Fault change positioning method, device, equipment, medium and program product

Abstract

The embodiment of the invention discloses a fault change positioning method and device, equipment, a medium and a program product. The method comprises the steps that text similarity calculation is carried out between description texts corresponding to all changes in a change set within a preset time period of complaint moments corresponding to user complaint texts and the user complaint texts, so that whether the changes in the change set are related to the user complaint texts or not is determined, and if yes, the user complaint texts in the user complaint texts are not related to the user complaint texts. And if so, determining the related change as a fault change corresponding to the user complaint text.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a fault change location method, device, equipment, medium and program product. Background Art

[0002] In order to fix existing vulnerabilities (bugs) and improve software performance, technicians need to frequently change the software. Software changes require changes to the system configuration or code, which can easily cause failures. After a failure occurs, it is necessary to quickly locate the change corresponding to the failure, that is, to determine which historical change caused the failure, so that technicians can quickly roll back the located changes to quickly stop the loss and restore the system to normal operation.

[0003] Currently, faults can be discovered through user complaints. However, the process of locating faults through user complaints is relatively long. After receiving user complaints, customer service needs to summarize information about similar faults in multiple user complaints to avoid false positives. After confirming that there is a technical problem, the technical staff is notified to check multiple changes one by one to determine which change caused the fault, that is, to determine the location of the faulty change. Summary of the invention

[0004] The embodiments of this specification provide a method, device, equipment, medium and program product for locating fault changes, which uses natural language technology to associate user complaint texts with each change through text similarity, and can process user complaint texts as long as they appear, thus achieving short-long text matching, greatly speeding up the location and repair of fault changes, and improving the efficiency of processing user complaint texts. The above technical solutions are as follows:

[0005] In a first aspect, an embodiment of this specification provides a fault change location method, including:

[0006] Get user complaint text;

[0007] Based on the complaint time of the user complaint text, determine a set of changes within a preset time period;

[0008] Get the description text corresponding to each change in the above change set;

[0009] Calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text;

[0010] If the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, the change is determined as a fault change corresponding to the user complaint text.

[0011] In a possible implementation, the description text includes at least one sentence;

[0012] The above calculation of the text similarity between the description text corresponding to each of the above changes and the above user complaint text includes:

[0013] Calculate the first word overlap score between the user complaint text and each sentence in the description text corresponding to each change; the first word overlap score is used to characterize the correlation between the user complaint text and each sentence in the description text corresponding to each change;

[0014] Extract N sentences with the highest first word overlap scores from the description text corresponding to each of the above changes; N is a positive integer;

[0015] The text similarity between the description text corresponding to each of the changes and the user complaint text is calculated based on the N sentences corresponding to each of the changes.

[0016] In a possible implementation, calculating the text similarity between the description text corresponding to each change and the user complaint text according to the N sentences corresponding to each change includes:

[0017] Calculate N text similarities between the N sentences corresponding to each of the changes and the user complaint text;

[0018] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the L highest text similarities among the above N text similarities; the above L is a positive integer greater than 0 and less than or equal to N.

[0019] In a possible implementation, determining the text similarity between the description text corresponding to each change and the user complaint text according to the L highest text similarities among the N text similarities includes:

[0020] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text.

[0021] In a possible implementation, the description text includes a plurality of sentences;

[0022] The above calculation of the text similarity between the description text corresponding to each of the above changes and the above user complaint text includes:

[0023] All sentences in the description text corresponding to each of the above changes are spliced ​​according to a first preset rule to obtain at least one target text corresponding to the above description text; the above target text includes at least one of the above sentences;

[0024] Calculating a second word overlap score between the user complaint text and each target text corresponding to the description text; the second word overlap score is used to characterize the correlation between the user complaint text and each target text corresponding to the description text;

[0025] Extracting M target texts with the highest second word overlap scores from the at least one target text corresponding to the description text; M is a positive integer;

[0026] The text similarity between the description text corresponding to each of the changes and the user complaint text is calculated based on the M target texts corresponding to the description text of each of the changes.

[0027] In a possible implementation, calculating the text similarity between the description text corresponding to each change and the user complaint text according to the M target texts corresponding to the description text of each change includes:

[0028] Calculate the M text similarities between the M target texts corresponding to each of the changes and the user complaint texts;

[0029] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the K highest text similarities among the above M text similarities; the above K is a positive integer greater than 0 and less than or equal to M.

[0030] In a possible implementation, determining the text similarity between the description text corresponding to each change and the user complaint text according to the K highest text similarities among the M text similarities includes:

[0031] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text.

[0032] In a possible implementation, after receiving the user complaint text and before calculating the text similarity between the description text corresponding to each change and the user complaint text, the method further includes:

[0033] Filter the user complaint text and the description text corresponding to each of the changes according to a preset stop word library and a second preset rule to obtain a filtered user complaint text and a filtered description text corresponding to each of the changes;

[0034] The above calculation of the text similarity between the description text corresponding to each of the above changes and the above user complaint text includes:

[0035] The text similarity between the description text corresponding to each filtered change and the filtered user complaint text is calculated.

[0036] In a possible implementation, after receiving the user complaint text and before calculating the text similarity between the description text corresponding to each change and the user complaint text, the method further includes:

[0037] Determine X historical user complaint texts complained before the complaint time of the above user complaint text; the above X is a positive integer;

[0038] Calculate the text similarity between the user complaint text and each of the X historical user complaint texts;

[0039] If there are Y historical user complaint texts and the above user complaint text, the text similarity is greater than or equal to the second preset threshold, then the above Y historical user complaint texts are spliced ​​with the above user complaint text according to the third preset rule to obtain the target complaint text; the above Y is a positive integer greater than 0 and less than or equal to X;

[0040] The above calculation of the text similarity between the description text corresponding to each of the above changes and the above user complaint text includes:

[0041] Calculate the text similarity between the description text corresponding to each of the above changes and the above target complaint text;

[0042] If the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, the change is determined as a fault change corresponding to the user complaint text, including:

[0043] If the text similarity between the description text corresponding to the change and the target complaint text is greater than or equal to the first preset threshold, the change is determined as the fault change corresponding to the user complaint text and the Y historical user complaint texts.

[0044] In a second aspect, an embodiment of the present specification provides a fault change location device, including:

[0045] Acquisition module, used to obtain user complaint text;

[0046] A first determination module, used to determine a set of changes within a preset time period based on the complaint time of the user complaint text;

[0047] The acquisition module is further used to obtain a description text corresponding to each change in the change set;

[0048] A first calculation module is used to calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text;

[0049] The second determination module is used to determine the change as a fault change corresponding to the user complaint text if the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold.

[0050] In a possible implementation, the description text includes at least one sentence;

[0051] The first calculation module includes:

[0052] A first calculation unit is used to calculate the first word overlap score between the user complaint text and each sentence in the description text corresponding to each change; the first word overlap score is used to characterize the correlation between the user complaint text and each sentence in the description text corresponding to each change;

[0053] A first extraction unit is used to extract N sentences with the highest first word overlap scores from the description text corresponding to each change, where N is a positive integer;

[0054] The second calculation unit is used to calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text according to the above N sentences corresponding to each of the above changes.

