem root cause as a wireless network optimization scheme, wherein the preset relation table comprises a plurality of reference problem root causes and the reference network optimization scheme corresponding to the reference problem root cause.

The preset relation table is configured in advance for the mapping relation between the reference problem root cause and the reference network optimization scheme.

For example, in the embodiment of the disclosure, after a network optimization engineer locates the root cause of a network problem, it may query relevant parameters of a cell or a base station according to its own experience, and formulate an adjustment policy for the relevant parameters to optimize the network quality. Aiming at the behavior formulated by the user proposal, the present disclosure designs the operation log data acquisition tables of the module according to two applications of 0-phase parameter query and parameter adjustment respectively as shown in table 6 and table 7, wherein table 6 is a user parameter query behavior log table, and table 7 is a user parameter adjustment behavior log table.

TABLE 6

TABLE 7

The reference problem root cause and the reference network optimization scheme refer to a network problem root cause which is configured in a preset relation table in advance to serve as a reference, and a network problem root cause optimization scheme suitable for the reference problem root cause.

That is, in the embodiment of the present disclosure, a preset relationship table may be preconfigured to determine a wireless network optimization scheme adapted to a root cause of a network problem, so as to effectively improve the acquisition efficiency of the wireless network optimization scheme and ensure the reliability of the obtained wireless network optimization scheme.

S407: and adjusting network parameters of the network quality degradation cell based on the wireless network optimization scheme at the target time point.

The target time point may refer to a time point when the wireless network optimization scheme is executed.

In the embodiment of the disclosure, when the network parameter adjustment is performed on the network quality degradation cell based on the wireless network optimization scheme at the target time point, the network quality degradation cell can be optimized in time, and the service quality of the cell is improved.

Optionally, in some embodiments, the reference network optimization scheme includes a plurality of candidate network optimization schemes, each candidate network optimization scheme has a corresponding priority value, when the network parameter adjustment is performed on the network quality degradation cell based on the wireless network optimization scheme at the target time point, the target network optimization scheme may be determined from the plurality of candidate network optimization schemes according to the priority value, and the network parameter adjustment is performed on the network quality degradation cell based on the target network optimization scheme at the target time point, thereby effectively improving the reliability of the obtained target network optimization scheme based on the priority value.

The candidate network optimization scheme refers to an optimization scheme which is contained in the reference network optimization scheme and is suitable for a network quality degradation cell.

In the embodiment of the disclosure, the priority value may be used to indicate the execution priority information corresponding to each candidate network optimization scheme, where the priority value may be determined according to the optimization effect corresponding to each candidate network optimization scheme.

For example, after a network optimization engineer formulates a network optimization policy, the present disclosure may automatically execute a network adjustment scheme and feed back an execution result, and the data table design of the execution situation of the optimization scheme collected by the present disclosure is shown in table 8:

TABLE 8

S408: first index data of a network quality degradation cell before a target time point and second index data after the target time point are acquired.

The first index data may be, for example, average index data 3 days before the target time point, or index data at any time before the target time point, which is not limited.

The second index data may be, for example, average index data 3 days after the target time point, or index data at any time after the target time point, which is not limited.

In the embodiment of the disclosure, the first index data and the second index data of the network quality degradation cell before and after the target time point are acquired.

S409: and determining evaluation information according to the first index data and the second index data.

In the embodiment of the disclosure, when determining the evaluation information according to the first index data and the second index data, the first index data and the second index data may be input into a pre-trained evaluation model to determine the evaluation information, or the evaluation information may be determined according to the first index data and the second index data based on a data shape combination method, which is not limited.

For example, in the embodiment of the disclosure, a network optimization engineer evaluates an optimization scheme, and an optimization scheme evaluation conclusion is made by querying the change condition of the optimization target index before and after optimization. The optimization scheme evaluation log table design for the collection of the present disclosure is shown in table 9:

TABLE 9

Optionally, in some embodiments, if the evaluation information does not meet the preset condition, updating a priority value of each candidate network optimization scheme, and determining a new target network optimization scheme, so that the new target network optimization scheme can be determined in time when the evaluation information does not meet the preset condition, so as to improve robustness of the optimization process.

