Process event response method based on large model

Abstract

The present invention relates to the field of artificial intelligence, and in particular to a process event response method based on a large model, comprising: detecting text quality parameters and video quality parameters of business information acquired each time based on a pre-trained model, and determining information quality parameters according to the text quality parameters and video quality parameters; determining information categories of corresponding business information according to the information quality parameters, and determining an analysis method of a process processing coefficient according to the information categories of the business information acquired each time, including determining the process processing coefficient according to a reference matching coefficient and a proportion of effective information; periodically determining whether to perform process scheduling for business information to be processed according to a processing coefficient difference value, and determining the business information to be scheduled and a target scheduling processing end based on a historical correlation coefficient and a node processing coefficient. The present invention improves the balance of task volume distribution, thereby improving the overall processing efficiency of business information.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a process event response method based on a large model. Background Art

[0002] The traditional way of assigning tasks for business processes is usually random assignment or assignment based on subjective factors of managers, resulting in uneven task assignment or poor processing efficiency. Therefore, business process processing based on artificial intelligence technology has been proposed, which can automatically process a large amount of repetitive work to improve work efficiency and provide more accurate guidance and suggestions. For example, for the customer service platform, the content quality of business information is closely related to the speed and effectiveness of the process processing end's judgment on product problems, which can easily affect the processing efficiency of business information. However, the existing task assignment schemes often fail to take into account the impact of the effectiveness of the business information content, and it is difficult to ensure the overall processing efficiency of the business information by the client. Therefore, how to effectively adjust the task assignment scheme according to the actual quality of the acquired business information to ensure the overall processing efficiency of the business information is an urgent problem to be solved by technical personnel in this field.

[0003] Chinese Patent Publication No. CN114897429A discloses a task allocation method, device, equipment and storage medium suitable for customer service communication, wherein the method includes: determining a first candidate department identification set based on a target communication text and a department keyword set corresponding to each department identification; determining a second candidate department identification set based on the target communication text and a preset first mapping table; combining the first candidate department identification set and the second candidate department identification set to obtain a responsible department identification set to be analyzed; determining a main responsible department identification and an associated responsible department identification set from the responsible department identification set to be analyzed based on the target communication text and the second mapping table; assigning tasks to the target communication text based on the main responsible department identification and the associated responsible department identification set to obtain a task allocation result, but the above scheme has the following problems: failing to judge the difficulty of task processing based on the effectiveness of the text content actually obtained, and making effective adjustments to the task allocation process based on this, thereby ensuring the overall task processing efficiency. Summary of the invention

[0004] To this end, the present invention provides a process event response method based on a large model to overcome the problem that the prior art fails to effectively determine the difficulty of analyzing the business information actually obtained, resulting in failure to effectively adjust the task allocation according to the difficulty of analyzing the business information, and further resulting in low overall task processing efficiency.

[0005] To achieve the above object, the present invention provides a process event response method based on a large model, comprising:

[0006] Detecting text quality parameters and video quality parameters of each acquired business information based on the pre-trained model, and determining information quality parameters according to the text quality parameters and the video quality parameters;

[0007] Determine the information category of the corresponding business information according to the information quality parameter, and determine the analysis method of the process processing coefficient according to the information category of the business information obtained each time;

[0008] The analysis method includes determining the process processing coefficient of the first type of business information according to the reference correlation validity coefficient and determining the process processing coefficient of the second type of business information according to the valid information ratio;

[0009] Periodically determine whether to perform process scheduling for the pending business information of each type of process processing end according to the processing coefficient difference value, and determine the pending business information and the target scheduling processing end based on the historical correlation coefficient and the node processing coefficient.

[0010] Further, information quality parameter detection is performed on each piece of business information, and the information category of the corresponding business information is determined according to the information quality parameter;

[0011] If the information quality parameter of a business information is greater than the preset information quality parameter, the business information is determined to be a type of business information;

[0012] If the information quality parameter of a service information is less than or equal to the preset information quality parameter, the service information is determined to be second-class service information.