[0055] In a possible implementation manner, the second computing unit is specifically configured to:

[0056] Calculate N text similarities between the N sentences corresponding to each of the changes and the user complaint text;

[0057] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the L highest text similarities among the above N text similarities; the above L is a positive integer greater than 0 and less than or equal to N.

[0058] In a possible implementation, when the second calculation unit determines the text similarity between the description text corresponding to each change and the user complaint text according to the L highest text similarities among the N text similarities, it is specifically used to:

[0059] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text.

[0060] In a possible implementation, the description text includes a plurality of sentences;

[0061] The first calculation module further includes:

[0062] A concatenation unit, configured to concatenate all the sentences in the description text corresponding to each of the above changes according to a first preset rule to obtain at least one target text corresponding to the above description text; the above target text includes at least one of the above sentences;

[0063] The third calculation unit is used to calculate the second word overlap score between the user complaint text and each target text corresponding to the description text; the second word overlap score is used to characterize the correlation between the user complaint text and each target text corresponding to the description text;

[0064] A second extraction unit is used to extract M target texts with the highest second word overlap scores from the at least one target text corresponding to the description text; the M is a positive integer;

[0065] The fourth calculation unit is used to calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text according to the above M target texts corresponding to the description text of each of the above changes.

[0066] In a possible implementation manner, the fourth computing unit is specifically configured to:

[0067] Calculate the M text similarities between the M target texts corresponding to each of the changes and the user complaint texts;

[0068] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the K highest text similarities among the above M text similarities; the above K is a positive integer greater than 0 and less than or equal to M.

[0069] In a possible implementation, when determining the text similarity between the description text corresponding to each change and the user complaint text according to the K highest text similarities among the M text similarities, the fourth calculation unit is specifically used to:

[0070] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text.

[0071] In a possible implementation, the above-mentioned fault change location device further includes:

[0072] A filtering module, used to filter the user complaint text and the description text corresponding to each of the changes according to a preset stop word library and a second preset rule to obtain the filtered user complaint text and the description text corresponding to each of the filtered changes;

[0073] The first calculation module is specifically used for:

[0074] The text similarity between the description text corresponding to each filtered change and the filtered user complaint text is calculated.

[0075] In a possible implementation, the above-mentioned fault change location device further includes:

[0076] The third determination module is used to determine X historical user complaint texts complained before the complaint time of the user complaint text; the above X is a positive integer;

[0077] A second calculation module is used to calculate the text similarity between the user complaint text and each of the X historical user complaint texts;

[0078] A splicing module, configured to splice the Y historical user complaint texts with the user complaint text according to a third preset rule to obtain a target complaint text if the text similarity between the Y historical user complaint texts and the user complaint text is greater than or equal to a second preset threshold; the Y is a positive integer greater than 0 and less than or equal to X;

[0079] The first calculation module is further used to calculate the text similarity between the description text corresponding to each change and the target complaint text;

[0080] The second determination module is also used to determine the change as a fault change corresponding to the user complaint text and the Y historical user complaint texts if the text similarity between the description text corresponding to the change and the target complaint text is greater than or equal to a first preset threshold.

[0081] In a third aspect, an embodiment of this specification provides an electronic device, including: a processor and a memory;

[0082] The processor is connected to the memory;

[0083] The above-mentioned memory is used to store executable program code;

[0084] The processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the method provided by the first aspect of the embodiment of this specification or any possible implementation of the first aspect.

[0085] In a fourth aspect, an embodiment of the present specification provides a computer storage medium, wherein the computer storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the method provided by the first aspect of the embodiment of the present specification or any possible implementation of the first aspect.

[0086] In a fifth aspect, an embodiment of the present specification provides a computer program product comprising instructions, which, when the above-mentioned computer program product runs on a computer or a processor, enables the above-mentioned computer or the above-mentioned processor to execute the fault change location method provided by the first aspect of the embodiment of the present specification or any possible implementation method of the first aspect.

[0087] The embodiment of this specification determines a set of changes within a preset time period based on the complaint time of the user complaint text, and then calculates the text similarity between the description text corresponding to each change in the above change set and the above user complaint text, and determines the changes whose text similarity with the above user complaint text is greater than or equal to a first preset threshold as the fault changes corresponding to the above user complaint text, thereby utilizing natural language technology to associate the user complaint text with each change through text similarity, so that the user complaint text can be processed as long as it appears, and can match short (user complaint text) - long (description text corresponding to the change) text, thereby greatly speeding up the location and repair of fault changes, reducing the troubleshooting time of technical personnel, and improving the efficiency of processing user complaint texts. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] In order to more clearly illustrate the technical solutions in the embodiments of this specification, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0089] Figure 1 A schematic diagram of the architecture of a fault change location system provided by an exemplary embodiment of this specification;

[0090] Figure 2A flowchart from a user discovering a fault to a technician repairing the fault provided in the related technology;

[0091] Figure 3 A flowchart of a fault change location method provided by an exemplary embodiment of this specification;

[0092] Figure 4 A schematic diagram of changes provided for an exemplary embodiment of this specification and description texts corresponding to the changes;

[0093] Figure 5 A schematic diagram of an implementation flow of a text similarity model provided as an exemplary embodiment of this specification;

[0094] Figure 6 A schematic diagram of the implementation flow of another text similarity model provided as an exemplary embodiment of this specification;

[0095] Figure 7 A flowchart of another fault change location method provided by an exemplary embodiment of this specification;

[0096] Figure 8 A schematic diagram of the structure of a fault change positioning device provided by an exemplary embodiment of this specification;

[0097] Fig. 9 The present invention is a schematic diagram of the structure of an electronic device provided by an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0098] The technical solutions in the embodiments of this specification will be described clearly and completely below in conjunction with the drawings in the embodiments of this specification.

[0099] The terms "first", "second", "third", etc. in this specification and claims and the above drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0100] Please refer to Figure 1 , Figure 1 The following is a schematic diagram of the architecture of a fault change positioning system provided by an exemplary embodiment of this specification. Figure 1 As shown, the fault change location system may include: a first terminal cluster, a second terminal 120, a server 130 and a third terminal cluster. Among them:

[0101] The first terminal cluster may be a user terminal, specifically including one or more user terminals, wherein the multiple user terminals may include user terminal 110a, user terminal 110b, user terminal 110c, etc. A user version of the software may be installed in the first terminal cluster to implement functions such as user online input of user complaint text or complaint voice. Any user terminal in the first terminal cluster may establish a data relationship with the network, and establish a data connection relationship with the second terminal 120 through the network, such as sending a user complaint text, initiating a complaint call, etc. Among them, any user terminal in the first terminal cluster may be, but is not limited to, a mobile phone, tablet computer, laptop computer, or other device with the user version of the software installed.

[0102] The second terminal 120 can be a customer service terminal, which can specifically include one or more customer service terminals. A customer service version of the software can be installed in the second terminal 120 to implement functions such as obtaining user complaint text or complaint voice online, answering complaint calls, and converting the content of the complaint call into complaint text. The second terminal 120 can establish a data relationship with the network, and establish a data connection relationship with any user terminal in the first terminal cluster through the network, such as receiving user complaint text, answering complaint calls, etc. The second terminal 120 can also establish a data connection relationship with the server 130 through the network, such as sending user complaint text or complaint calls, etc. Among them, the second terminal 120 can be, but is not limited to, a mobile phone, tablet computer, laptop computer and other devices installed with the customer service version of the software.