Optionally, in some embodiments, when determining the evaluation information according to the first index data and the second index data, a difference value of at least one index to be optimized may be determined according to the first index data and the second index data, a weight value of each index to be optimized may be determined, a first index improvement score of the network quality degradation cell after optimization may be obtained by calculating according to the difference value and the weight value, a second index improvement score of a neighboring cell of the network quality degradation cell after optimization of the network quality degradation cell may be obtained, and the first index improvement score and the second index improvement score may be input into a pre-trained optimization scheme evaluation model to obtain evaluation information output by the optimization scheme evaluation model, thereby, in a process of determining the evaluation information, an index change condition of the neighboring cell may be combined, and the comprehensiveness of the obtained evaluation information may be effectively improved, so as to avoid a situation of causing neighbor cell index degradation when optimizing the network quality of the local cell.

The difference value of the index to be optimized can be used for describing the difference condition of the index to be optimized before and after optimization. The weight value can be used for indicating the influence degree information of each index to be optimized on the evaluation information.

Wherein the indicator improvement score may be used to indicate an improvement in the network after optimization.

That is, in the embodiment of the present disclosure, after determining the network problem root cause of the network quality degradation cell according to the user analysis behavior data, the reference problem root cause that is matched with the network problem root cause may be determined from the preset relationship table, and the reference network optimization scheme corresponding to the matched reference problem root cause is used as the wireless network optimization scheme, where the preset relationship table includes a plurality of reference problem root causes and the reference network optimization scheme corresponding to the reference problem root cause, the network parameter adjustment is performed on the network quality degradation cell based on the wireless network optimization scheme at the target time point, the first index data of the network quality degradation cell before the target time point and the second index data after the target time point are obtained, and the evaluation information is determined according to the first index data and the second index data, thereby, the wireless network optimization scheme may be rapidly and accurately determined based on the preset relationship table, and the self-evaluation of the wireless network optimization scheme may be implemented, and the robustness of the network optimization may be improved.

According to the method, the system and the device, the optimization evaluation indexes are further extracted, in order to fully evaluate the network quality, other performance indexes are added and the indexes corresponding to the adjacent cells are acquired, corresponding weights are set according to the importance degree of the evaluation indexes, then the weighted TOPSIS algorithm is utilized to calculate the comprehensive score of the network indexes of the optimization target cell and the comprehensive score of the network performance indexes of the adjacent cells, and then the historical evaluation result data of the optimization scheme is combined, and the intelligent evaluation model of the optimization scheme is built by utilizing classification algorithms such as lightgbm and the like, so that the self evaluation of the network optimization scheme is realized. The specific implementation process is shown in fig. 4b, and fig. 4b is a rationality evaluation flow chart of the bad cell optimization scheme proposed according to the present disclosure.

For example: and acquiring index average values of the optimized cells before and after optimization for 3 days, wherein the index average values of the optimized cells after optimization for 3 days are as follows, the index average values of the cells before optimization for 460-00-XXXXX-XX are as follows, the radio connection rate average value of the cells before optimization for 3 days is 93.41, the switching success rate is 92.88, the radio connection drop rate is 0.42, the RRC average connection number is 6.87, the radio connection rate is 99.67, the switching success rate is 99.02, the radio connection drop rate is 0.01, and the RRC average connection number is 15.27. The method comprises the steps of calculating difference values before and after optimization of each index, defining the types of the indexes, such as a wireless connection rate, a switching success rate and an RRC average connection number as maximum indexes, wherein the wireless disconnection rate is minimum index, giving weights of different indexes according to service experience, such as the wireless connection rate is 0.3, the switching success rate is 0.2, the RRC average connection number is 0.2, the wireless disconnection rate is 0.3, calculating an index improvement degree comprehensive score after optimization of the cell by adopting a weighted TOPSIS algorithm to be 0.982, and similarly, collecting average values of 3 days before and after optimization of the neighbor cell and calculating an index improvement degree comprehensive score before and after optimization of the neighbor cell to be 0.834. And then, constructing a classification label by combining with a historical quality difference cell optimization evaluation result (pass/fail), and constructing an intelligent evaluation model of the optimization scheme by adopting a lightgbm class classification algorithm to realize the self-evaluation of the network optimization scheme.

