Early warning method, device, equipment, medium and program product

Abstract

The invention discloses an early warning method and device, equipment, a medium and a program product. The method comprises the steps that in the process that a first object provides business service for a second object, at least one kind of physiological indication data of the second object is acquired; based on the at least one kind of physiological indication data, determining experience data of the second object for the business service; determining early warning information for the business service based on the experience data so as to warn the first object to adjust at least one to-be-adjusted behavior; wherein the to-be-adjusted behavior is associated with the process of the business service. By means of the method, real-time accurate tracking of the service quality of the first object can be achieved in the business service process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to a pre-warning method and device, equipment, medium and program product. BACKGROUND

[0002] In actual application, after the business service ends, the service provider determines the link and process to be optimized or improved in the business service process by the video data collected in the business process or the service quality score fed back by the customer. However, the above method lacks accuracy and has obvious hysteresis, and cannot meet the optimization and improvement needs of the business service process. SUMMARY

[0003] Based on the above technical problems, the present application provides a pre-warning method, device, equipment, medium and program product.

[0004] The technical scheme provided by the present application is as follows:

[0005] The present application provides a pre-warning method, which comprises the following steps:

[0006] In the process that a first object provides a business service for a second object, at least one physiological indication data of the second object is acquired;

[0007] Based on the at least one physiological indication data, experience data of the second object for the business service is determined;

[0008] Based on the experience data, pre-warning information for the business service is determined to pre-warn the first object to adjust at least one behavior to be adjusted; wherein the behavior to be adjusted is associated with the process of the business service.

[0009] In some embodiments, after the pre-warning information for the business service is determined based on the experience data, the method further comprises:

[0010] A candidate feature set associated with the first object in a specified period is determined; wherein the specified period comprises at least one period in the process of the business service; the features in the candidate feature set are at least associated with the behavior of the first object in the specified period;

[0011] Based on the features in the candidate feature set, a target feature set is determined; wherein the degree of association between the features in the target feature set and the pre-warning information is greater than or equal to a degree threshold;

[0012] The factor associated with the target feature set is determined as a target factor;

[0013] The target factor is output for the first object to adjust the behavior to be adjusted.

[0014] In some embodiments, the determining the target feature set based on the features in the candidate feature set comprises:

[0015] The features in the candidate feature set are sorted based on the degree of association between the features in the candidate feature set and the early warning information, to obtain a sorting result.

[0016] The target feature set comprises at least one feature in the sorting result.

[0017] In some embodiments, the determining the target feature set based on the features in the candidate feature set comprises:

[0018] The historical sorting parameters and / or historical feedback data are obtained.

[0019] The degree of association between the features in the candidate feature set and the early warning information is determined.

[0020] The features in the candidate feature set are sorted based on the degree of association, the historical sorting parameters and / or the historical feedback data, to obtain the target feature set.

[0021] In some embodiments, the behavior of the first object in the specified period of time comprises a body behavior and / or a language behavior of the first object in the specified period of time; the candidate feature set is further associated with the first object and an environment in which the second object is located in the specified period of time; the determining the target feature set based on the features in the candidate feature set comprises:

[0022] Feature extraction is performed on behavior data corresponding to the body behavior and / or the language behavior, to obtain a body feature corresponding to the body behavior and / or a language feature corresponding to the language behavior.

[0023] An environment parameter of the environment in which the first object and the second object are located in the specified period of time is collected.

[0024] The environment parameter is processed to obtain an environment feature of the environment.

[0025] The features in the candidate feature set are screened based on a matching degree between the environment feature, the body feature and / or the language feature and a reference environment feature, a reference body feature and / or a reference language feature, respectively, to obtain the target feature set.

[0026] In some embodiments, the determining the experience data of the second object for the business service based on the at least one physiological indication data comprises:

[0027] quantify the at least one physiological indicator data to obtain a first quantification result;

[0028] collect a pupil video including a pupil region of the second object;

[0029] analyze the pupil video to determine a pupil change state of the pupil region;

[0030] quantify the pupil change state to obtain a second quantification result;

[0031] determine the experience data based on the first quantification result and the second quantification result.

[0032] In some embodiments, the at least one physiological indicator data includes at least one of a heart rate, a blood pressure, and a skin conductance level of the second object; and the quantifying the at least one physiological indicator data to obtain a first quantification result includes:

[0033] obtain an initial heart rate, an initial blood pressure, and an initial skin conductance level at an initial time; wherein the initial time includes a historical time of a current time;

[0034] determine a heart rate threshold, a blood pressure threshold, and a conductance threshold;

[0035] quantify the heart rate based on the initial heart rate and the heart rate threshold to obtain a heart rate quantification result, quantify the blood pressure based on the initial blood pressure and the blood pressure threshold to obtain a blood pressure quantification result, and quantify the skin conductance level based on the initial skin conductance level and the conductance threshold to obtain a skin conductance quantification result;

[0036] determine the first quantification result based on at least one of the heart rate quantification result, the blood pressure quantification result, and the skin conductance quantification result.

[0037] In some embodiments, the determining the experience data based on the first quantification result and the second quantification result includes:

[0038] determine a weight parameter;

[0039] process the first quantification result and the second quantification result based on a weight value in the weight parameter to obtain the experience data.

[0040] In some embodiments, the determining the warning information for the business service based on the experience data includes:

[0041] obtain a time threshold corresponding to a current time period;

[0042] If the experience data is less than or equal to the time period threshold, the warning information is determined.

[0043] In some embodiments, before obtaining the time period threshold corresponding to the current time period, the method further includes:

[0044] Obtain a collection of historical experience data associated with historical time periods prior to the current time period;

[0045] The historical experience data is processed by a prediction model to obtain a prediction result for the current time period;

[0046] The prediction results are statistically averaged to obtain the time period threshold corresponding to the current time period.

[0047] In some embodiments, before obtaining the time period threshold corresponding to the current time period, the method further includes:

[0048] Obtain the historical threshold corresponding to the historical time periods preceding the current time period;

[0049] The historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

[0050] In some embodiments, adjusting the historical threshold to obtain a time period threshold corresponding to the current time period includes:

[0051] Obtain prediction results obtained by making predictions based on historical experience data; wherein, the historical experience data is associated with historical physiological indicator data within a historical period;

[0052] If the difference between the prediction result and the experience data is greater than or equal to the first threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

[0053] In some embodiments, adjusting the historical threshold to obtain a time period threshold corresponding to the current time period includes:

[0054] If, within the time period corresponding to the first time point and the second time point, the amount of historical experience data associated with the historical threshold is greater than or equal to the second threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period; wherein, the first time point is used to characterize the time at which the historical threshold is determined; the second time point includes future times of the first time point.

[0055] In some embodiments, adjusting the historical threshold to obtain a time period threshold corresponding to the current time period includes:

[0056] Obtain historical experience data associated with the historical threshold;

[0057] If the fluctuation between the experience data and the historical experience data is greater than or equal to the third threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

[0058] This application embodiment also provides an early warning device, the early warning device comprising:

[0059] The acquisition module is used to acquire at least one physiological indicator data of the second object during the process of the first object providing business services to the second object;

[0060] The determination module is used to determine the experience data of the second object for the business service based on the at least one physiological indicator data;

[0061] The early warning module is used to determine early warning information for the business service based on the experience data, so as to warn the first object to adjust at least one behavior to be adjusted; wherein the behavior to be adjusted is associated with the process of the business service.

[0062] This application embodiment also provides an early warning device, which includes a processor and a memory; wherein, the memory stores a computer program; when the computer program is executed by the processor, it can implement the early warning method as described above.

[0063] This application also provides a computer-readable storage medium storing a computer program; when the computer program is executed by a processor of an electronic device, it can implement the early warning method as described above.

[0064] This application also provides a computer program product, which includes a computer program; when the computer program is executed by the processor of an electronic device, it can implement the early warning method as described above.

[0065] The early warning method provided in this application, during the process of a first object providing business services to a second object, acquires at least one physiological indicator data of the second object. This improves the correlation between the at least one physiological indicator data and the business services provided by the first object, enabling the at least one physiological indicator data to accurately, comprehensively, and in real time reflect the second object's perception of the business services. Furthermore, based on the at least one physiological indicator data, the second object's experience data regarding the business services is determined. This not only improves the real-time nature of the experience data but also enhances its accuracy and comprehensiveness. On this basis, early warning information for the business services is determined based on the experience data to warn the first object to adjust at least one behavior to be adjusted, with the behavior to be adjusted being correlated with the process of the business services. This not only enables real-time early warning of at least one behavior to be adjusted by the first object but also provides data basis for the first object to adjust the behavior to be adjusted in a timely manner, thereby improving the pertinence and real-time nature of the first object's adjustment of the behavior to be adjusted. This, in turn, provides data basis for the adjustment and optimization of business services, meeting the needs for adjusting and optimizing business services.

[0066] When the first object is a service personnel and the second object is a customer, and with the customer's permission, the early warning method provided in this application embodiment can not only determine the early warning information associated with the business service process in real time and accurately, but also provide real-time and accurate data assurance for adjusting the behavior of service personnel in the business service process, improving the level of business service, and improving the quality of business service in a targeted manner. Attached Figure Description

[0067] Figure 1 A flowchart illustrating the early warning method provided in this application embodiment;

[0068] Figure 2 This is a flowchart illustrating the process of determining experience data provided in an embodiment of this application.