[0013] Further, the information quality parameter of each service information is determined according to the text quality parameter and the video quality parameter;

[0014] Extract keywords from the text information in the business information and determine the text quality parameters based on the number of keywords and the completeness of the association between keywords;

[0015] Determine video quality parameters according to the shooting stability and video clarity of the video information in the business information;

[0016] The information quality parameter is positively correlated with the text quality parameter and the video quality parameter respectively.

[0017] Furthermore, an analysis method of the process processing coefficient is determined according to the information category of each business information;

[0018] For a single business information,

[0019] If the information category is a type of business information, the process processing coefficient of the business information is determined according to the reference correlation validity coefficient;

[0020] If the information category is Class II business information, the process processing coefficient of the business information shall be determined based on the proportion of valid information.

[0021] Further, when determining the process processing coefficient of each type of business information according to the reference correlation validity coefficient, extract the target keyword of the text information of each type of business information, and determine the correlation keyword group based on the target keyword;

[0022] The correlation effectiveness coefficient of each correlation keyword group is detected, and the reference correlation effectiveness coefficient of any one type of business information is the average value of the correlation effectiveness coefficients of each correlation keyword group in the text information of the one type of business information;

[0023] The process processing coefficient of each type of business information is positively correlated with the corresponding reference correlation validity coefficient.

[0024] Further, the correlation validity coefficient of each associated keyword group is determined according to the effective keyword proportion of the corresponding associated keyword group, and the effective keyword is determined according to the effective reference value of each keyword, and the keyword with an effective reference value greater than the preset effective reference value is recorded as the effective keyword;

[0025] The correlation effectiveness coefficient of each associated keyword group is positively correlated with the effective keyword proportion of the corresponding associated keyword group.

[0026] Furthermore, when determining the process processing coefficient of the second type of business information according to the proportion of effective information, effective information in the text information and the video information of the second type of business information is extracted respectively, and the process processing coefficient of the second type of business information is determined according to the proportion of effective information;

[0027] The process processing coefficient is negatively correlated with the proportion of valid information.

[0028] Furthermore, the effective information ratio is determined based on the effective text ratio and the effective video ratio;

[0029] The effective text ratio of each second-category business information is the ratio of the number of words of the keywords in the text information of the corresponding second-category business information;

[0030] The effective video ratio of each second-category business information is the ratio of the duration of effective video information in the video information of the corresponding second-category business information, and the effective video information is the associated video information whose video quality parameter is greater than the preset video quality parameter.

[0031] Furthermore, the category of each process processing end is determined based on the proportion of key business information;

[0032] The process processing end whose key business information ratio is greater than the preset key business information ratio is recorded as a first-class process processing end;

[0033] The process processing end whose key business information ratio is less than or equal to the preset key business information ratio is recorded as the second-class process processing end;

[0034] Key business information refers to business information whose process processing coefficient is greater than the preset process processing coefficient.

[0035] Furthermore, if the processing coefficient difference value is greater than the preset processing coefficient difference value, process scheduling is performed for the to-be-processed business information of each type of process processing end;

[0036] For any first-class process processing end, determine the business information to be scheduled according to the historical correlation coefficient of each business information to be processed;

[0037] For any business information to be scheduled, the scheduling priority coefficient of the corresponding process processing end is set according to the node processing coefficient of each second-class process processing end and the historical correlation coefficient between each second-class process processing end and the business information to be scheduled, so as to determine the target scheduling processing end;

[0038] The scheduling priority coefficient is positively correlated with the node processing coefficient and the historical correlation coefficient respectively.

[0039] Compared with the prior art, the beneficial effect of the present invention lies in that, in the technical scheme of the present invention, the category of business information is determined according to the information quality parameters of the business information, and a targeted analysis difficulty determination strategy is determined according to the category of the business information to obtain the process processing coefficient of each business information, and the task allocation of each process processing end is effectively adjusted in real time based on the process processing coefficient of the business information to ensure the balance of the task amount allocated to each process processing end, thereby effectively ensuring the overall efficiency of the actual processing process for the business information.

