An accident monitoring system and method based on semantic analysis of group chat content

Through an accident monitoring system based on semantic analysis of group chat content, semantic analysis and accident knowledge base are used to solve the problem of unclear user feedback in the e-commerce system, the accurate positioning and rapid resolution of accidents are achieved, and the user experience and problem resolution are improved. efficiency.

CN114330361BActive Publication Date: 2024-11-261DRUG TECH (SHANGHAI) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111509155.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-10
Publication Date
2024-11-26
Estimated Expiration
2041-12-10

AI Technical Summary

Technical Problem

The detection and monitoring methods of the existing e-commerce system are not perfect enough, resulting in unclear feedback on the problems encountered by users when shopping online, making it difficult to accurately locate the cause of the accident, resulting in poor user experience and tedious communication and time-consuming technicians.

Method used

Design an accident monitoring system based on semantic analysis of group chat content. By obtaining user chat records, semantic analysis, matching accident objects and types, evaluating accident reliability, outputting early warning information, and using a pre-established accident knowledge base to assist technicians quickly Positioning and solving problems.

Benefits of technology

It realizes accurate positioning of user feedback and identification of accident causes, reduces the number of communication between technicians and users, improves problem-solving efficiency, and enhances user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114330361B_ABST
    Figure CN114330361B_ABST
Patent Text Reader

Abstract

The present invention discloses an accident monitoring system and method based on semantic analysis of group chat content, which is applicable to online shopping platforms and belongs to the field of e-commerce, and includes: an accident knowledge base, an acquisition unit, a parsing unit, a prejudgment unit, and a monitoring management unit. The beneficial effects of the present invention are: the present invention can accurately locate the accident by analyzing the records of user feedback, avoid the technicians' incorrect judgment of the cause of the accident due to the user's unclear expression or wrong understanding of the accident, and obtain the various causes that may cause the accident from the existing accident knowledge base, assist the technicians to repair as soon as possible, solve the user's problems, and enhance the user experience; at the same time, the technicians do not need to communicate with the feedback users one by one, saving time and improving the efficiency of problem solving.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of e-commerce, and in particular to an accident monitoring system and method based on semantic analysis of group chat content. Background Art

[0002] The customer service chat group is established to promptly resolve problems or troubles encountered by users in the customer service chat group when shopping online, so as to better serve users and enable them to have a better experience when shopping online. The customer service chat group is equipped with customer service personnel. Usually, when users encounter problems or have unclear points when shopping online, they can raise questions through the customer service chat group, and the customer service personnel will respond and handle them in a timely manner.

[0003] At present, the detection and monitoring methods of e-commerce systems are not perfect, and there are loopholes or bugs that escape to online shopping platforms. Therefore, when users find abnormalities in the business system, they usually report the problem in the customer service chat group. However, due to the imperfect data model and the non-technical background of users or customer service personnel, they cannot clearly understand the actual problem of the feedback. The content of the feedback will mislead technical personnel to locate the cause of the accident and fail to solve the problem in time, resulting in a poor user experience. Technical support personnel are usually required to communicate with the feedback users one by one by phone or group chat, which is cumbersome and time-consuming. Therefore, in response to the above problems, it is urgent to design an accident monitoring system and method based on semantic analysis of group chat content to meet the needs of actual use. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an accident monitoring system and method based on semantic analysis of group chat content.

[0005] The technical problem solved by the present invention can be achieved by adopting the following technical solutions:

[0006] The present invention provides an accident monitoring system based on semantic analysis of group chat content, which is applicable to an online shopping platform; the accident monitoring system is connected to a pre-established accident knowledge base, the accident knowledge base is used to store historical accident data, the historical accident data includes accident objects, object relationships between the accident objects, accident types formed by the accident objects and the object relationships, and multiple accident causes corresponding to the accident types;

[0007] The accident monitoring system comprises:

[0008] an acquisition unit, configured to acquire the user's chat records and corresponding time from a customer service chat group of the online shopping platform;

