Loss processing evaluation method and training method and device of loss parameter prediction model

Abstract

The invention provides a loss processing evaluation method and a training method and device of a loss parameter prediction model, and relates to the technical field of artificial intelligence such as deep learning, and the method comprises the steps: obtaining a to-be-evaluated first loss event and a first loss stopping strategy of the first loss event; acquiring a first loss parameter of the first loss event under the first stop loss strategy; obtaining a trained target loss parameter prediction model, and obtaining a second loss parameter of the first loss event through the target loss parameter prediction model; and obtaining a first target stop loss evaluation result of the first loss event according to the first loss parameter and the second loss parameter.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing, and in particular to the field of artificial intelligence technologies such as deep learning. Background Art

[0002] During the daily operation of the software system, there may be events that cause loss of system resources. Among them, the events that cause loss of system resources can be promptly stopped. In this scenario, the stop-loss processing effect of the event can be analyzed and evaluated.

[0003] In the related art, the loss problems can be divided into levels, and the loss problems at each level after the division are quantitatively counted to achieve the purpose of evaluation, but the evaluation accuracy is poor. Summary of the invention

[0004] The present disclosure proposes a loss processing evaluation method, a loss parameter prediction model training method and a device.

[0005] According to the first aspect of the present disclosure, a loss handling assessment method is proposed, including: obtaining a first loss event to be assessed and a first stop-loss strategy for the first loss event; obtaining a first loss parameter of the first loss event under the first stop-loss strategy; obtaining a trained target loss parameter prediction model, and obtaining a second loss parameter of the first loss event through the target loss parameter prediction model; and obtaining a first target stop-loss assessment result of the first loss event based on the first loss parameter and the second loss parameter.

[0006] According to the second aspect of the present disclosure, a training method for a loss parameter prediction model is proposed, including: obtaining a candidate loss parameter prediction model to be trained; obtaining a sample loss event, a sample loss event bucket to which the sample loss event belongs, a sample bucketing strategy corresponding to the sample loss event bucket, and a sample resource loss threshold value corresponding to the sample loss event bucket; generating training samples for the candidate loss parameter prediction model based on the sample loss event, the sample bucketing strategy, and the sample resource loss threshold value; performing model training on the candidate loss parameter prediction model based on the training samples until the training is completed, to obtain a trained target loss parameter prediction model, wherein the target loss parameter prediction model is used to implement the loss processing evaluation method proposed in the first aspect above.

[0007] According to a third aspect of the present disclosure, a loss processing evaluation device is proposed, including: a first acquisition module, used to obtain a first loss event to be evaluated and a first stop-loss strategy for the first loss event; a second acquisition module, used to obtain a first loss parameter of the first loss event under the first stop-loss strategy; a prediction module, used to obtain a trained target loss parameter prediction model, and obtain the second loss parameter of the first loss event through the target loss parameter prediction model; an evaluation module, used to obtain a first target stop-loss evaluation result of the first loss event based on the first loss parameter and the second loss parameter.

[0008] According to a fourth aspect of the present disclosure, a training device for a loss parameter prediction model is proposed, comprising: a third acquisition module, used to acquire a candidate loss parameter prediction model to be trained; a fourth acquisition module, used to acquire a sample loss event, a sample loss event bucket to which the sample loss event belongs, a sample bucketing strategy corresponding to the sample loss event bucket, and a sample resource loss threshold value corresponding to the sample loss event bucket; a generation module, used to generate training samples of the candidate loss parameter prediction model according to the sample loss event, the sample bucketing strategy and the sample resource loss threshold value; a training module, used to perform model training on the candidate loss parameter prediction model based on the training samples until the training is completed to obtain a trained target loss parameter prediction model, wherein the target loss parameter prediction model is used to implement the above-mentioned loss processing evaluation device.

[0009] According to the fifth aspect of the present disclosure, an electronic device is proposed, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the loss processing assessment method proposed in the first aspect and / or the training method of the loss parameter prediction model proposed in the second aspect.

