An alarm information extraction method, device and computer readable storage medium

Abstract

This application discloses an alarm information extraction method, the method comprising: determining a first pattern string of sample multi-valued information elements of a sample alarm text; wherein the sample multi-valued information elements include key information items and value information items composed of multiple key-value pairs; the first pattern string is used to characterize the overall structure of the value information items in the sample multi-valued information elements; clustering the first pattern string to obtain a pattern set, and determining the set features of the pattern set; wherein the set features can identify the pattern set; generating a feature pattern library based on the set features of the pattern set and the first pattern string of the pattern set; and analyzing and processing the alarm text to be processed and the feature pattern library to obtain the target multi-valued information elements of the alarm text to be processed. This application also discloses an alarm information extraction device and a computer-readable storage medium.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a method, device and computer-readable storage medium for extracting alarm information. Background Technology

[0002] In network management fault management systems, it is crucial to process a large number of received alarm texts quickly and effectively. Alarm texts contain multiple key-value pairs, each consisting of a key and a value. Furthermore, alarm texts generated by different devices have different formats and contain different key information items. Therefore, effectively extracting information from different alarm texts is essential. Currently, a data type table storing the association between key information items and data types is constructed. After receiving real-time alarm data, the key-value pairs are extracted based on the key information items in this data type table. The extracted key-value pairs are then validated against the data type in the table. If the validation passes, the extracted key-value pairs are recorded, thus retrieving the data from the real-time alarm data. However, this method primarily extracts value information items composed of single information items. When extracting value information items composed of multiple key-value pairs, errors may occur in the extracted alarm information. Summary of the Invention

[0003] To address the aforementioned technical problems, this application aims to provide an alarm information extraction method, device, and computer-readable storage medium, which solves the problem in related technologies where alarm information extracted from value information items composed of multiple key-value pairs is erroneous, thereby improving the accuracy and efficiency of information extraction.

[0004] The technical solution of this application is implemented as follows:

[0005] A method for extracting alarm information, the method comprising:

[0006] A first pattern string is determined for the multi-valued information elements of the sample alarm text; wherein, the multi-valued information elements include key information items and value information items composed of multiple key-value pairs; the first pattern string is used to characterize the overall structure of the value information items in the multi-valued information elements.

[0007] Clustering the first pattern string yields a pattern set, and the set features of the pattern set are determined; wherein the set features can identify the pattern set.

[0008] Based on the set features of the pattern set and the first pattern string of the pattern set, a feature pattern library is generated;

[0009] The alarm text to be processed and the feature pattern library are analyzed and processed to obtain the target multi-valued information elements of the alarm text to be processed.

[0010] In the above scheme, the first pattern string for determining the multi-valued information elements of the sample alarm text includes:

[0011] Determine the key-value separator and key-value pair separator for the multi-value information elements of the sample; wherein, the key-value separator is the separator between key information items and value information items, and the key-value pair separator is the separator between key-value pairs;

[0012] The first pattern string is determined based on the key-value separator, the key-value pair separator, and the key information item of the sample multi-value information element.

[0013] In the above scheme, determining the first pattern string based on the key-value separator, the key-value pair separator, and the key information item of the sample multi-value information element includes:

[0014] The first symbol is obtained by sequentially determining the non-repeating delimiters from the key-value delimiters;

[0015] The second symbol is obtained by sequentially determining the non-repeating delimiters from the delimiters between the key-value pairs;

[0016] Based on the first symbol, the second symbol, and the key information item of the sample multi-valued information element, obtain the second pattern string of the sample multi-valued information element;

[0017] The first pattern string is obtained based on the association between the second pattern string and the pattern string.

[0018] In the above scheme, obtaining the second pattern string of the sample multi-valued information element based on the first symbol, the second symbol, and the key information item of the sample multi-valued information element includes:

[0019] Based on the structure of the first symbol, the second symbol, and the key-value pairs, a matching pattern is constructed;

[0020] Based on the matching pattern and the key information item of the sample multi-value information element, the second pattern string is obtained.

[0021] In the above scheme, determining the second pattern string based on the matching pattern and the key information item of the sample multi-value information element includes:

[0022] Based on the matching pattern, key-value pairs are extracted from the value information items of the sample multi-value information elements as target key-value pairs;

[0023] If the target key-value pair is the same as the key-value pair in the value information item of the sample multi-value information element, the second pattern string is determined based on the matching pattern and the key information item of the sample multi-value information element.

[0024] In the above scheme, the step of clustering the first pattern string to obtain a pattern set includes:

[0025] Based on the first pattern string, construct a pattern tree for the multi-valued information elements of the sample;

[0026] Based on the information of the nodes in the pattern tree, the similarity of the first pattern string is obtained;

[0027] The first pattern string is clustered based on its similarity to obtain the pattern set.

[0028] In the above scheme, determining the set features of the pattern set includes:

[0029] Determine the key information items in the pattern set;

[0030] The key information items that can identify the pattern set are determined as the set features of the pattern set.

[0031] In the above scheme, generating a feature pattern library based on the set features of the pattern set and the first pattern string of the pattern set includes:

[0032] The feature pattern library is generated based on the correspondence between the set features of the pattern set and the first pattern string of the pattern set.

[0033] In the above scheme, the analysis and processing of the alarm text to be processed and the feature pattern library to obtain the target multi-valued information elements of the alarm text to be processed includes:

[0034] The alarm text to be processed is matched with the features of the feature pattern library.

[0035] The set of features from the successfully matched feature pattern library is used as the target set features, and the alarm text to be processed is matched with the first pattern string of the pattern set corresponding to the target set features.

[0036] If pattern matching is successful, the target multi-value information element of the alarm text to be processed is obtained by extracting the first pattern string of the pattern set corresponding to the successfully matched target set features.

[0037] The method in the above scheme further includes:

[0038] In the event of feature matching failure, a third pattern string of the target multi-value information elements of the alarm text to be processed is obtained; wherein, the third pattern string is used to characterize the overall structure of the value information items in the alarm text to be processed;

[0039] Add the third pattern string of the target multi-valued information element to the target pattern set;

[0040] Determine the set characteristics of the target pattern set;

[0041] The feature pattern library is updated based on the set features of the target pattern set and the third pattern string of the target pattern set.