[0103] The third terminal cluster can be a technician terminal, specifically including one or more technician terminals, wherein the multiple technician terminals may include a technician terminal 140a, a technician terminal 140b, a technician terminal 140c, etc. A technician version of the software can be installed in the third terminal cluster to implement functions such as technicians receiving fault changes online, troubleshooting and repairing fault changes, etc. Any technician terminal in the third terminal cluster can establish a data relationship with the network, and establish a data connection relationship with the server 130 through the network, such as sending a description text corresponding to the change, receiving a confirmed fault change, etc. Among them, any technician terminal in the third terminal cluster can be, but is not limited to, a mobile phone, a tablet computer, a laptop computer, and other devices with the technician version of the software installed.

[0104] The server 130 may be a server that can provide multiple fault change positioning, and may receive data such as user complaint texts sent by the network or the second terminal 120, and receive description texts corresponding to the changes sent by any technician terminal in the third terminal cluster, and store the description texts corresponding to the changes sent by all technician terminals in the database, and then determine the fault change corresponding to the user complaint text according to the user complaint text and the description text corresponding to each change in the change set within the preset time period at the complaint moment of the user complaint text. The server 130 may also send the user complaint text to any technician terminal in the third terminal cluster or the technician terminal of the fault change corresponding to the user complaint text. The server 130 may be, but is not limited to, a hardware server, a virtual server, a cloud server, etc.

[0105] The network may be a medium providing a communication link between the second terminal 120 and any user terminal in the first terminal cluster, between the second terminal 120 and the server 130, and between the server 130 and any technician terminal in the third terminal cluster, or may be the Internet including network equipment and transmission media, but is not limited thereto. The transmission medium may be a wired link (for example, but not limited to, coaxial cable, optical fiber, and digital subscriber line (DSL), etc.) or a wireless link (for example, but not limited to, wireless fidelity (WIFI), Bluetooth, and mobile device network, etc.).

[0106] Understandably, Figure 1 The number of the first terminal cluster, the second terminal 120, the server 130, and the third terminal cluster in the fault change location system shown is only an example. In a specific implementation, the fault change location system may include any number of user terminals, customer service terminals, technician terminals, and servers. This specification embodiment does not specifically limit this. For example, but not limited to, the second terminal 120 may be a second terminal cluster composed of multiple second terminals (customer service terminals), and the server 130 may be a server cluster composed of multiple servers.

[0107] Next, combine Figure 1 , which introduces the overall process from the user discovering the fault to the technician repairing the fault in the relevant technology. For details, please refer to Figure 2 , which is a flowchart provided in the related art from the user discovering the fault to the technician repairing the fault. Figure 2As shown, if a software failure occurs (201, failure occurs), the customer service will receive a user's telephone complaint and / or online complaint (202, customer complaint), and then the customer service will need to manually submit the received user complaint call and / or online complaint text (203, complaint reporting). Since a complaint call or complaint text may be a false alarm, multiple customer service or on-site general control personnel need to summarize the complaint calls and / or complaint texts of similar failures (204, complaint summary), and after confirming that it is a technical problem, notify the technical staff, and the technical staff will pull a large group to check one by one (205, pull a group to check), so as to locate the specific location of the fault change (206, locate the fault change), and the technical staff corresponding to the above fault change will roll back and repair (207, roll back & repair). Since there are a large number of changes and a large number of corresponding technical personnel, it not only takes time to wait for multiple complaints of similar failures to appear (204, complaint summary) before starting to process complaints that are confirmed to be technical problems, but also requires a lot of time to determine which change caused the failure. In other words, the existing process from user discovery of the failure to technical personnel repairing the failure is too long, the workload of customer service and technical personnel is extremely large, and they cannot handle every user complaint call or complaint text in a timely manner.

[0108] The embodiments of this specification use machine learning natural language processing technology to associate the text corresponding to the user's complaint phone number and the complaint text (short text) with the relevant description (long text) when the technician changes the system or code, etc., in time sequence, that is, to achieve text similarity calculation of short-long text, and confirm whether the user complaint is related to a certain change. As a result, as soon as the change is released, the technician can immediately perceive how many related user complaints there are, which greatly speeds up the location of the fault change and can handle each user complaint in a timely manner. That is, there is no need to perform step 204 (complaint summary) to confirm whether the user complaint is a technical problem. If so, there is no need to perform step 205 (group investigation) to allow the machine to quickly determine the corresponding fault change, thereby notifying the technician corresponding to the above-mentioned fault change to roll back and repair it in time.

[0109] Next, combine Figure 1-Figure 2 , introduces the fault change location method provided by the embodiment of this specification. For details, please refer to Figure 3 , which is a flow chart of a fault change location method provided by an exemplary embodiment of this specification. Figure 3 As shown, the fault change location method includes the following steps:

[0110] Step 302, obtaining the user complaint text.

[0111] Specifically, after receiving a user complaint call, the customer service end may first convert the voice data of the user complaint call into a user complaint text, and then send the complaint text to the server, so that the server can obtain the user complaint text. The customer service end may also directly send the voice data of the user complaint call to the server, and the server converts the voice data into the user complaint text. The server may also directly receive the user complaint text sent by the customer service end or the user end through the network, etc. This specification does not limit this.

[0112] Step 304: Determine a set of changes within a preset time period based on the complaint time of the user complaint text.

[0113] Specifically, the user complaint text includes the corresponding complaint time. The above complaint time is used to characterize the time when the user starts to make a complaint call or starts to send an online complaint text. Since the release time corresponding to different changes may also be different, and many changes may occur every day, from a chronological point of view, the changes before the complaint time of the user complaint text and close to the complaint time of the above user complaint text are more likely to be correlated with the user complaint text. Therefore, according to the complaint time of the user complaint text, the set of changes within the preset time period of the above complaint time can be determined. The above preset time period can be 1 minute, 8 hours, 3 days, 1 month, etc., and this specification does not limit this. The above change set includes all changes within the preset time period before the complaint time of the user complaint text.

[0114] Exemplarily, if the preset time period is 1 day and the complaint time of the user complaint text is 00:00:00 on January 26, 2022, all changes published between 00:00:00 on January 25, 2022 and 00:00:00 on January 26, 2022 can be determined as a set of changes within the preset time period.

[0115] Step 306: Obtain the description text corresponding to each change in the change set.

[0116] Specifically, the server can search for the description text corresponding to each change in the above change set from the description texts corresponding to all pre-stored changes according to the name corresponding to each change in the change set. The changes may include but are not limited to code changes, configuration changes, system changes, interface changes, etc.

