Risk early warning method, device and system for service order and electronic equipment

Abstract

The invention provides a risk early warning method and device for a service order, electronic equipment and a storage medium, and relates to the technical field of computers. The method comprises the steps of obtaining customer information, service information and service provider information in response to received order data for a target service; extracting features of corresponding dimensions from the customer information, the service information and the service provider information to obtain multi-dimensional features; performing risk prediction on the multi-dimensional features by using a risk prediction model to obtain a prediction result of each risk type; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network; and generating a risk early warning event of the current service order based on the prediction result so as to carry out risk early warning. According to the method and the device, the problems of poor customer experience and low service quality caused by an early warning mode after a service risk occurs in related technologies are solved to a certain extent.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and in particular, to a method, an apparatus, a system, and an electronic device for risk warning of service orders. Background Art

[0002] With the refined requirements for services, there are more and more scenarios of on-site services. However, such scenarios that require coordination of multiple parties of resources are more likely to cause customer complaints and service timeouts. Most of the existing solutions for such scenarios monitor the service process through multi-party status monitoring and timeout warnings. However, this method can only give warnings after service risks occur, resulting in poor customer experience and affecting service quality.

[0003] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0004] The purpose of the embodiments of the present disclosure is to provide a method, an apparatus, an electronic device, and a storage medium for risk warning of service orders, thereby solving to a certain extent the problems of poor customer experience and low service quality caused by the warning method after service risks occur in the related art.

[0005] According to a first aspect of the present disclosure, there is provided a method for risk warning of service orders, including: in response to receiving order data for a target service, obtaining customer information, service information, and service provider information; respectively extracting features of corresponding dimensions from the customer information, service information, and service provider information to obtain multi-dimensional features; using a risk prediction model to perform risk prediction on the multi-dimensional features to obtain prediction results of each risk type; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network; generating a risk warning event for the current service order based on the prediction results to perform risk warning.

[0006] Optionally, the service information includes service recipient information, service execution condition information, order assignment information, order rejection information, and service risk data; obtaining customer information, service information, and service provider information includes: in response to the operation platform generating a service order for a target service, scraping customer information and service recipient information from the order data of the service order; in response to the operation platform generating a service execution reservation form for the service order, scraping service execution condition information from the reservation data corresponding to the reservation form; in response to the order assignment of the operation platform for the service execution reservation form or the order rejection by the service provider, obtaining the order assignment information or order rejection information for the service execution reservation form; obtaining service provider information and service risk data from the static database, where the service risk data includes first risk data corresponding to the service order type and second risk data corresponding to the service recipient type.

[0007] Optionally, using a risk prediction model to perform risk prediction on multi-dimensional features includes: inputting the multi-dimensional features into the risk prediction model to output probability values corresponding to each risk type, where the risk prediction model is obtained by training a Catboost classifier on simulated risk data and real order data.

[0008] Optionally, the method further includes: determining the candidate risk order quantity and model prediction performance corresponding to each candidate threshold of the risk prediction model; determining the expected risk order push quantity within the target period according to the model prediction performance requirement and the operation processing capacity within the target period; matching the expected risk order push quantity among the candidate risk order quantities, and using the candidate threshold corresponding to the successfully matched candidate risk order quantity as the target threshold corresponding to the target period; determining the prediction results of each risk type within the target period according to the comparison results of the probability values corresponding to each risk type and the target threshold.

[0009] Optionally, extracting features of corresponding dimensions from customer information, service information, and service provider information includes: respectively extracting corresponding attribute fields from customer information, service information, and service provider information according to preset target attributes to obtain multi-dimensional features.

[0010] Optionally, before extracting the multi-dimensional features, the method further includes: respectively performing data cleaning on customer information, service information, and service provider information, where the data cleaning includes at least one of data deduplication, missing value processing, outlier exclusion, and data standardization.

[0011] Optionally, the method further includes: within the target period, pushing a risk warning event to the operation platform according to the prediction results determined by the target threshold.

[0012] Optionally, the method further includes: determining the risk level of each service provider according to the risk warning event for each service provider; adjusting the order assignment strategy of the order assignment system according to the risk levels of each service provider.

[0013] According to a second aspect of the present disclosure, there is provided a risk warning device for service orders, the device comprising: an acquisition module configured to acquire customer information, service information, and service provider information in response to receiving order data for a target service; a feature extraction module configured to extract features in corresponding dimensions from the customer information, service information, and service provider information respectively to obtain multi-dimensional features; a prediction module configured to perform risk prediction on the multi-dimensional features by using a risk prediction model to obtain prediction results of various risk types; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network; and a warning module configured to generate a risk warning event for the current service order based on the prediction results to perform risk warning.