In this embodiment, by determining the target network optimization scheme from the plurality of candidate network optimization schemes according to the priority value, the network parameter adjustment is performed on the network quality degradation cell based on the target network optimization scheme at the target time point, so that the reliability of the obtained target network optimization scheme can be effectively improved based on the priority value. If the evaluation information does not meet the preset conditions, updating the priority value of each candidate network optimization scheme, and determining a new target network optimization scheme, so that the new target network optimization scheme can be determined in time when the evaluation information does not meet the preset conditions, and the robustness of the optimization process is improved. The method comprises the steps of determining a difference value of at least one index to be optimized according to first index data and second index data, determining a weight value of each index to be optimized, calculating to obtain a first index improvement score of a network quality degradation cell after optimization according to the difference value and the weight value, obtaining a second index improvement score of a neighboring cell of the network quality degradation cell after network quality degradation cell optimization, and inputting the first index improvement score and the second index improvement score into a pre-trained optimization scheme evaluation model to obtain evaluation information output by the optimization scheme evaluation model, wherein the change condition of the index of the neighboring cell can be combined in the process of determining the evaluation information, and the comprehensiveness of the obtained evaluation information can be effectively improved to avoid the condition of neighbor cell index degradation when the network quality of the cell is optimized. Determining a reference problem root cause matched with a network problem root cause from a preset relation table, and taking a reference network optimization scheme corresponding to the matched reference problem root cause as a wireless network optimization scheme, wherein the preset relation table comprises a plurality of reference problem root causes and the reference network optimization scheme corresponding to the reference problem root cause, performing network parameter adjustment on a network quality degradation cell based on the wireless network optimization scheme at a target time point, acquiring first index data of the network quality degradation cell before the target time point and second index data after the target time point, and determining evaluation information according to the first index data and the second index data, thereby quickly and accurately determining the wireless network optimization scheme based on the preset relation table, realizing self-evaluation of the wireless network optimization scheme, and improving the robustness of the network optimization.

In the embodiment of the disclosure, after the multiple tables are obtained, the optimization worksheet basic information application table, the user query log application table, the user data analysis behavior log table, the user parameter adjustment execution condition result table and the user evaluation optimization scheme behavior log table may be associated. And obtaining log data of the whole-flow operation behaviors of a network engineer for analysis, scheme formulation, evaluation and the like aiming at the network problems. The table design is shown in table 10:

table 10

The method and the system learn different network environments (including network types, home areas, coverage scenes, coverage types, equipment manufacturers, problem types and the like) by utilizing an AI algorithm based on a network optimization behavior log table, and learn the optimal network optimization schemes of various network problem root causes according to the network optimization schemes and the effect scores of the optimization methods under the positioned network root causes. The wireless network optimization behavior data acquisition method can intelligently sense network quality, intelligently locate network problem reasons, intelligently formulate a network optimization scheme, monitor the network optimization scheme in real time and intelligently evaluate the network optimization scheme, and change manual processing to update each task node model continuously by a sample with the latest data of each full-service operation process when the output optimal scheme cannot meet the network optimization target, so that full-automatic intelligent closed-loop management of the wireless network optimization engineering bill is achieved, a specific implementation process is shown in fig. 4c, and fig. 4c is a full-automatic intelligent closed-loop management flow chart of the wireless network optimization engineering bill according to the disclosure.

For example, for a network type of 5G, a home area of urban area, a coverage scene of a bottom resident, a coverage type of indoor, a device manufacturer of manufacturer D, a problem type of coverage problem, the network problem root is a weak coverage, and the historical optimization scheme includes: the HW5G minimum receiving level value is adjusted to be 60 from 80, the maximum power is adjusted to be 210 from 130, the electronic downtilt angle is adjusted to be 13 from 9, and the like. Based on the improvement scores of indexes such as coverage rate before and after historical optimization, searching an optimal optimization scheme under the network environment through a reinforcement learning algorithm: the HW5G minimum received level value is adjusted from-80 to-60. The network performance index is monitored every day to find that the coverage rate index of the 5G cell 460-00-xxxxx-xx in the residential compartment at the bottom layer of a certain urban area is abnormal, the equipment manufacturer is manufacturer D, the network root is further positioned because of weak coverage, the lowest receiving level value of the optimal optimizing scheme HW5G for recommending history learning is adjusted to be 60 from 80, the system automatically executes the optimizing scheme according to the recommended scheme and automatically evaluates after 3 days of executing the scheme, the evaluation is finished, and if the evaluation is finished, the TOP2 scheme is executed, until the TOP3 scheme does not pass the evaluation and then is manually processed. And finally, expanding the manual processing process data into a new learning sample.