[0117] For example, Figure 4As shown, the change 400 includes a configuration change 410, a code change 420, and a system change 430. The description text corresponding to the configuration change 410 includes a parameter description 411, an interface description 412, and a system description 413 corresponding to this change written by a technical staff, and a requirement description 414 corresponding to this change written by a non-technical staff such as a product manager. The description text corresponding to the code change 420 includes a code note 421, an iteration description 422, and a system description 423 corresponding to this change written by a technical staff, and a requirement description 424 corresponding to this change written by a non-technical staff such as a product manager. The description text corresponding to the system change 430 includes a system description 431 corresponding to this change written by a technical staff, and a requirement description 432 corresponding to this change written by a non-technical staff such as a product manager.

[0118] Step 308: Calculate the text similarity between the description text corresponding to each change and the user complaint text.

[0119] Specifically, the description text corresponding to each change and the user complaint text can be directly input into the BERT text similarity model, so as to output the text similarity between the description text corresponding to each change and the user complaint text. The above BERT text similarity model is trained by the description texts corresponding to multiple changes, multiple user complaint texts, and the text similarity between the description text corresponding to each change and each of the user complaint texts.

[0120] Optionally, in order to improve the accuracy of the above-mentioned BERT text similarity model, the user complaint text and the description text corresponding to each change can be first filtered according to the preset stop word library according to the second preset rule to obtain the filtered user complaint text and the description text corresponding to each filtered change, and then the text similarity between the description text corresponding to each filtered change and the filtered user complaint text is calculated. The above-mentioned second preset rule can be to filter out the words in the user complaint text and the description text that are the same as the stop words in the preset stop word library. The above-mentioned stop words are used to represent words that are irrelevant to the description of faults and changes, such as thank you, OK, etc. In order to avoid the software-related descriptive words in the description text corresponding to each change being filtered out, which would lead to a lower accuracy rate in locating the fault change, the above-mentioned second preset rule may also be to filter out stop words in the user complaint text and the description text according to a preset stop word library, while retaining the software-related descriptions in the above-mentioned description text, such as the description of professional terms such as configuration and system in the description text. That is, in order to avoid certain words in the software-related description that are the same as the stop words in the preset stop word library from being filtered out, a protection field may be added to it, and this manual does not limit this.

[0121] Step 310: If the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, the change is determined as a fault change corresponding to the user complaint text.

[0122] Specifically, the change corresponding to the description text whose text similarity with the user complaint text is greater than or equal to the first preset threshold can be determined as the fault change corresponding to the above-mentioned user complaint text, that is, it may be due to the above-mentioned fault change that the above-mentioned user encounters problems when using the software, thereby initiating a complaint, that is, generating the above-mentioned user complaint text. The above-mentioned first preset threshold can be 0.1, 0.9, etc., and this specification does not limit this. The fault change corresponding to the above-mentioned user complaint text includes at least one change. After determining at least one fault change related to the above-mentioned user complaint text, the server can send the above-mentioned user complaint text and its corresponding at least one fault change to the technical personnel corresponding to each of the above-mentioned fault changes for troubleshooting, rollback and repair.

[0123] The embodiment of this specification determines a set of changes within a preset time period based on the complaint time of the user complaint text, and then calculates the text similarity between the description text corresponding to each change in the above change set and the above user complaint text, and determines the change whose text similarity with the above user complaint text is greater than or equal to a first preset threshold as the fault change corresponding to the above user complaint text, thereby utilizing natural language technology to associate the user complaint text with each change through text similarity, so that the user complaint text can be processed as long as it appears, and can match the short (user complaint text) - long (description text corresponding to the change) text, so that the technical staff only needs to troubleshoot, roll back and repair the fault changes associated with the above user complaint text, and does not need to troubleshoot all the changes, which greatly speeds up the location and repair of the fault changes, reduces the troubleshooting time of the technical staff, and improves the efficiency of processing the user complaint text.

[0124] The traditional BERT text similarity model (BERT model) can only process short texts with a length of less than 512. The length of user complaint texts is generally less than 512, while the length of the description text corresponding to each change is much longer than 512. In order to enable the BERT text similarity model to calculate the similarity of short and long texts, it is generally adopted to compress long texts or extract keywords or key sentences including keywords in long texts for matching. However, the above methods will result in the loss of a large amount of meaningful text in the description text corresponding to the change, thereby affecting the accuracy of the BERT text similarity model in calculating the similarity of short and long texts, resulting in low accuracy in locating fault changes. In order to solve the above problems, we will combine Figure 1-Figure 4, introduces a specific implementation process of the above step 308 in the fault change location method provided in the embodiment of this specification, which calculates the text similarity between the description text corresponding to each change and the user complaint text. For details, please refer to Figure 5 , which is a schematic diagram of the implementation flow of a text similarity model provided by an exemplary embodiment of this specification. Figure 5 As shown in the figure, the implementation process of the text similarity model includes the following steps:

[0125] Step 502 , calculating the first word overlap score between the user complaint text and each sentence in the description text corresponding to each change.

[0126] Specifically, the description text corresponding to the change includes at least one sentence. The description text corresponding to each change can be segmented according to periods, and then the user complaint text and each sentence in the description text corresponding to each change text are scored for word overlap, that is, the score of the same or similar words between the user complaint text and each sentence in the description text corresponding to each change is calculated, that is, the first word overlap score. The above first word overlap score is used to characterize the correlation between the user complaint text and each sentence in the description text corresponding to each change.

[0127] Step 504: extract N sentences with the highest first word overlap scores from the description text corresponding to each change.

[0128] Specifically, after calculating the first word overlap score between the user complaint text and each sentence in the description text corresponding to each change, N sentences with the highest first word overlap scores in the description text corresponding to each change can be extracted. The above N is a positive integer.

[0129] For example, five sentences with the highest first word overlap scores may be extracted from the description text corresponding to each change.

[0130] Step 506 , calculating the text similarity between the description text corresponding to each change and the user complaint text based on the N sentences corresponding to each change.

[0131] Optionally, the N text similarities between the N sentences with the highest first word overlap scores corresponding to each change and the user complaint text can be calculated first, that is, the N sentences with the highest first word overlap scores corresponding to each change and the user complaint text can be input into the BERT text similarity model respectively, so as to output the N text similarities between the N sentences corresponding to each change and the user complaint text. Then, the text similarity between the description text corresponding to each change and the user complaint text is determined based on the L highest text similarities among the N text similarities. The above L is a positive integer greater than 0 and less than or equal to N.

[0132] Exemplarily, the average of the L highest text similarities among the N text similarities corresponding to each of the above changes can be calculated, and the above average value can be determined as the text similarity between the description text corresponding to each of the above changes and the above user complaint text; or the L highest text similarities among the N text similarities corresponding to each of the above changes can be weightedly summed to obtain the text similarity between the description text corresponding to each of the above changes and the above user complaint text, which is not limited to the embodiments of this specification.

[0133] Optionally, after calculating the N text similarities between the N sentences with the highest first word overlap scores corresponding to each change and the user complaint text, the text similarity between the description text corresponding to each change and the user complaint text can be determined based on the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text. That is, the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text can be weightedly fused to obtain the text similarity between the description text corresponding to each change and the user complaint text. The above L is a positive integer greater than 0 and less than or equal to N.