[0009] a parsing unit, connected to the acquisition unit, configured to perform semantic parsing on the chat record and output a parsing result;

[0010] an accident location unit, connected to the analysis unit, for matching the accident object and the accident type of the current accident corresponding to the chat record in the accident knowledge base according to the analysis result, locating the occurrence time of the accident, and finally outputting an accident location information;

[0011] a prediction unit, connected to the accident location unit, for obtaining the frequency of feedback from multiple users in the customer service chat group on the current accident within a predetermined time, evaluating the reliability of the current accident according to the feedback frequency, and outputting a prediction result;

[0012] A monitoring management unit is respectively connected to the accident location unit and the prediction unit, and is used to output an early warning information according to the prediction result when the reliability level of the current accident meets a preset condition. The early warning information includes the accident location information and the reliability level corresponding to the current accident.

[0013] Preferably, the parsing unit comprises:

[0014] a feature extraction module, used to extract feature keywords of the chat record, wherein the feature keywords include entity keywords and entity relationship keywords;

[0015] a first query module, connected to the feature extraction module, for querying the accident object corresponding to the entity keyword from the accident knowledge base according to the extracted entity keyword as a target object;

[0016] a second query module, connected to the feature extraction module and the first query module, respectively, for querying and obtaining all the matching object relationships as target object relationships from the accident knowledge base according to the extracted entity relationship keywords and the target object obtained by querying, and then obtaining the accident type constituted by the target object and the target object relationship;

[0017] The output module is used to include the target object, the accident type and the time corresponding to the chat record obtained by the query in the analysis result output.

[0018] Preferably, for each of the analysis results, the accident location unit uses the analysis result as a reference analysis result, and performs secondary location on the accident object and the occurrence time of the accident according to the accident type in the reference analysis result;

[0019] Each of the users corresponds to a plurality of associable objects, wherein the associable objects represent objects that the user has permission to use on the online shopping platform;

[0020] The accident location unit comprises:

[0021] A first positioning module, configured to determine the accident object corresponding to the current accident according to the multiple associable objects and the benchmark analysis results respectively corresponding to all the users who generated chat records in the customer service chat group within a preset time, and output the accident object and the accident type in the benchmark analysis results in a first positioning information;

[0022] a second positioning module, connected to the first positioning module, configured to determine the occurrence time of the accident corresponding to the accident type according to the analysis results of the plurality of users corresponding to the accident object determined by the first positioning module, and output a second positioning information;

[0023] The accident location information includes the first location information and the second location information.

[0024] Preferably, the first positioning module includes:

[0025] A selection submodule, configured to process the obtained repetition rates of the associable objects and sort them from high to low, and select a plurality of the associable objects with the highest ranking of the repetition rates as the accident objects corresponding to the currently occurring accident and output them;

[0026] A matching submodule, connected to the selection submodule, configured to match and output at least one type of accident cause corresponding to the accident type from the accident knowledge base according to the accident type in the benchmark analysis result;

[0027] The first positioning information includes a plurality of the accident objects, the accident type in the benchmark analysis result, and at least one type of the accident cause corresponding to the accident type.

[0028] Preferably, the second positioning module includes the earliest accident occurrence time in the analysis results as the occurrence time of the current accident in the second positioning information and outputs it based on the multiple analysis results.

[0029] Preferably, the prediction unit includes:

[0030] A calculation module, used for obtaining a reliability value of the current accident according to the predetermined time and the frequency of the user's feedback on the current accident within the predetermined time;

[0031] A judgment module, connected to the calculation module, for matching the reliability values ​​according to a plurality of preset accident reliability value intervals, and determining the reliability degree of the current accident according to the matching result;

[0032] Each of the accident reliability value intervals corresponds to a preset reliability level.

[0033] Preferably, the calculation module calculates the reliability value of the current accident according to the following formula:

[0034] n = tan (m / t);

[0035] in,

[0036] t represents the predetermined time;

[0037] m represents the frequency of all the users reporting the current accident within the predetermined time;

[0038] n represents the reliability value of the current accident.