[0069] Figure 3 A schematic diagram illustrating the service quality feedback optimization process provided in this application embodiment;

[0070] Figure 4 This is a schematic diagram of the structure of the early warning device provided in the embodiments of this application;

[0071] Figure 5 This is a schematic diagram of the structure of the early warning device provided in the embodiment of this application. Detailed Implementation

[0072] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0073] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0074] The service quality management system of a business service provider identifies areas or processes that need improvement after the business service process is completed, based on the results of customer-completed service quality-related questionnaires or business quality follow-up interviews.

[0075] However, while the above methods can track service quality, they are executed after the service has ended, resulting in significant lag and insufficient real-time performance. Furthermore, the accuracy of customer feedback data obtained through questionnaires and follow-up interviews is insufficient, making it impossible to obtain accurate service quality data. Consequently, these methods cannot provide analytical basis for service improvement and may lead to customer complaints or loss of customer resources.

[0076] To address the aforementioned technical challenges, related technologies also include an early warning system installed in unmanned service halls. During service operations, this system collects video data from customers via environmental acquisition units, including cameras. This video data is then sent to a central processing unit for analysis. The central processing unit combines this data with a database of special service group customers and the energy internet marketing service system to determine whether the video data indicates normal or abnormal situations. Abnormal situations are then sent to the operation display unit for display, and an audible and visual alarm is issued when necessary to alert staff. This early warning system can promptly alert staff when customers encounter problems but do not proactively seek help, enabling staff to proactively assist customers.

[0077] However, the video data collected by the aforementioned early warning system is easily affected by actual factors such as sound, light, and obstructions in the customer's environment. Therefore, the accuracy of the abnormal situations identified by the above scheme is insufficient.

[0078] Related technologies also provide a business early warning method based on an adaptive model. This method involves the field of artificial intelligence and can be applied to the financial sector. The method first obtains first data to be processed, which is associated with the financial business processing procedure, based on preset data indicators. After pre-analyzing the first data, it is input into a pre-trained intelligent recognition model to obtain normal and abnormal data. Then, based on the normal and abnormal data, early warning processing is performed on the second data to be processed. This method can achieve early warning processing of financial business data through an intelligent recognition model.

[0079] However, the first data to be processed is similar to the video data collection process in the aforementioned scheme. Therefore, the abnormal data identified by the above business early warning method cannot accurately reflect the need for business process improvement.

[0080] Based on the above technical problems, embodiments of this application provide an early warning method, device, equipment, medium, and program product.

[0081] Figure 1 This is a flowchart illustrating the early warning method provided in the embodiments of this application, as shown below. Figure 1 As shown, the method may include the following steps:

[0082] Step 101: During the process of the first object providing business services to the second object, at least one physiological indicator data of the second object is obtained.

[0083] In one implementation, the first object may be a provider of business services; for example, the first object may have vital signs, such as a staff member in a business service hall; for example, the first object may include an electronic device, such as a robot with service provision functions.

[0084] In one implementation, the second object may include a person receiving business services; for example, the second object may include at least one customer.

[0085] In one implementation, business services may include services of at least one business type, such as finance, securities, education, and healthcare.

[0086] In one implementation, a business service may include at least one step, stage, or process.

[0087] In one implementation, the customer represented by the second object and the staff represented by the first object may be in the same physical space. For example, the customer and the staff may both be located in the same business hall or meeting room set up by the service provider.

[0088] In one implementation, the customer represented by the second object and the staff represented by the first object can be located in different physical spaces. For example, the customer can be in a first location and the staff can be in a second location. The staff can establish a communication connection with the second device associated with the customer through the first device and provide online remote business services to the customer through the communication connection.

[0089] In one implementation, the degree of correlation between at least one physiological indicator data and the emotional state of the second object can be greater than or equal to a preset degree. Therefore, by using at least one physiological indicator, the emotional state of the second object during the business service process, or the experience and feelings of the second object regarding the business service process, can be determined.

[0090] In one implementation, at least one physiological indicator may include at least one of the second subject's heart rate, blood pressure, and body temperature.

[0091] In one implementation, at least one physiological indicator data can be obtained through any of the following methods:

[0092] At least one physiological indicator data of the second object is collected by at least one sensor device installed in the second device.

[0093] During the process of the first object providing business services to the second object, at least one physiological indicator data can be collected in real time continuously or at specified time intervals through at least one sensor device.

[0094] Step 102: Based on at least one physiological indicator data, determine the experience data of the second object for the business service.

[0095] In one implementation, experience data can characterize the second object's emotional state of satisfaction or dissatisfaction with at least one aspect or step of the business service.

[0096] In one implementation, experience data can be determined in any of the following ways:

[0097] Based on the fluctuation state of at least one physiological indicator data during at least one time period in the business service process, the experience data of the second object for the business service is determined; for example, if the fluctuation state is greater than or equal to a preset state, the experience data of the second object for the business service can be determined to be unsatisfactory, and if the fluctuation state is less than the preset state, the experience data of the second object for the business service can be determined to be satisfactory.

[0098] The system collects audio and facial expression data output by the second object during the business service process. If the fluctuation of the audio data, facial expression data, and at least one physiological indicator data during the business service process is greater than or equal to a preset state, it can be determined that the second object's experience data regarding the business service is unsatisfactory.

[0099] Step 103: Based on experience data, determine early warning information for business services, and adjust at least one behavior to be adjusted for the first target of the early warning.

[0100] Among them, the behaviors to be adjusted are related to the process of business services.

[0101] In one implementation, the warning information includes at least the unsatisfactory experience outcome represented by the experience data.

[0102] In one implementation, the warning information can be presented in the form of text, voice, video, images, and flashing indicator lights, or it can be presented in the form of shaking of the warning device; wherein, the warning device may include a first device associated with the first object.

[0103] In one implementation, the behavior to be adjusted may be associated with the language or actions of the first object.

[0104] In one implementation, the behavior to be adjusted is associated with a business service process, which may include at least one of the following:

[0105] The types and / or number of behaviors to be adjusted associated with different processes of the same business service can be different;

[0106] When the service objects associated with the same business service change, the corresponding behaviors to be adjusted for the business service can be different;

[0107] The types and / or number of behaviors to be adjusted may differ for different business services.

[0108] For example, when the business service is a business processing service, the behavior to be adjusted in the process of obtaining user needs can be different from the behavior to be adjusted in the business processing process.

[0109] For example, when the business service is a business processing service, the behavior to be adjusted can vary based on factors such as the health status and / or education level of the user requesting the business processing service.

[0110] For example, the types and / or quantities of behaviors to be adjusted for business processing services may differ from those for business consultation services.

[0111] In one implementation, the warning information can be determined in any of the following ways:

[0112] Based on the level of dissatisfaction represented by the experience data, the warning level corresponding to the warning information is determined; for example, the warning level can be directly proportional to the level of dissatisfaction.

[0113] The warning level is determined based on the number of unsatisfactory steps or processes contained in the experience data; for example, the warning level may be proportional to the number of steps or processes.

[0114] In one implementation, after the warning information is determined, it can be sent to the first device associated with the first object, so that the first device can output the warning information in real time, thereby realizing real-time warning of the behavior to be adjusted of the first object.

[0115] In this embodiment of the application, the process of determining experience data and determining warning information can be implemented by an edge computing (MEC) device set in the environment where the first object is located.

[0116] As can be seen from the above, the early warning method provided in this application, during the process of the first object providing business services to the second object, acquires at least one physiological indicator data of the second object. This improves the correlation between the at least one physiological indicator data and the business services provided by the first object, enabling the at least one physiological indicator data to accurately, comprehensively, and in real time reflect the second object's perception of the business services. Furthermore, based on the at least one physiological indicator data, the experience data of the second object regarding the business services is determined. This not only improves the real-time nature of the experience data but also enhances its accuracy and comprehensiveness. On this basis, early warning information for the business services is determined based on the experience data to warn the first object to adjust at least one behavior to be adjusted, and the behavior to be adjusted is associated with the process of the business services. This not only enables real-time early warning of at least one behavior to be adjusted by the first object but also provides data basis for the first object to adjust the behavior to be adjusted in a timely manner. This improves the pertinence and real-time nature of the first object's adjustment of the behavior to be adjusted, thereby providing data basis for the adjustment and optimization of business services and meeting the needs for adjusting and optimizing business services.

[0117] When the first object is a service personnel and the second object is a customer, and with the customer's permission, the early warning method provided in this application embodiment can not only determine the early warning information associated with the business service process in real time and accurately, but also provide real-time and accurate data assurance for adjusting the behavior of service personnel in the business service process, improving the level of business service, and improving the quality of business service in a targeted manner.

[0118] Based on the foregoing embodiments, in the early warning method provided in this application embodiment, after determining the early warning information for the business service based on experience data, the following steps can also be performed:

[0119] Step A1: Determine the set of candidate features associated with the first object within the specified time period.

[0120] The specified time period includes at least one time period during the business process; the features in the candidate feature set are at least associated with the behavior of the first object during the specified time period.