[0014] According to a third aspect of the present disclosure, there is provided a risk warning system for service orders, comprising the risk warning device of any one of the above embodiments.

[0015] According to a fourth aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program, which when executed by a processor, implements the method of any one of the above embodiments.

[0016] According to a fifth aspect of the present disclosure, there is provided an electronic device, comprising: one or more processors; and a storage device for one or more programs, which when executed by the one or more processors, cause the one or more processors to execute the method of any one of the above embodiments.

[0017] The exemplary embodiments of the present disclosure may have some or all of the following beneficial effects:

[0018] In the risk warning method for service orders provided by the exemplary embodiment of the present disclosure, on the one hand, by using a risk prediction model to perform risk prediction on multi-dimensional features, risk factors that may affect service quality in a service order can be known in advance, realizing pre-discovery warning and processing of risks, and thus to a certain extent avoiding the occurrence of abnormal service order execution phenomena and improving the customer experience. On the other hand, the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data generated based on a trained generative adversarial network can be used to train the machine learning model to obtain the risk prediction model, improving the prediction accuracy of the risk prediction model; furthermore, by acquiring customer information, service information, and service provider information and extracting corresponding multi-dimensional features, comprehensive risk assessment of service orders can be performed from multiple dimensions such as customers, services, and service providers, ensuring the reliability of the prediction results.

[0019] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and do not limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0021] Figure 1 FIG. 1 schematically shows one of the flow diagrams of the risk warning method for service orders according to an embodiment of the present disclosure.

[0022] Figure 2 FIG. 2 schematically shows the schematic diagram of the training process of the risk prediction model according to an embodiment of the present disclosure.

[0023] Figure 3 FIG. 3 schematically shows another flow diagram of the risk warning process for service orders according to an embodiment of the present disclosure.

[0024] Figure 4 FIG. 4 schematically shows the structural block diagram of the risk warning device for service orders according to an embodiment of the present disclosure.

[0025] Figure 5 FIG. 5 shows a block diagram of an electronic device suitable for implementing the embodiments of the present disclosure. DETAILED DESCRIPTION

[0026] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more complete and will fully convey the concept of the example embodiments to those skilled in the art. The features, structures, or characteristics described may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of the present disclosure. However, those skilled in the art will recognize that the technical solutions of the present disclosure can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be used. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring the various aspects of the present disclosure.

[0027] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0028] The risk warning method for service orders of the present disclosure can be applied to a risk prevention and control platform for service orders, which can communicate with the operation platform of service orders to obtain real-time order data and historical order data. Specifically, the execution entity of the risk prevention and control platform can be various terminals or servers, or a system including terminals and servers.

[0029] The terminal can be a laptop computer, a tablet computer, a desktop computer, a set-top box, various mobile terminals such as mobile phones, tablets or wearable devices, etc. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud storage, etc., or a node in a blockchain. This example is not limited thereto.

[0030] Next, in combination with specific embodiments, the risk warning method for service orders disclosed in the embodiments of this specification will be introduced.

[0031] Refer to Figure 1 As shown, a risk warning method for service orders in an exemplary implementation manner provided by the present disclosure, which can be a risk prevention and control platform, may include the following steps:

[0032] Step S110, in response to receiving order data for a target service, obtain customer information, service information, and service provider information;

[0033] Step S120, respectively extract features of corresponding dimensions from the customer information, service information, and service provider information to obtain multi-dimensional features;

[0034] Step S130, use a risk prediction model to perform risk prediction on the multi-dimensional features to obtain prediction results of each risk type; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network;

[0035] Step S140, generate a risk warning event for the current service order based on the prediction results to perform risk warning.

[0036] In the risk early warning method of the service order provided by the present exemplary embodiment, on the one hand, by using the risk prediction model to perform risk prediction on multi-dimensional features, the risk factors that may affect the service quality of the service order can be obtained in advance, realizing the pre-discovery warning and processing of risks, and thus to a certain extent avoiding the occurrence of abnormal service order execution phenomena and improving the customer experience. On the other hand, the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data. The machine learning model can be trained based on the simulated risk data generated by the trained generative adversarial network to obtain the risk prediction model, improving the prediction accuracy of the risk prediction model. Moreover, by obtaining customer information, service information, and service provider information and extracting the corresponding multi-dimensional features, a comprehensive risk assessment of the service order can be carried out from multiple dimensions such as customers, services, and service providers, ensuring the reliability of the prediction results.