[0010] According to the sixth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is proposed, wherein the computer instructions are used to enable the computer to execute the loss processing assessment method proposed in the first aspect and / or the loss parameter prediction model training method proposed in the second aspect.

[0011] According to the seventh aspect of the present disclosure, a computer program product is proposed, including a computer program, which, when executed by a processor, implements the loss processing evaluation method proposed in the first aspect and / or the training method of the loss parameter prediction model proposed in the second aspect.

[0012] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0014] Figure 1 A schematic diagram of a flow chart of a loss handling and assessment method according to an embodiment of the present disclosure;

[0015] Figure 2 A schematic diagram of a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure;

[0016] Figure 3 A schematic diagram of a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure;

[0017] Figure 4 A schematic diagram of a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure;

[0018] Figure 5 A schematic diagram of a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure;

[0019] Figure 6 A schematic diagram of a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure;

[0020] Figure 7 A flowchart of a method for training a loss parameter prediction model according to an embodiment of the present disclosure;

[0021] Figure 8 It is a structural schematic diagram of a loss handling and assessment device according to an embodiment of the present disclosure;

[0022] Fig. 9 A schematic diagram of the structure of a training device for a loss parameter prediction model according to an embodiment of the present disclosure;

[0023] Fig.10 A schematic block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0025] Data processing is a basic part of system engineering and automatic control. Data is a form of expression of facts, concepts or instructions, which can be processed by manual or automatic devices. After data is interpreted and given a certain meaning, it becomes information. Data processing is the collection, storage, retrieval, processing, transformation and transmission of data. The basic purpose of data processing is to extract and derive valuable and meaningful data for certain specific people from a large amount of data that may be disorganized and difficult to understand.

[0026] Deep learning (DL) refers to machine learning based on deep neural network models and methods. It is developed based on statistical machine learning, artificial neural network and other algorithm models, combined with the development of contemporary big data and large computing power. The most important technical feature of deep learning is the ability to automatically extract features. Neural network algorithms, computing power and data are the three elements for deep learning.

[0027] Artificial Intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. Artificial Intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. Since the birth of artificial intelligence, the theory and technology of artificial intelligence have become increasingly mature, and the application field has continued to expand. It can be imagined that the technological products brought by artificial intelligence in the future will be the "container" of human wisdom. Artificial Intelligence can simulate the information process of human consciousness and thinking.

[0028] Figure 1 FIG. 1 is a flow chart of a loss handling and assessment method according to an embodiment of the present disclosure. Figure 1 As shown, the method includes:

[0029] S101, obtaining a first loss event to be evaluated and a first stop-loss strategy for the first loss event.

[0030] During the operation of the system, some abnormal situations may occur, resulting in a certain degree of resource loss in the system. In this scenario, the event that causes the system resource loss can be marked as a system loss event.

[0031] In the disclosed embodiment, the system's loss events can be stop-loss processed based on a preset loss processing strategy, and the stop-loss processing results can be evaluated. In this scenario, the loss event that currently needs to be evaluated for the stop-loss processing results can be marked as the first loss event to be evaluated, and the stop-loss strategy adopted when the first loss event is stop-loss processed can be marked as the first stop-loss strategy for the first loss event.

[0032] S102, obtaining a first loss parameter of a first loss event under a first stop-loss strategy.

[0033] In the disclosed embodiment, the relevant parameters of resource loss caused by a loss event occurring in the system may be marked as loss parameters of the loss event, wherein the loss parameters may include the duration of the loss event, the loss rate of resource loss caused by the loss event, and the amount of resource loss, etc., which are not specifically limited here.

[0034] Optionally, parameters related to resource loss caused by stop-loss processing of the first loss event based on the first stop-loss strategy may be marked as first loss parameters of the first loss event.

[0035] It can be understood that the first loss parameter is a parameter related to the actual resource loss caused by the occurrence of the first loss event.

[0036] S103, obtaining a trained target loss parameter prediction model, and obtaining a second loss parameter of the first loss event through the target loss parameter prediction model.