[0040] Furthermore, the present invention determines the category of corresponding business information based on information quality parameters. The information quality parameters can characterize the proportion of effective information in the acquired text information that can assist the process processing end in analyzing the cause of the problem and the difficulty of obtaining effective information based on video information, so as to preliminarily judge the analysis difficulty of each business information, and determine a targeted analysis difficulty determination strategy, thereby improving the accuracy of the determined process processing coefficient, and further improving the effectiveness of subsequent task volume adjustment.

[0041] Furthermore, in the present invention, for business information with larger information quality parameters, the associated effectiveness coefficient of each associated keyword group is detected, and the process processing coefficient of each type of business information is determined based on the reference associated effectiveness coefficient. By dividing the keywords into different associated keyword groups and determining the effectiveness of each keyword based on the abnormal position corresponding to each associated keyword group, the process processing coefficient determined in this way can effectively indicate the effectiveness of the text information. The present invention improves the accuracy of the determined process processing coefficient.

[0042] Furthermore, in the present invention, for business information with smaller information quality parameters, the effective information ratio of this category of business information is determined based on the effective text ratio and the effective video ratio, and the process processing coefficient is determined based on this. Since there is less information in this category of business information that can assist the client in problem analysis and it is difficult to determine the matching coefficient of each keyword, the process processing coefficient is determined based on the keyword ratio in the text information and the ratio of videos that can assist keyword analysis and are clearly shot, thereby ensuring the efficiency and accuracy of the process of determining the process processing coefficient.

[0043] Furthermore, the present invention periodically determines whether to perform process scheduling for the business information to be processed of each type of process processing end based on the difference value of the analysis difficulty of each process processing end, so that when the task distribution is unbalanced, the business information processed by each process processing end can be adjusted in a timely and effective manner, the business information to be scheduled is determined according to the historical correlation coefficient, and the process processing end corresponding to the business information to be scheduled is selected according to the node processing coefficient and the historical correlation coefficient. The present invention improves the balance of the task amount allocated to each process processing end, thereby improving the overall efficiency of the actual processing process for the business information. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of a process event response method based on a large model of the present invention;

[0045] Figure 2 A flowchart of the present invention for determining the information category of corresponding business information according to information quality parameters;

[0046] Figure 3 A flow chart of an analysis method for determining a process processing coefficient according to information categories of business information of the present invention;

[0047] Figure 4 This is a flow chart of the present invention for determining whether to perform process scheduling for the pending business information of each type of process processing end according to the analysis difficulty difference value. DETAILED DESCRIPTION

[0048] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0049] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0050] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0051] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0052] See also Figures 1 to 4 As shown, the present invention provides a process event response method based on a large model, comprising:

[0053] Detecting text quality parameters and video quality parameters of each acquired business information based on the pre-trained model, and determining information quality parameters according to the text quality parameters and the video quality parameters;

[0054] Determine the information category of the corresponding business information according to the information quality parameter, and determine the analysis method of the process processing coefficient according to the information category of the business information obtained each time;

[0055] The analysis method includes determining the process processing coefficient of the first type of business information according to the reference correlation validity coefficient and determining the process processing coefficient of the second type of business information according to the valid information ratio;

[0056] Periodically determine whether to perform process scheduling for the pending business information of each type of process processing end according to the processing coefficient difference value, and determine the pending business information and the target scheduling processing end based on the historical correlation coefficient and the node processing coefficient.