[0037] The following describes each step of the present disclosure in more detail.

[0038] In step S110, in response to receiving the order data of the service order for the target service, customer information, service information, and service provider information are obtained.

[0039] In the present exemplary embodiment, the operation platform can generate a service order in response to the customer's order placement operation on the client side; the order data of the service order can be uploaded by the operation platform or called by the risk prevention and control platform. This exemplary embodiment does not limit this data acquisition method. The target service refers to a service that requires service personnel to arrive at a specified place to perform corresponding services. For example, services such as installation, repair, and cleaning of products (such as furniture and household appliances). The order data of the service order may include information related to the order itself, customer information corresponding to the order, and service information corresponding to the order. Information related to the order itself may include information such as order number, order identifier, order type (such as urgent order or ordinary order), order placement time, and order timeliness; customer information corresponding to the order may include basic customer information (such as name, contact information, user name, address, etc.), customer type (such as individual customer or enterprise customer), customer level (such as member or non-member), customer's historical complaint situation (such as historical complaint rate or historical complaint quantity), and so on.

[0040] In the present exemplary embodiment, the service information corresponding to an order may include service object information (such as product information corresponding to the service, such as product type / category, product name, manufacturer, model, etc.), service execution condition information (such as service time, service location, service execution environment, estimated service duration, service required equipment, etc.), dispatch information (such as dispatch order number, dispatch time, dispatch object, service timeliness, etc.), rejection information (such as rejection trigger, rejection source, rejection time, rejection reason, rejection number), and service risk data (such as first risk data corresponding to the service order type and second risk data corresponding to the service object type). The service risk data refers to data that may bring service risks, such as service timeout data, service quality unqualified data, service personnel not implementing service data, customer false reservation data, or false feedback data, that is, data that cannot complete the service on time and with guaranteed quality. The service provider information refers to information related to the service provider. For example, when the service is an appliance installation service, the service provider is the appliance installer. The service provider information may include the basic information of the service provider, service information, customer complaint information, and service information and customer complaint information of the service network to which the service provider belongs; the service information may include service duration, service on-time rate, service quality, service quantity, and fluctuations in service quality / quantity within a certain period of time; the customer complaint information may include the number of customer complaints, average daily customer complaint volume, fluctuations in customer complaint volume, customer complaint types, customer complaint reasons, and other information within a certain period of time. In this example, the customer information and service information can be obtained through online real-time and offline methods, and the service provider information can be obtained through offline methods.

[0041] Exemplarily, the service information includes service object information, service execution condition information, dispatch information, rejection information, and service risk data; obtaining the customer information, service information, and service provider information may include the following steps:

[0042] First step, in response to the operation platform generating a service order for a target service, capture the customer information and service object information from the order data of the service order.

[0043] In the present exemplary embodiment, the operation platform generates a corresponding service order in response to the customer's order placement operation for the target service. The risk prevention and control platform can monitor the orders of the operation platform in real time. When a new order is generated, capture the customer information and service object information from the order data of the service order.

[0044] Second step, in response to the operation platform generating a service execution reservation order for the service order, capture the service execution condition information from the reservation data corresponding to the reservation order.

[0045] In the present exemplary embodiment, the operation platform may generate a service execution reservation form in response to a service reservation operation of a customer at the user terminal, and a binding relationship may be established between the service execution reservation form and the corresponding service order. The reservation data may include a reservation number, a service type, a reserved time for service execution, and service execution condition information (such as information on reserved service duration, service execution location, environment, etc.).

[0046] In the third step, in response to the dispatching of the service execution reservation form by the operation platform or the rejection of the form by the service provider, obtain the dispatching information or rejection information for the service execution reservation form.

[0047] In the present exemplary embodiment, the operation platform may screen out a list of candidate service personnel meeting the conditions based on the service execution location, reserved time, and service type in the service execution reservation form, and then perform dispatching based on the priorities of the service personnel in the list of candidate service personnel, that is, send a dispatching task to the terminal corresponding to the determined service personnel, and the service personnel may accept or reject the dispatching task. The priority of the service personnel may be determined based on influencing factors such as the historical customer complaint situation, service quality, service overtime situation, rejection rate, etc. of the service personnel, and different weight values may be set for each influencing factor to calculate the corresponding priority coefficient. Service personnel with a low customer complaint rate, high service quality, low overtime rate, and low rejection rate may also be set with a higher dispatching priority, and the present example does not limit this.