[0121] In one implementation, the behavior of the first object during a specified time period may include at least some of the behavior of the first object during the specified time period; for example, the at least some behavior may include at least one type of behavior of the first object; for example, the at least some behavior may include the facial expression change behavior of the first object.

[0122] In one implementation, the specified time period may be associated with the rate of change of at least one physiological indicator data; for example, the specified time period may include the time period corresponding to at least one physiological indicator data whose rate of change is greater than or equal to a preset threshold.

[0123] In one implementation, the candidate feature set may further include the environmental features of the second object, or the first object and the environment in which the second object is located; for example, environmental parameters of the environment can be collected and analyzed to determine the environmental features; for example, the environmental parameters may include at least one of ambient temperature, humidity, brightness, noise intensity and spatial enclosure degree.

[0124] In one implementation, the candidate feature set can be determined in the following way:

[0125] Feature extraction results are obtained by performing feature extraction on the behavior of the first object within a specified time period, and the feature extraction results are determined as a candidate feature set.

[0126] Step A2: Determine the target feature set based on the features in the candidate feature set.

[0127] Among them, the degree of correlation between the features in the target feature set and the early warning information is greater than or equal to the degree threshold.

[0128] In one implementation, if the degree of correlation between the features in the target feature set and the early warning information is greater than or equal to a degree threshold, it can characterize that there is a causal relationship between the features in the target feature set and the early warning information.

[0129] In one implementation, the target feature set can be determined in the following way:

[0130] The matching relationship between the features in the candidate feature set and the standard features corresponding to the standard business service is determined, and the set of features in the candidate feature set that do not match the standard features is determined as the target feature set. For example, the matching relationship between the features in the candidate feature set and the standard features can be identified by an anomaly detection model. The anomaly detection model can determine the above matching relationship by statistical methods including box plot analysis, clustering algorithms, or machine learning-based anomaly detection algorithms.

[0131] Step A3: Identify the factors associated with the target feature set as target factors.

[0132] In one implementation, the target factors may include a list of behavioral factors performed by the first subject during a specified time period that may cause the rate of change of at least one physiological indicator data to be greater than or equal to a preset threshold.

[0133] In one implementation, the target factor may include behavioral or environmental characteristics that deviate significantly from the standard characteristics. The factors corresponding to the aforementioned behavioral and environmental characteristics may be key factors that lead to a decline in the quality of the business service process.

[0134] In one implementation, the target factor can be determined in the following way:

[0135] Based on the process of determining the candidate feature set, the correlation between the behavior of the first object in the specified time period and the features in the candidate feature set is determined. Based on the correlation and the features in the target feature set, the behavior of the first object in the specified time period is filtered, and the list of filtered behaviors is determined as the target factors.

[0136] Step A4: Output the target factors so that the first target can adjust the behavior to be adjusted.

[0137] In one implementation, the target factor can be output in at least one form of image, text, voice and video, so that the first object can obtain the target factor in real time and intuitively, thereby providing a basis for the first object to adjust the behavior to be adjusted.

[0138] In one implementation, after the device associated with the first object receives the target factor, it can output the target factor through a real-time pop-up window, so that the service personnel represented by the first object can take timely and targeted measures to adjust the operation to be adjusted, thereby providing data basis for improving the quality of business services.

[0139] In one implementation, the target factor can be sent to the first device associated with the first object via a real-time communication mechanism including WebSocket and / or Message Queuing Telemetry Transport (MQTT).

[0140] In one implementation, the target factors may also include suggestions for improvement of the behavior to be adjusted, so as to facilitate the first target to quickly understand the problem and take effective measures.

[0141] It should be noted that after the first object adjusts its behavior according to the target factors, the changes in the experience data of the second object can be continuously monitored, and the process of determining the target factors can be continuously optimized based on the changes in the experience data to improve the targeting and real-time nature of the target factors. Furthermore, in the process of optimizing the process of determining the target factors, historical experience data can also be mined and analyzed to help discover potential problems in the business service process.

[0142] As can be seen from the above, the early warning method provided in this application determines a set of candidate features associated with a first object within a specified time period, and the specified time period includes at least one time period during the continuous process of business services. The features in the candidate feature set are associated with the behavior of the first object within the specified time period. Thus, the candidate feature set can achieve a comprehensive representation of the behavior of the first object within the specified time period. Furthermore, based on the features in the candidate feature set, a target feature set with a correlation degree greater than or equal to a threshold degree between the features and the early warning information is determined, realizing accurate screening and filtering of features in the candidate feature set at the level of early warning information. On this basis, the factors associated with the target feature set, i.e., target factors, are determined and output for the first object to adjust the behavior to be adjusted, thereby improving the correlation degree between the target factors and the early warning information, thus improving the pertinence of the first object's adjustment of the behavior to be adjusted, and further improving the accuracy and efficiency of business service optimization and improvement.

[0143] Based on the foregoing embodiments, the early warning method provided in this application, which determines the target feature set based on features in the candidate feature set, can be implemented through the following methods:

[0144] Based on the degree of correlation between the features in the candidate feature set and the early warning information, the features in the candidate feature set are ranked to obtain the ranking result; the target feature set is determined to include at least one feature in the ranking result.

[0145] In one implementation, the degree of association can be determined in the following way:

[0146] By using feature importance determination methods including correlation analysis, mutual information, and random forest, we analyze the behavioral features, environmental features, and early warning information associated with the behavior of the first object in the candidate feature set, thereby determining the weight of each feature in the candidate feature set on the impact of the decline in business service quality represented by the early warning information, and then determining the degree of correlation based on the magnitude of the above weights.

[0147] In one implementation, features in the candidate feature set can be sorted based on the strength or magnitude of the correlation to obtain a sorting result. For example, in the actual sorting process, the influence of the behavioral features of the first object on the experience data can be greater than the influence of environmental features on the experience data. For instance, the influence of a customer service representative's impatience on the customer's experience data can be greater than the influence of a slight change in the temperature of the customer's environment on the experience data.

[0148] In one implementation, the target feature set can be determined in the following way:

[0149] The target feature set is determined by identifying at least one feature in the ranking results whose correlation degree is greater than or equal to the degree threshold.

[0150] As can be seen from the above, the early warning method provided in this application sorts the features in the candidate feature set based on the degree of correlation between the features in the candidate feature set and the early warning information, thereby enhancing the strength of the correlation between the features in the sorted result and the early warning information. On this basis, by determining that the target feature set includes at least one feature in the sorted result, the causal correlation between the features in the target feature set and the early warning information can be improved.

[0151] Based on the foregoing embodiments, the early warning method provided in this application, which determines the target feature set based on features in the candidate feature set, can also be implemented through the following steps:

[0152] Step B1: Obtain historical sorting parameters and / or historical feedback data.

[0153] In one implementation, the historical ranking parameters may include the ranking criteria or ranking strategy on which the features in the historical candidate feature set are ranked; for example, the ranking criteria or ranking strategy may include the ranking parameters of the time and / or spatial dimensions referenced when ranking the features in the historical candidate feature set.

[0154] In one implementation, historical feedback data may include feedback data obtained within at least one historical period that adjusts, modifies, or updates historical sorting parameters; exemplarily, historical feedback data may be provided by any party to the business service; exemplarily, the business service participants may include a first object and a second object.

[0155] In one implementation, the type of business service associated with the historical sorting parameters and historical feedback data can be the same as the type of business service currently in which the first object and the second object are located.

[0156] Step B2: Determine the degree of correlation between the features in the candidate feature set and the early warning information.

[0157] Step B3: Based on the degree of correlation, historical ranking parameters and / or historical feedback data, filter and rank the features in the candidate feature set to obtain the target feature set.

[0158] In one implementation, the target feature set can be determined in the following way:

[0159] The features in the candidate feature set are initially screened and sorted according to their correlation to obtain an initial result; then, based on historical sorting parameters and / or historical feedback data, the order in the initial result is adjusted or corrected, and the result of the adjustment or correction is determined as the target feature set.

[0160] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining historical ranking parameters and / or historical feedback data, and determining the degree of correlation between the features in the candidate feature set and the early warning information, the features in the candidate feature set are ranked based on the degree of correlation, historical ranking parameters, and / or historical feedback data to obtain the target feature set. Thus, in the process of ranking the features in the candidate feature set, not only is the degree of correlation between the features in the candidate feature set and the early warning information during the current business service process taken into account, but historical ranking parameters and / or historical feedback data are also incorporated. This improves the stability and accuracy of ranking the features in the candidate feature set, thereby enabling the target feature set to meet the optimization needs of the business service.

[0161] Based on the foregoing embodiments, in the early warning method provided in this application, the behavior of the first object during a specified time period includes the physical behavior and / or language behavior of the first object during the specified time period; the candidate feature set is also associated with the environment in which the first object and the second object are located during the specified time period.

[0162] In one embodiment, physical behavior may include at least one of the following: whether the first object moves at least one of its body parts during a specified time period, the actions performed by the first object's limbs, and the frequency of occurrence of the aforementioned actions.