[0037] Optionally, in order to evaluate the effect of stop-loss processing on the first loss event based on the first stop-loss strategy, relevant loss parameters of resource losses that may be caused by the first loss event can be predicted, and the predicted parameters can be compared with the first loss parameters actually caused by the first loss event. Through the results of the comparative analysis, the stop-loss processing effect of the first loss event is evaluated.

[0038] In the disclosed embodiment, the trained model can be used to predict the relevant parameters of the resource loss that may be caused by the first loss event, wherein the trained model can be marked as a target loss parameter prediction model.

[0039] In this scenario, relevant information about the first loss event and the first loss parameter can be obtained, and input data of the target loss parameter prediction model can be extracted from it. The target loss parameter prediction model can be used to predict the loss parameters of the resource loss that may be caused by the first loss event based on the input data, and then the predicted loss parameters of the first loss event output by the model are obtained and marked as the second loss parameters of the first loss event.

[0040] It can be understood that the second loss parameter is a parameter related to the upper limit of the loss of resources that may be caused by the first loss event obtained by the target loss parameter prediction model.

[0041] S104: Obtain a first target stop loss assessment result of the first loss event according to the first loss parameter and the second loss parameter.

[0042] In the disclosed embodiment, a preset comparative analysis strategy between the actual first loss parameter and the predicted second loss parameter can be obtained, and the first loss parameter and the second loss parameter can be comparatively analyzed according to the strategy, so as to obtain an evaluation result of the stop-loss effect of the first loss event based on the first stop-loss strategy according to the result of the comparative analysis.

[0043] Among them, the assessment result can be marked as the first target stop loss assessment result of the first loss event.

[0044] Optionally, the actual amount of resource loss of the first loss event can be obtained from the first loss parameter, and the predicted amount of resource loss of the first loss event can be obtained from the second loss parameter, and the two can be compared and analyzed. When the actual amount of resource loss is less than the predicted amount of resource loss, it can be determined that the stop-loss processing effect of the first loss event based on the first stop-loss strategy meets the expected stop-loss processing requirements, and further it can be determined that the evaluation result of the stop-loss processing effect of the first loss event is up to standard.

[0045] In this scenario, the achievement of the target can be determined as the first target stop-loss assessment result of the first loss event.

[0046] The loss processing assessment method proposed in the present disclosure obtains a first loss event and a first stop-loss strategy for the first loss event to obtain a first loss parameter of the first loss event under the first stop-loss strategy, obtains a trained target loss parameter prediction model to obtain a second loss parameter corresponding to the first loss event, and then obtains a first target stop-loss assessment result of the first loss event based on the first loss parameter and the second loss parameter. In the present disclosure, the predicted second loss parameter of the first loss event is obtained according to the trained target loss parameter prediction model, which improves the acquisition efficiency and accuracy of the predicted loss parameter. In the scenario where the stop-loss effect is evaluated based on the predicted second loss parameter and the first loss parameter, the assessment accuracy and efficiency of the first target stop-loss assessment result are improved, which provides accurate data support for downstream tasks and optimizes the evaluation method of the stop-loss effect of the loss event.

[0047] In the above embodiment, the acquisition of the first target stop loss evaluation result can be combined with Figure 2 Further understanding, Figure 2 FIG. 1 is a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure. Figure 2 As shown, the method includes:

[0048] S201, obtaining a first loss event to be evaluated and a first stop-loss strategy for the first loss event.

[0049] Optionally, the third loss speed of each second stop-loss strategy in the preset second stop-loss strategy set is obtained, and the fourth loss speed of the target loss event bucket to which the first loss event belongs is obtained.

[0050] In the disclosed embodiment, there is a correlation between the loss speed of a loss event and the stop-loss strategy. In this scenario, the first stop-loss strategy for the first loss event can be obtained based on the correlation.

[0051] Optionally, a preset stop-loss strategy in the system may be obtained and marked as a second stop-loss strategy, thereby obtaining a second stop-loss strategy set consisting of a plurality of second stop-loss strategies.

[0052] In this scenario, the loss speed of each second stop-loss strategy can be obtained based on the correlation between the stop-loss strategy and the loss speed, and marked as the third loss speed of each second stop-loss strategy.