[0048] In the present exemplary embodiment, the dispatching task includes dispatching information, and the dispatching information may include all or part of the information in the reservation data corresponding to the service execution reservation form, and may also include information such as the dispatching time and service precautions, and the present example does not limit this. In response to the operation platform receiving a rejection message, obtain the rejection information, and the rejection information may include information such as the reason for rejection, rejection time, and rejection source, and the present example does not limit this.

[0049] In the fourth step, obtain service provider information and service risk data from the static database, where the service risk data includes first risk data corresponding to the service order type and second risk data corresponding to the service object type.

[0050] In the present exemplary embodiment, the static database may include an offline service provider information library and an offline historical service risk database. The service provider information library stores service provider information, and the historical service risk database stores first risk data corresponding to each service order type and second risk data corresponding to the service object type. The first risk data and the second risk data may include corresponding customer complaint information.

[0051] In the above embodiments, the model side can obtain online data and offline data related to service orders in real time, evaluate order risks using the online data and offline data. Offline features can be calculated and statistically analyzed through Spark, and the offline data can be updated by scheduling the data of Spark. This not only considers the historical data features of each dimension but also combines the real-time situation of the order, mining historical patterns while conforming to the actual situation. Sensitive information in customer information and service provider information can be encrypted before transmission.

[0052] In step S120, corresponding dimensional features are respectively extracted from customer information, service information, and service provider information to obtain multi-dimensional features.

[0053] In the present exemplary embodiment, feature extraction can be performed based on preset rules, or by means such as principal component analysis or key information extraction. This example does not limit this. Exemplarily, target attributes for feature extraction can be respectively set for customer information, service information, and service provider information. According to each target attribute, the corresponding attribute fields are respectively extracted from customer information, service information, and service provider information to obtain multi-dimensional features. Customer, service, and service provider can be respectively used as a feature dimension for feature extraction at the first-level dimension, or the attributes corresponding to each first-level dimension can be used as a feature dimension for feature extraction at the second-level dimension. This example does not limit this. Feature extraction can be performed in units of service order identifiers.

[0054] In some embodiments, before extracting multi-dimensional features, data cleaning can also be respectively performed on customer information, service information, and service provider information. Data cleaning includes at least one of data deduplication, missing value processing, outlier exclusion, and data standardization.

[0055] In the present exemplary embodiment, data cleaning refers to removing data errors, outliers, and noise in the data to make the data more accurate, reliable, and easy to analyze. Data deduplication refers to deleting duplicate data, which can reduce the amount of data processing. Missing value processing can be respectively performed for categorical and numerical types, and each type of missing data is replaced with appropriate data according to the distribution characteristics of the data.

[0056] In some embodiments, data evaluation can also be performed after data cleaning to ensure the accuracy and integrity of the data, and data verification (determining whether the data belongs to the corresponding business scenario) can be performed to ensure that the data conforms to the business scenario.

[0057] In step S130, a risk prediction model is used to perform risk prediction on the multi-dimensional features to obtain prediction results for each risk type.

[0058] In the present exemplary embodiment, the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network. The multi-dimensional features can be vectorized and then input into the risk prediction model for risk prediction. Specifically, the features of each dimension can be arranged in sequence to form a feature vector.

[0059] Exemplarily, before risk prediction, the model needs to be trained first. The training process of the model is as Figure 2 shown. First, a training sample set can be obtained. For example, historical log data for the target service in the past month / several months can be obtained, and the historical log data is subjected to data cleaning (such as at least one of data deduplication, missing value processing, outlier exclusion, and data standardization). The generative adversarial network is trained using the training sample set. The generative adversarial network includes a generator and a discriminator. The generator learns the features of the input training samples and generates forged data similar to the training samples; the discriminator determines whether the input data is real data or forged data and predicts the authenticity of the forged data. The generator continuously learns to produce forged data closer to the real data, while the discriminator deceives the generator by judging the difference between the forged data and the real data. The two networks continuously learn and compete with each other until the difference between the forged data generated by the generator and the real data is continuously less than a preset value, completing the training process of the generative adversarial network. Using the trained generative adversarial network, simulated risk data similar to it is generated based on the input features corresponding to the training samples (real risk data); using the simulated risk data and real order data (such as real risk data and normal order data), a catboost classification model (risk prediction model) is constructed, and the classification model parameters are updated by calculating the loss function until the classification model converges, completing the training of the classification model. The catboost classification model is a machine learning model based on gradient boosting decision trees and is a binary classification model based on decision trees. A corresponding classifier can be constructed for each risk type to complete the classification of all risk types. The above generative adversarial network and classification model can be updated regularly (i.e., retrained, such as once a week), and by retraining the model, features with higher correlation with risks are captured to ensure the prediction accuracy of the model.