[0057] Among them, the pre-trained model in the present invention is used to determine the information quality parameters of each business information. The user can determine the keywords according to the actual processing needs and historical processing records, and complete the training of the pre-trained model to ensure that it effectively screens the keyword content in the business information. There are several process processing terminals in the present invention, which are used to process the business information to give the corresponding abnormal cause determination results. The business information in the present invention is feedback information of the goods that provide customer service. The business information includes text information and video information. The categories of keywords in the text information in the present invention include but are not limited to: abnormal phenomena, abnormal positions and abnormal duration. The goods that provide customer service in the present invention are washing machines. The abnormal phenomena in the present invention include: water leakage, abnormal noise, abnormal jitter and control panel failure. The abnormal positions are not limited to: bottom plate, water pipe interface, drum shaft and control panel. The user can pre-set the keyword content according to the actual processing scenario. In addition, the keyword content can be updated according to the partial text information of the business information that assists in analyzing the business information when each process processing terminal processes different business information in the historical processing records. How to pre-train the model and how to determine the keyword content are easy to understand for technicians in this field and will not be repeated here.

[0058] The present invention uses several historical processing records, and any one of the historical processing records records the information quality parameters of each business information in the distribution processing process of the business information at least once, the video quality parameters of each valid video information, the valid reference value, the process processing coefficient, the processing coefficient difference value and the historical correlation coefficient, and each historical monitoring record corresponds to a qualified mark, which records whether the overall efficiency requirements of the business processing meet the user needs.

[0059] Specifically, information quality parameter detection is performed on each business information, and the information category of the corresponding business information is determined according to the information quality parameter;

[0060] If the information quality parameter of a business information is greater than the preset information quality parameter, the business information is determined to be a type of business information;

[0061] If the information quality parameter of a service information is less than or equal to the preset information quality parameter, the service information is determined to be second-class service information.

[0062] Among them, the value of the preset information quality parameter can be determined by the user according to the actual work scenario. For example, the user can set it according to the historical evaluation record. The higher the user's requirement for the overall efficiency of business processing, the larger the value of the preset information quality parameter. A method for setting the value of the preset information quality parameter is provided, and the minimum value of the information quality parameter of a class of business information in the historical processing records that meets the user's requirement for the overall efficiency of business processing is recorded as the preset information quality parameter.

[0063] Specifically, the information quality parameter of each service information is determined based on the text quality parameter and the video quality parameter;

[0064] Extract keywords from the text information in the business information and determine the text quality parameters based on the number of keywords and the completeness of the association between keywords;

[0065] Determine video quality parameters according to the shooting stability and video clarity of the video information in the business information;

[0066] The information quality parameter is positively correlated with the text quality parameter and the video quality parameter respectively.

[0067] Among them, for any business information obtained at one time, the information quality parameter is the sum of the text quality parameter and the video quality parameter, the text quality parameter of the text information of the business information = the product of the number of keywords and the corresponding influence coefficient + the product of the association integrity and the corresponding influence coefficient, the association integrity between keywords = the number of associated keywords / the number of keywords, the historical processing record in the present invention respectively records the abnormal cause determination results for each abnormal position, and the keywords existing in the abnormal cause determination results of each abnormal position in the historical processing record are recorded as the associated keywords of the corresponding abnormal position, and the user can provide a number of keywords and the value of the influence coefficient corresponding to the association integrity according to the actual processing scenario, the value of the influence coefficient corresponding to the number of keywords and the value of the influence coefficient corresponding to the association integrity is 0.3, and the value of the influence coefficient corresponding to the association integrity is 0.7;

[0068] The video quality parameter of each video information is the natural logarithm of the product of the shooting stability and video clarity of the corresponding video information. Several video frames are randomly extracted from the video information, and the shooting stability of the video information is determined according to the jitter index of each extracted video frame. The shooting stability is negatively correlated with the average value of the jitter index of each extracted video frame. For any video frame, several reference points are randomly selected, and the offset of each reference point between the video frame and the next adjacent video frame of the video frame in the video information is calculated. The jitter index of the video frame is the average value of the offset of each reference point. The video clarity of the video information is determined according to the resolution of each extracted video frame. The video clarity is positively correlated with the average value of the resolution of each extracted video frame. How to determine the resolution of each video frame and how to determine the offset of reference points between video frames are contents that are easy to understand for technicians in this field and will not be elaborated here.