[0042] The method in the above scheme further includes:

[0043] In the event that pattern matching fails, the third pattern string of the target multi-value information element of the alarm text to be processed is obtained;

[0044] Add the third pattern string to the pattern set corresponding to the target set feature.

[0045] An alarm information extraction device, the device comprising: a processor, a memory, and a communication bus;

[0046] The communication bus is used to realize the communication connection between the processor and the memory;

[0047] The processor is used to execute the alarm information extraction program in the memory to implement the steps of the alarm information extraction method described above.

[0048] A computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement the steps of the alarm information extraction method described above.

[0049] The alarm information extraction method, device, and computer-readable storage medium provided in the embodiments of this application can determine the first pattern string of sample multi-valued information elements of sample alarm text, cluster the first pattern string to obtain a pattern set, determine the set features of the pattern set, and generate a feature pattern library based on the set features and the first pattern string of the pattern set. Thus, the alarm text to be processed can be analyzed and processed based on the feature pattern library to obtain the target multi-valued information elements of the alarm text to be processed. This solves the problem in related technologies that the extracted alarm information will be incorrect when extracting value information items composed of multiple key-value pairs, and improves the accuracy and efficiency of information extraction. Attached Figure Description

[0050] Figure 1 A flowchart illustrating an alarm information extraction method provided for an embodiment of this application;

[0051] Figure 2 A schematic diagram of a sample alarm text provided for an embodiment of this application;

[0052] Figure 3 A flowchart illustrating another alarm information extraction method provided for an embodiment of this application;

[0053] Figure 4 A schematic diagram of a pattern tree structure provided for an embodiment of this application;

[0054] Figure 5 A flowchart illustrating another alarm information extraction method provided for an embodiment of this application;

[0055] Figure 6 This is a schematic diagram of the structure of an alarm information extraction device provided in an embodiment of this application. Detailed Implementation

[0056] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0057] It should be understood that the phrases "embodiments of this application" or "foreign embodiments" throughout the specification mean that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, "embodiments of this application" or "in the foreign embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0058] Unless otherwise specified, any step in the embodiments of this application performed by the electronic device may be executed by the processor of the electronic device. It is also worth noting that the embodiments of this application do not limit the order in which the electronic device performs the following steps. Furthermore, the methods used to process data in different embodiments may be the same or different methods. It should also be noted that any step in the embodiments of this application can be executed independently by the electronic device; that is, when the electronic device performs any step in the following embodiments, it may not depend on the execution of other steps.

[0059] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0060] This application provides an alarm information extraction method, which can be applied to alarm information extraction devices. (Refer to...) Figure 1 As shown, the method includes the following steps:

[0061] Step 101: Determine the first pattern string of the multi-valued information element of the sample alarm text.

[0062] The sample multi-value information element includes key information items and value information items composed of multiple key-value pairs. The first pattern string is used to characterize the overall structure of the value information items in the sample multi-value information element.

[0063] In this embodiment of the application, the sample alarm text can be a plurality of pre-acquired alarm texts, and each of these alarm texts can include one or more sample multi-valued information elements. For example... Figure 2 As shown, Figure 2 The sample alarm text shown includes multiple single-value information elements and one multi-value information element. Figure 2 The alarm occurrence time shown is a single-value information element, where the key information item is: alarm occurrence time; and the value information item is: 2021-03-19 17:32:02; Figure 2 The alarmLocation(Adinfo) shown is a multi-valued information element. The key information item is: alarmLocation(Adinfo), and the value information items are: Aid: 620797954#1232; AlarmCode33: 052xxxxxx4; Remark; :PVFlag VNF NF=5f7efobe-486f-2032, IP=2409 802. It can be seen that the value information item of alarmLocation(Adinfo) consists of Aid, AlarmCode, and Remark. The Remark is composed of PVFlag VNF NF and IP, meaning that the value information item of alarmLocation(Adinfo) is composed of multiple key-value pairs. Optionally, before determining the first pattern string of the multi-valued information element in the sample alarm text, the sample alarm text can be preprocessed. For example, alarm texts with empty (or null) multi-valued information element value information items can be filtered out to improve the speed of subsequent processing.

[0064] Step 102: Cluster the first pattern string to obtain a pattern set, and determine the set characteristics of the pattern set.

[0065] Among them, set features can identify a set of patterns.

[0066] In this embodiment of the application, the first pattern strings are clustered, that is, the first pattern strings with strong similarity are clustered together, so as to discover the common features of the sample alarm texts, which helps to determine the set features of the pattern set in the subsequent process, and thus helps to extract information from the alarm text. Determining the set features of the pattern set is to improve the matching speed, so as to find the corresponding pattern set based on the set features and improve the alarm information extraction rate.

[0067] Step 103: Generate a feature pattern library based on the set features of the pattern set and the first pattern string of the pattern set.

[0068] In this embodiment of the application, the feature pattern library is generated based on the set features of the pattern set and the first pattern string of the pattern set. The pattern set is obtained by clustering the first pattern string of the multi-valued information elements of the sample alarm text. Therefore, the multi-valued information elements in the alarm text can be extracted based on the feature pattern library.

[0069] Step 104: Analyze and process the alarm text to be processed and the feature pattern library to obtain the target multi-valued information elements of the alarm text to be processed.

[0070] In this embodiment, the target multi-valued information element is the multi-valued information element in the alarm text to be processed. After obtaining the alarm text to be processed, the alarm text to be processed can be analyzed and processed with the feature pattern library obtained by the first pattern string of sample multi-valued information elements based on the sample alarm text to obtain the target multi-valued information element of the alarm text to be processed.

[0071] The alarm information extraction method provided in this application embodiment can determine the first pattern string of sample multi-valued information elements of sample alarm text, cluster the first pattern string to obtain a pattern set, determine the set features of the pattern set, and generate a feature pattern library based on the set features and the first pattern string of the pattern set. In this way, the alarm text to be processed can be analyzed and processed based on the feature pattern library to obtain the target multi-valued information elements of the alarm text to be processed. This solves the problem in related technologies that the alarm information extracted when extracting value information items composed of multiple key-value pairs will be incorrect, and improves the accuracy and efficiency of information extraction.