In summary, as shown in fig. 4d, fig. 4d is a flowchart illustrating the operation of the network optimization full-service operation module according to the present disclosure, wherein specific operation steps of each model may be referred to the description of the above embodiments, and are not repeated herein.

Fig. 5 is a schematic structural diagram of a data processing apparatus for wireless network optimization according to an embodiment of the present disclosure.

As shown in fig. 5, the data processing apparatus 50 for wireless network optimization includes:

a first obtaining module 501, configured to obtain an initial network problem worksheet, where the initial network problem worksheet includes: a plurality of initial texts, wherein the initial texts are used for describing network problems;

the processing module 502 is configured to pre-process at least a portion of the initial text in the initial network problem work order to obtain a target network problem work order, where the target network problem work order includes: the type of problem;

a second obtaining module 503, configured to obtain user query behavior data and user analysis behavior data, where the user query behavior data is data obtained by performing query processing on network performance indicators related to a problem type by a user, and the user analysis behavior data is behavior data that is used to represent a decision that affects a user analysis network anomaly;

The identifying module 504 is configured to identify and obtain a network anomaly identification result according to the user query behavior data, where the network anomaly identification result at least includes; a network quality degradation cell;

a first determining module 505, configured to determine a root cause of a network problem of a network quality degradation cell according to user analysis behavior data;

the second determining module 506 is configured to determine a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the root cause of the network problem.

In some embodiments of the present disclosure, the processing module 502 is specifically configured to:

identifying stop words and/or synonyms and/or paraphrasing words and/or invalid information contained in each initial text;

determining whether each initial text contains effective information so as to obtain a determination result;

determining text similarity values between different initial texts, wherein the different initial texts are initial texts containing effective information and initial texts not containing effective information;

and preprocessing at least part of the initial text in the initial network problem work order according to the stop words and/or the synonyms and/or the paraphrasing and/or the invalid information, the determination result and the text similarity value to obtain the target network problem work order.

In some embodiments of the present disclosure, the processing module 502 is further configured to:

deleting the stop words and/or invalid information contained in the initial text in the initial network problem work order;

combining synonyms and/or near-meaning words in an initial text in an initial network problem work order;

and performing target processing on the initial text which does not contain the effective information in the initial network problem work order according to the text similarity value and the initial text which contains the effective information, wherein the initial network problem work order obtained by processing is used as a target network problem work order.

if the text similarity value is larger than the similarity threshold value, filling the effective information into an initial text which does not contain the effective information;

and if the text similarity value is smaller than or equal to the similarity threshold value, acquiring processing experience data of the network problem, analyzing from the processing experience data to obtain effective information matched with the problem type of the initial network problem work order, and filling the matched effective information into the initial text which does not contain the effective information.

In some embodiments of the present disclosure, the second obtaining module 503 is specifically configured to:

acquiring a user query log table, wherein the user query log table is used for describing information related to query processing, and the information related to query processing comprises at least one of a work order identifier, query time, query field name and query condition of a target network problem work order;

Analyzing first content information corresponding to at least one of a work order identifier, query time, a query field name and a query condition from a user query log table, and taking the first content information as user query behavior data;

the method comprises the steps of obtaining a user analysis behavior log table, wherein the user analysis behavior log table is used for describing analysis behavior related information, the analysis behavior related information comprises at least one of a work order identifier, operation time, an analysis field name and a target query field name, and the target query field name is a query field name corresponding to the analysis field name;

and analyzing second content information corresponding to at least one of the work order identification, the operation time, the analysis field name and the target query field name from the user analysis behavior log table, and taking the second content information as user analysis behavior data.

In some embodiments of the present disclosure, the identification module 504 is specifically configured to:

identifying and obtaining a plurality of abnormal indexes according to the user query behavior data;

acquiring an index dynamic early warning threshold value related to each abnormal index;

selecting at least part of abnormal indexes from the plurality of abnormal indexes according to the dynamic early warning thresholds of the plurality of indexes;

determining a network quality degradation cell according to the abnormal fluctuation days corresponding to at least part of the abnormal indexes;

And taking the abnormal index, the index dynamic early warning threshold value and the network quality degradation cell as network abnormal recognition results.