[0163] In one implementation, the first data containing body language can be obtained in the following way:

[0164] First data is obtained by continuously acquiring video of the service personnel represented by the first object during business service through an image acquisition device associated with the first object; for example, the working mode of the image acquisition device can be pre-adjusted to improve the clarity of the first data, thereby improving the subtlety and accuracy of body behavior recognition; for example, when the environment where the first object and the second object are located is a business hall, the above-mentioned first data can be acquired through a panoramic camera installed in the service window; for example, the above-mentioned video can be acquired for at least a portion of the time period before the warning information is determined, wherein the length of the above-mentioned at least a portion of the time period can be 5 minutes.

[0165] In one implementation, language behavior may include the content of the language output by the first object within a specified time period and whether the content is repeated.

[0166] In one implementation, the second data containing language behavior can be obtained in the following way:

[0167] With the permission of the second party, the second data is obtained by collecting call recordings or chat logs between the first and second parties and organizing these recordings or logs; alternatively, the second data is obtained by collecting audio data output by the first party during the continuous operation of the business service through an audio acquisition device associated with both the first and second parties. For example, the second data may include text data and / or audio data. It should be noted that the integrity and authenticity of the second data must be ensured during the above process.

[0168] Accordingly, determining the target feature set based on the features in the candidate feature set can be achieved through the following steps:

[0169] Step C1: Extract features from the behavioral data corresponding to body language and / or language behavior to obtain body features corresponding to body language and / or language features corresponding to language behavior.

[0170] In one embodiment, the limb features may include at least one of the semantic information expressed by the limb behavior of the first object, the range of motion of the first object's limbs, the direction of motion, and the frequency of motion.

[0171] In one implementation, the language features may include at least one of the semantics of the language output by the first object, the timbre, loudness, and speech rate of the first object during the output of the language.

[0172] In one implementation, the first data, including body movements, can be processed using a Convolutional Neural Network (CNN) or a 3D CNN to obtain body features; the second data, including language behaviors, can be processed using a Recurrent Neural Network (RNN) or a Transformer model to obtain language features.

[0173] Step C2: Collect environmental parameters of the environment in which the first object and the second object are located within a specified time period; process the environmental parameters to obtain the environmental characteristics of the environment.

[0174] Step C3: Based on environmental features, body features, and / or language features, and the degree of matching between these features and the baseline environmental features, baseline body features, and / or baseline language features, the features in the candidate feature set are filtered to obtain the target feature set.

[0175] In one implementation, the baseline physical characteristics may include a set of characteristics that a first object, such as a service personnel, should possess in the physical behavior of a standard business service process.

[0176] In one implementation, the baseline language features may include a set of features that a first object, such as a business person, should possess in the language behavior during a standard business service process.

[0177] In one implementation, the target feature set can be obtained in the following way:

[0178] The environmental features that do not match the baseline environmental features are identified as the first feature, the limb features that do not match the baseline limb features are identified as the second feature, and the language features that do not match the baseline language features are identified as the third feature. Then, the set of the first feature, the second feature, and / or the third feature is identified as the target feature set.

[0179] For example, the process of determining the target feature set can also be implemented by a recognition model that includes CNN and RNN. In order to improve the accuracy of the target feature set, the CNN and RNN in the initial state can be trained in advance to build a recognition model.

[0180] For example, when the first object is a service worker and the second object is a customer, the CNN and RNN in the initial state can be trained using the following process:

[0181] First, during the process of service personnel providing business services, body video data including the service personnel's physical behavior is collected through cameras set up in connection with the service personnel, and audio data including the service personnel's language behavior is collected through audio acquisition devices set up in connection with the service personnel. Additionally, a set of indicator data including at least one physiological indicator data of the customer can also be collected.

[0182] Secondly, experts and experienced service personnel, based on business service standards, correlate, classify, and label the body video data, language audio data, and indicator data sets to determine the third data that meets the service standards and the fourth data that does not meet the service standards from the above various data. They can also determine the fifth data that the fluctuation state of the indicator data in the physiological indicator data set has an impact greater than or equal to the degree threshold. For example, the above data can cover aspects such as the frequency of service personnel's smiles, speech speed, usage of specific words, and waiting time at the service window.

[0183] For example, preprocessing operations can also be performed on the third to fifth data to remove noise and outliers, and normalization operations can be performed on the third and fifth data to obtain the first to third samples, thereby improving the data quality of the first to third samples; wherein, the preprocessing operations on the video data contained in the third to fifth data may include image frame extraction and background removal processing, and the preprocessing operations on the text data contained in the third to fifth data may include text segmentation processing.

[0184] For example, after determining the first to third samples, features can be extracted from the first to third samples using computer vision technology and natural language processing technology, thereby obtaining sample features including the service personnel's body posture features, movement amplitude, movement direction, word usage frequency, and emotional tendency.

[0185] For example, statistical analysis and machine learning algorithms, including recursive feature elimination, feature selection based on correlation coefficients, and principal component analysis, can be used to screen sample features and determine the most discriminative feature subset among the sample features. The features in the feature subset can explain the differences between different data in the first to third samples to the greatest extent. Therefore, the features in the feature subset can be called key features.

[0186] For example, after determining the key features, a recognition model can be constructed; wherein, the recognition model may include a CNN and an RNN obtained by training the initial state CNN and the initial state RNN with the first to third samples and the key features.

[0187] For example, the identification model obtained in the above manner can automatically capture and track the differences and correlations between standard and non-standard business services, as well as between the physical and verbal behaviors of service personnel and the physiological indicators of customers.

[0188] For example, after constructing the recognition model, the performance of the recognition model can be evaluated through methods such as cross-validation and A / B testing to improve the recognition accuracy of the recognition model and the stability of various data processing obtained in the actual business service process.

[0189] As can be seen from the above, in the early warning method provided in this application embodiment, the behavior of the first object within a specified time period includes the physical behavior and / or language behavior of the first object within the specified time period. Furthermore, the candidate feature set is also associated with the environment in which the first object and the second object are located within the specified time period. The environmental features are obtained by processing the environmental parameters of the environment in which the first object and the second object are located within the specified time period. By extracting features from the behavioral data corresponding to the physical behavior and / or language behavior, physical features corresponding to the physical behavior and / or language features corresponding to the language behavior are obtained. Thus, through the above operations, the richness of the features included in the business service process can be improved, thereby enhancing the comprehensiveness and completeness of the features of the business service process. Moreover, by filtering the features in the candidate feature set based on the matching degree between environmental features, physical features, and / or language features and the baseline environmental features, baseline physical features, and / or baseline language features, respectively, a target feature set is obtained. This not only achieves accurate and comprehensive filtering of features in the candidate feature set but also improves the comprehensiveness and accuracy of features in the target feature set.

[0190] Based on the foregoing embodiments, the early warning method provided in this application determines the experience data of a second object regarding business services based on at least one physiological indicator data, which can be achieved through... Figure 2 The process shown is implemented. Figure 2 This is a schematic diagram of the process for determining experience data provided in an embodiment of this application, such as... Figure 2 As shown, the process may include the following steps:

[0191] Step 201: Quantify at least one physiological indicator data to obtain the first quantification result.

[0192] In one implementation, the first quantization result can be obtained in the following way:

[0193] Based on the predetermined first quantization interval set and the magnitude of the data in at least one physiological indicator data, the interval in which the at least one physiological indicator data is located in the first quantization interval set is determined, and the set of interval values ​​of the at least one physiological indicator data in the first quantization interval set is determined as the first quantization result; for example, the above interval values ​​may include the median of any interval in the first quantization interval set; for example, each interval in the first quantization interval can be determined according to the actual quantization requirements of the physiological indicator data.

[0194] Step 202: Collect pupil video including the pupil area of ​​the second object, analyze the pupil video, and determine the pupil change state of the pupil area.

[0195] In one implementation, with the permission of the second object, a pupil video of the pupil area of ​​the second object can be captured by an image acquisition device associated with the second object; for example, if the environment where the first object and the second object are located is a business hall, the pupil video can be captured by a capture camera set outside the service window of the business hall.

[0196] It should be noted that the behavioral video containing the first subject's physical and / or verbal behavior, as well as the pupil video, and at least one physiological indicator data, can be correlated with the experience data in the temporal and spatial dimensions, thereby improving the matching degree between the first subject's physical and / or verbal behavior and the second subject's sensory changes in the temporal and spatial dimensions.

[0197] In one embodiment, pupil changes may include pupil diameter increasing or decreasing.

[0198] In one implementation, a neural network model capable of extracting features from the pupil region can be used to extract features from the pupil video, thereby obtaining the pupil change state of the pupil region within at least one time period.

[0199] In one implementation, changes in pupil size can reflect changes in the emotions and attitudes of the customer represented by the second object during the business service process.

[0200] Step 203: Quantify the pupil change state to obtain the second quantification result.

[0201] In one implementation, the second quantization result can be obtained in any of the following ways:

[0202] Determine the interval in the second quantization interval set where the pupil change state is located according to the size of the pupil diameter or area represented by the pre-determined second quantization interval set and the pupil change state, and determine the set of interval values of the interval where the pupil change state is located in the second quantization interval set as the second quantization result; Exemplarily, the above interval values may include the median of any interval in the second quantization interval set; Exemplarily, each interval in the second quantization interval may be determined according to the actual quantization requirements of the pupil change state.