[0072] Based on the foregoing embodiments, this application provides a method for extracting alarm information, referring to... Figure 3 As shown, the method includes the following steps:

[0073] Step 201: The alarm information extraction device determines the key-value separator and key-value pair separator of the multi-value information element in the sample.

[0074] Here, the key-value separator is the separator between key information items and value information items, and the key-value pair separator is the separator between key-value pairs.

[0075] In this embodiment, the alarm information extraction device can determine the key-value separator and key-value pair separator of the multi-value information element in the sample, so as to subsequently obtain the first pattern string of the multi-value information element in the sample. For example... Figure 2The sample alarm text shown may include a multi-valued information element, which includes key information items, value information items, separators between key-value pairs, and separators between key-value pairs. For example, the multi-valued information element k can be (k1)(:)(v1)(;)(k2)(:)(v2)(;)(k3)(:)((k4)(=)(v4)(,)(k5)(=)(v5)), then the separator between key-value pairs in k is {:::==}, and the separator between key-value pairs in k is {;;,}.

[0076] Step 202: The alarm information extraction device determines the first pattern string based on the key information items of the key-value separator, the key-value pair separator, and the multi-value information element of the sample.

[0077] Step 202 can be implemented in the following way:

[0078] Step 202a: The alarm information extraction device sequentially determines the first symbol from the key-value separators, identifying the separators with non-repeating symbols.

[0079] In this embodiment, the alarm information extraction device sequentially determines the first symbol from the key-value separators, ensuring that the symbols are not repeated, so that the first pattern string can be obtained subsequently. For example, the sample multi-value information element k can be (k1)(:)(v1)(;)(k2)(:)(v2)(;)(k3)(:)((k4)(=)(v4)(,)(k5)(=)(v5)), then the key-value separators contained in k are {:::==}, and the first symbol obtained from the key-value separators is {:=}.

[0080] Step 202b: The alarm information extraction device sequentially determines the second symbol from the key-value pair separators, ensuring that the symbols are not repeated.

[0081] In this embodiment of the application, the alarm information extraction device sequentially determines the second symbol from the key-value pair separators, ensuring that the symbols are not repeated, so that the first pattern string can be obtained subsequently. For example, the sample multi-value information element k can be (k1)(:)(v1)(;)(k2)(:)(v2)(;)(k3)(:)((k4)(=)(v4)(,)(k5)(=)(v5)), then the key-value pair separators contained in k are {;;,}, and the second symbol obtained from the key-value pair separators is {;,}.

[0082] Step 202c: The alarm information extraction device obtains the second pattern string of the sample multi-value information element based on the key information items of the first symbol, the second symbol, and the sample multi-value information element.

[0083] In this embodiment of the application, the second pattern string is a partial structure of the sample multi-value information element. The first symbol is determined from the key-value separator, and the second symbol is determined from the key-value pair separator. Therefore, based on the first symbol, the second symbol, and the key information item, the partial structure of the sample multi-value information element (i.e., the second pattern string) can be obtained, and the first pattern string can be obtained based on the second pattern string and the association between the second pattern strings.

[0084] The operation in step 202c can be implemented in the following way:

[0085] Step A: The alarm information extraction device constructs a matching pattern based on the structure of the first symbol, the second symbol, and key-value pairs.

[0086] In this embodiment, the alarm information extraction device constructs a matching pattern based on the structure of a first symbol, a second symbol, and key-value pairs. The process can be as follows: First, the first determined symbol is used as the first target symbol; second, the first determined symbol is used as the second target symbol; a matching pattern is constructed based on the structure of the first target symbol, the second target symbol, and key-value pairs; if the matching pattern verification is successful, step B is executed: the second symbol already used as the second target symbol is deleted; the next determined second symbol is used as the new second target symbol; the process returns to the step of constructing the matching pattern based on the first target symbol and the second target symbol, and subsequent steps; if the matching pattern verification fails, the next determined second symbol is used as the new second target symbol; the process returns to the step of constructing the matching pattern based on the first target symbol and the second target symbol, and subsequent steps; after traversing all the second symbols, the next determined first symbol is used as the new first target symbol; the first determined second symbol is obtained from the remaining second symbols as the target symbol; the process returns to the step of constructing the matching pattern based on the structure of the first target symbol, the second target symbol, and key-value pairs, and subsequent steps.

[0087] In this embodiment, the alarm information extraction device first constructs a matching pattern for the outermost key-value pair of the multi-value information element of the sample based on the structure of the first symbol, the second symbol, and the key-value pair. Then, it constructs a matching pattern for the key-value pair in the outermost value information item of the multi-value information element of the sample, and so on, until all the first symbols and the second symbols are traversed, that is, all matching patterns that can obtain the second pattern string are obtained, and then the first pattern string can be obtained based on the second pattern string.

[0088] Step B: The alarm information extraction device determines the second pattern string based on the key information items of the matching pattern and sample multi-value information elements.

[0089] In this embodiment of the application, the alarm information extraction device determines a second pattern string based on the matching pattern and the key information item of the sample multi-valued information element, which can characterize part of the structure of the sample multi-valued information element. Furthermore, the second pattern string can be used to obtain a first pattern string that characterizes the overall structure of the sample multi-valued information element.

[0090] Step B can be performed as follows: the alarm information extraction device extracts key-value pairs from the value information items of the sample multi-value information element based on the matching pattern as the target key-value pairs; if the target key-value pairs are the same as the key-value pairs in the value information items of the sample multi-value information element, a second pattern string is obtained based on the matching pattern and the key information items of the sample multi-value information element.

[0091] For example, to obtain Figure 2 The second pattern string of alarmLocation(Adinfo) is shown, and the matching pattern is (k)(:)(v)(;); where k and v in the matching pattern can both be [\\u4e00-\\u9fa5]|\\s|\\S+, to indicate that k and v in the matching pattern can both be character sequences of any length containing Chinese characters. The operation of extracting key-value pairs as target key-value pairs from the value information items of multi-valued information elements of samples based on matching patterns is as follows: First, three key-value pairs are obtained: (Aid)(:)(620797954#1232)(;), (AlarmCode)(:)(33052xxxxxx4)(;), and (Remark)(:)(PVFlag VNF NF=5f7efobe-486f-2032,IP=2409 802)(;). Then, the three key-value pairs obtained based on the matching pattern and their acquisition order are concatenated to obtain the target pattern string: (Aid)(:)(620797954#1232)(;)(AlarmCode)(:)(33052xxxxxx4)(;)(Remark)(:)(PVFlag VNF NF=5f7efobe-486f-2032,IP=2409 If the target pattern string obtained is the same as the key-value pair in the value information item of the sample multi-value information element, then the verification result representing the successful verification of the matching pattern is obtained. Then, based on the matching pattern and the key information item of the sample multi-value information element, the second pattern string is obtained: (Aid)(:)(v)(;)(AlarmCode)(:)(v)(;)(Remark)(:)(v)(;).