[0060] Exemplarily, the multi-dimensional features are input into the risk prediction model, and the probability values corresponding to each risk type are output. The risk prediction model is obtained by training a Catboost classifier on simulated risk data and real order data.

[0061] In this exemplary embodiment, the risk types may include service timeout, false reservation, the service personnel not going to the reservation service execution location, unqualified service quality, false feedback, etc. The probability values corresponding to each risk type are output by the risk prediction model, and based on the comparison result between the probability value and the target threshold, it is determined whether each risk type exists in the current service order (prediction result). The target threshold can be set according to the actual situation.

[0062] In some embodiments, the target threshold can be determined through the following steps:

[0063] First step, determine the number of candidate risk orders corresponding to the risk prediction model under each candidate threshold and the model prediction performance.

[0064] In this exemplary embodiment, the candidate thresholds can be set according to the actual situation. For example, the candidate thresholds can be set to 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, etc. The target threshold can be set to each candidate threshold, and the number of risk orders predicted on the current day under each candidate threshold is determined. The model prediction performance can be characterized by one or more of the model prediction performance indicators such as the prediction precision rate, recall rate, and accuracy rate of the risk prediction model, and the prediction performance indicators of the current risk prediction model can be determined through the model testing process.

[0065] Second step, according to the model prediction performance requirements and the operation processing capacity within the target period, determine the expected risk order push volume within the target period.

[0066] In this exemplary embodiment, the model prediction performance requirements refer to the requirements for each prediction performance indicator of the risk prediction model. For example, the model prediction performance requirements are that the precision rate is greater than a1 and the recall rate is greater than a2. The operation processing capacity within the target period can be determined based on the time for operation personnel to process risk orders within the target period. The longer the processing time, the lower the processing capacity, and the two have an inverse correlation relationship. Therefore, the operation processing capacity can be determined through the inverse correlation relationship formula of the processing time.

[0067] Third step, match the expected risk order push volume among the numbers of candidate risk orders, and use the candidate threshold corresponding to the number of candidate risk orders with successful matching as the target threshold corresponding to the target period.

[0068] In this exemplary embodiment, the expected risk order push volume refers to the preset risk order push volume within the target period. The range where the expected risk order push volume falls into the number of candidate risk orders is determined to be a successful match. For example, the expected risk order push volume is 1500, and the number of candidate risk orders corresponding to the candidate threshold 0.85 within one day is 1000 - 2000, then the two match successfully, and the candidate threshold 0.85 corresponding to the number of candidate risk orders with successful matching is used as the target threshold corresponding to that day.

[0069] Step 4: Determine the prediction results for each risk type based on the comparison results between the probability values corresponding to the respective risk types and the target thresholds.

[0070] In the present exemplary embodiment, a situation where the probability value output by the classifier is greater than the target threshold can be determined as the existence of the corresponding risk type. The target threshold is determined for the target time period, and within different time periods, the target threshold can be different or the same, and it is necessary to determine the target thresholds and prediction results corresponding to each time period according to the above steps.

[0071] The above embodiments can ensure the model prediction performance, significantly reduce the customer complaint risk of the target service, and improve the service quality; at the same time, it can ensure that the pushed orders can be processed in a timely manner, ensure the timely handling of risks, and further improve the service quality.

[0072] In step S140, generate a risk warning event for the current service order based on the prediction result for risk warning.

[0073] In the present exemplary embodiment, the prediction result may include the risk type of the order. Generate a corresponding risk warning event according to the risk type of the order, so that the operation personnel can learn about the order risk in advance and intervene and handle it in advance to reduce the risk occurrence rate. Exemplarily, the operation personnel can verify the risk warning event and intervene and handle it when it is confirmed that the risk exists. The intervention and handling may include reassigning the order, contacting the customer for communication and settlement and soothing the customer, cooperating with the service network personnel for handling, order follow-up, etc.

[0074] In some embodiments, within the target time period, a risk warning event can be pushed to the operation platform according to the prediction result determined by the target threshold.

[0075] In the present exemplary embodiment, within the target time period (such as one day), a risk warning event can be generated according to the risk order corresponding to the prediction result determined by the target threshold, and the risk warning event can be pushed to the operation platform to perform advance intervention and handling on the risk order, improve the service quality, and reduce the customer complaint rate. The number of risk orders in different time periods can be different or the same, and it is specifically determined according to the corresponding target threshold determined.