[0203] Determine the diameter difference between the pupil diameter represented by the pupil change state and the initial pupil diameter, and determine the ratio of the diameter difference to the initial pupil diameter as the pupil diameter ratio, and then determine the second quantization result according to the relationship between the pupil diameter ratio and the pupil threshold; where the initial pupil diameter may include the pupil diameter of the second object at the moment when the service is started, and the initial pupil diameter may continuously capture the pupil image of the user at the moment when the service is started, and determine the average value of the pupil diameters in three consecutive pupil images as the initial pupil diameter; the pupil threshold may change or be adjusted according to the environment where the second object is located, the gender of the second object, and the duration of the service.

[0204] Exemplarily, the pupil diameter ratio can be calculated by formula (1):

[0205] L1 = |w1n - w1| / w1 * 100% (1)

[0206] Where, L1 may be the pupil diameter ratio, w1n may be the pupil diameter represented by the pupil change state, and w1 may represent the initial pupil diameter.

[0207] Exemplarily, the pupil threshold can be denoted as N1. If L1 ≤ N1, the pupil diameter can be quantified as a neutral emotion, and its value can be O1. If N1 < L1 < N1 + 3, the pupil diameter can be quantified as a satisfied emotion, and its value can be P1. If L1 ≥ N1 + 3, the pupil diameter can be quantified as a dissatisfied emotion, and its value can be Q1.

[0208] Step 204, Determine the experience data based on the first quantization result and the second quantization result.

[0209] In one implementation, the experience data can be determined in the following manner:

[0210] Sum the first quantization result and the second quantization result to obtain a sum result, and determine the sum result as the experience data.

[0211] As can be seen from the above, the early warning method provided in this application obtains a first quantification result by quantifying at least one physiological indicator data. Thus, the first quantification result enables a concise representation of at least one physiological indicator data. Furthermore, by collecting pupil videos including the pupil region of a second object and analyzing the pupil videos to determine the pupil change state of the pupil region, real-time and accurate tracking of the pupil change state of the second object can be achieved during business services. Based on this, the pupil change state is quantified to obtain a second quantification result, which simplifies the representation of the pupil change state. On the other hand, by determining experience data based on the first and second quantification results, not only can the efficiency of the experience data determination process be improved, but also the comprehensiveness of the experience data can be enhanced.

[0212] Based on the foregoing embodiments, in the early warning method provided in this application, at least one physiological indicator data includes at least one of the second subject's heart rate, blood pressure, and skin conductance level.

[0213] In one embodiment, after the customer represented by the second object sits in a seat equipped with at least one type of sensor device, with the permission of the second object, at least one physiological indicator data can be continuously collected through at least one sensor device during the business service process; for example, the heart rate, blood pressure and skin conductance levels of the second object can be collected by a heart rate sensor module, a blood pressure sensor module and a skin conductance sensor module installed in the seat, respectively.

[0214] In one embodiment, after at least one sensor module in the seat collects at least one physiological indicator data and pupil video, the at least one physiological indicator data and pupil video can be sent to an MEC device set in the environment where the seat is located via a wireless communication module. The MEC device can then process the physiological indicator data and pupil video through its storage unit, image recognition unit, analysis unit, data recording unit, and computing unit. For example, the environment where the seat is located may include the business hall or meeting room where the first object and the second object are located.

[0215] In one implementation, data such as the heart rate, blood pressure, and skin conductance of the second subject can be used to reflect the physiological state and comfort level of the customer represented by the second subject during the business service process in real time.

[0216] Accordingly, quantifying at least one physiological indicator data to obtain a first quantification result can be achieved through the following steps:

[0217] Step D1: Obtain the initial heart rate, initial blood pressure, and initial skin conductance at the initial moment.

[0218] Among them, the initial moment includes the historical moments of the current moment.

[0219] In one implementation, the initial moment may include the moment when the business service process is started.

[0220] In one implementation, the initial heart rate, initial blood pressure, and initial skin conductance level can be collected through at least one sensor device at the initial moment.

[0221] Step D2: Determine the heart rate threshold, blood pressure threshold, and conductance threshold.

[0222] In one implementation, the heart rate threshold, blood pressure threshold, and conductance threshold may be related to at least one of the gender of the second object, the duration of the business service, the type of the business service, and the environmental parameters of the environment where the second object is located.

[0223] Step D3: Quantify the heart rate based on the initial heart rate and the heart rate threshold to obtain a heart rate quantization result, quantify the blood pressure based on the initial blood pressure and the blood pressure threshold to obtain a blood pressure quantization result, and quantify the skin conductance level based on the initial skin conductance level and the conductance threshold to obtain a skin conductance quantization result.

[0224] In one implementation, the heart rate quantization result can be obtained in the following manner:

[0225] Determine the heart rate difference between the heart rate and the initial heart rate, and determine the ratio of the heart rate difference to the initial heart rate as the heart rate ratio. Then, determine the heart rate quantization result according to the relationship between the heart rate ratio and the heart rate threshold; Exemplarily, the heart rate ratio can be calculated by Equation (2):

[0226] L2 = |w2n - w2| / w2 * 100% (2)

[0227] Among them, L2 can be the heart rate ratio, w2n can be the heart rate at the current moment, and w2 can represent the initial heart rate.

[0228] Exemplarily, the heart rate threshold can be denoted as N2. If L2 ≤ N2, the heart rate can be quantified as a neutral emotion, and its value can be O2. If N2 < L2 < N2 + 3, the heart rate can be quantified as a satisfied emotion, and its value can be P2; and if L2 ≥ N2 + 3, the heart rate can be quantified as a dissatisfied emotion, and its value can be Q2.

[0229] In one implementation, the blood pressure quantization result can be obtained in the following manner:

[0230] Determine the blood pressure difference between the current blood pressure and the initial blood pressure, and determine the ratio of the blood pressure difference to the initial blood pressure as the blood pressure ratio. Then, determine the blood pressure quantization result according to the relationship between the blood pressure ratio and the blood pressure threshold. Exemplarily, the blood pressure ratio can be calculated by Equation (3):

[0231] L3 = |w3n - w3| / w3 * 100% (3)

[0232] where L3 can be the blood pressure ratio, w3n can be the blood pressure at the current moment, and w3 can represent the initial blood pressure.

[0233] Exemplarily, the blood pressure threshold can be denoted as N3. If L3 ≤ N3, the blood pressure can be quantified as a neutral emotion, and its value can be O3. If N3 < L3 < N3 + 5, the blood pressure can be quantified as a satisfied emotion, and its value can be P3. If L3 ≥ N3 + 5, the blood pressure can be quantified as a dissatisfied emotion, and its value can be Q3.

[0234] In one implementation, the skin conductance quantization result can be obtained through the following method:

[0235] Determine the skin conductance difference between the current skin conductance level and the initial skin conductance level, and determine the ratio of the skin conductance difference to the initial skin conductance level as the skin conductance ratio. Then, determine the skin conductance quantization result according to the relationship between the skin conductance ratio and the skin conductance threshold. Exemplarily, the skin conductance ratio can be calculated by Equation (4):

[0236] L4 = |w4n - w4| / w4 * 100% (4)

[0237] where L4 can be the skin conductance ratio, w4n can be the skin conductance level at the current moment, and w4 can represent the initial skin conductance level.

[0238] Exemplarily, the skin conductance threshold can be denoted as N4. If L4 ≤ N4, the skin conductance can be quantified as a neutral emotion, and its value can be O4. If N4 < L4 < N4 + 9, the skin conductance can be quantified as a satisfied emotion, and its value can be P4. If L4 ≥ N4 + 9, the skin conductance can be quantified as a dissatisfied emotion, and its value can be Q4.

[0239] Step D4: Determine the first quantization result based on at least one of the heart rate quantization result, the blood pressure quantization result, and the skin conductance quantization result.

[0240] In one implementation, the first quantization result can be determined according to at least one of the heart rate quantization result, the blood pressure quantization result, and the skin conductance quantization result.

[0241] It should be noted that, in the embodiments of this application, after acquiring at least one physiological indicator data and pupil video, the MEC device can first perform a cleaning operation on the above-mentioned data to remove outliers and noise data contained in the above-mentioned data; then, the data after removing outliers and noise data is converted into a format to improve the efficiency of subsequent quantization of the data.

[0242] For example, the process of quantifying at least one physiological indicator data and the pupil diameter characterized by pupil change state can be implemented by a quantization model set in the data analysis unit of the MEC device; and, in order to improve quantization accuracy, the quantization model of the initial state can be trained in advance to obtain the quantization model; for example, the quantization model can be constructed based on machine learning or deep learning technology.

[0243] For example, firstly, a sufficient amount of physiological indicator data, such as a data volume greater than or equal to 20,000, and pupil diameter data need to be collected as training samples. At the same time, the training samples are cleaned to remove outliers, missing values, and noise to obtain sample data. Then, the sample data is quantized using the method provided in the aforementioned embodiment to obtain sample quantization results. Then, a quantization model, including a random forest model, is used to train based on the sample data and sample quantization results to determine the correspondence between different physiological indicator data and emotional states, thereby obtaining the quantization model.