[0092] Step 202d: The alarm information extraction device obtains the first pattern string based on the association between the second pattern string and the pattern string.

[0093] In this embodiment, the alarm information extraction device first obtains the outermost second pattern string of the sample multi-value information element based on the matching pattern and the key information item of the sample multi-value information element. Then, it obtains the second pattern string in the value information item of the outermost key-value pair, and so on, until the innermost second pattern string is obtained. After obtaining all the second pattern strings, the alarm information extraction device obtains the first pattern string based on all the second pattern strings and the association relationship between the pattern strings.

[0094] For example, if the sample multi-valued information element k is (k1)(:)(v1)(;)(k2)(:)(v2)(;)(k3)(:)((k4)(=)(v4)(,)(k5)(=)(v5)), then the first symbol in k is {:=}, and the second symbol in k is {;,}. First, the first determined symbol ":" is taken as the first target symbol, and the first determined symbol ";" is taken as the second target symbol. Based on the structure of ":", ";" and key-value pairs, a matching pattern is constructed, resulting in the matching pattern (k)(:)(v)(;), where k and v can be character sequences of arbitrary length. This matching pattern is then verified, i.e., the target pattern string is obtained according to this matching pattern. Three key-value pairs are obtained: (k1)(:)(v1)(;), (k2)(:)(v2)(;), and (k3)(:)((k4)(=)(v4)(,)(k5)(=)(v5))(;). Then, these three key-value pairs are... The target pattern string is obtained by concatenation: (k1)(:)(v1)(;)(k2)(:)(v2)(;)(k3)(:)((k4)(=)(v4)(,)(k5)(=)(v5)). If the target pattern string is the same as this multi-valued information element, then the verification result representing the successful verification of the matching pattern is obtained. Based on the matching pattern (k)(:)(v)(;) and the outermost key information item of this sample multi-valued information element, the outermost second pattern string (k1)(:)(v)(;)(k2)(:)(v)(;)(k3)(:)(v)(;) is obtained. Thus, the extraction of the outermost second pattern string is completed.Next, the second symbol ";" that has been used as the second target symbol is deleted. At this point, only {,} remain as the second symbol. The key-value pairs in the outermost value information item are then extracted, specifically from the key-value pairs in k3. The previously determined second symbol "," is obtained from the remaining second symbols and used as the new second target symbol. A matching pattern is constructed based on ":" and ",", resulting in the matching pattern (k)(:)(v)(,). This matching pattern is then verified, i.e., the target pattern string is obtained according to this matching pattern. The obtained target pattern string is (k3)(:)((k4)). The target pattern string obtained is different from the key-value pairs in k3 of the sample multi-value information element, resulting in a verification result indicating that the matching pattern verification failed. After traversing the second symbol of the separator between key-value pairs, the next determined first symbol "=" is taken as the new first target symbol. A matching pattern is constructed based on the structure of ":", "," and key-value pairs, and the constructed matching pattern is (k)(=)(v)(,). The matching pattern is verified, that is, the target pattern string is obtained according to the matching pattern. First, two key-value pairs are obtained: (k4)(=)(v4). The matching pattern (k5) is then concatenated with the key-value pairs obtained based on the matching pattern and their acquisition order. The resulting target pattern string is (k4) (k5) (v4) (,) (k5) (v5). Since the target pattern string matches the key-value pairs in k3 of this multi-value information element, a verification result indicating successful matching pattern verification is obtained. Based on the matching pattern (k) (k) (v) (,) and the key information item in the outermost value information item of the sample multi-value information element, the second pattern string (k4) (k5) (v5) in the outermost value information item is obtained. v)(,)(k5)(=)(v), delete the second symbol "," that has been used as the second target symbol. At this point, all the second symbols have been deleted, thus completing the extraction of the second pattern string in the outermost value information item. With the extraction of the second pattern strings in both the outermost and innermost layers completed, the first pattern string (k1)(:)(v)(;)(k2)(:)(v)(;)(k3)(:)((k4)(=)(v)(,)(k5)(=)(v)) is obtained based on the relationship between the second pattern string and the pattern string.

[0095] Step 203: The alarm information extraction device constructs a pattern tree of sample multi-value information elements based on the first pattern string.

[0096] In this embodiment of the application, since the outermost second pattern string is obtained first when the first pattern string is obtained, and then the second pattern string of the outermost value information item is obtained, and so on until the innermost second pattern string is obtained, a pattern tree of sample multi-value information elements can be constructed based on the process of obtaining the first pattern string. For example, the first pattern string of the multi-valued information element of the sample alarm text is (k1)(:)(v)(;)(k2)(:)(v)(;)(k3)(:)((k4)(=)(v)(,)(k5)(=)(v)). As can be seen from step 202d above, the outermost second pattern string obtained first is (k1)(:)(v)(;)(k2)(:)(v)(;)(k3)(:)(v)(;). Then, the key-value pairs in the outermost value information item are (k4)(=)(v4) and (k5)(=)(v5). Then, each key-value pair in the outermost second pattern string is taken as a node of the second layer, and each key-value pair in the obtained outermost value information item is taken as a node of the third layer. The nodes of the third layer are connected to the corresponding nodes of the previous layer, and the nodes of the same layer are sorted from left to right in the pattern string. The root node is represented as p, thus obtaining... Figure 4 The pattern tree shown is given; where h is the depth of the pattern tree.

[0097] Step 204: The alarm information extraction device obtains the similarity of the first pattern string based on the information of the nodes in the pattern tree.