[0039] Preferably, it also includes:

[0040] A plurality of clients are remotely connected to the monitoring management unit respectively, and each of the clients is carried by a technician to receive the warning information.

[0041] Preferably, the acquisition unit acquires the chat record of the user and the corresponding time by calling a calling interface of the customer service chat group.

[0042] The present invention also provides an accident monitoring method based on semantic analysis of group chat content, including the accident monitoring system based on semantic analysis of group chat content as described above, the accident monitoring method comprising:

[0043] Step S1, obtaining the user's chat history and corresponding time from the customer service chat group of the online shopping platform;

[0044] Step S2, performing semantic analysis on the chat record and outputting the analysis result;

[0045] Step S3, according to the analysis result, matching the accident object and the accident type of the current accident corresponding to the chat record in the accident knowledge base, locating the occurrence time of the accident, and finally outputting accident location information;

[0046] Step S4, obtaining the frequency of feedback of the current accident by multiple users in the customer service chat group within a predetermined time, evaluating the reliability of the current accident according to the feedback frequency, and outputting a prediction result;

[0047] Step S5, according to the prediction result, when the reliability level of the current accident meets a preset condition, output a warning information, the warning information including the accident location information and the reliability level corresponding to the current accident.

[0048] The beneficial effects of the present invention are:

[0049] The present invention can accurately locate accidents by analyzing records of user feedback, avoid technicians' erroneous judgment of the cause of the accident due to unclear expressions of users or wrong cognition of the accident, and obtain various possible causes of the accident from an existing accident knowledge base, so as to assist technicians in repairing and solving users' problems as quickly as possible and enhance user experience. At the same time, technicians do not need to communicate with users who have given feedback one by one about the problem, which saves time and improves the efficiency of problem solving. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural block diagram of an accident monitoring system based on semantic analysis of group chat content in the present invention;

[0051] Figure 2 It is a structural block diagram of a specific embodiment of the parsing unit in the present invention;

[0052] Figure 3 It is a structural block diagram of a specific embodiment of the accident location unit in the present invention;

[0053] Figure 4 It is a structural block diagram of a specific embodiment of the first positioning module in the present invention;

[0054] Figure 5 It is a structural block diagram of a specific embodiment of the prediction unit in the present invention;

[0055] Figure 6 It is a flow chart of an accident monitoring method based on semantic analysis of group chat content in the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0058] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0059] The present invention provides an accident monitoring system based on semantic analysis of group chat content, which is applicable to online shopping platforms; it belongs to the field of e-commerce, such as Figure 1 As shown, the accident monitoring system is connected to a pre-established accident knowledge base 1, and the accident knowledge base 1 is used to store historical accident data, and the historical accident data includes accident objects, object relationships between accident objects, accident types composed of accident objects and object relationships, and multiple accident causes corresponding to the accident types;

[0060] The accident monitoring system includes:

[0061] An acquisition unit 2, used to acquire the user's chat record and corresponding time from the customer service chat group of the online shopping platform;

[0062] A parsing unit 3, connected to the acquisition unit 2, for semantically parsing the chat records and outputting the parsing results;

[0063] An accident location unit 4, connected to the parsing unit 3, is used to match the accident object and accident type of the current accident corresponding to the chat record in the accident knowledge base 1 according to the parsing result, locate the occurrence time of the accident, and finally output an accident location information;

[0064] A prediction unit 5, connected to the accident location unit 4, is used to obtain the frequency of multiple users in the customer service chat group reporting the current accident within a predetermined time, evaluate the reliability of the current accident according to the feedback frequency, and output a prediction result;

[0065] A monitoring management unit 6 is connected to the accident location unit 4 and the prediction unit 5 respectively, and is used to output an early warning information according to the prediction result when the reliability level of the current accident meets a preset condition. The early warning information includes the accident location information and reliability level corresponding to the current accident.