[0244] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining the initial heart rate, initial blood pressure, and initial skin conductance level at the initial moment, and determining the heart rate threshold, blood pressure threshold, and conductance threshold, the heart rate is quantified based on the initial heart rate and heart rate threshold to obtain a heart rate quantification result. Blood pressure is quantified based on the initial blood pressure and blood pressure threshold to obtain a blood pressure quantification result. Skin conductance level is quantified based on the initial skin conductance level and conductance threshold to obtain a skin conductance quantification result. This allows the quantification results of the heart rate, blood pressure, and skin conductance level of the second object to take into account the initial heart rate, initial blood pressure, initial skin conductance level, heart rate threshold, blood pressure threshold, and conductance threshold, thereby improving the accuracy of the above quantification operations and thus improving the accuracy of the heart rate quantification result, blood pressure quantification result, and skin conductance quantification result. Furthermore, based on at least one of the heart rate quantification result, blood pressure quantification result, and skin conductance quantification result, a first quantification result is determined, which improves the accuracy of the first quantification result.

[0245] Based on the foregoing embodiments, the early warning method provided in this application, which determines experience data based on the first quantification result and the second quantification result, can be implemented in the following way:

[0246] Determine the weight parameters; based on the weight values ​​in the weight parameters, process the first quantization result and the second quantization result to obtain the experience data.

[0247] In one implementation, the number of weight values ​​included in the weight parameters can be the same as the number of quantization results included in the first quantization result and the second quantization result; for example, the experience data G can be calculated using equation (5):

[0248] G=L1*r+L2*s+L3*t+L4*u (5)

[0249] Wherein, {r,s,t,u} can be four weight values ​​included in the weight parameters, but they must satisfy the condition r+s+u+t=1, and the values ​​of the four weight values ​​can be flexibly adjusted to reflect the degree of contribution of their respective quantization results to the experience data; at the same time, the value ranges of the quantization results in the first quantization result and the second quantization result can be L1∈{O1,P1,Q1}, L2∈{O2,P2,Q2}, L3∈{O3,P3,Q3} and L4∈{O4,P4,Q4}, respectively.

[0250] As can be seen from the above, the early warning method provided in this application, after determining the weight parameters, processes the first quantization result and the second quantization result based on the weight values ​​in the weight parameters to obtain experience data. Thus, through the above operations, not only can the comprehensiveness and completeness of the data contained in the experience data be improved, but the computational load in the experience data acquisition process can also be reduced, thereby improving the efficiency of the experience data acquisition process.

[0251] In the case where the first object is customer service personnel and the second object is customers, in this embodiment of the application, by processing at least one physiological indicator of the customer to determine experience data, the shortcomings of related technologies in determining the accuracy of customer emotional changes solely through facial images can be overcome. Furthermore, by quantifying and weighting the customer's at least one physiological indicator data, the comprehensiveness and accuracy of the experience data can be improved. Moreover, through the above process, subtle changes in customer emotions can be captured, thereby enabling timely detection of problems in the business service process and providing data guidance for improving the service quality of customer service personnel.

[0252] Based on the foregoing embodiments, the early warning method provided in this application, which determines early warning information for business services based on experience data, can be implemented in the following ways:

[0253] Obtain the time period threshold corresponding to the current time period; if the experience data is less than or equal to the time period threshold, determine the warning information.

[0254] Accordingly, if the experience data exceeds the time period threshold, the operation of determining the warning information can be skipped.

[0255] In one implementation, the current time period may include the period between the start of the business service and the time when the fluctuation of at least one physiological indicator data is greater than or equal to a preset threshold.

[0256] In one implementation, the time period threshold may be adjusted or varied with at least one of the time period in which the business service takes place, the type of business service, and the gender of the customer represented by the second object.

[0257] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining the time period threshold corresponding to the current time period, if the experience data is less than or equal to the time period threshold, then an early warning information is determined. Thus, through the above operations, not only is control over whether to determine early warning information achieved, improving the controllability of early warning information determination, but also, by determining whether to determine early warning information based on the time period threshold corresponding to the current time period, the flexibility and versatility of the early warning information determination operation can be improved.

[0258] Based on the foregoing embodiments, before obtaining the time period threshold corresponding to the current time period in the early warning method provided in this application embodiment, the following steps may also be performed:

[0259] Step E1: Obtain a collection of historical experience data associated with historical time periods prior to the current time period.

[0260] In one implementation, historical experience data may include experience data determined based on at least one historical physiological indicator data of at least one second object during at least one historical period and during the provision of at least one type of business service.

[0261] In one implementation, during the process of collecting historical physiological indicator data and determining historical experience data, the distribution of historical experience data can be statistically analyzed to determine its mean, mode, and other statistical quantities. Furthermore, historical experience data can be saved using a rolling time window, for example, with a one-day time granularity. When historical experience data for a new day is acquired, the historical experience data for the earliest day can be automatically discarded. This can improve the status update of historical experience data in the time dimension and reduce the storage and processing burden caused by the large amount of historical experience data.

[0262] In one implementation, the length of the historical period can be preset; for example, the length of the historical period can be one month or one week.

[0263] Step E2: Process the historical experience data using a prediction model to obtain the prediction results for the current time period.

[0264] In one implementation, the prediction model may include a Long Short-Term Memory (LSTM) model.

[0265] In one implementation, the historical experience data can contain multiple data points. Thus, by processing the historical experience data through a prediction model, multiple prediction results can be obtained.

[0266] Step E3: Statistically average the prediction results to obtain the time period threshold corresponding to the current time period.

[0267] In one implementation, the time period threshold for the current time period can be obtained in the following way:

[0268] The prediction results are statistically averaged to obtain a statistical average result. Then, based on the service quality requirements corresponding to the business services, the statistical average result is adjusted to obtain the time period threshold corresponding to the current time period. For example, if the service quality represented by the service quality requirements is greater than or equal to the quality threshold, the statistical average result can be reduced. For example, 75% of the statistical average result can be determined as the time period threshold to facilitate the timely triggering of the early warning information determination process. Conversely, if the service quality represented by the service quality requirements is less than the quality threshold, the statistical average result can be increased. For example, 125% of the statistical average result can be used as the time period threshold to reduce the probability of false triggering of early warning information.

[0269] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining a set of historical experience data associated with historical time periods before the current time period, the historical experience data is processed by a prediction model to obtain a prediction result for the current time period. In this way, the prediction result can reflect the changing state of historical experience data from a time dimension. Furthermore, by statistically averaging the prediction result, the time period threshold corresponding to the current time period can be obtained, which can reduce the negative impact of data fluctuations in the prediction result on the time period threshold, thereby improving the accuracy of the time period threshold and also improving the forward-looking nature of the time period threshold.

[0270] Based on the foregoing embodiments, the early warning method provided in this application can further perform the following operations before obtaining the time period threshold corresponding to the current time period:

[0271] Obtain the historical threshold corresponding to the historical time period before the current time period; adjust the historical threshold to obtain the time period threshold corresponding to the current time period.

[0272] In one implementation, the time period threshold corresponding to the current time period can be obtained in the following way:

[0273] Based on the degree of difference between the service quality requirements corresponding to the current time period and the historical quality requirements corresponding to the historical threshold, the historical threshold is adjusted up or down, and the adjusted or lowered historical threshold is determined as the time period threshold corresponding to the current time period.

[0274] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining the historical threshold corresponding to the historical time period before the current time period, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period. In this way, through the above operations, the correlation between the historical threshold and the time period threshold in the time dimension is realized, and the probability of false or frequent early warnings caused by excessive differences between the time period threshold and the historical threshold can also be reduced.

[0275] Based on the foregoing embodiments, the early warning method provided in this application adjusts the historical threshold to obtain the time period threshold corresponding to the current time period, which can be achieved in the following ways:

[0276] Obtain the prediction results obtained from historical experience data. If the difference between the prediction results and the experience data is greater than or equal to the first threshold, adjust the historical threshold to obtain the time period threshold corresponding to the current time period.

[0277] Among them, historical experience data is linked to historical physiological indicator data within historical periods.

[0278] Accordingly, if the difference between the prediction result and the experience data is less than the first threshold, the operation of adjusting the historical threshold is not performed, or the historical threshold can be determined as the time period threshold for the current time period.

[0279] In one implementation, if the difference between the prediction result and the experience data is greater than or equal to a first threshold, it indicates that the prediction accuracy of the prediction model used to predict the historical experience data has deviated, and the prediction model needs to be adjusted and optimized.

[0280] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining the prediction result obtained from the prediction based on historical experience data, if the difference between the prediction result and the experience data is greater than or equal to a first threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period, and the historical experience data is associated with the historical physiological indicator data within the historical time period. Thus, through the above method, flexible control is achieved in the operation of adjusting the historical threshold to obtain the time period threshold; and, by determining whether to adjust the historical threshold based on whether the difference between the prediction result and the experience data is greater than the first threshold, strict control is achieved in the historical experience data and the time dimension of the experience data in relation to the historical experience data.

[0281] Based on the foregoing embodiments, the early warning method provided in this application, which adjusts historical thresholds to obtain time-period thresholds corresponding to the current time period, can also be implemented in the following ways:

[0282] If, within the time period corresponding to the first moment and the second time period, the amount of historical experience data associated with the historical threshold is greater than or equal to the second threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

[0283] The first moment is used to characterize the moment when the historical threshold is determined; the second moment includes future moments of the first moment.