[0053] In the disclosed embodiment, the first loss event exists in a loss event bucket to which it belongs, wherein the loss event bucket can be marked as the target loss event bucket to which the first loss event belongs.

[0054] The acquisition of the target loss event bucket can be understood in combination with the following contents:

[0055] Optionally, a preset candidate loss event bucket set and a first loss dimension and a first loss speed of each of the candidate loss event bucket sets are obtained, and a second loss dimension and a second loss speed of the first loss event are obtained.

[0056] In the disclosed embodiment, when a loss event causes resource loss, the lost resource has a resource dimension to which it belongs and a loss rate of the resource, wherein the resource dimension can be marked as the loss dimension of the loss event, and the loss rate of the resource can be marked as the loss rate of the loss event.

[0057] It should be noted that the loss dimension of a loss event may include the loss dimension corresponding to the loss of economic resources, or may include the loss dimension corresponding to the loss of user resources, which is not specifically limited here.

[0058] For example, when the resource loss caused by any loss event is a capital loss of the system, the capital loss dimension can be determined as the loss dimension of the loss event, and the speed of the capital loss can be determined as the loss speed of the loss event.

[0059] In this scenario, the loss dimension and loss speed corresponding to the first loss event can be obtained and marked as the second loss speed and the second loss dimension of the first loss event, respectively.

[0060] In the disclosed embodiment, loss events in the system can be bucketed according to a preset bucketing strategy, wherein a corresponding bucketing strategy can be constructed according to the loss speed and loss dimension of the loss event, and each historical loss event in the system can be bucketed based on the bucketing strategy, thereby obtaining a plurality of loss event buckets, wherein the plurality of loss event buckets can be marked as a plurality of candidate loss event buckets, thereby obtaining a candidate loss event bucket set consisting of the plurality of candidate loss event buckets.

[0061] It should be noted that, for any candidate loss event bucket, all loss events included in the bucket belong to the same loss dimension, and the loss speeds of all loss events included in the bucket belong to the same loss speed range.

[0062] Furthermore, the loss dimension corresponding to the loss event included in any candidate loss event bucket is obtained, and the loss dimension is marked as the first loss dimension of the candidate loss event bucket, and the loss speed of each loss event included in the candidate loss event bucket is obtained, and then the loss speed corresponding to the candidate loss event bucket is obtained based on each loss speed as the first loss speed.

[0063] The first loss speed may be a loss speed interval including each loss speed, or may be a specific speed value obtained by performing algorithm processing on each loss speed, which is not specifically limited here.

[0064] Optionally, for any candidate loss event bucket, in response to a first loss dimension corresponding to the candidate loss event bucket matching a second loss dimension and a first loss speed matching a second loss speed, the candidate loss event bucket is determined to be a target loss event bucket for the first loss event.

[0065] In the disclosed embodiment, the first loss dimension of each candidate loss event bucket can be obtained and compared with the second loss dimension of the first loss event, and the first loss speed of each candidate loss event bucket can be compared with the second loss speed of the first loss event, so as to identify whether the first loss dimension matches the second loss dimension, and whether the first loss speed matches the second loss speed.

[0066] Optionally, each first loss dimension may be compared with the second loss dimension respectively. When any first loss dimension is the same as the second loss dimension, it can be determined that the first loss dimension matches the second loss dimension.

[0067] In addition, each first loss speed can be compared with the second loss speed respectively. When the second loss speed falls within the speed value interval corresponding to any first loss speed, or when the second loss speed falls within the speed value error interval corresponding to any first loss speed, it can be determined that the first loss speed matches the second loss speed.

[0068] Furthermore, a candidate loss event bucket whose first loss dimension matches the second loss dimension and whose first loss speed matches the second loss speed is obtained, and marked as the target loss event bucket to which the first loss event belongs, and the first loss speed corresponding to the candidate loss event bucket is marked as the fourth loss speed of the target loss event bucket to which the first loss event belongs.

[0069] Optionally, for any second stop-loss strategy, in response to the fourth loss rate matching the third loss rate of the second stop-loss strategy, the second stop-loss strategy is determined to be the first stop-loss strategy adopted for the first loss event.