[0066] Users will have a series of problems when shopping. At this time, users will usually give feedback to customer service staff in the customer service chat group, but the customer service staff is limited, and both users and customer service staff are non-technical backgrounds, resulting in inefficient problem solving or failure to fundamentally solve the problem, resulting in a poor user experience.

[0067] Specifically, in this embodiment, in order to solve the above problems, the present invention establishes an accident knowledge base 1 in advance based on historical accident data, and the accident knowledge base 1 stores accident objects in historical accidents, object relationships between accident objects, accident types composed of accident objects and object relationships, multiple accident causes corresponding to accident types, and possible solutions for each accident cause;

[0068] The present invention obtains chat records of user feedback in a customer service chat group by calling an API interface of an operating system corresponding to an online shopping platform, performs semantic analysis on the feedback chat records one by one, converts natural language into machine language through natural language conversion and vocabulary mapping, accurately locates accidents, and predicts whether the occurrence of accidents is reliable. When it is predicted that the reliability of the occurrence of a certain accident currently meets preset conditions, an early warning is issued, and the positioning information is sent to technical personnel for maintenance. Without increasing the number of technical personnel, user problems can be quickly solved, thereby improving user experience.

[0069] It should be noted that, in this embodiment, professional knowledge training is required for users and customer service personnel to ensure the accuracy of user feedback issues.

[0070] As a preferred embodiment, the accident monitoring system, wherein: Figure 2 As shown, the parsing unit 3 includes:

[0071] A feature extraction module 31, used to extract feature keywords of chat records, wherein the feature keywords include entity keywords and entity relationship keywords;

[0072] A first query module 32, connected to the feature extraction module 31, is used to query the accident object corresponding to the entity keyword from the accident knowledge base 1 as the target object according to the extracted entity keyword;

[0073] a second query module 33, connected to the feature extraction module 31 and the first query module 32, for querying and obtaining all matching object relationships as target object relationships from the accident knowledge base 1 according to the extracted entity relationship keywords and the target object obtained by the query, and then obtaining the accident type composed of the target object and the target object relationship;

[0074] The output module 34 is used to include the target object, accident type and time corresponding to the chat record obtained by the query in the analysis result output.

[0075] Specifically, in this embodiment, the parsing unit 3 includes a feature extraction module 31, a first query module 32, a second query module 33, and an output module 34; feature keywords are extracted from a chat record of a certain user, for example, the chat record is "the product cannot be added to the shopping cart, solve it quickly", then the extracted entity keywords include "shopping cart", "product" and the corresponding "a certain user", and the entity relationship keywords include "cannot be added", and the corresponding "shopping cart" is queried from the accident knowledge base 1 according to the keyword "shopping cart" as the target object, and then according to the extracted keyword "cannot be added" and the "shopping cart" object obtained by the query, a second query is performed from the accident knowledge base 1 to obtain the object relationship of all matching keywords "cannot be added" and "shopping cart" objects, and then "add product abnormality" is obtained as the accident type, and the final output parsing result is {"time":"XX:XX","object":"shopping cart","accident type":"add product abnormality"}.

[0076] As a preferred embodiment, the accident monitoring system, wherein: Figure 3 As shown, for each analysis result, the accident location unit 4 uses the analysis result as a reference analysis result, and performs secondary location on the accident object and the occurrence time of the accident according to the accident type in the reference analysis result;

[0077] Each user corresponds to a plurality of associable objects, and the associable objects represent objects that the user has the right to use on the online shopping platform;

[0078] The accident location unit 4 comprises:

[0079] A first positioning module 41, configured to determine the accident object corresponding to the current accident according to a plurality of associable objects and benchmark analysis results respectively corresponding to all users who generate chat records in the customer service chat group within a preset time, and output the accident object and the accident type in the benchmark analysis result in a first positioning information;

[0080] a second positioning module 42, connected to the first positioning module 41, for determining the occurrence time of the accident corresponding to the accident type according to the analysis results of multiple users corresponding to the accident object determined by the first positioning module 41, and outputting a second positioning information;

[0081] The accident location information includes first location information and second location information.