[0076] In some embodiments, the risk warning events of each service provider can be used to classify the risk levels or dispatch priorities of each service provider, and then the dispatch strategy of the dispatch system can be adjusted according to the risk levels of each service provider. For example, orders with higher risks or rejected orders can be preferentially dispatched to service providers with higher priorities to improve the completion efficiency of service orders.

[0077] For example, such as Figure 3As shown, the risk warning method for service orders can be completed through interaction between the service operation platform and the risk prevention and control platform, and this example does not limit this.

[0078] Step S301, the risk prevention and control platform monitors the order data of the operation platform. In this example, in response to a customer's order placement operation for a target service, the operation platform generates a service order, and the service order can also be generated by the order system and sent to the operation platform.

[0079] Step S302, in response to monitoring the service order, obtain customer information and service information from the order data.

[0080] Step S303, in response to monitoring the service execution reservation form of the operation platform, the risk prevention and control platform grabs service execution condition information from the reservation data corresponding to the reservation form;

[0081] Step S304, in response to the dispatching of the service execution reservation form by the operation platform or the rejection of the order by the service provider, the risk prevention and control platform obtains the dispatching information or rejection information for the service execution reservation form;

[0082] Step S305, the risk prevention and control platform obtains service provider information and service risk data from the static database.

[0083] Step S306, the risk prevention and control platform performs data cleaning on the data obtained in S301 - S305 above.

[0084] Step S307, the risk prevention and control platform extracts features of corresponding dimensions from the cleaned data to obtain multi-dimensional features.

[0085] Step S308, the risk prevention and control platform inputs the multi-dimensional features into the risk prediction model for risk prediction to obtain an output probability value.

[0086] Step S309, the risk prevention and control platform determines the target threshold corresponding to the target period according to the model prediction performance and the operation processing capacity within the target period.

[0087] Step S310, the risk prevention and control platform determines the prediction result according to the output probability value and the target threshold.

[0088] Step S311, the risk prevention and control platform generates a risk warning event for the current service order based on the prediction result.

[0089] Step S312, the risk prevention and control platform pushes the risk warning event to the operation platform.

[0090] In this example, the risk prediction model is a trained catboot classification model. Real-time order data can be obtained through a real-time interface. The big data model side can be set for each model in the process, or it can be set within the risk prevention and control platform. This example does not make any restrictions on this.

[0091] The specific details of each step in this embodiment have been described in detail in the foregoing embodiments, so they will not be repeated here.

[0092] This disclosure designs a pre-positioned prediction and early warning scheme for risk links for target service business scenarios that need to go to a specified location. For risk types such as service timeout, false reservation, false feedback, non-standard installation quality, and insecurity, data mining is carried out from multiple dimensions such as customer characteristics, service characteristics, and service providers. Early warning and identification of scenarios where customer complaints may occur in the future are carried out and pushed to the operation platform in real time, providing risk warnings for each entity involved in the service provider, realizing the pre-knowledge and prevention and control of customer complaint risks. By pre-intervening and processing risks, the occurrence of customer complaints for on-site services is effectively reduced, and the user experience and service quality are improved.

[0093] Furthermore, in the implementation manner of this example, a risk early warning device 400 for service orders is also provided. The risk early warning device 400 for service orders can be applied to a risk prevention and control platform. Refer to Figure 4 As shown, the risk early warning device 400 for service orders may include: an acquisition module 410, configured to acquire customer information, service information, and service provider information in response to receiving order data for a target service. A feature extraction module 420, configured to extract features of corresponding dimensions from the customer information, service information, and service provider information respectively to obtain multi-dimensional features. A prediction module 430, configured to perform risk prediction on the multi-dimensional features by using a risk prediction model to obtain prediction results of each risk type; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network. An early warning module 440, configured to generate a risk early warning event for the current service order based on the prediction results for risk early warning.

[0094] In an exemplary embodiment of the present disclosure, the service information includes service object information, service execution condition information, order assignment information, order rejection information, and service risk data; the acquisition module is further configured to:

[0095] In response to the operation platform generating a service order for a target service, grab customer information and service object information from the order data of the service order;

[0096] In response to the operation platform generating a service execution reservation form for the service order, grab service execution condition information from the reservation data corresponding to the reservation form;

[0097] In response to the dispatching of a reservation order for service execution by the operation platform or the rejection of an order by the service provider, obtain the dispatching information or rejection information for the reservation order for service execution;

[0098] Obtain service provider information and service risk data from the static database, where the service risk data includes first risk data corresponding to the service order type and second risk data corresponding to the service object type.