[0098] In this embodiment, the alarm information extraction device can calculate the similarity of the first pattern string based on node information such as the content of the nodes, the order of the nodes, and the level of the nodes in the pattern tree. When calculating the similarity of the first pattern string, the influence of nodes at different levels on the similarity of the first pattern string is considered, and different weights are assigned to nodes at different levels. The weight of a node at a level farther away from the root node is smaller.

[0099] In one possible implementation, the first pattern string Exp i and the first pattern string Exp j The similarity calculation formula is: Sim(Exp i ,Exp j ) = Sim(p_root i ,p_root j Among them, p_root i For the first pattern string Exp i The root node, p_root j For the first pattern string Exp j The root node;

[0100] The similarity of the root node of the first pattern string is the sum of the similarities of all its nodes. The first pattern string Expi and the first pattern string Exp j The similarity calculation formula for nodes in the same position is: Sim(e hi ,e' hi )=(h-1) / h*SimNode(e hi ,e' hi )+1 / h*IDS(e hi ,e' hi ), where h is the depth of the node, i is the position of the node at the same level, and the positions of the nodes at each level are sorted from left to right starting from 1, and e hi For the first pattern string Exp i The node in, e' hi For the first pattern string Exp j The node in SimNode(e hi ,e' hi ) represents the similarity of node content; IDS(e hi ,e' hi The similarity between the direct descendant nodes of two nodes is denoted as .

[0101] First pattern string Exp i and the first pattern string Exp j The formula for calculating the similarity of node content in the code is: Where s1 is the first pattern string Exp i The key-value separator in the string, s2 is the first pattern string Exp i The key-value separator in the string, k is the first pattern string Exp i In the key information item, s'1 is the first pattern string Exp j The key-value separator in the string, s'2 is the first pattern string Exp j The key-value separator in the string, k' is the first pattern string Exp j Key information items in;

[0102] First pattern string Exp i and the first pattern string Exp j The formula for calculating the similarity of direct descendant nodes in the model is:

[0103]

[0104] Among them, e hi and e' hi When all nodes have direct descendant nodes, use formula (1) to calculate IDS(e hi ,e' hi );e hi and e' hiWhen there is a node without direct descendant nodes, formula (2) is used to calculate IDS(e hi ,e' hi ); when both e hi and e' hi have no direct descendant nodes, formula (1) is used to calculate IDS(e hi ,e' hi ).

[0105] Step 205: The alarm information extraction device clusters the first pattern strings based on the similarity of the first pattern strings to obtain the pattern set.

[0106] In the embodiment of the present application, the alarm information extraction device can cluster the first pattern strings based on the similarity of the first pattern strings by using a bottom-up hierarchical clustering algorithm. The bottom-up hierarchical clustering algorithm can visually see the clustering result in each clustering iteration process, and adopts the similarity calculation method of the first pattern strings described in step 204. In the clustering process, the first pattern strings with relatively high similarity but not completely identical are clustered together, so that each pattern set will contain first pattern strings with relatively high similarity but not completely the same.

[0107] Exemplarily, after clustering the first pattern strings by the alarm information extraction device, the obtained multiple pattern sets can be represented as Clusters = {c1, c2, c3,..., c k}, where k is the number of ethnic groups, 0 < k < n (n is the sample space number); c i is the i-th pattern set, and c i = (P, ExpSet), P is the sample number included in the pattern set, and ExpSet is the first pattern string included in the pattern set.

[0108] In the embodiment of the present application, after clustering the first pattern strings by the alarm information extraction device, the first pattern strings with relatively high similarity are clustered in the same pattern set. However, based on the existing clustering results, it is difficult to directly apply them to alarm information extraction. In a feasible way, the alarm information extraction device can obtain the set features of each pattern set through steps 206 - step 207.

[0109] Step 206: The alarm information extraction device determines the key information items in the pattern set.

[0110] In the embodiment of the present application, in addition to multi-valued information elements, the sample alarm text also contains many single-valued information elements, such as Figure 2The alarm title, alarm type, network management alarm level, alarm standard name, manufacturer, device type, data source, and alarm object type are shown. Each single-value information element can characterize some attributes of the sample alarm text, and these attributes are simple key-value pair structures, which are easy to use as matching rules. Therefore, based on these key information items, a set of features that can identify the set of patterns can be obtained.

[0111] Step 207: The alarm information extraction device determines the key information items that can identify the pattern set as the set feature of the pattern set.

[0112] Among them, set features are key information items that can identify a set of patterns.

[0113] In this embodiment of the application, the set feature of the pattern set can be a key information item that can identify the pattern set and the value corresponding to the key information item, or it can be multiple key information items that can identify the pattern set and the value corresponding to each of the multiple key information items, as long as the determined set feature can identify the pattern set.

[0114] Among them, the operation of step 207 can be: (1) The alarm information extraction device preprocesses the single-value data elements in the sample alarm texts in each mode set. The preprocessing can be: performing necessary resource backfilling and normalization on the value data items in each single-value information element to avoid outliers interfering with the extracted set features; (2) Define the feature set X = (x1, x2, ..., x...) of the alarm text. n ) T n is the number of features; each feature x i Represented as x i =(x i1 ,x i2 ,...,x im ), m is the dimension of the feature, and x is the dimension of each key. i Each information item corresponds to a feature. A feature is initialized for each pattern set. The matrix, The maximum dimension of all features. If a feature value appears in the pattern set, the corresponding position in the matrix is ​​set to 1; otherwise, it is set to 0. Each value of the matrix is ​​stored bit by bit; (3) For a certain feature x j If a certain pattern set c i Includes feature x j If the corresponding feature value is not included in other pattern sets, then this feature x j and the feature x j The corresponding values ​​in this pattern set can all be the feature rules of this pattern set; (4) For each remaining pattern set, the set features can be obtained through the following code:

[0115]

[0116] After obtaining feature S, each feature item in S and its corresponding value constitute the set features of the pattern set. That is, both the determined key information item and its value information item can be considered set features of the pattern set. The relationship between the values ​​of multiple value information items corresponding to a key information item is "OR," while the relationship between each key information item is "AND."

[0117] Step 208: The alarm information extraction device generates a feature pattern library based on the correspondence between the set features of the pattern set and the first pattern string of the pattern set.