[0069] Specifically, the analysis method of the process processing coefficient is determined according to the information category of each business information;

[0070] For a single business information,

[0071] If the information category is a type of business information, the process processing coefficient of the business information is determined according to the reference correlation validity coefficient;

[0072] If the information category is Class II business information, the process processing coefficient of the business information shall be determined based on the proportion of valid information.

[0073] Specifically, when determining the process processing coefficient of each type of business information according to the reference correlation validity coefficient, extract the target keyword of the text information of each type of business information, and determine the correlation keyword group based on the target keyword;

[0074] The correlation effectiveness coefficient of each correlation keyword group is detected, and the reference correlation effectiveness coefficient of any one type of business information is the average value of the correlation effectiveness coefficients of each correlation keyword group in the text information of the one type of business information;

[0075] The process processing coefficient of each type of business information is positively correlated with the corresponding reference correlation validity coefficient.

[0076] Specifically, the correlation effectiveness coefficient of each associated keyword group is determined according to the effective keyword proportion of the corresponding associated keyword group, and the effective keyword is determined according to the effective reference value of each keyword, and the keyword with an effective reference value greater than the preset effective reference value is recorded as the effective keyword;

[0077] The correlation effectiveness coefficient of each associated keyword group is positively correlated with the effective keyword proportion of the corresponding associated keyword group.

[0078] Among them, for any type of text information of business information, the target keyword is a keyword in the text information whose category is abnormal position, and the target keywords indicating the same abnormal position are divided into the same associated keyword group. For any associated keyword group, the associated keywords of the abnormal position represented by the associated keyword group are included in the associated keyword group. The effective reference value of any keyword is the number of associated keywords of the abnormal position represented by the associated keyword group in the historical processing record. The value of the preset effective reference value can be determined by the user according to the actual work scenario. For example, the user can set it according to the historical evaluation record. The higher the user's overall efficiency requirement for business processing, the larger the value of the preset effective reference value. A method for determining the preset effective reference value is provided, and the minimum value of the effective reference value of each effective keyword in the historical processing record that meets the user's overall efficiency requirement for business processing is recorded as the preset effective reference value. The effective keyword ratio = the number of effective keywords in the associated keyword group / the number of keywords in the associated keyword group.

[0079] Specifically, when determining the process processing coefficient of the second type of business information according to the proportion of effective information, the effective information in the text information and the video information of the second type of business information are extracted respectively, and the process processing coefficient of the second type of business information is determined according to the proportion of effective information;

[0080] The process processing coefficient is negatively correlated with the proportion of valid information.

[0081] Specifically, the effective information ratio is determined based on the effective text ratio and the effective video ratio;

[0082] The effective text ratio of each second-category business information is the ratio of the number of words of the keywords in the text information of the corresponding second-category business information;

[0083] The effective video ratio of each second-category business information is the ratio of the duration of effective video information in the video information of the corresponding second-category business information, and the effective video information is the associated video information whose video quality parameter is greater than the preset video quality parameter.

[0084] Wherein, for any one or two types of business information, the effective information includes keywords in the text information and effective video information in the video information, the effective information ratio is the sum of the text effective ratio and the video effective ratio, the text effective ratio = the sum of the number of words of each keyword in the text information of the two types of business information / the number of words in the text information of the two types of business information, the video effective ratio = the sum of the duration of each effective video information in the video information of the two types of business information / the duration of the video information of the two types of business information;

[0085] Target keywords are extracted from the text information of the two types of business information, and the video information of the two types of business information is divided to obtain several video information segments of the same video length. If there is an associated image frame in a video information segment, the video information segment is recorded as the associated video information, and the image frame containing the abnormal position indicated by any target keyword is recorded as the associated image frame. For example, if there is a target keyword in the text information indicating that the abnormal position is a control panel, the image frame containing the control panel is the associated image frame. How to determine whether each image frame contains the abnormal position indicated by any target keyword is easy to understand for those skilled in the art and will not be elaborated here.