[0099] In an exemplary embodiment of the present disclosure, the prediction module 430 is further configured to: input multi-dimensional features into a risk prediction model and output probability values corresponding to each risk type, where the risk prediction model is obtained by training a Catboost classifier on simulated risk data and real order data.

[0100] In an exemplary embodiment of the present disclosure, the prediction module 430 is further configured to: determine the candidate risk order quantity and model prediction performance corresponding to each candidate threshold of the risk prediction model; determine the expected risk order push quantity within the target period according to the model prediction performance requirement and the operation processing capacity within the target period; match the expected risk order push quantity among the candidate risk order quantities, and use the candidate threshold corresponding to the successfully matched candidate risk order quantity as the target threshold corresponding to the target period; determine the prediction results of each risk type within the target period according to the comparison results between the probability values corresponding to each risk type and the target threshold.

[0101] In an exemplary embodiment of the present disclosure, the feature extraction module 420 is further configured to extract corresponding attribute fields from customer information, service information, and service provider information respectively according to a preset target attribute to obtain multi-dimensional features.

[0102] In an exemplary embodiment of the present disclosure, before extracting multi-dimensional features, the device 400 further includes a preprocessing module for respectively performing data cleaning on customer information, service information, and service provider information, and the data cleaning includes at least one of data deduplication, missing value processing, outlier exclusion, and data standardization.

[0103] In an exemplary embodiment of the present disclosure, the device 400 further includes: a push module for pushing risk warning events to the operation platform according to the prediction results determined by the target threshold within the target period.

[0104] In an exemplary embodiment of the present disclosure, the device 400 further includes a policy adjustment module for determining the risk level of each service provider according to the risk warning events for each service provider; adjusting the dispatching policy of the dispatching system according to the risk levels of each service provider.

[0105] The specific details of each module or unit in the risk warning device for the above service order have been described in detail in the risk warning method for the corresponding service order, so they will not be elaborated here.

[0106] As another aspect, the present application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or may exist alone without being assembled into the electronic device. When the one or more programs carried by the computer-readable medium are executed by an electronic device, the electronic device implements the method in the following embodiments. For example, the electronic device may implement the various steps as Figures 1-3 shown, etc.

[0107] It should be noted that the computer-readable medium shown in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program may be used by or in combination with an instruction execution system, apparatus, or device. In the present disclosure, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries the computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted by any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0108] Next, refer to Figure 5 to describe the electronic device 500 according to this embodiment of the present disclosure. Figure 5 The electronic device 500 shown is only an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0109] As Figure 5 shown, the electronic device 500 is presented in the form of a general-purpose computing device. The components of the electronic device 500 may include, but are not limited to: at least one of the above-mentioned processing units 510, at least one of the above-mentioned storage units 520, a bus 530 connecting different system components (including the storage unit 520 and the processing unit 510), and a display unit 540.

[0110] Among them, the storage unit stores program code, and the program code can be executed by the processing unit 510, so that the processing unit 510 executes the steps according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section of this specification.

[0111] The storage unit 520 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 5201 and / or a cache storage unit 5202, and may further include a read-only storage unit (ROM) 5203.

[0112] The storage unit 520 may also include a program / utility 5204 having a set (at least one) of program modules 5205. Such program modules 5205 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0113] The bus 530 may represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0114] The electronic device 500 can also communicate with one or more external devices 570 (such as a keyboard, a pointing device, a Bluetooth device, etc.), can also communicate with one or more devices that enable a user to interact with the electronic device 500, and / or can communicate with any device that enables the electronic device 500 to communicate with one or more other electronic devices (such as a router, a modem, etc.). Such communication can be carried out through an input / output (I / O) interface 550. And, the electronic device 500 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 560. As shown in the figure, the network adapter 560 communicates with other modules of the electronic device 500 through the bus 530. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the electronic device 500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RA identification systems, tape drives, and data backup storage systems, etc.

[0115] In an exemplary embodiment, a computer program product is further provided, including one or more instructions that can be executed by a processor of a computer device to complete the risk warning method for service orders provided in each of the above embodiments.

[0116] Through the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described herein can be implemented by software or by a combination of software and necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable an electronic device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0117] In addition, the above drawings are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, rather than for limiting purposes. It is easy to understand that the processes shown in the above drawings do not indicate or limit the chronological order of these processes. Additionally, it is also easy to understand that these processes can be executed synchronously or asynchronously in, for example, multiple modules.