[0118] In this embodiment of the application, the alarm information extraction device can use an XML-based representation method to generate a feature pattern library based on the correspondence between the set features of the pattern set and the first pattern string of the pattern set. Of course, it can also generate a feature pattern library based on the correspondence between the set features of the pattern set and the first pattern string of the pattern set through a rule representation method. This embodiment of the application does not limit this.

[0119] For example, the specific implementation of generating a feature pattern library using an XML-based representation method is as follows:

[0120]

[0121]

[0122] in,<rule_pattern> It is a feature-pattern library used to store the set features of a pattern set, the first pattern string in the pattern set, and the correspondence between the set features and the first pattern string. <key>For key information items of multi-valued data elements, such as Figure 1 The sample alarm text shown contains alarmLocation(Addinfo). <cluster>For each pattern set, there is a set feature and the first pattern string in the pattern set identified by the set feature, where id is the pattern set identifier.<match_rule> Nodes represent set features, and relations represent the relationships between set features, usually expressed as "and". <tkey>This is the value of the key information item, where name is the key data item. <value>The values ​​of a key data item in the pattern set are represented by "or". The relationship between multiple values ​​of the same key information item is "or".<pattern_str> Used to store the first pattern string contained in the pattern set. Each <pattern>Let be a first pattern string, weight be the weight of the first pattern string, the weight is determined according to the proportion of a certain type of first pattern string in the pattern set, and value be the specific first pattern string.

[0123] Step 209: The alarm information extraction device performs feature matching between the alarm text to be processed and the feature pattern library.

[0124] In this embodiment of the application, for the alarm text to be processed, feature matching is first performed with the features of the feature pattern library, that is, according to step 208. <cluster>internal <match>Perform feature matching. If the match is successful, proceed to steps 210-213 after step 209; if the match fails, proceed to steps 214-217 after step 209.

[0125] Step 210: The alarm information extraction device takes the set features of the successfully matched feature pattern library as the target set features, and performs pattern matching between the alarm text to be processed and the first pattern string of the pattern set corresponding to the target set features.

[0126] It should be noted that if the pattern matching is successful, step 211 is executed after step 210; if the pattern matching fails, steps 212 to 213 are executed after step 210.

[0127] Step 211: If the pattern matching is successful, the alarm information extraction device extracts the target multi-value information element of the alarm text to be processed based on the first pattern string of the pattern set corresponding to the feature of the successfully matched target set.

[0128] In this embodiment, when pattern matching is successful, the target multi-value information elements of the alarm text to be processed are extracted based on the first pattern string of the pattern set corresponding to the set features of the successfully matched feature pattern library. For example, if the id of the first pattern string of the pattern set corresponding to the set features of the successfully matched feature pattern library is 2, it means that the multi-value information elements in the alarm text to be processed are similar to the first pattern string with id 2. Then, the first pattern string with id 2 is extracted to obtain the target multi-value information elements of the alarm text to be processed.

[0129] Step 212: If pattern matching fails, the alarm information extraction device obtains the third pattern string of the target multi-value information element of the alarm text to be processed.

[0130] In this embodiment of the application, if pattern matching fails, it means that there is no first pattern string in the first pattern string of the pattern set corresponding to the target set feature that is similar to the target multi-value information element of the alarm text to be processed. In this case, the third pattern string of the target multi-value information element of the alarm text to be processed is obtained to update the first pattern string of the pattern set corresponding to the target set feature.

[0131] Step 213: The alarm information extraction device adds the third pattern string to the set of patterns corresponding to the successfully matched feature pattern library.

[0132] In this embodiment of the application, after obtaining the third pattern string, the alarm information extraction device can directly add the third pattern string to the pattern set corresponding to the target pattern set in order to update the feature pattern library.

[0133] Step 214: In the event of feature matching failure, the alarm information extraction device obtains the third pattern string of the target multi-value information element of the alarm text to be processed.

[0134] The third pattern string is used to characterize the overall structure of the value information item of the target multi-valued information element in the alarm text to be processed.

[0135] In this embodiment, feature matching failure indicates that the current feature pattern library does not contain the key information item in the alarm text to be processed, nor does it contain the first pattern library corresponding to the target multi-value information element of the alarm text to be processed. In this case, the alarm information extraction device adds the alarm text to be processed to the self-learning process, that is, it updates the feature pattern library based on the third pattern string of the target multi-value information element of the alarm text to be processed and the key information item in the alarm text to be processed.

[0136] Step 215: The alarm information extraction device adds the third pattern string of the target multi-value information element to the target pattern set.

[0137] In this embodiment of the application, the target pattern set is a new pattern set. The third pattern string of the target multi-valued information element is added to the target pattern set, that is, the third pattern string is added to the new pattern set to update the feature pattern library.

[0138] Step 216: The alarm information extraction device determines the set characteristics of the target pattern set.

[0139] In this embodiment, the key information item that can identify the alarm text to be processed can be determined as the set feature of the target pattern set, so as to match the third pattern string based on the set feature.

[0140] Step 217: The alarm information extraction device updates the feature pattern library based on the set features of the target pattern set and the third pattern string of the target pattern set.

[0141] In this embodiment, the alarm information extraction device updates the feature pattern library based on the set features of the target pattern set and the third pattern string of the target pattern set, which can continuously expand the feature pattern library to facilitate the subsequent analysis and processing of new alarm texts to be processed.

[0142] like Figure 5 As shown in the embodiments of this application, the alarm information extraction method first determines the first pattern string of sample multi-valued information elements of the sample alarm text, clusters the first pattern string to obtain a pattern set, and determines the set features of the pattern set. Then, a feature pattern library is generated based on the set features of the pattern set and the first pattern string of the pattern set. Thus, when obtaining the alarm text to be processed, the alarm text to be processed is matched with the set features of the feature pattern library. If the feature matching is successful, the alarm text is matched with the first pattern string of the pattern set corresponding to the successfully matched set features, thereby obtaining the target multi-valued information elements of the alarm text to be processed.