[0086] The video quality parameters of each associated video information of the two types of business information are detected, and the video quality parameters of each associated video information are the natural logarithm of the product of the shooting stability and the video clarity of the corresponding associated video information. A number of video frames are randomly extracted for the associated video information, and the shooting stability of the associated video information is determined according to the jitter index of each extracted video frame, and the shooting stability is negatively correlated with the average value of the jitter index of each extracted video frame. The video clarity of the associated video information is determined according to the resolution of each extracted video frame, and the video clarity is positively correlated with the average value of the resolution of each extracted video frame. The value of the preset video quality parameter can be determined by the user according to the actual working scenario. For example, the user can set it according to the historical evaluation record. The higher the user's overall efficiency requirement for business processing, the larger the value of the preset video quality parameter. A method for determining the value of the preset video quality parameter is provided, and the minimum value of the video quality parameter of each valid video information in the historical processing record that meets the user's overall efficiency requirement for business processing is recorded as the preset video quality parameter.

[0087] Specifically, the category of each process processing end is determined based on the proportion of key business information;

[0088] The process processing end whose key business information ratio is greater than the preset key business information ratio is recorded as a first-class process processing end;

[0089] The process processing end whose key business information ratio is less than or equal to the preset key business information ratio is recorded as the second-class process processing end;

[0090] Key business information refers to business information whose process processing coefficient is greater than the preset process processing coefficient.

[0091] Among them, the present invention applies a cyclic processing monitoring cycle, and the duration of the processing monitoring cycle can be determined by the user. The higher the user's requirements for the overall efficiency of business processing, the shorter the duration of the processing monitoring cycle. A processing monitoring cycle is provided, and the processing monitoring cycle is 15 minutes. At the end of each processing monitoring cycle, the category of each process processing end is determined according to the proportion of key business information, and the reference process processing coefficient of the pending business information of each process processing end is obtained to determine the processing coefficient difference value of each process processing end, and whether to perform process scheduling for the pending business information of each type of process processing end is determined according to the processing coefficient difference value;

[0092] For any process processing end, the proportion of key business information = the number of key business information in the business information to be analyzed at the process processing end / the number of business information in the business information to be analyzed at the process processing end. The business information to be analyzed is a collection of business information that needs to be processed but has not been processed by each process processing end when determining the category of each process processing end. The values ​​of the preset key business information proportion and the preset process processing coefficient can be determined by the user according to the actual work scenario. For example, the user can set them according to historical evaluation records. The higher the user's overall efficiency requirement for business processing, the smaller the value of the preset key business information proportion, and the smaller the value of the preset process processing coefficient. A method for determining the value of the preset key business information proportion is provided, and the minimum value of the key business information proportion of a type of process processing end that meets the user's overall efficiency requirement for business processing is recorded as the preset key business information proportion. A method for determining the value of the preset process processing coefficient is provided, and the minimum value of the process processing coefficient of the key business information that meets the user's overall efficiency requirement for business processing is recorded as the preset process processing coefficient.

[0093] Specifically, if the processing coefficient difference value is greater than the preset processing coefficient difference value, process scheduling is performed for the to-be-processed business information of each type of process processing end;

[0094] For any first-class process processing end, determine the business information to be scheduled according to the historical correlation coefficient of each business information to be processed;

[0095] For any business information to be scheduled, the scheduling priority coefficient is set according to the node processing coefficient of each second-class process processing end and the historical correlation coefficient between each second-class process processing end and the business information to be scheduled, so as to determine the target scheduling processing end;

[0096] The scheduling priority coefficient is positively correlated with the node processing coefficient and the historical correlation coefficient respectively.