[0134] The embodiment of this specification combines the L highest text similarities among the N text similarities corresponding to the N sentences with the highest first word overlap scores in each change description text, and the L first word overlap scores between the L sentences corresponding to the above L highest text similarities and the above user complaint text to jointly determine the text similarity between the description text corresponding to each of the above changes and the above user complaint text, thereby further improving the accuracy of calculating text similarity, thereby further improving the accuracy of fault change locating.

[0135] Since everyone has different habits of writing description texts, sometimes they write periods and sometimes they write semicolons, and the sentence lengths are inconsistent. If the description text corresponding to each change is divided according to the periods as mentioned above, and then the text similarity between the description text corresponding to each change and the user complaint text is calculated based on certain sentences in the description text corresponding to each change, the text semantics will be cut due to the extra periods when writing, and each description text corresponding to the change includes many sentences, that is, each of them needs to calculate the first word overlap score many times, and the amount of calculation is particularly large, resulting in the efficiency of fault change location is not high enough, the accuracy of text similarity calculation is not high enough, and the fault change location is not accurate enough. In order to solve the above problems, we will combine Figure 1-Figure 4, introduces another specific implementation process of the above step 308 in the fault change location method provided in the embodiment of this specification, which is to calculate the text similarity between the description text corresponding to each change and the user complaint text. For details, please refer to Figure 6 , which is a schematic diagram of the implementation flow of another text similarity model provided by an exemplary embodiment of this specification. Figure 6 As shown in the figure, the implementation process of the text similarity model includes the following steps:

[0136] Step 602: All sentences in the description text corresponding to each change are concatenated according to a first preset rule to obtain at least one target text corresponding to the description text.

[0137] Specifically, the description text corresponding to each change includes multiple sentences. All sentences in the description text corresponding to each change can be spliced ​​according to the first preset rule to obtain at least one target text corresponding to the description text. The target text includes at least one sentence in the description text. The first preset rule can be to splice multiple sentences in the description text corresponding to each change in the order in the description text until the length of the spliced ​​target text is closest to or equal to the preset length. The sentence is a short sentence with a length less than a preset threshold. The preset threshold can be any value greater than 1 and less than or equal to 512, such as 256, 200, 512, etc., which is not limited to this in this specification. The first preset rule can also be to splice the sentences or characters in the description text in the order in the description text until the length of the spliced ​​target text is closest to or equal to the preset length, which is not limited to this in this specification embodiment. The preset length can be the maximum text length 512 trained by the BERT text similarity model, or it can be any length less than 512, such as 256, 400, etc., which is not limited to this in this specification embodiment.

[0138] Step 604 , calculating the second word overlap score between the user complaint text and each target text corresponding to the description text.

[0139] Specifically, the word overlap score can be performed between the user complaint text and each target text in at least one target text corresponding to each description text, that is, the score of the same or similar words between the user complaint text and each target text in at least one target text corresponding to each description text, i.e., the second word overlap score, is calculated. The second word overlap score is used to characterize the correlation between the user complaint text and each target text corresponding to the description text.

[0140] Step 606: extract M target texts with the highest second word overlap scores from at least one target text corresponding to the description text.

[0141] Specifically, after calculating the second word overlap score between the user complaint text and each target text in at least one target text corresponding to each changed description text, M target texts with the highest second word overlap score in at least one target text corresponding to each changed description text can be extracted. The above M is a positive integer, and the above M is less than or equal to the total number of all target texts corresponding to the description text.

[0142] For example, if 20 target texts are obtained by concatenating sentences in a description text corresponding to a certain change, three target texts with the highest second word overlap scores may be extracted from the 20 target texts corresponding to the description text.

[0143] Step 608 , calculating the text similarity between the description text corresponding to each change and the user complaint text according to the M target texts corresponding to the description text of each change.

[0144] Optionally, the M text similarities between the M target texts with the highest second word overlap scores corresponding to each description text and the user complaint text can be calculated first, that is, the M target texts with the highest second word overlap scores corresponding to each description text can be input into the BERT text similarity model together with the user complaint text, thereby outputting the M text similarities between the M target texts corresponding to each description text and the user complaint text. Then, the text similarity between the description text corresponding to each change and the user complaint text is determined based on the K highest text similarities among the M text similarities. The above K is a positive integer greater than 0 and less than or equal to M.

[0145] Exemplarily, the average of the K highest text similarities among the M text similarities corresponding to each of the above changes can be calculated, and the above average value can be determined as the text similarity between the description text corresponding to each of the above changes and the above user complaint text; or the K highest text similarities among the M text similarities corresponding to each of the above changes can be weightedly summed to obtain the text similarity between the description text corresponding to each of the above changes and the above user complaint text, which is not limited to the embodiments of this specification.

[0146] Optionally, after calculating the M text similarities between the M target texts with the highest second word overlap scores corresponding to each change and the user complaint text, the text similarity between the description text corresponding to each change and the user complaint text can be determined based on the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text. That is, the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text can be weightedly fused to obtain the text similarity between the description text corresponding to each change and the user complaint text. The above K is a positive integer greater than 0 and less than or equal to M.

[0147] The embodiments of this specification do not segment according to the original periods, but instead calculate the second word overlap score by splicing the sentences in the description text corresponding to the change to a length close to or equal to the maximum length trained by the BERT text similarity model, and calculate the text similarity between the description text corresponding to each change and the user complaint text based on the M target texts with the highest second word overlap scores corresponding to the description text of each change. This avoids the situation where text semantics are segmented due to multiple periods when writing, and reduces the amount of calculation for calculating the second word overlap score. This not only improves the efficiency of locating fault changes, but also can further improve the accuracy of calculating text similarity, thereby further improving the accuracy of fault change locating.

[0148] Since the unsupervised BERT text similarity model does not miss any possibility, that is, it tries to associate user complaint text with changes, and a user's description may not be comprehensive, the technicians who monitor the market need to check a large number of fault changes that are suspected to be associated with multiple similar but different user complaint texts every day. Figure 1-Figure 2 , introduces the fault change location method provided by the embodiment of this specification. For details, please refer to Figure 7 , which is a flow chart of another fault change location method provided by an exemplary embodiment of this specification. Figure 7 As shown, the fault change location method includes the following steps:

[0149] Step 702, obtaining the user complaint text.

[0150] Specifically, step 702 is consistent with step 302 and will not be repeated here.

[0151] Step 704: Determine a set of changes within a preset time period based on the complaint time of the user complaint text.

[0152] Specifically, step 704 is consistent with step 304 and will not be repeated here.

[0153] Step 706: Obtain the description text corresponding to each change in the change set.

[0154] Specifically, step 706 is consistent with step 306 and will not be repeated here.

[0155] Step 708: Determine X historical user complaint texts complained before the complaint time of the user complaint text.

[0156] Specifically, X historical user complaint texts that are before the complaint time of the current user complaint text and whose complaint time is closest to the complaint time of the user complaint text may be found from all the acquired user complaint texts, where X is a positive integer.

[0157] Exemplarily, the above X can be 50, 100, 200, etc., which is not limited in the embodiments of this specification.

[0158] Step 710: Calculate the text similarity between the user complaint text and each of the X historical user complaint texts.

[0159] Specifically, the user complaint text and X historical user complaint texts may be input into the BERT text similarity model trained in the above embodiment, thereby outputting the text similarity between the user complaint text and each of the X historical user complaint texts.