[0118] It should be noted that although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that these steps must be executed in that specific order, or that all the steps shown must be executed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution, etc., all of which should be regarded as part of the present disclosure.

[0119] It should be understood that the present disclosure disclosed and defined herein extends to all alternative combinations of two or more separate features mentioned or apparent in the text and / or drawings. All these different combinations constitute multiple alternative aspects of the present disclosure. The embodiments of the present specification illustrate the best mode known for implementing the present disclosure and will enable those skilled in the art to utilize the present disclosure.

Claims

1. A risk warning method for service orders, characterized in that, it includes: Upon receiving order data for a target service, obtain customer information, service information, and service provider information; Extract features of corresponding dimensions from the customer information, the service information, and the service provider information respectively to obtain multi-dimensional features; Use a risk prediction model to perform risk prediction on the multi-dimensional features to obtain prediction results for each risk type; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network; Generate a risk warning event for the current service order based on the prediction results to perform risk warning.

2. The risk warning method for service orders according to claim 1, characterized in that, the service information includes service object information, service execution condition information, dispatch information, rejection information, and service risk data; The obtaining of customer information, service information, and service provider information includes: Upon the operation platform generating a service order for a target service, extract the customer information and the service object information from the order data of the service order; Upon the operation platform generating a service execution reservation order for the service order, extract service execution condition information from the reservation data corresponding to the reservation order; Upon the operation platform dispatching for the service execution reservation order or the service provider rejecting it, obtain dispatch information or rejection information for the service execution reservation order; Obtain the service provider information and the service risk data from a static database, and the service risk data includes first risk data corresponding to the service order type and second risk data corresponding to the service object type.

3. The risk warning method for service orders according to claim 1, characterized in that, The using of the risk prediction model to perform risk prediction on the multi-dimensional features includes: Input the multi-dimensional features into the risk prediction model and output probability values corresponding to each risk type, and the risk prediction model is obtained by training a Catboost classifier with simulated risk data and real order data.

4. The risk warning method for service orders according to claim 3, characterized in that, The method further includes: Determine the candidate risk order quantity and model prediction performance corresponding to the risk prediction model under each candidate threshold; Determine the expected risk order push quantity within the target period according to the model prediction performance requirement and the operation processing capacity within the target period; Match the expected risk order push quantity among the candidate risk order quantities, and use the candidate threshold corresponding to the successfully matched candidate risk order quantity as the target threshold corresponding to the target period; Determine the prediction results for each risk type within the target period according to the comparison results between the probability values corresponding to each risk type and the target threshold.

5. The risk warning method for service orders according to claim 2, characterized in that, The extracting of features of corresponding dimensions from the customer information, the service information, and the service provider information includes: Extract corresponding attribute fields from the customer information, the service information, and the service provider information respectively according to the preset target attributes, so as to obtain multi-dimensional features.

6. The risk warning method for a service order according to any one of claims 1 to 5, wherein, before extracting the multi-dimensional features, the method further includes: performing data cleaning on the customer information, the service information, and the service provider information respectively, and the data cleaning includes at least one of data deduplication, missing value processing, outlier exclusion, and data standardization.

7. The risk warning method for a service order according to claim 4, wherein, the method further includes: pushing the risk warning event to the operation platform according to the prediction result determined by the target threshold within the target time limit.

8. The risk warning method for a service order according to claim 1, wherein, the method further includes: determining the risk level of each service provider according to the risk warning event for each service provider; adjusting the dispatching strategy of the dispatching system according to the risk levels of each service provider.

9. A risk warning device for a service order, wherein, the device includes: an acquisition module, configured to acquire customer information, service information, and service provider information in response to receiving order data for a target service; a feature extraction module, configured to extract features in corresponding dimensions from the customer information, the service information, and the service provider information respectively, so as to obtain multi-dimensional features; a prediction module, configured to perform risk prediction on the multi-dimensional features by using a risk prediction model to obtain prediction results of each risk type; the risk prediction model is obtained by training a machine learning model based on simulated risk data and real order data, and the simulated risk data is generated based on a trained generative adversarial network; a warning module, configured to generate a risk warning event for the current service order based on the prediction result for risk warning.

10. A risk warning system for a service order, wherein, it includes the risk warning device according to claim 9.

11. An electronic device, wherein, it includes: one or more processors; and a storage device, configured to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method according to any one of claims 1 to 8.

Images