[0082] Specifically, in this embodiment, since the user is not a technical person, the problem expressed by the user may be different from the fundamental problem. For example, "the product cannot be added to the shopping cart" may be due to a problem with the shopping cart, a network problem, or the user's account is blacklisted due to some bad credit behavior and cannot use the shopping cart permission, or other reasons. If only the object and accident type in the analysis result are used, the problem may not be effectively solved in the end. In this case, it is necessary to conduct a secondary positioning of the accident. According to all users who generate chat records in the customer service chat group within a preset time, the preset time is a shorter time limit of about 5 minutes. In a short period of time, the problems reported by these users are relatively concentrated. According to these users, The objects with permission to use are used to determine the accident objects that may correspond to the current accident and the time of occurrence of the accident. For example, there are four persons, A, B, C, and D, and the online shopping platform includes multiple objects, such as A, B, C, D, and E. Among them, the associable objects of A are A, B, and E, the associable objects of B are A, D, and E, the associable objects of C are A, B, and C, and the associable objects of D are A, C, and E. The first positioning module 41 outputs the accident objects as A and E. The second positioning module 42 determines the time of occurrence of the accident according to the analysis results of A, B, and D corresponding to the accident objects A and E through the "time" in the analysis results, and assists technical personnel in repairing the problem through further accident location information.

[0083] As a preferred embodiment, the accident monitoring system, wherein: Figure 4 As shown, the first positioning module 41 includes:

[0084] The selection submodule 411 is used to process the obtained repetition rates of the associable objects and sort them from high to low, and select a plurality of associable objects with the highest repetition rates as the accident objects corresponding to the current accident;

[0085] The matching submodule 412 is connected to the selection submodule 411 and is used to match and output at least one type of accident cause corresponding to the accident type from the accident knowledge base 1 according to the accident type in the benchmark analysis result;

[0086] The first positioning information includes multiple accident objects, accident types in the benchmark analysis results, and at least one type of accident cause corresponding to the accident type.

[0087] Specifically, in this embodiment, the repetition rate of the associable objects obtained by the selection submodule 411 is sorted from high to low, and the repetition rate is sorted from high to low as A, E, B, C, and D. The accident objects that may correspond to the current accident are A and E. The matching submodule 412 matches the accident type in the benchmark analysis result from the accident knowledge base 1 to obtain at least one type of accident cause corresponding to the accident type and the output of the solution corresponding to each type of accident cause.

[0088] As a preferred embodiment, in the accident monitoring system, the second positioning module 42 includes the earliest accident occurrence time in the analysis results as the occurrence time of the current accident in the second positioning information for output based on multiple analysis results.

[0089] As a preferred embodiment, the accident monitoring system, wherein: Figure 5 As shown, the prediction unit 5 includes:

[0090] A calculation module 51 is used to obtain a reliability value of the current accident according to a predetermined time and a frequency of the current accident reported by users within the predetermined time, wherein the predetermined time is a relatively short time, such as 3 minutes, 5 minutes, or 10 minutes;

[0091] The judgment module 52 is connected to the calculation module 51 and is used to match the reliability values ​​according to a plurality of preset accident reliability value intervals, and determine the reliability degree of the current accident according to the matching result;

[0092] Each accident reliability value interval corresponds to a preset reliability level, wherein the preset intervals include a first preset interval [a, b], a second preset interval [c, d], and a third preset interval [e, f], corresponding to high, medium, and low reliability, respectively. The values ​​of a, b, c, d, e, and f can be set according to different time periods, such as holidays, online shopping promotion festivals, or other reasons. For example, the first preset interval is [10, +∞], the second preset interval is [5, 10], and the third preset interval is [0, 5].

[0093] As a preferred embodiment, in the accident monitoring system, the calculation module 51 calculates the reliability value of the current accident according to the following formula:

[0094] n = tan (m / t);

[0095] in,

[0096] t represents the scheduled time;

[0097] m represents the frequency of current accidents reported by all users within the scheduled time;

[0098] n represents the reliability value of the current accident.