[0097] Wherein, for any process processing end, the pending business information is business information that needs to be processed by the process processing end but has not been processed, the reference process processing coefficient is the average value of the process processing coefficients of each pending business information of the process processing end, and the processing coefficient difference value = , n is the number of process processing terminals, is the reference process processing coefficient of the i-th process processing end, is the average value of the reference process processing coefficient of each process processing end; the value of the preset processing coefficient difference value can be determined by the user according to the actual work scenario, for example, the user can set it according to the historical evaluation record, the higher the user's overall efficiency requirement for business processing, the smaller the value of the preset processing coefficient difference value is, and a method for determining the preset processing coefficient difference value is provided, the historical processing record that has not been processed for process scheduling of the business information to be processed is recorded as a difference reference record, and the maximum value of the processing coefficient difference value measured each time in the difference reference record that meets the user's overall efficiency requirement for business processing is recorded as the preset processing coefficient difference value;

[0098] For any one type of process processing end, the to-be-processed business information whose historical correlation coefficient is less than the preset historical correlation coefficient is recorded as the to-be-scheduled business information of the one type of process processing end. The historical correlation coefficient between any to-be-processed business information and the corresponding process processing end is positively correlated with the number of historical processing times of the relevant business information of the to-be-processed business information by the process processing end. The relevant business information is the business information in which the name of the abnormal phenomenon in the text information of the to-be-processed business information exists in the text information. The number of historical processing times is the number of times the process processing end processes the relevant business information in the historical processing record. The value of the preset historical correlation coefficient can be determined by the user according to the actual work scenario. For example, the user can set it according to the historical evaluation record. The higher the user's overall efficiency requirement for business processing, the larger the value of the preset historical correlation coefficient. A method for setting the value of the preset historical correlation coefficient is provided. The maximum value of the historical correlation coefficient of the to-be-scheduled business information in the historical processing record that meets the user's overall efficiency requirement for business processing is recorded as the preset historical correlation coefficient.

[0099] For any business information to be scheduled, the node processing coefficients of other second-class process processing terminals and the historical correlation coefficients between them and the business information to be scheduled are detected to determine the scheduling priority coefficients of each second-class process processing terminal, and the business information to be scheduled is scheduled to the business information to be processed of the second-class process processing terminal with a larger scheduling priority coefficient, and the second-class process processing terminal is recorded as the target scheduling processing terminal. For any second-class process processing terminal, the scheduling priority coefficient is the natural logarithm of the product of the node processing coefficient and the historical correlation coefficient. The node processing coefficient is the average time used by the second-class process processing terminal to process each business information. The historical correlation coefficient between the second-class process processing terminal and the business information to be scheduled is positively correlated with the number of historical processing times of the second-class process processing terminal for the related business information of the business information to be scheduled.

[0100] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A process event response method based on a large model, characterized in that: include: Detecting text quality parameters and video quality parameters of each acquired business information based on the pre-trained model, and determining information quality parameters according to the text quality parameters and the video quality parameters; Determine the information category of the corresponding business information according to the information quality parameter, and determine the analysis method of the process processing coefficient according to the information category of the business information obtained each time; The analysis method includes determining the process processing coefficient of the first type of business information according to the reference correlation validity coefficient and determining the process processing coefficient of the second type of business information according to the valid information ratio; Periodically determine whether to perform process scheduling for the pending business information of each type of process processing end according to the processing coefficient difference value, and determine the pending business information and the target scheduling processing end based on the historical correlation coefficient and the node processing coefficient; Determine the analysis method of process processing coefficient according to the information category of each business information; For a single business information, If the information category is a type of business information, the process processing coefficient of the business information is determined according to the reference correlation validity coefficient; If the information category is Class II business information, the process processing coefficient of the business information is determined based on the proportion of valid information; For any text information of a type of business information, the target keyword is a keyword of abnormal position in the text information, and the target keywords indicating the same abnormal position are classified into the same associated keyword group. For any associated keyword group, the associated keywords of the abnormal position represented by the associated keyword group are classified into the associated keyword group together, and the effective reference value of any keyword is the number of times the associated keyword of the abnormal position represented by the associated keyword group is in the historical processing record; Extract target keywords from the text information of the two types of business information, and divide the video information of the two types of business information to obtain a number of video information segments with the same video length. If there is an associated image frame in a video information segment, the video information segment is recorded as the associated video information, and the image frame containing the abnormal position indicated by any target keyword is recorded as the associated image frame; If the processing coefficient difference value is greater than the preset processing coefficient difference value, process scheduling is performed for the pending business information of each type of process processing end; For any first-class process processing end, determine the business information to be scheduled according to the historical correlation coefficient of each business information to be processed; For any business information to be scheduled, the scheduling priority coefficient of the corresponding process processing end is set according to the node processing coefficient of each second-class process processing end and the historical correlation coefficient between each second-class process processing end and the business information to be scheduled, so as to determine the target scheduling processing end; The scheduling priority coefficient is positively correlated with the node processing coefficient and the historical correlation coefficient respectively.