[0284] Accordingly, if the amount of historical experience data is less than the second threshold within the time period corresponding to the first and second moments, then the historical threshold can be left unchanged, or the historical threshold can be set as the current threshold.

[0285] In one implementation, the MEC device can collect historical experience data in real time and count the amount of historical experience data in real time. If the amount of historical experience data is detected to be greater than or equal to a second threshold, an adjustment operation on the historical threshold can be initiated.

[0286] In one implementation, the second moment may include any moment during the current service process.

[0287] As can be seen from the above, in the early warning method provided in this application embodiment, if the amount of historical experience data associated with the historical threshold is greater than or equal to a second threshold within the time period between the first time corresponding to the determination time of the historical threshold and a future time of the first time, then the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period. Thus, through the above process, strict control over whether to adjust the historical threshold is achieved in terms of the amount of historical experience data; and, when historical experience data is associated with the historical threshold, judging the amount of historical experience data can improve the accuracy, diversity, and comprehensiveness of the historical threshold, thereby enabling precise adjustment of the historical threshold.

[0288] Based on the foregoing embodiments, the early warning method provided in this application, which adjusts historical thresholds to obtain time-period thresholds corresponding to the current time period, can also be implemented in the following ways:

[0289] Obtain historical experience data associated with historical thresholds; if the fluctuation between the experience data and historical experience data is greater than or equal to the third threshold, adjust the historical thresholds to obtain the time period threshold corresponding to the current time period.

[0290] Correspondingly, if the fluctuation between the experience data and the historical experience data is less than the third threshold, then the historical threshold need not be adjusted.

[0291] In one implementation, if the fluctuation between the experience data and the historical experience data is greater than or equal to a third threshold, it can indicate that the quality of the business service has fluctuated significantly.

[0292] As can be seen from the above, in the early warning method provided in this application embodiment, after obtaining historical experience data associated with historical thresholds, if the fluctuation between the experience data and historical experience data is greater than or equal to a third threshold, the historical threshold is adjusted to obtain a time-period threshold corresponding to the current time period. Thus, through the above operations, real-time monitoring of the fluctuation of experience data and historical experience data is achieved, thereby realizing real-time monitoring of the service quality of business services. Furthermore, by determining whether to adjust the historical threshold based on whether the fluctuation between historical experience data and historical experience data is greater than the third threshold, targeted adjustment of the historical threshold is achieved from the perspective of service quality of business services. This improves the accuracy of historical threshold adjustment and also enhances the correlation between historical thresholds, time-period thresholds, and changes in the service quality of business services.

[0293] It should be noted that, in the embodiments of this application, the adjustment of the historical threshold does not have to be performed at a fixed time period, but can be flexibly adjusted as the business service type changes and the business service standard is adjusted.

[0294] In this application embodiment, there is a strong correlation between the service quality of business services and customer emotions and seasonal cycles, so historical thresholds can also be adjusted according to seasonal cycles.

[0295] It should be noted that the early warning method provided in this application embodiment can be applied to multiple fields such as medical care, education, and human resources to monitor and manage the service quality of patients, students, or employees. At the same time, this solution can also be combined with big data analysis technology to deeply explore the causes and trends of service quality fluctuations in various business services, thereby providing data basis for improving service quality.

[0296] Figure 3 A schematic diagram of the service quality feedback optimization process provided in the embodiments of this application is shown below. Figure 3 As shown, the process may include the following steps:

[0297] Step 301: Collect the customer's heart rate, blood pressure and skin conductance data in real time through the seat equipped with sensors, record the customer's behavior and pupil changes through the camera to capture the customer's emotions and attitudes, and transmit all collected data to the MEC device in real time through the wireless communication module.

[0298] For example, the customer can be the second object in the foregoing embodiments; the skin conductance data can be the skin conductance level; and the pupil change can be the pupil change state in the foregoing embodiments.

[0299] Step 302: After receiving the data, MEC first performs a cleaning operation to remove outliers and noisy data, and then converts the data format to meet the requirements of subsequent analysis algorithms.

[0300] Step 303: The image recognition unit performs frame-by-frame analysis of the pupil video data to determine the pupil diameter.

[0301] For example, the pupil video data can be the pupil video in the foregoing embodiments.

[0302] Step 304: The analysis unit combines physiological indicator data and pupil diameter to analyze the customer's perception data.

[0303] Step 305: Confirm and output alarm information to customer service personnel to prompt them to adjust the behavior to be adjusted.

[0304] For example, the alarm information can be the warning information in the aforementioned embodiments, and the customer service personnel can be the first object.

[0305] Step 306: Feedback and optimization of service quality.

[0306] For example, service quality can be fed back and optimized using the same methods as those described in the foregoing embodiments.

[0307] For example, after step 306 is completed, step 304 can be executed.

[0308] Through the above process, not only is real-time analysis of customer experience data achieved, but also closed-loop real-time optimization of service quality is realized.

[0309] Based on the foregoing embodiments, this application also provides an early warning device. Figure 4 This is a schematic diagram of the structure of the early warning device provided in the embodiments of this application, such as... Figure 4 As shown, the early warning device 4 may include:

[0310] The acquisition module 401 is used to acquire at least one physiological indicator data of the second object during the process of the first object providing business services to the second object;

[0311] The determination module 402 is used to determine the experience data of the second object for the business service based on at least one physiological indicator data.

[0312] The early warning module 403 is used to determine early warning information for business services based on experience data, so as to warn the first object to adjust at least one behavior to be adjusted; wherein, the behavior to be adjusted is associated with the process of the business service.

[0313] In some embodiments, the determining module 402 is configured to determine a set of candidate features associated with a first object within a specified time period; wherein the specified time period includes at least one time period during the continuous process of business services; and the features of the candidate feature set are at least associated with the behavior of the first object within the specified time period.

[0314] The determination module 402 is used to determine the target feature set based on the features in the candidate feature set; determine the factors associated with the target feature set as target factors; output the target factors for the first object to adjust the behavior to be adjusted; wherein the degree of association between the features in the target feature set and the early warning information is greater than or equal to the degree threshold.

[0315] In some embodiments, the determining module 402 is used to sort the features in the candidate feature set based on the degree of correlation between the features in the candidate feature set and the early warning information, and obtain a sorting result;

[0316] The target feature set is determined to include at least one feature from the ranking results.

[0317] In some embodiments, the acquisition module 401 is used to acquire historical sorting parameters and / or historical feedback data;

[0318] The determination module 402 is used to determine the degree of correlation between the features in the candidate feature set and the early warning information; based on the degree of correlation, historical ranking parameters and / or historical feedback data, the features in the candidate feature set are ranked to obtain the target feature set.

[0319] In some embodiments, the behavior of the first object during a specified time period includes the physical behavior and / or language behavior of the first object during the specified time period; the candidate feature set is also associated with the environment in which the first object and the second object are located during the specified time period; the determining module 402 is used to extract features from the behavioral data corresponding to the physical behavior and / or language behavior to obtain physical features corresponding to the physical behavior and / or language features corresponding to the language behavior.

[0320] The acquisition module 401 is used to collect environmental parameters of the environment in which the first object and the second object are located within a specified time period.

[0321] The determination module 402 is used to process environmental parameters to obtain environmental features of the environment; based on the degree of matching between environmental features, limb features and / or language features and benchmark environmental features, benchmark limb features and / or benchmark language features, respectively, the features in the candidate feature set are filtered to obtain the target feature set.

[0322] In some embodiments, the determining module 402 is used to quantify at least one physiological indicator data to obtain a first quantification result;

[0323] Acquisition module 401 is used to acquire pupil video including the pupil region of the second object;

[0324] The determination module 402 is used to analyze the pupil video and determine the pupil change state in the pupil region; quantify the pupil change state to obtain a second quantization result; and determine the experience data based on the first quantization result and the second quantization result.

[0325] In some embodiments, at least one physiological indicator data includes at least one of the second subject's heart rate, blood pressure, and skin conductance level; quantifying at least one physiological indicator data to obtain a first quantification result includes:

[0326] The acquisition module 401 is used to acquire the initial heart rate, initial blood pressure, and initial skin conductance level at an initial moment; wherein, the initial moment includes the historical moment of the current moment;

[0327] The determining module 402 is used to determine a heart rate threshold, a blood pressure threshold, and a conductance threshold; quantify the heart rate based on the initial heart rate and the heart rate threshold to obtain a heart rate quantification result; quantify the blood pressure based on the initial blood pressure and the blood pressure threshold to obtain a blood pressure quantification result; quantify the skin conductance level based on the initial skin conductance level and the conductance threshold to obtain a skin conductance quantification result; and determine a first quantification result based on at least one of the heart rate quantification result, the blood pressure quantification result, and the skin conductance quantification result.

[0328] In some embodiments, the determining module 402 is used to determine weight parameters; based on the weight values ​​in the weight parameters, the first quantization result and the second quantization result are processed to obtain experience data.

[0329] In some embodiments, the acquisition module 401 is used to acquire the time period threshold corresponding to the current time period;

[0330] The determination module 402 is used to determine the warning information if the experience data is less than or equal to the time period threshold.