[0143] For example, if the constructed feature pattern library is as follows, where the key information item of the multi-valued information element is alarmLocation(Addinfo), the set feature of this multi-valued information element is alarm type, the value corresponding to the alarm type is XXX or YYY, and the pattern string corresponding to the alarm type is (Aid):([\u4e00-\u9fa5]|\s|\S+); (AlarmCode):([\u4e00-\u9fa5]|\s|\S+); (Remark):([\u4e00-\u9fa5]|\s|\S+). The alarm text to be processed is... Figure 1 The alarm text shown is: Figure 1 If the alarm text includes the alarm type "XXX", it can be successfully matched with the features in the feature pattern library shown below. Then, pattern matching will proceed. Figure 1 The multi-valued information element shown is Aid: 620797954#1232; AlarmCode33: 052xxxxxx4; Remark: PVFlag VNF NF=5f7efobe-486f-2032, IP=2409 802. It successfully matches the pattern string corresponding to the following alarm type feature set feature. Therefore, the target multi-valued information element obtained according to this feature pattern library is: Aid: 620797954#1232; AlarmCode: 33052xxxxxx4; Remark: PV Flag VNF NF=5f7efobe-486f-2032, IP=2409 802.

[0144]

[0145] Based on the foregoing embodiments, embodiments of this application provide an alarm information extraction device, which can be applied to... Figure 1 , 3 In the alarm information extraction and determination method provided in the embodiment corresponding to 5, refer to Figure 6 As shown, the alarm information extraction device 3 may include: a processor 31, a memory 32, and a communication bus 33, wherein:

[0146] Communication bus 33 is used to realize the communication connection between processor 31 and memory 32;

[0147] The processor 31 is used to execute the alarm information extraction program in the memory 32 to perform the following steps:

[0148] The first pattern string is used to determine the multi-valued information elements of the sample alarm text; wherein, the multi-valued information elements include key information items and value information items composed of multiple key-value pairs; the first pattern string is used to characterize the overall structure of the value information items in the multi-valued information elements.

[0149] Clustering the first pattern string yields a pattern set, and the set features of the pattern set are determined; wherein, the set features can identify the pattern set;

[0150] A feature pattern library is generated based on the set features of the pattern set and the first pattern string of the pattern set;

[0151] The alarm text to be processed and the feature pattern library are analyzed and processed to obtain the target multi-valued information elements of the alarm text to be processed.

[0152] In other embodiments of this application, processor 31 is used to execute the alarm information extraction program in memory 32 to determine the first pattern string of sample multi-valued information elements of sample alarm text, in order to implement the following steps:

[0153] Determine the key-value separator and key-value pair separator for multi-value information elements in the sample; where the key-value separator is the separator between key information items and value information items, and the key-value pair separator is the separator between key-value pairs;

[0154] The first pattern string is determined based on the key information items of the key-value separator, the key-value pair separator, and the multi-value information element of the sample.

[0155] In other embodiments of this application, the processor 31 is used to execute the alarm information extraction program in the memory 32 to determine the first pattern string based on the key information items of the key-value separator, the key-value pair separator, and the sample multi-value information element, so as to implement the following steps:

[0156] The first symbol is obtained by sequentially identifying non-repeating delimiters from the key-value delimiters;

[0157] The second symbol is obtained by sequentially identifying non-repeating delimiters from the delimiters between key-value pairs;

[0158] Based on the key information items of the first symbol, the second symbol, and the multi-valued information elements of the sample, the second pattern string of the multi-valued information elements of the sample is obtained;

[0159] Based on the relationship between the second pattern string and the pattern string, the first pattern string is obtained.

[0160] In other embodiments of this application, the processor 31 is configured to execute a key information item in the memory 32 based on the first symbol, the second symbol, and the sample multi-value information element, and obtain a second pattern string of the sample multi-value information element to implement the following steps:

[0161] Based on the structure of the first symbol, the second symbol, and key-value pairs, a matching pattern is constructed.

[0162] The second pattern string is determined based on the matching pattern and the key information item of the multi-valued information element of the sample.

[0163] In other embodiments of this application, the processor 31 is used to execute the alarm information extraction program in the memory 32 based on the key information item of the matching pattern and sample multi-value information element to determine the second pattern string in order to implement the following steps:

[0164] Based on the matching pattern, key-value pairs are extracted from the value information items of multi-value information elements in the sample as target key-value pairs;

[0165] If the target key-value pair is the same as the key-value pair in the value information item of the sample multi-value information element, a second pattern string is obtained based on the matching pattern and the key information item of the sample multi-value information element.

[0166] In other embodiments of this application, the first processor 31 is used to execute the alarm information extraction program in the first memory 32 to cluster the first pattern string to obtain a pattern set, so as to implement the following steps:

[0167] Based on the first pattern string, construct a pattern tree for the multi-valued information elements of the sample;

[0168] Based on the information of the nodes in the pattern tree, obtain the similarity of the first pattern string;

[0169] Clustering is performed on the first pattern string based on its similarity to obtain a pattern set.

[0170] In other embodiments of this application, the processor 31 is used to execute the set features of the determination pattern set of the alarm information extraction program in the memory 32 to implement the following steps:

[0171] Identify the key information items in the pattern set;

[0172] Key information items that can identify a set of patterns are defined as set features of the pattern set.

[0173] In other embodiments of this application, processor 31 is used to execute the alarm information extraction program in memory 32 based on the set features of the pattern set and the first pattern string of the pattern set to generate a feature pattern library, in order to implement the following steps:

[0174] A feature pattern library is generated based on the correspondence between the set features of the pattern set and the first pattern string of the pattern set.

[0175] In other embodiments of this application, the processor 31 is used to execute the alarm information extraction program in the memory 32 to analyze and process the alarm text to be processed and the feature pattern library, to obtain the target multi-value information elements of the alarm text to be processed, so as to achieve the following steps:

[0176] The alarm text to be processed is matched with the features of the feature pattern library.

[0177] The set of features from the successfully matched feature pattern library is used as the target set features, and the alarm text to be processed is matched with the first pattern string of the pattern set corresponding to the target set features.

[0178] If pattern matching is successful, the target multi-value information element of the alarm text to be processed is extracted based on the first pattern string of the pattern set corresponding to the feature of the successfully matched target set.

[0179] In other embodiments of this application, the processor 31 may also perform the following steps when executing the alarm information extraction program in the memory 32:

[0180] In the event of feature matching failure, the third pattern string of the target multi-value information element of the alarm text to be processed is obtained; wherein, the third pattern string is used to characterize the overall structure of the value information item in the alarm text to be processed;

[0181] Add the third pattern string of the target multi-valued information element to the target pattern set;

[0182] Determine the set characteristics of the target pattern set;

[0183] The feature pattern library is updated based on the set features of the target pattern set and the third pattern string of the target pattern set.