In some embodiments of the present disclosure, the identification module 504 is further configured to:

acquiring network state data related to each anomaly index;

and determining an index dynamic early warning threshold value related to the corresponding abnormal index according to the network state data.

analyzing the network state data to obtain a quarter-digit number and a three-quarter-digit number of the corresponding abnormal index in the preset time length;

calculating the bit distance of the abnormality index according to the quarter bit number and the three-quarter bit number;

and determining an index dynamic early warning threshold corresponding to the abnormal index according to the quarter fraction number and the fraction distance.

In some embodiments of the present disclosure, the first determining module 505 is specifically configured to:

acquiring initial index data corresponding to a target query field name;

carrying out characteristic engineering treatment on the initial index data to obtain target index data;

and inputting target index data into a pre-trained quality difference cell root cause positioning model, and obtaining a network problem root cause of a network quality degradation cell output by the quality difference cell root cause positioning model.

In some embodiments of the present disclosure, the bad cell root cause positioning model is trained based on:

constructing an initial classification model;

obtaining a plurality of sample data, wherein the sample data comprises: sample index data, and problem root cause tag data corresponding to the sample index data;

and training an initial classification model based on the sample index data and the problem root cause label data to obtain a quality difference cell root cause positioning model.

In some embodiments of the present disclosure, the second determining module 506 is specifically configured to:

determining a reference problem root cause matched with a network problem root cause from a preset relation table, and taking a reference network optimization scheme corresponding to the matched reference problem root cause as a wireless network optimization scheme, wherein the preset relation table comprises a plurality of reference problem root causes and the reference network optimization scheme corresponding to the reference problem root cause;

network parameter adjustment is carried out on the network quality degradation cell based on the wireless network optimization scheme at the target time point;

acquiring first index data of a network quality degradation cell before a target time point and second index data of the network quality degradation cell after the target time point;

and determining evaluation information according to the first index data and the second index data.

In some embodiments of the present disclosure, the reference network optimization scheme includes a plurality of candidate network optimization schemes, each candidate network optimization scheme having a corresponding priority value;

wherein, the second determining module 506 is further configured to:

determining a target network optimization scheme from a plurality of candidate network optimization schemes according to the priority value;

and adjusting network parameters of the network quality degradation cell based on the target network optimization scheme at the target time point.

In some embodiments of the present disclosure, the second determining module 506 is further configured to:

and if the evaluation information does not meet the preset conditions, updating the priority value of each candidate network optimization scheme, and determining a new target network optimization scheme.

determining a difference value of at least one index to be optimized according to the first index data and the second index data;

determining a weight value of each index to be optimized;

according to the difference value and the weight value, calculating to obtain a first index improvement score of the network quality degradation cell after optimization;

acquiring a second index improvement score of a neighboring cell of the network quality degradation cell after the network quality degradation cell is optimized;

And inputting the first index improvement score and the second index improvement score into a pre-trained optimization scheme evaluation model to obtain evaluation information output by the optimization scheme evaluation model.

It should be noted that the foregoing explanation of the data processing method for wireless network optimization is also applicable to the data processing apparatus for wireless network optimization of the present embodiment, and will not be repeated here.

FIG. 6 illustrates a block diagram of an exemplary computer device suitable for use in implementing embodiments of the present disclosure. The computer device 12 shown in fig. 6 is merely an example and should not be construed as limiting the functionality and scope of use of the disclosed embodiments.

As shown in FIG. 6, the computer device 12 is in the form of a general purpose computing device. Components of computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, a bus 18 that connects the various system components, including the system memory 28 and the processing units 16.

Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, and a local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include industry Standard architecture (Industry Standard Architecture; hereinafter ISA) bus, micro channel architecture (Micro Channel Architecture; hereinafter MAC) bus, enhanced ISA bus, video electronics standards Association (Video Electronics Standards Association; hereinafter VESA) local bus, and peripheral component interconnect (Peripheral Component Interconnection; hereinafter PCI) bus.

Computer device 12 typically includes a variety of computer system readable media. Such media can be any available media that is accessible by computer device 12 and includes both volatile and nonvolatile media, removable and non-removable media.

Memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (Random Access Memory; hereinafter: RAM) 30 and/or cache memory 32. The computer device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from or write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, commonly referred to as a "hard disk drive").