[0331] In some embodiments, the acquisition module 401 is used to acquire a set of historical experience data associated with historical time periods before the current time period;

[0332] The determination module 402 is used to process historical experience data through a prediction model to obtain prediction results for the current time period;

[0333] The determination module 402 is used to perform statistical averaging on the prediction results to obtain the time period threshold corresponding to the current time period.

[0334] In some embodiments, the acquisition module 401 is used to acquire historical thresholds corresponding to historical time periods before the current time period;

[0335] The determination module 402 is used to adjust the historical threshold to obtain the time period threshold corresponding to the current time period.

[0336] In some embodiments, the acquisition module 401 is used to acquire the prediction result obtained by predicting based on historical experience data; wherein the historical experience data is associated with historical physiological indicator data within a historical period.

[0337] The determination module 402 is used to adjust the historical threshold if the difference between the prediction result and the experience data is greater than or equal to the first threshold, so as to obtain the time period threshold corresponding to the current time period.

[0338] In some embodiments, the determining module 402 is configured to adjust the historical threshold if the amount of historical experience data associated with the historical threshold is greater than or equal to the second threshold within the time period corresponding to the first time period and the second time period, thereby obtaining a time period threshold corresponding to the current time period; wherein the first time period is used to characterize the time at which the historical threshold is determined; and the second time period includes a future time of the first time period.

[0339] In some embodiments, the acquisition module 401 is used to acquire historical experience data associated with historical thresholds;

[0340] The determination module 402 is used to adjust the historical threshold if the fluctuation between the experience data and the historical experience data is greater than or equal to the third threshold, so as to obtain the time period threshold corresponding to the current time period.

[0341] This application also provides an early warning device. Figure 5 This is a schematic diagram of the structure of the early warning device provided in the embodiments of this application, such as... Figure 5 As shown, the early warning device 5 includes a processor 501 and a memory 502; wherein, the memory 502 stores a computer program; when the computer program is executed by the processor 501, it can implement the early warning method as described above.

[0342] This application also provides a computer-readable storage medium storing a computer program; when the computer program is executed by a processor of an electronic device, it can implement the early warning method as described above.

[0343] This application also provides a computer program product, which includes a computer program; when the computer program is executed by the processor of an electronic device, it can implement the early warning method as described above.

[0344] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0345] The methods disclosed in the various method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict.

[0346] The features disclosed in the various product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0347] The features disclosed in the various method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0348] It should be noted that the aforementioned computer-readable storage media can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc.; or it can be various electronic devices that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0349] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0350] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0351] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware nodes. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0352] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0353] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0354] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0355] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. An early warning method, characterized in that, The method includes: In the process of the first object providing business services to the second object, at least one physiological indicator data of the second object is obtained; Based on the at least one physiological indicator data, determine the experience data of the second object for the business service; Based on the experience data, early warning information is determined for the business service to warn the first object to adjust at least one behavior to be adjusted; wherein, the behavior to be adjusted is associated with the process of the business service.

2. The method according to claim 1, characterized in that, After determining the warning information for the service based on the experience data, the method further includes: Determine a set of candidate features associated with the first object within a specified time period; wherein the specified time period includes at least one time period during the continuous process of the business service; the features of the candidate feature set are at least associated with the behavior of the first object within the specified time period; Based on the features in the candidate feature set, a target feature set is determined; wherein the correlation between the features in the target feature set and the warning information is greater than or equal to a degree threshold. Factors associated with the target feature set are identified as target factors; The target factors are output so that the first object can adjust the behavior to be adjusted.

3. The method according to claim 2, characterized in that, Determining the target feature set based on the features in the candidate feature set includes: Based on the degree of correlation between the features in the candidate feature set and the early warning information, the features in the candidate feature set are sorted to obtain a sorting result; The target feature set is determined to include at least one feature from the sorting results.

4. The method according to claim 2, characterized in that, Determining the target feature set based on the features in the candidate feature set includes: Obtain historical sorting parameters and / or historical feedback data; Determine the degree of correlation between the features in the candidate feature set and the early warning information; Based on the degree of correlation, the historical ranking parameters, and / or the historical feedback data, the features in the candidate feature set are ranked to obtain the target feature set.

5. The method according to claim 2, characterized in that, The behavior of the first object during the specified time period includes the physical and / or verbal behavior of the first object during the specified time period; the candidate feature set is also associated with the environment in which the first object and the second object are located during the specified time period. Determining the target feature set based on the features in the candidate feature set includes: Feature extraction is performed on the behavioral data corresponding to the physical behavior and / or the language behavior to obtain the physical features corresponding to the physical behavior and / or the language features corresponding to the language behavior; Collect environmental parameters of the environment in which the first object and the second object are located during the specified time period; The environmental parameters are processed to obtain the environmental characteristics of the environment; Based on the environmental features, the body features, and / or the language features, and the degree of matching between them and the baseline environmental features, baseline body features, and / or baseline language features, the features in the candidate feature set are filtered to obtain the target feature set.

6. The method according to claim 1, characterized in that, The step of determining the experience data of the second object for the business service based on the at least one physiological indicator data includes: The at least one physiological indicator data is quantified to obtain a first quantification result; Acquire pupil video including the pupil region of the second object; The pupil video is analyzed to determine the pupil change state in the pupil region; The pupil change state is quantified to obtain a second quantization result; The experience data is determined based on the first quantization result and the second quantization result.

7. The method according to claim 6, characterized in that, The at least one physiological indicator data includes at least one of the second subject's heart rate, blood pressure, and skin conductance level; the quantification of the at least one physiological indicator data to obtain a first quantification result includes: The initial heart rate, initial blood pressure, and initial skin conductance level are obtained at an initial moment; wherein, the initial moment includes historical moments of the current moment; Determine the heart rate threshold, blood pressure threshold, and conductance threshold; The heart rate is quantified based on the initial heart rate and the heart rate threshold to obtain a heart rate quantification result; the blood pressure is quantified based on the initial blood pressure and the blood pressure threshold to obtain a blood pressure quantification result; and the skin conductance level is quantified based on the initial skin conductance level and the conductance threshold to obtain a skin conductance quantification result. The first quantification result is determined based on at least one of the heart rate quantification result, the blood pressure quantification result, and the skin conductance quantification result.

8. The method according to claim 6, characterized in that, The step of determining the experience data based on the first quantization result and the second quantization result includes: Determine the weighting parameters; Based on the weight values ​​in the weight parameters, the first quantization result and the second quantization result are processed to obtain the experience data.

9. The method according to claim 1, characterized in that, The determination of early warning information for the service based on the experience data includes: Get the time period threshold corresponding to the current time period; If the experience data is less than or equal to the time period threshold, the warning information is determined.

10. The method according to claim 9, characterized in that, Before obtaining the time period threshold corresponding to the current time period, the method further includes: Obtain a collection of historical experience data associated with historical time periods prior to the current time period; The historical experience data is processed by a prediction model to obtain a prediction result for the current time period; The prediction results are statistically averaged to obtain the time period threshold corresponding to the current time period.

11. The method according to claim 9, characterized in that, Before obtaining the time period threshold corresponding to the current time period, the method further includes: Obtain the historical threshold corresponding to the historical time periods preceding the current time period; The historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

12. The method according to claim 11, characterized in that, The step of adjusting the historical threshold to obtain the time period threshold corresponding to the current time period includes: Obtain prediction results obtained by making predictions based on historical experience data; wherein, the historical experience data is associated with historical physiological indicator data within a historical period; If the difference between the prediction result and the experience data is greater than or equal to the first threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

13. The method according to claim 11, characterized in that, The step of adjusting the historical threshold to obtain the time period threshold corresponding to the current time period includes: If, within the time period corresponding to the first time point and the second time point, the amount of historical experience data associated with the historical threshold is greater than or equal to the second threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period; wherein, the first time point is used to characterize the time at which the historical threshold is determined; the second time point includes future times of the first time point.

14. The method according to claim 11, characterized in that, The step of adjusting the historical threshold to obtain the time period threshold corresponding to the current time period includes: Obtain historical experience data associated with the historical threshold; If the fluctuation between the experience data and the historical experience data is greater than or equal to the third threshold, the historical threshold is adjusted to obtain the time period threshold corresponding to the current time period.

15. An early warning device, characterized in that, The early warning device includes: The acquisition module is used to acquire at least one physiological indicator data of the second object during the process of the first object providing business services to the second object; The determination module is used to determine the experience data of the second object for the business service based on the at least one physiological indicator data; The early warning module is used to determine early warning information for the business service based on the experience data, so as to warn the first object to adjust at least one behavior to be adjusted; wherein the behavior to be adjusted is associated with the process of the business service.

16. An early warning device, characterized in that, The early warning device includes a processor and a memory; wherein the memory stores a computer program; when the computer program is executed by the processor, it can implement the early warning method as described in any one of claims 1 to 14.

17. A computer-readable storage medium, characterized in that, The storage medium stores a computer program; when the computer program is executed by the processor of the electronic device, it can implement the early warning method as described in any one of claims 1 to 14.

18. A computer program product, characterized in that, The program product includes a computer program; when the computer program is executed by the processor of an electronic device, it is capable of implementing the early warning method as described in any one of claims 1 to 14.

Images