[0099] As a preferred embodiment, the accident monitoring system further includes:

[0100] Multiple clients (01, 02...03) are remotely connected to the monitoring management unit 6 respectively. Each client (01, 02...03) is carried by a technician to receive early warning information so that the technician can perform maintenance in time according to the accident location information and reliability level corresponding to the current accident in the early warning information.

[0101] As a preferred embodiment, in the accident monitoring system, the acquisition unit 2 acquires the user's chat history and the corresponding time by calling a calling interface of the customer service chat group.

[0102] The present invention also provides an accident monitoring method based on semantic analysis of group chat content, including the accident monitoring system based on semantic analysis of group chat content as described above, such as Figure 6 As shown, the accident monitoring method includes:

[0103] Step S1, obtaining the user's chat history and corresponding time from the customer service chat group of the online shopping platform;

[0104] Step S2, semantically analyzing the chat records and outputting the analysis results;

[0105] Step S3, according to the analysis result, the accident object and accident type of the current accident corresponding to the chat record are matched in the accident knowledge base 1, and the occurrence time of the accident is located, and finally an accident location information is output;

[0106] Step S4, obtaining the frequency of feedback from multiple users in the customer service chat group about the current accident within a predetermined time, evaluating the reliability of the current accident according to the feedback frequency, and outputting a prediction result;

[0107] Step S5, according to the prediction result, when the reliability level of the current accident meets a preset condition, output a warning information, the warning information includes the accident location information and reliability level corresponding to the current accident.

[0108] The beneficial effects of the present invention are:

[0109] The present invention can accurately locate accidents by analyzing records of user feedback, avoid technicians' erroneous judgment of the cause of the accident due to unclear expressions of users or wrong cognition of the accident, and obtain various possible causes of the accident from the existing accident knowledge base 1, so as to assist technicians in repairing and solving users' problems as soon as possible and enhance user experience; at the same time, technicians do not need to communicate with users who have given feedback one by one about the problem, which saves time and improves the efficiency of problem solving.

[0110] The above description is only a preferred embodiment of the present invention, and does not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. An accident monitoring system based on semantic analysis of group chat content, suitable for online shopping platforms; characterized in that: The accident monitoring system is connected to a pre-established accident knowledge base, the accident knowledge base is used to store historical accident data, the historical accident data includes accident objects, object relationships between the accident objects, accident types formed by the accident objects and the object relationships, and multiple accident causes corresponding to the accident types; The accident monitoring system comprises: an acquisition unit, configured to acquire the user's chat records and corresponding time from a customer service chat group of the online shopping platform; a parsing unit, connected to the acquisition unit, configured to perform semantic parsing on the chat record and output a parsing result; an accident location unit, connected to the analysis unit, for matching the accident object and the accident type of the current accident corresponding to the chat record in the accident knowledge base according to the analysis result, locating the occurrence time of the accident, and finally outputting an accident location information; a prediction unit, connected to the accident location unit, for obtaining the frequency of feedback from multiple users in the customer service chat group on the current accident within a predetermined time, evaluating the reliability of the current accident according to the feedback frequency, and outputting a prediction result; a monitoring management unit, connected to the accident location unit and the prejudgment unit, respectively, for outputting a warning message according to the prejudgment result when the reliability of the current accident meets a preset condition, wherein the warning message includes the accident location information and the reliability corresponding to the current accident; The parsing unit comprises: a feature extraction module, used to extract feature keywords of the chat record, wherein the feature keywords include entity keywords and entity relationship keywords; a first query module, connected to the feature extraction module, for querying the accident object corresponding to the entity keyword from the accident knowledge base according to the extracted entity keyword as a target object; a second query module, connected to the feature extraction module and the first query module, respectively, for querying and obtaining all the matching object relationships as target object relationships from the accident knowledge base according to the extracted entity relationship keywords and the target object obtained by querying, and then obtaining the accident type constituted by the target object and the target object relationship; An output module, configured to include the target object, the accident type, and the time corresponding to the chat record obtained by the query in the analysis result output; For each of the analysis results, the accident location unit uses the analysis result as a reference analysis result, and performs secondary location on the accident object and the occurrence time of the accident according to the accident type in the reference analysis result; Each of the users corresponds to a plurality of associable objects, wherein the associable objects represent objects that the user has permission to use on the online shopping platform; The accident location unit comprises: A first positioning module, configured to determine the accident object corresponding to the current accident according to the multiple associable objects and the benchmark analysis results respectively corresponding to all the users who generated chat records in the customer service chat group within a preset time, and output the accident object and the accident type in the benchmark analysis results in a first positioning information; a second positioning module, connected to the first positioning module, configured to determine the occurrence time of the accident corresponding to the accident type according to the analysis results of the plurality of users corresponding to the accident object determined by the first positioning module, and output a second positioning information; The accident location information includes the first location information and the second location information; The first positioning module includes: A selection submodule, configured to process the obtained repetition rates of the associable objects and sort them from high to low, and select a plurality of the associable objects with the highest ranking of the repetition rates as the accident objects corresponding to the currently occurring accident and output them; A matching submodule, connected to the selection submodule, configured to match and output at least one type of accident cause corresponding to the accident type from the accident knowledge base according to the accident type in the benchmark analysis result; The first positioning information includes a plurality of the accident objects, the accident type in the benchmark analysis result, and at least one type of the accident cause corresponding to the accident type.