Although not shown in fig. 6, a magnetic disk drive for reading from and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from or writing to a removable non-volatile optical disk (e.g., a compact disk read only memory (Compact Disc Read Only Memory; hereinafter CD-ROM), digital versatile read only optical disk (Digital Video Disc Read Only Memory; hereinafter DVD-ROM), or other optical media) may be provided. In such cases, each drive may be coupled to bus 18 through one or more data medium interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to carry out the functions of the various embodiments of the disclosure.

A program/utility 40 having a set (at least one) of program modules 42 may be stored in, for example, memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment. Program modules 42 generally perform the functions and/or methods in the embodiments described in this disclosure.

The computer device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), one or more devices that enable a person to interact with the computer device 12, and/or any devices (e.g., network card, modem, etc.) that enable the computer device 12 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 22. Moreover, the computer device 12 may also communicate with one or more networks such as a local area network (Local Area Network; hereinafter LAN), a wide area network (Wide Area Network; hereinafter WAN) and/or a public network such as the Internet via the network adapter 20. As shown, network adapter 20 communicates with other modules of computer device 12 via bus 18. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with computer device 12, including, but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.

The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, for example, implementing the data processing method for wireless network optimization mentioned in the foregoing embodiment.

To achieve the above-described embodiments, the present disclosure also proposes a non-transitory computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a data processing method for wireless network optimization as proposed in the foregoing embodiments of the present disclosure.

To achieve the above embodiments, the present disclosure also proposes a computer program product which, when executed by an instruction processor in the computer program product, performs a data processing method for wireless network optimization as proposed by the previous embodiments of the present disclosure.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This disclosure is intended to cover any adaptations, uses, or adaptations of the disclosure following the general principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It is to be understood that the present disclosure is not limited to the precise arrangements and instrumentalities shown in the drawings, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

It should be noted that in the description of the present disclosure, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Furthermore, in the description of the present disclosure, unless otherwise indicated, the meaning of "a plurality" is two or more.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and further implementations are included within the scope of the preferred embodiment of the present disclosure in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the embodiments of the present disclosure.

It should be understood that portions of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.

Those of ordinary skill in the art will appreciate that all or a portion of the steps carried out in the method of the above-described embodiments may be implemented by a program to instruct related hardware, where the program may be stored in a computer readable storage medium, and where the program, when executed, includes one or a combination of the steps of the method embodiments.

Furthermore, each functional unit in the embodiments of the present disclosure may be integrated in one processing module, or each unit may exist alone physically, or two or more units may be integrated in one module. The integrated modules may be implemented in hardware or in software functional modules. The integrated modules may also be stored in a computer readable storage medium if implemented in the form of software functional modules and sold or used as a stand-alone product.

The above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, or the like.

In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present disclosure. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Although embodiments of the present disclosure have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the present disclosure, and that variations, modifications, alternatives, and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the present disclosure.

Claims

1. A data processing method for wireless network optimization, the method comprising:

2. The method of claim 1, wherein preprocessing at least a portion of the initial text in the initial network problem ticket to obtain a target network problem ticket comprises:

determining whether each initial text contains effective information or not to obtain a determination result;

determining a text similarity value between different initial texts, wherein the different initial texts are initial texts containing effective information and initial texts not containing effective information;

and preprocessing at least part of the initial texts in the initial network problem work order according to the stop words and/or synonyms and/or near-meaning words and/or invalid information, the determination result and the text similarity value to obtain the target network problem work order.

3. The method according to claim 2, wherein preprocessing at least part of the initial text in the initial network problem work order according to the stop word and/or synonym and/or paraphrasing and/or invalidation information, the determination result and the text similarity value to obtain the target network problem work order comprises:

Deleting the stop words and/or the invalid information contained in the initial text in the initial network problem work order;

merging the synonyms and/or the paraphrasing in the initial text in the initial network question work order;

and performing target processing on the initial text which does not contain effective information in the initial network problem work order according to the text similarity value and the initial text which contains the effective information, wherein the initial network problem work order obtained by processing is used as the target network problem work order.