[0160] Step 712: If the text similarity between the Y historical user complaint texts and the user complaint text is greater than or equal to the second preset threshold, the Y historical user complaint texts are concatenated with the user complaint text according to a third preset rule to obtain a target complaint text.

[0161] Specifically, all the historical user complaint texts among the X historical user complaint texts whose text similarity with the user complaint text is greater than or equal to the second preset threshold, that is, Y historical user complaint texts, can be spliced ​​with the user complaint text according to the third preset rule to obtain the target complaint text. The above-mentioned third preset rule can be to splice the above-mentioned Y historical user complaint texts with the user complaint text in the order of the complaint time, or to splice the above-mentioned Y historical user complaint texts with the user complaint text in any order, or to filter out stop words in the above-mentioned Y historical user complaint texts and the user complaint text, and then splice them in any order, etc., which is not limited in the embodiments of this specification. The above-mentioned Y is a positive integer greater than 0 and less than or equal to X. The above-mentioned second preset threshold can be 0.6, 0.9, etc., which is not limited in the embodiments of this specification.

[0162] Step 714, calculating the text similarity between the description text corresponding to each change and the target complaint text.

[0163] Specifically, the description text corresponding to each change and the target complaint text can be directly input into the BERT text similarity model, so as to output the text similarity between the description text corresponding to each change and the target complaint text. The above BERT text similarity model is trained by the description texts corresponding to multiple changes, multiple user complaint texts, and the text similarity between the description text corresponding to each change and each of the user complaint texts.

[0164] Step 716: If the text similarity between the description text corresponding to the change and the target complaint text is greater than or equal to the first preset threshold, the change is determined as a fault change corresponding to the user complaint text and Y historical user complaint texts.

[0165] Specifically, the change corresponding to the description text whose text similarity with the target complaint text is greater than or equal to the first preset threshold can be determined as the target text, i.e., the fault change corresponding to the user complaint text and Y historical user complaint texts. That is, it may be due to the fault change that the users corresponding to the user complaint text and the Y historical user complaint texts have problems when using the software, thereby initiating a complaint, i.e., generating the user complaint text and the Y historical user complaint texts. The first preset threshold can be 0.1, 0.9, etc., which is not limited in this specification. The fault changes corresponding to the user complaint text and the Y historical user complaint texts include at least one change, and the number of the fault changes is less than or equal to the number of fault changes determined only based on the user complaint text. After determining at least one fault change related to the user complaint text, the server can send the user complaint text and the Y historical user complaint texts and at least one fault change corresponding thereto to the technicians corresponding to each of the fault changes for troubleshooting, rollback and repair.

[0166] In the embodiment of this specification, Y historical user complaint texts complained before the complaint time of this user complaint text and whose text similarity with the user complaint text is greater than or equal to the second preset threshold are spliced ​​with the above-mentioned user complaint text into a target complaint text according to a third preset rule, and the fault changes corresponding to the above-mentioned user complaint text and the above-mentioned X historical user complaint texts are determined by calculating the text similarity between the description text corresponding to each of the above-mentioned changes and the above-mentioned target complaint text. This can avoid the problem of inaccurate positioning due to incomplete description of a user or too many fault changes suspected to be associated with the user complaint text, thereby improving the accuracy of fault change positioning, and can process multiple "similar" user complaint texts at the same time, greatly reducing the workload of technical personnel.

[0167] Please refer to Figure 8 , Figure 8 A fault change location device is provided as an exemplary embodiment of this specification. The fault change location device 800 includes:

[0168] An acquisition module 810 is used to acquire a user complaint text;

[0169] A first determining module 820, configured to determine a set of changes within a preset time period based on the complaint time of the user complaint text;

[0170] The acquisition module 810 is further used to acquire a description text corresponding to each change in the change set;

[0171] A first calculation module 830 is used to calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text;

[0172] The second determination module 840 is used to determine the change as a fault change corresponding to the user complaint text if the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold.

[0173] In a possible implementation, the description text includes at least one sentence;

[0174] The first calculation module 830 includes:

[0175] A first calculation unit is used to calculate the first word overlap score between the user complaint text and each sentence in the description text corresponding to each change; the first word overlap score is used to characterize the correlation between the user complaint text and each sentence in the description text corresponding to each change;

[0176] A first extraction unit is used to extract N sentences with the highest first word overlap scores from the description text corresponding to each change, where N is a positive integer;

[0177] The second calculation unit is used to calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text according to the above N sentences corresponding to each of the above changes.

[0178] In a possible implementation manner, the second computing unit is specifically configured to:

[0179] Calculate N text similarities between the N sentences corresponding to each of the changes and the user complaint text;

[0180] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the L highest text similarities among the above N text similarities; the above L is a positive integer greater than 0 and less than or equal to N.

[0181] In a possible implementation, when the second calculation unit determines the text similarity between the description text corresponding to each change and the user complaint text according to the L highest text similarities among the N text similarities, it is specifically used to:

[0182] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text.

[0183] In a possible implementation, the description text includes a plurality of sentences;

[0184] The first calculation module 830 further includes:

[0185] A concatenation unit, configured to concatenate all the sentences in the description text corresponding to each of the above changes according to a first preset rule to obtain at least one target text corresponding to the above description text; the above target text includes at least one of the above sentences;

[0186] The third calculation unit is used to calculate the second word overlap score between the user complaint text and each target text corresponding to the description text; the second word overlap score is used to characterize the correlation between the user complaint text and each target text corresponding to the description text;

[0187] A second extraction unit is used to extract M target texts with the highest second word overlap scores from the at least one target text corresponding to the description text; the M is a positive integer;

[0188] The fourth calculation unit is used to calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text according to the above M target texts corresponding to the description text of each of the above changes.

[0189] In a possible implementation manner, the fourth computing unit is specifically configured to:

[0190] Calculate the M text similarities between the M target texts corresponding to each of the changes and the user complaint texts;

[0191] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the K highest text similarities among the above M text similarities; the above K is a positive integer greater than 0 and less than or equal to M.

[0192] In a possible implementation, when determining the text similarity between the description text corresponding to each change and the user complaint text according to the K highest text similarities among the M text similarities, the fourth calculation unit is specifically used to:

[0193] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text.

[0194] In a possible implementation, the fault change location device 800 further includes:

[0195] A filtering module, used to filter the user complaint text and the description text corresponding to each of the changes according to a preset stop word library and a second preset rule to obtain the filtered user complaint text and the description text corresponding to each of the filtered changes;

[0196] The first calculation module 830 is specifically used for:

[0197] The text similarity between the description text corresponding to each filtered change and the filtered user complaint text is calculated.