[0070] In the disclosed embodiment, the third loss rate and the fourth loss rate of any second stop-loss strategy can be algorithmically processed based on the matching degree acquisition algorithm in the relevant technology, and then whether the third loss rate and the fourth loss rate match each other can be obtained based on the result of the algorithm processing.

[0071] Optionally, when it is identified that the third loss rate matches the fourth loss rate, it can be determined that the second stop-loss strategy corresponding to the third loss rate is the first stop-loss strategy adopted when performing stop-loss processing on the first loss event.

[0072] S202, obtaining a first loss parameter of a first loss event under a first stop-loss strategy.

[0073] Optionally, the event perception time and event location time of the first loss event are obtained, and stop-loss processing is performed on the loss event according to the first stop-loss strategy to obtain the stop-loss processing duration of the first loss event under the first stop-loss strategy; and, based on the event perception time, the event location time and the stop-loss processing duration, the loss duration of the first loss event is obtained.

[0074] In the disclosed embodiment, before the first loss event is stopped based on the first stop-loss strategy, loss event perception and loss event location need to be performed, wherein the time when the first loss event is perceived to have occurred in the system can be marked as the event perception time of the first loss event, and the location time of determining the location of the first loss event in the system can be marked as the event location time of the first loss event.

[0075] Furthermore, after sensing and locating the first loss event, stop-loss processing can be performed on the first loss event based on the first stop-loss strategy, wherein the duration used for the process of stop-loss processing on the first loss event based on the first stop-loss strategy can be marked as the stop-loss processing duration corresponding to the first loss event.

[0076] In this scenario, the event perception time, event location time, and stop-loss processing time corresponding to the first loss event can be sorted in order from early to late to obtain the earliest time and the latest time. The duration obtained based on the earliest time and the latest time as endpoints can be determined as the loss duration of the first loss event.

[0077] In the scenario where the first loss event is based on the processing order from perception to positioning to stop loss, the event perception event of the first loss event can be used as the earliest time endpoint, and the latest time in the stop loss processing duration can be used as the latest time endpoint, so as to obtain the loss duration of the first loss event.

[0078] Optionally, a first loss parameter of the first loss event under the first stop-loss strategy is obtained according to the loss duration and the second loss speed of the first loss event.

[0079] In the disclosed embodiment, in a scenario where the first loss parameter is the amount of resource loss, the loss duration and the second loss speed of the first loss event can be algorithmically processed according to a preset resource loss algorithm, and the amount of resource loss of the system caused by the first loss event under the first stop-loss strategy can be obtained according to the result of the algorithm processing as the first loss parameter.

[0080] The loss duration may be multiplied by the second loss speed, and the product of the two may be determined as the first loss parameter.

[0081] S203, obtaining a trained target loss parameter prediction model, and obtaining a second loss parameter of the first loss event through the target loss parameter prediction model.

[0082] Optionally, the target loss event bucket to which the first loss event belongs, the bucketing strategy corresponding to the target loss event bucket, and the resource loss threshold value corresponding to the target loss event bucket are obtained, and a target input data set of a target loss parameter prediction model is generated based on the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter.

[0083] Among them, the bucketing strategy for the target loss event may include a bucketing strategy determined based on the loss speed and loss dimension of the loss event, or a bucketing strategy set based on other dimensional information of the loss event, which is not specifically limited here.

[0084] In the embodiments of the present disclosure, there is a corresponding resource loss threshold value in the loss events included in each candidate loss event bucket. It can be understood that, for any candidate loss event bucket, the upper limit value of the amount of resource loss that may be caused by each loss event included in the bucket can be evaluated and calculated, and the upper limit value of the amount of resource loss shared by each loss event included in the bucket can be obtained based on the results of the evaluation and calculation.

[0085] Further, when the candidate loss event bucket serves as the target loss event bucket to which the first loss event belongs, the upper limit value may be determined as the resource loss threshold value of the target loss event bucket.

[0086] Optionally, an initial input data set is generated according to the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter, and redundant abnormal data filtering processing is performed on the initial input data set to obtain a processed candidate input data set.