[0184] In other embodiments of this application, the processor 31 may also perform the following steps when executing the alarm information extraction program in the memory 32:

[0185] In the event of a pattern matching failure, obtain the third pattern string of the target multi-valued information element of the alarm text to be processed;

[0186] Add the third pattern string to the pattern set corresponding to the feature of the target set.

[0187] It should be noted that a detailed description of the steps performed by the processor can be found in [reference needed]. Figure 1 , 3 The alarm information extraction method provided in the embodiment corresponding to 5 will not be described again here.

[0188] The alarm information extraction device provided in this application generates a feature pattern library based on the first pattern string of sample multi-value information elements of the sample alarm text. By analyzing and processing the alarm text to be processed and the feature pattern library, the target multi-value information elements of the alarm text to be processed can be obtained. This solves the problem in related technologies where the extracted alarm information is incorrect when extracting value information items composed of multiple key-value pairs, and improves the accuracy and efficiency of information extraction.

[0189] Based on the foregoing embodiments, embodiments of this application provide a computer-readable storage medium storing one or more programs, which can be executed by one or more processors to implement... Figure 1 , 3 The steps of the alarm information extraction method provided in the embodiment corresponding to 5.

[0190] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0191] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0192] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0193] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0194] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application.< / match> < / cluster> < / pattern> < / value> < / tkey> < / cluster> < / key>

Claims

1. A method for extracting alarm information, characterized in that, The method includes: A first pattern string is determined for the multi-valued information elements of the sample alarm text; wherein, the multi-valued information elements include key information items and value information items composed of multiple key-value pairs; the first pattern string is used to characterize the overall structure of the value information items in the multi-valued information elements. Clustering the first pattern string yields a pattern set, and the set features of the pattern set are determined; wherein the set features can identify the pattern set. Based on the set features of the pattern set and the first pattern string of the pattern set, a feature pattern library is generated; The alarm text to be processed and the feature pattern library are analyzed and processed to obtain the target multi-valued information elements of the alarm text to be processed.

2. The method according to claim 1, characterized in that, The first pattern string of the sample multi-valued information elements of the determined sample alarm text includes: Determine the key-value separator and key-value pair separator for the multi-value information elements of the sample; wherein, the key-value separator is the separator between key information items and value information items, and the key-value pair separator is the separator between key-value pairs; The first pattern string is determined based on the key-value separator, the key-value pair separator, and the key information item of the sample multi-value information element.

3. The method according to claim 2, characterized in that, Determining the first pattern string based on the key information items of the key-value separator, the key-value pair separator, and the sample multi-value information element includes: The first symbol is obtained by sequentially determining the non-repeating delimiters from the key-value delimiters; The second symbol is obtained by sequentially determining the non-repeating delimiters from the delimiters between the key-value pairs; Based on the first symbol, the second symbol, and the key information item of the sample multi-valued information element, obtain the second pattern string of the sample multi-valued information element; The first pattern string is obtained based on the association between the second pattern string and the pattern string.

4. The method according to claim 3, characterized in that, The step of obtaining the second pattern string of the sample multi-valued information element based on the key information item of the first symbol, the second symbol, and the sample multi-valued information element includes: Based on the structure of the first symbol, the second symbol, and the key-value pairs, a matching pattern is constructed; The second pattern string is determined based on the matching pattern and the key information item of the sample multi-value information element.

5. The method according to claim 4, characterized in that, Determining the second pattern string based on the key information item of the matching pattern and the sample multi-value information element includes: Based on the matching pattern, key-value pairs are extracted from the value information items of the sample multi-value information elements as target key-value pairs; If the target key-value pair is the same as the key-value pair in the value information item of the sample multi-value information element, the second pattern string is determined based on the matching pattern and the key information item of the sample multi-value information element.

6. The method according to claim 1, characterized in that, The step of clustering the first pattern string to obtain a pattern set includes: Based on the first pattern string, construct a pattern tree for the multi-valued information elements of the sample; Based on the information of the nodes in the pattern tree, the similarity of the first pattern string is obtained; The first pattern string is clustered based on its similarity to obtain the pattern set.

7. The method according to claim 1, characterized in that, The determination of the set features of the pattern set includes: Determine the key information items in the pattern set; The key information items that can identify the pattern set are determined as the set features of the pattern set.

8. The method according to claim 1, characterized in that, The step of generating a feature pattern library based on the set features of the pattern set and the first pattern string of the pattern set includes: The feature pattern library is generated based on the correspondence between the set features of the pattern set and the first pattern string of the pattern set.

9. The method according to claim 1, characterized in that, The process involves analyzing the alarm text to be processed and the feature pattern library to obtain target multi-valued information elements of the alarm text to be processed, including: The alarm text to be processed is matched with the features of the feature pattern library. The set of features from the successfully matched feature pattern library is used as the target set features, and the alarm text to be processed is matched with the first pattern string of the pattern set corresponding to the target set features. If pattern matching is successful, the target multi-value information element of the alarm text to be processed is obtained by extracting the first pattern string of the pattern set corresponding to the successfully matched target set features.

10. The method according to claim 9, characterized in that, The method further includes: In the event of feature matching failure, a third pattern string of the target multi-value information elements of the alarm text to be processed is obtained; wherein, the third pattern string is used to characterize the overall structure of the value information items in the alarm text to be processed; Add the third pattern string of the target multi-valued information element to the target pattern set; Determine the set characteristics of the target pattern set; The feature pattern library is updated based on the set features of the target pattern set and the third pattern string of the target pattern set.

11. The method according to claim 9, characterized in that, The method further includes: In the event that pattern matching fails, the third pattern string of the target multi-value information element of the alarm text to be processed is obtained; Add the third pattern string to the pattern set corresponding to the target set feature.

12. An alarm information extraction device, characterized in that, The device includes: a processor, a memory, and a communication bus; The communication bus is used to realize the communication connection between the processor and the memory; The processor is used to execute an alarm information extraction program in the memory to implement the steps of the alarm information extraction method as described in any one of claims 1 to 11.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the alarm information extraction method as described in any one of claims 1 to 11.

Images