[0198] In a possible implementation, the fault change location device 800 further includes:

[0199] The third determination module is used to determine X historical user complaint texts complained before the complaint time of the user complaint text; the above X is a positive integer;

[0200] A second calculation module is used to calculate the text similarity between the user complaint text and each of the X historical user complaint texts;

[0201] A splicing module, configured to splice the Y historical user complaint texts with the user complaint text according to a third preset rule to obtain a target complaint text if the text similarity between the Y historical user complaint texts and the user complaint text is greater than or equal to a second preset threshold; the Y is a positive integer greater than 0 and less than or equal to X;

[0202] The first calculation module 830 is further used to calculate the text similarity between the description text corresponding to each change and the target complaint text;

[0203] The second determination module 840 is also used to determine the change as a fault change corresponding to the user complaint text and the Y historical user complaint texts if the text similarity between the description text corresponding to the change and the target complaint text is greater than or equal to a first preset threshold.

[0204] The division of each module in the above-mentioned fault change location device is only for illustration. In other embodiments, the fault change location device can be divided into different modules as needed to complete all or part of the functions of the above-mentioned fault change location device. The implementation of each module in the fault change location device provided in the embodiment of this specification can be in the form of a computer program. The computer program can be run on a terminal or a server. The program modules constituted by the computer program can be stored in the memory of the terminal or the server. When the computer program is executed by the processor, all or part of the steps of the fault change location method described in the embodiment of this specification are implemented.

[0205] See also Fig. 9 , Fig. 9 This is a schematic diagram of the structure of an electronic device provided by an exemplary embodiment of this specification. Fig. 9 As shown, the electronic device 900 may include: at least one processor 910 , at least one communication bus 920 , a user interface 930 , at least one network interface 940 , and a memory 950 .

[0206] The communication bus 920 may be used to realize the connection and communication among the above-mentioned components.

[0207] The user interface 930 may include a display screen (Display) and a camera (Camera), and the optional user interface may also include a standard wired interface and a wireless interface.

[0208] The network interface 940 may optionally include a Bluetooth module, a Near Field Communication (NFC) module, a Wireless Fidelity (Wi-Fi) module, and the like.

[0209] Among them, the processor 910 may include one or more processing cores. The processor 910 uses various interfaces and lines to connect various parts within the entire electronic device 900, and executes various functions and processes data of the routing electronic device 900 by running or executing instructions, programs, code sets or instruction sets stored in the memory 950, and calling data stored in the memory 950. Optionally, the processor 910 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 910 can integrate one or more combinations of a processor (Central Processing Unit, CPU), a graphics processor (Graphics Processing Unit, GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 910, but may be implemented separately through a chip.

[0210] The memory 950 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 950 includes a non-transitory computer-readable medium. The memory 950 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 950 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as an acquisition function, a calculation function, a determination function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 950 may optionally be at least one storage device located away from the aforementioned processor 910. As Fig. 9 As shown, the memory 950 as a computer storage medium may include an operating system, a network communication module, a user interface module, and program instructions.

[0211] Specifically, the processor 910 may be used to call the program instructions stored in the memory 950, and specifically perform the following operations:

[0212] Get the user complaint text.

[0213] Based on the complaint time of the user complaint text, a set of changes within a preset time period is determined.

[0214] Get the description text corresponding to each change in the above changeset.

[0215] Calculate the text similarity between the description text corresponding to each of the above changes and the above user complaint text.

[0216] If the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, the change is determined as a fault change corresponding to the user complaint text.

[0217] In some possible embodiments, the above description text includes at least one sentence.

[0218] When the processor 910 calculates the text similarity between the description text corresponding to each change and the user complaint text, it is specifically used to perform:

[0219] Calculate the first word overlap score between the user complaint text and each sentence in the description text corresponding to each change; the first word overlap score is used to characterize the correlation between the user complaint text and each sentence in the description text corresponding to each change.

[0220] N sentences with the highest first word overlap scores are extracted from the description text corresponding to each of the above changes; the above N is a positive integer.

[0221] The text similarity between the description text corresponding to each of the changes and the user complaint text is calculated based on the N sentences corresponding to each of the changes.

[0222] In some possible embodiments, when the processor 910 calculates the text similarity between the description text corresponding to each change and the user complaint text according to the N sentences corresponding to each change, it is specifically configured to execute:

[0223] Calculate the N text similarities between the N sentences corresponding to each of the changes and the user complaint text.

[0224] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the L highest text similarities among the above N text similarities; the above L is a positive integer greater than 0 and less than or equal to N.

[0225] In some possible embodiments, when the processor 910 determines the text similarity between the description text corresponding to each change and the user complaint text according to the L highest text similarities among the N text similarities, it is specifically configured to execute:

[0226] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text.

[0227] In some possible embodiments, the above description text includes multiple sentences.

[0228] When the processor 910 calculates the text similarity between the description text corresponding to each change and the user complaint text, it is specifically used to perform:

[0229] All sentences in the description text corresponding to each of the above changes are spliced ​​according to a first preset rule to obtain at least one target text corresponding to the above description text; the above target text includes at least one of the above sentences.

[0230] The second word overlap score between the user complaint text and each target text corresponding to the description text is calculated; the second word overlap score is used to characterize the correlation between the user complaint text and each target text corresponding to the description text.

[0231] M target texts with the highest second word overlap scores are extracted from the at least one target text corresponding to the description text; M is a positive integer.

[0232] The text similarity between the description text corresponding to each of the changes and the user complaint text is calculated based on the M target texts corresponding to the description text of each of the changes.

[0233] In some possible embodiments, when the processor 910 calculates the text similarity between the description text corresponding to each change and the user complaint text according to the M target texts corresponding to the description text of each change, it is specifically configured to execute:

[0234] Calculate the M text similarities between the M target texts corresponding to each of the above changes and the user complaint texts.

[0235] The text similarity between the description text corresponding to each of the above changes and the above user complaint text is determined based on the K highest text similarities among the above M text similarities; the above K is a positive integer greater than 0 and less than or equal to M.

[0236] In some possible embodiments, when the processor 910 determines the text similarity between the description text corresponding to each change and the user complaint text according to the K highest text similarities among the M text similarities, it is specifically configured to execute:

[0237] The text similarity between the description text corresponding to each change and the user complaint text is determined based on the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text.

[0238] In some possible embodiments, after receiving the user complaint text and before calculating the text similarity between the description text corresponding to each change and the user complaint text, the processor 910 is further configured to execute:

[0239] The user complaint text and the description text corresponding to each of the changes are filtered according to a preset stop word library and a second preset rule to obtain filtered user complaint text and filtered description text corresponding to each of the changes.

[0240] When the processor 910 calculates the text similarity between the description text corresponding to each change and the user complaint text, it is specifically used to perform:

[0241] The text similarity between the description text corresponding to each filtered change and the filtered user complaint text is calculated.

[0242] In some possible embodiments, after receiving the user complaint text and before calculating the text similarity between the description text corresponding to each change and the user complaint text, the processor 910 is further configured to execute:

[0243] Determine X historical user complaint texts complained before the complaint time of the user complaint text; the X is a positive integer.

[0244] The text similarity between the user complaint text and each of the X historical user complaint texts is calculated.

[0245] If there are Y historical user complaint texts whose text similarity with the above user complaint text is greater than or equal to the second preset threshold, the above Y historical user complaint texts are spliced ​​with the above user complaint text according to the third preset rule to obtain the target complaint text; the above Y is a positive integer greater than 0 and less than or equal to X.