[0087] In the embodiment of the present disclosure, an algorithm can be constructed based on the data set in the relevant technology, and the first loss event and the acquired bucket strategy, resource loss threshold value and first loss parameter can be algorithmically processed, and the data set obtained according to the algorithm processing results can be marked as the initial input data set.

[0088] Optionally, the initial input data set may include redundant data or abnormal data. In this scenario, the redundant data and abnormal data included therein may be marked as redundant abnormal data. Furthermore, based on the data filtering algorithm in the relevant technology, the redundant abnormal data in the initial input data set is filtered, and the data set obtained after the filtering process is marked as a candidate input data set.

[0089] Optionally, a data input format of a target loss parameter prediction model is obtained, and candidate input data sets are integrated and processed according to the data input format to obtain a processed target input data set.

[0090] In the disclosed embodiment, loss parameters can be predicted based on a candidate input data set using a trained target loss parameter prediction model, wherein the target loss parameter prediction model has a corresponding input data format limitation. In this scenario, the data input format of the target loss parameter prediction model can be obtained, and the format dimensions of each item of data in the candidate input data set can be sorted according to the format information, and the data set obtained after sorting can be determined as the target input data set of the target loss parameter prediction model.

[0091] Optionally, a second loss parameter of the first loss event is obtained according to a target loss parameter prediction model and a target input data set.

[0092] Among them, the target input data set can be input into the target loss parameter prediction model, and the characteristic information of each data in the target input data set can be extracted through the target loss parameter prediction model, and the loss parameters can be predicted based on the extracted characteristic information, so as to obtain the relevant parameters of the resource loss that may be caused by the first loss event, and output them as the second loss parameters.

[0093] As an example, Figure 3 As shown, the target input data set can be obtained based on the first loss event, the bucketing strategy of the target loss event bucket to which the first loss event belongs, the resource loss threshold of the target loss event bucket, and the first loss parameter corresponding to the first loss event, and the target input data set is input into Figure 3 The target loss parameter prediction model is shown by Figure 3 The target loss parameter prediction model shown obtains the second loss parameter corresponding to the first loss event.

[0094] S204: Obtain a first target stop loss assessment result of the first loss event according to the first loss parameter and the second loss parameter.

[0095] Optionally, in response to the first loss parameter matching the second loss parameter, a first target stop-loss assessment result of the first loss event is determined to be meeting the stop-loss target.

[0096] In the disclosed embodiment, when the comparison relationship between the first loss parameter and the second loss parameter satisfies a preset matching condition, it can be determined that the first loss parameter matches the second loss parameter.

[0097] As an example, for the loss duration in the first loss parameter and the loss duration in the second loss parameter, when the loss duration in the first loss parameter is less than or equal to the loss duration in the second loss parameter, it can be determined that the comparison relationship between the first loss parameter and the second loss parameter meets the preset matching condition, and further it can be determined that the first loss parameter matches the second loss parameter.

[0098] Furthermore, when it is identified that the first loss parameter matches the second loss parameter, it can be determined that the stop-loss processing effect of the first loss event has reached expectations, and the stop-loss compliance can be determined as the first target stop-loss assessment result of the first loss event.

[0099] Optionally, in response to the first loss parameter not matching the second loss parameter, it is determined that the first target stop loss assessment result of the first loss event is that the stop loss fails to meet the standard.

[0100] In the disclosed embodiment, when the comparison relationship between the first loss parameter and the second loss parameter does not satisfy a preset matching condition, it can be determined that the first loss parameter and the second loss parameter do not match.

[0101] As an example, for the loss duration in the first loss parameter and the loss duration in the second loss parameter, when the loss duration in the first loss parameter is greater than the loss duration in the second loss parameter, it can be determined that the comparison relationship between the first loss parameter and the second loss parameter does not meet the preset matching conditions, and further it can be determined that the first loss parameter and the second loss parameter do not match.

[0102] Furthermore, when it is identified that the first loss parameter does not match the second loss parameter, it can be determined that the stop-loss processing effect of the first loss event has not met expectations, and the failure to meet the stop-loss standard can be determined as the first target stop-loss assessment result of the first loss event.