2. The accident monitoring system based on semantic analysis of group chat content according to claim 1 is characterized in that: The second positioning module includes the earliest accident occurrence time in the analysis results as the occurrence time of the current accident in the second positioning information and outputs it according to the plurality of analysis results.

3. The accident monitoring system based on semantic analysis of group chat content according to claim 1 is characterized in that: The prediction unit comprises: A calculation module, used for obtaining a reliability value of the current accident according to the predetermined time and the frequency of the user's feedback on the current accident within the predetermined time; A judgment module, connected to the calculation module, for matching the reliability values ​​according to a plurality of preset accident reliability value intervals, and determining the reliability degree of the current accident according to the matching result; Each of the accident reliability value intervals corresponds to a preset reliability level.

4. The accident monitoring system based on semantic analysis of group chat content according to claim 3 is characterized in that: The calculation module calculates the reliability value of the current accident according to the following formula: ; in, t represents the predetermined time; m represents the frequency of all the users reporting the current accident within the predetermined time; n represents the reliability value of the current accident.

5. The accident monitoring system based on semantic analysis of group chat content according to claim 1 is characterized in that: Also includes: A plurality of clients are remotely connected to the monitoring management unit respectively, and each of the clients is carried by a technician to receive the warning information.

6. The accident monitoring system based on semantic analysis of group chat content according to claim 1 is characterized in that: The acquisition unit acquires the chat record of the user and the corresponding time by calling a calling interface of the customer service chat group.

7. An accident monitoring method based on semantic analysis of group chat content, characterized in that: The accident monitoring system based on semantic analysis of group chat content as claimed in any one of claims 1 to 6, wherein the accident monitoring method comprises: Step S1, obtaining the user's chat history and corresponding time from the customer service chat group of the online shopping platform; Step S2, performing semantic analysis on the chat record and outputting the analysis result; Step S3, according to the analysis result, matching the accident object and the accident type of the current accident corresponding to the chat record in the accident knowledge base, locating the occurrence time of the accident, and finally outputting accident location information; Step S4, obtaining the frequency of feedback of the current accident by multiple users in the customer service chat group within a predetermined time, evaluating the reliability of the current accident according to the feedback frequency, and outputting a prediction result; Step S5, according to the prediction result, when the reliability level of the current accident meets a preset condition, output a warning information, the warning information including the accident location information and the reliability level corresponding to the current accident.