2. The process event response method based on a large model according to claim 1 is characterized in that: Perform information quality parameter detection on each business information, and determine the information category of the corresponding business information according to the information quality parameter; If the information quality parameter of a business information is greater than the preset information quality parameter, the business information is determined to be a type of business information; If the information quality parameter of a service information is less than or equal to the preset information quality parameter, the service information is determined to be second-class service information.

3. The process event response method based on a large model according to claim 2 is characterized in that: The information quality parameter of each service information is determined according to the text quality parameter and the video quality parameter; Extract keywords from the text information in the business information and determine the text quality parameters based on the number of keywords and the completeness of the association between keywords; Determine video quality parameters according to the shooting stability and video clarity of the video information in the business information; The information quality parameter is positively correlated with the text quality parameter and the video quality parameter respectively.

4. The process event response method based on a large model according to claim 3 is characterized in that: When determining the process processing coefficient of each type of business information according to the reference correlation validity coefficient, extracting the target keywords of the text information of each type of business information, and determining the associated keyword group based on the target keywords; The correlation effectiveness coefficient of each correlation keyword group is detected, and the reference correlation effectiveness coefficient of any one type of business information is the average value of the correlation effectiveness coefficients of each correlation keyword group in the text information of the one type of business information; The process processing coefficient of each type of business information is positively correlated with the corresponding reference correlation validity coefficient.

5. The process event response method based on a large model according to claim 4 is characterized in that: The correlation effectiveness coefficient of each associated keyword group is determined according to the effective keyword proportion of the corresponding associated keyword group, and the effective keyword is determined according to the effective reference value of each keyword, and the keyword with an effective reference value greater than the preset effective reference value is recorded as the effective keyword; The correlation effectiveness coefficient of each associated keyword group is positively correlated with the effective keyword proportion of the corresponding associated keyword group.

6. The process event response method based on a large model according to claim 5 is characterized in that: When determining the process processing coefficient of the second type of business information according to the proportion of effective information, the effective information in the text information and the video information of the second type of business information are extracted respectively, and the process processing coefficient of the second type of business information is determined according to the proportion of effective information; The process processing coefficient is negatively correlated with the proportion of valid information.

7. The process event response method based on a large model according to claim 6 is characterized in that: The effective information ratio is determined based on the effective text ratio and the effective video ratio; The effective text ratio of each second-category business information is the ratio of the number of words of the keywords in the text information of the corresponding second-category business information; The effective video ratio of each second-category business information is the ratio of the duration of effective video information in the video information of the corresponding second-category business information, and the effective video information is the associated video information whose video quality parameter is greater than the preset video quality parameter.

8. The process event response method based on a large model according to claim 7 is characterized in that: If the processing coefficient difference value is greater than the preset processing coefficient difference value, process scheduling is performed for the pending business information of each type of process processing end; For any first-class process processing end, determine the business information to be scheduled according to the historical correlation coefficient of each business information to be processed; For any business information to be scheduled, the scheduling priority coefficient of the corresponding process processing end is set according to the node processing coefficient of each second-class process processing end and the historical correlation coefficient between each second-class process processing end and the business information to be scheduled, so as to determine the target scheduling processing end; The scheduling priority coefficient is positively correlated with the node processing coefficient and the historical correlation coefficient respectively.

Images