[0103] The loss processing assessment method proposed in the present invention obtains multi-dimensional information of the first loss event to obtain a target input data set, and then outputs a predicted second loss parameter based on the target input data set through a target loss parameter prediction model, thereby improving the prediction accuracy and prediction efficiency of the loss parameter, thereby improving the assessment accuracy and assessment efficiency of the first target stop loss assessment result, and providing accurate data support for downstream tasks.

[0104] In the above embodiment, after obtaining the first target stop loss evaluation result, Figure 4 Further understanding, Figure 4 FIG. 1 is a flow chart of a loss handling and assessment method according to another embodiment of the present disclosure. Figure 4 As shown, the method includes:

[0105] S401, obtaining a second loss event set in a target loss event bucket to which a first loss event belongs, and obtaining a second target stop-loss assessment result of each second loss event in the second loss event set.

[0106] In the disclosed embodiment, the loss events included in the target loss event bucket to which the first loss event belongs may be marked as second loss events, and the set of the multiple second loss events may be marked as a second loss event set in the target loss event bucket.

[0107] Optionally, the stop-loss processing effect evaluation can be performed on each second loss event respectively based on the first stop-loss strategy, and the evaluation result of the stop-loss processing effect of each second loss event can be marked as the second target stop-loss evaluation result of each second loss event.

[0108] It should be noted that the process of obtaining the second target stop-loss assessment result can be understood in conjunction with the process of obtaining the first target stop-loss assessment result proposed in the above embodiment, and will not be repeated here.

[0109] S402: According to each second target stop-loss evaluation result, identify whether there is a need to update the first stop-loss strategy corresponding to the target loss event bucket.

[0110] In the disclosed embodiment, each second loss event adopts the first stop-loss strategy to perform stop-loss processing. In this scenario, it is possible to identify whether the currently adopted first stop-loss strategy is suitable for each second loss event included in the target loss event bucket based on the stop-loss processing effect of each second loss event.

[0111] Optionally, from each second target stop loss assessment result, a third target stop loss assessment result set in which the stop loss standard is met, and a fourth target stop loss assessment result set in which the stop loss standard is not met are obtained, and the stop loss compliance rate of the target loss event bucket is obtained based on the third target stop loss assessment result set and the fourth target stop loss assessment result set.

[0112] In the disclosed embodiment, some of the evaluation results indicating that the stop loss standard is met and some of the evaluation results indicating that the stop loss standard is not met can be obtained in each second target stop loss evaluation result, wherein the some of the evaluation results that meet the stop loss standard can be marked as the third target stop loss evaluation result in each second target stop loss evaluation result, and the set consisting of the some of the third target stop loss evaluation results can be marked as the third target stop loss evaluation result set in the second target stop loss evaluation result set.

[0113] Furthermore, some of the evaluation results that do not meet the stop loss standard are marked as the fourth target stop loss evaluation results in each second target stop loss evaluation result, and the set consisting of this part of the fourth target stop loss evaluation results is marked as the fourth target at least evaluation result set in the second target stop loss evaluation results.

[0114] In this scenario, based on the compliance rate algorithm in the relevant technology, the third target stop loss assessment result set and the fourth target stop loss assessment result can be algorithmically processed, and the compliance probability of the stop loss processing of the target loss event bucket can be obtained according to the results of the algorithm processing, which serves as the stop loss compliance rate of the target loss event bucket.

[0115] Among them, the number of elements in the third target stop loss assessment result set can be obtained and used as the divisor, and the number of elements in the target loss event bucket can be used as the dividend to obtain the quotient of the two, and the quotient is determined as the stop loss compliance rate of the target loss event bucket.

[0116] In addition, the number of elements in the fourth target stop loss assessment result set can be obtained and used as the divisor, and the number of elements in the target loss event bucket can be obtained and used as the dividend to obtain the quotient between the two, and the quotient can be used as the subtrahend and 1 as the minuend to obtain the difference between the two, and the difference can be determined as the stop loss compliance rate of the target loss event bucket.