4. The method of claim 3, wherein said targeting initial text that does not include valid information in said initial network question work order based on said text similarity value and said initial text that includes valid information comprises:

if the text similarity value is larger than a similarity threshold value, filling the effective information into the initial text which does not contain the effective information;

5. The method of claim 1, wherein the obtaining user query behavior data and user analysis behavior data comprises:

acquiring a user query log table, wherein the user query log table is used for describing information related to query processing, and the information related to query processing comprises at least one of a work order identifier, query time, query field name and query condition of the target network problem work order;

analyzing first content information corresponding to at least one of the work order identifier, the query time, the query field name and the query condition from the user query log table, and taking the first content information as the user query behavior data;

acquiring a user analysis behavior log table, wherein the user analysis behavior log table is used for describing analysis behavior related information, the analysis behavior related information comprises at least one of a work order identifier, operation time, an analysis field name and a target query field name, and the target query field name is the query field name corresponding to the analysis field name;

and analyzing second content information corresponding to at least one of the work order identification, the operation time, the analysis field name and the target query field name from the user analysis behavior log table, and taking the second content information as the user analysis behavior data.

6. The method of claim 1, wherein the identifying network anomaly identification results based on the user query behavior data comprises:

selecting at least part of abnormal indexes from the plurality of abnormal indexes according to a plurality of index dynamic early warning thresholds;

and taking the abnormal index, the index dynamic early warning threshold value and the network quality degradation cell as the network abnormal recognition result.

7. The method of claim 6, wherein said obtaining an indicator dynamic pre-warning threshold associated with each of said anomaly indicators comprises:

acquiring network state data related to each abnormal index;

8. The method of claim 7, wherein determining an indicator dynamic pre-warning threshold associated with the respective anomaly indicator based on the network status data comprises:

Analyzing the network state data to obtain a quarter-bit number and a three-quarter-bit number of the corresponding abnormal index in a preset duration;

calculating the bit distance of the abnormal index according to the quarter bit number and the three-quarter bit number;

and determining an index dynamic early warning threshold corresponding to the abnormal index according to the quarter quantile and the quantile distance.

9. The method of claim 5, wherein said determining a network problem root cause for the network quality degradation cell based on the user analysis behavior data comprises:

acquiring initial index data corresponding to the target query field name;

performing characteristic engineering processing on the initial index data to obtain target index data;

and inputting the target index data into a pre-trained quality difference cell root cause positioning model, and obtaining the network problem root cause of the network quality degradation cell output by the quality difference cell root cause positioning model.

10. The method of claim 9, wherein the bad cell root cause positioning model is trained based on:

constructing an initial classification model;

and training the initial classification model based on the sample index data and the problem root cause label data to obtain the quality difference cell root cause positioning model.

11. The method of claim 1, wherein the determining a wireless network optimization scheme and evaluation information corresponding to the wireless network optimization scheme according to the network problem root cause comprises:

determining a reference problem root cause matched with the network problem root cause from a preset relation table, and taking a reference network optimization scheme corresponding to the matched reference problem root cause as the wireless network optimization scheme, wherein the preset relation table comprises a plurality of reference problem root causes and the reference network optimization scheme corresponding to the reference problem root cause;

network parameter adjustment is carried out on the network quality degradation cell based on the wireless network optimization scheme at a target time point;

acquiring first index data of the network quality degradation cell before the target time point and second index data of the network quality degradation cell after the target time point;

And determining the evaluation information according to the first index data and the second index data.

12. The method of claim 11, wherein the reference network optimization scheme comprises a plurality of candidate network optimization schemes, each of the candidate network optimization schemes having a corresponding priority value;

wherein the performing network parameter adjustment on the network quality degradation cell based on the wireless network optimization scheme at the target time point includes:

determining a target network optimization scheme from the plurality of candidate network optimization schemes according to the priority value;

and carrying out network parameter adjustment on the network quality degradation cell based on the target network optimization scheme at the target time point.

13. The method as recited in claim 12, further comprising:

and if the evaluation information does not meet the preset condition, updating the priority value of each candidate network optimization scheme, and determining a new target network optimization scheme.

14. The method of claim 11, wherein the determining the evaluation information from the first index data and the second index data comprises:

Determining a weight value of each index to be optimized;

and inputting the first index improvement score and the second index improvement score into a pre-trained optimization scheme evaluation model to obtain the evaluation information output by the optimization scheme evaluation model.

15. A data processing apparatus for wireless network optimization, the apparatus comprising:

16. A computer device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-14.

17. A non-transitory computer readable storage medium storing computer instructions, wherein the computer instructions are for causing the computer to perform the method of any one of claims 1-14.

18. A computer program product comprising a computer program which, when executed by a processor, implements the steps of the method according to any one of claims 1-14.