[0246] When the processor 910 calculates the text similarity between the description text corresponding to each change and the user complaint text, it is specifically used to perform:

[0247] Calculate the text similarity between the description text corresponding to each of the above changes and the above target complaint text.

[0248] When the processor 910 determines the change as a fault change corresponding to the user complaint text if the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, the processor 910 is specifically configured to execute:

[0249] If the text similarity between the description text corresponding to the change and the above-mentioned target complaint text is greater than or equal to the first preset threshold, the above-mentioned change is determined as the fault change corresponding to the above-mentioned user complaint text and the above-mentioned Y historical user complaint texts.

[0250] The embodiments of this specification also provide a computer-readable storage medium, which stores instructions, and when the instructions are executed on a computer or a processor, the computer or the processor executes one or more steps in the above embodiments. If the components of the above fault change location device are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium.

[0251] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The above computer program product includes one or more computer instructions. When the above computer program instructions are loaded and executed on a computer, the above process or function according to the embodiment of this specification is generated in whole or in part. The above computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above computer instructions can be stored in a computer-readable storage medium or transmitted by the above computer-readable storage medium. The above computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode. The above computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, data center, etc. that contains one or more available media integrated. The above-mentioned available media can be magnetic media (for example, floppy disks, hard disks, tapes), optical media (for example, digital versatile discs (DVD)), or semiconductor media (for example, solid state disks (SSD)), etc.

[0252] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes. In the absence of conflict, the technical features in this embodiment and the implementation scheme can be combined arbitrarily.

[0253] The embodiments described above are merely preferred embodiments of this specification and are not intended to limit the scope of this specification. Without departing from the design spirit of this specification, various modifications and improvements made to the technical solutions of this specification by ordinary technicians in this field should fall within the scope of protection determined by the claims.

[0254] The above is a description of a specific embodiment of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A fault change location method, the method comprising: Get user complaint text; Determining a set of changes within a preset time period based on the complaint time of the user complaint text; Obtaining description text corresponding to each change in the change set; Calculating the text similarity between the description text corresponding to each change and the user complaint text; If the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, the change is determined as a fault change corresponding to the user complaint text.

2. The method of claim 1, wherein the description text comprises at least one sentence; The calculating the text similarity between the description text corresponding to each change and the user complaint text includes: Calculating a first word overlap score between the user complaint text and each sentence in the description text corresponding to each change; the first word overlap score is used to characterize the correlation between the user complaint text and each sentence in the description text corresponding to each change; Extracting N sentences with the highest overlap scores of the first word from the description text corresponding to each change; wherein N is a positive integer; The text similarity between the description text corresponding to each change and the user complaint text is calculated based on the N sentences corresponding to each change.

3. The method according to claim 2, wherein the calculating the text similarity between the description text corresponding to each change and the user complaint text according to the N sentences corresponding to each change comprises: Calculating N text similarities between the N sentences corresponding to each change and the user complaint text respectively; The text similarity between the description text corresponding to each change and the user complaint text is determined according to the L highest text similarities among the N text similarities; L is a positive integer greater than 0 and less than or equal to N.

4. The method according to claim 3, wherein determining the text similarity between the description text corresponding to each change and the user complaint text according to the L highest text similarities among the N text similarities comprises: The text similarity between the description text corresponding to each change and the user complaint text is determined based on the L highest text similarities among the N text similarities and the L first word overlap scores between the L sentences corresponding to the L highest text similarities and the user complaint text.

5. The method of claim 1, wherein the description text comprises a plurality of sentences; The calculating the text similarity between the description text corresponding to each change and the user complaint text includes: splicing all the sentences in the description text corresponding to each change according to a first preset rule to obtain at least one target text corresponding to the description text; The target text includes at least one of the sentences; Calculating a second word overlap score between the user complaint text and each target text corresponding to the description text; the second word overlap score is used to characterize the correlation between the user complaint text and each target text corresponding to the description text; Extracting M target texts with the highest overlap scores of the second word from the at least one target text corresponding to the description text; wherein M is a positive integer; The text similarity between the description text corresponding to each change and the user complaint text is calculated based on the M target texts corresponding to the description text of each change.

6. The method according to claim 5, wherein the step of calculating the text similarity between the description text corresponding to each change and the user complaint text according to the M target texts corresponding to the description text of each change comprises: Calculating M text similarities between the M target texts corresponding to each change and the user complaint text respectively; The text similarity between the description text corresponding to each change and the user complaint text is determined according to the K highest text similarities among the M text similarities; K is a positive integer greater than 0 and less than or equal to M.

7. The method according to claim 6, wherein determining the text similarity between the description text corresponding to each change and the user complaint text according to the K highest text similarities among the M text similarities comprises: The text similarity between the description text corresponding to each change and the user complaint text is determined based on the K highest text similarities among the M text similarities and the K second word overlap scores between the K target texts corresponding to the K highest text similarities and the user complaint text.

8. The method according to claim 1, after receiving the user complaint text and before calculating the text similarity between the description text corresponding to each change and the user complaint text, the method further comprises: Filter the user complaint text and the description text corresponding to each change according to a preset stop word library and a second preset rule to obtain a filtered user complaint text and a filtered description text corresponding to each change; The calculating the text similarity between the description text corresponding to each change and the user complaint text includes: The text similarity between the description text corresponding to each filtered change and the filtered user complaint text is calculated.

9. The method according to claim 1, after receiving the user complaint text and before calculating the text similarity between the description text corresponding to each change and the user complaint text, the method further comprises: Determine X historical user complaint texts complained before the complaint time of the user complaint text; X is a positive integer; Calculating the text similarity between the user complaint text and each of the X historical user complaint texts; If there are Y historical user complaint texts whose text similarity with the user complaint text is greater than or equal to the second preset threshold, the Y historical user complaint texts are concatenated with the user complaint text according to a third preset rule to obtain a target complaint text; Y is a positive integer greater than 0 and less than or equal to X; The calculating the text similarity between the description text corresponding to each change and the user complaint text includes: Calculating the text similarity between the description text corresponding to each change and the target complaint text; If the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold, determining the change as a fault change corresponding to the user complaint text includes: If the text similarity between the description text corresponding to the change and the target complaint text is greater than or equal to a first preset threshold, the change is determined as a fault change corresponding to the user complaint text and the Y historical user complaint texts.

10. A fault change positioning device, the device comprising: Acquisition module, used to obtain user complaint text; A first determination module, configured to determine a set of changes within a preset time period based on the complaint time of the user complaint text; The acquisition module is also used to obtain a description text corresponding to each change in the change set; A calculation module, used for calculating the text similarity between the description text corresponding to each change and the user complaint text; The second determination module is used to determine the change as a fault change corresponding to the user complaint text if the text similarity between the description text corresponding to the change and the user complaint text is greater than or equal to a first preset threshold.

11. An electronic device, comprising: Processor and memory; The processor is connected to the memory; The memory is used to store executable program code; The processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the method according to any one of claims 1 to 9.

12. A computer storage medium storing a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the method steps according to any one of claims 1 to 9.

13. A computer program product comprising instructions, which, when executed on a computer or a processor, enables the computer or the processor to execute the fault change location method according to any one of claims 1 to 9.

Images