[0117] As an example, Figure 5 As shown, it can be Figure 5 The first loss parameter and the second loss parameter are compared and evaluated to obtain Figure 5 The target stop-loss effect evaluation results are shown, and based on the same method, the evaluation results of each loss event in the target loss event bucket are obtained, and then the stop-loss compliance rate of the target loss event bucket is calculated.

[0118] Optionally, it is identified whether there is a need to update the first stop-loss strategy based on the stop-loss compliance rate.

[0119] In the disclosed embodiment, a preset stop loss compliance rate threshold value can be obtained, and the stop loss compliance rate can be compared with the compliance rate threshold value. When the stop loss compliance rate is greater than or equal to the compliance rate threshold value, it can be determined that the current first stop loss strategy can meet the stop loss processing requirements of each loss event included in the target loss event bucket, and further it can be determined that there is no policy update requirement for the first stop loss strategy in this scenario.

[0120] Accordingly, when the stop-loss compliance rate is less than the compliance rate threshold, it can be determined that the current first stop-loss strategy cannot meet the stop-loss processing requirements of each loss event included in the target loss event bucket, and further it can be determined that the first stop-loss strategy in this scenario needs to be adjusted and optimized, that is, there is a need for a strategy update for the first stop-loss strategy in the current scenario.

[0121] S403, in response to identifying that there is a need to update the first stop-loss strategy, the first stop-loss strategy is updated to obtain an updated third stop-loss strategy, and the third stop-loss strategy is used as a new first stop-loss strategy for the target loss event bucket.

[0122] In the disclosed embodiment, when it is identified that there is a need to update the first stop-loss strategy, partial information that needs to be adjusted and updated for the first stop-loss strategy can be obtained based on the second target stop-loss effect evaluation results of some loss events in the target loss event bucket that do not meet the stop-loss standards, and the first stop-loss strategy can be adjusted and updated based on the partial information, and the updated strategy can be marked as the third stop-loss strategy.

[0123] In this scenario, the third stop-loss strategy can be used as the stop-loss strategy for stop-loss processing of each loss event in the target loss event bucket. It can be understood that after a new loss event occurs, when the loss event belongs to the target loss event bucket, the stop-loss processing of the new loss event can be performed according to the third stop-loss strategy.

[0124] Further, the third stop-loss strategy is marked as a new first stop-loss strategy for the target loss event bucket.

[0125] In the disclosed embodiment, the stop-loss processing of the first loss event can be composed of multiple processing stages. In this scenario, the various stop-loss stages of the first loss event under the first stop-loss strategy can be obtained, as well as the first stage strategy of the first stop-loss strategy in each stop-loss stage.

[0126] Optionally, the various processing stages required for the first stop-loss strategy to perform stop-loss processing on the first loss event can be marked as stop-loss stages of the first loss event, wherein each stop-loss stage can include an event perception stage, an event location stage, and a loss repair stage of the first loss event, etc., which are not specifically limited here.

[0127] In the disclosed embodiment, each stop loss stage has its own processing strategy, which can be marked as the first stage strategy of each stop loss stage. For example, for the event perception stage, the processing strategy composed of various technical means required for event perception can be marked as the first stage strategy of the stop loss stage.

[0128] Optionally, the first-stage evaluation results of each stop-loss stage are obtained, and it is identified whether there are second-stage evaluation results that do not meet the stop-loss standards in each first-stage evaluation result.

[0129] In the disclosed embodiments, the stop-loss processing effect of each stop-loss stage can be evaluated in stages, and the evaluation results can be marked as the first-stage evaluation results of each stop-loss stage. Optionally, among the first-stage evaluation results of each stop-loss stage, there may be evaluation results that do not meet the processing standards. These evaluation results that do not meet the stop-loss standards can be marked as the second-stage evaluation results in each first-stage evaluation result.

[0130] Optionally, in response to identifying that there is a second-stage evaluation result in each first-stage evaluation result, an abnormal stop-loss stage corresponding to the second-stage evaluation result is obtained, and a first-stage strategy corresponding to the abnormal stop-loss stage is updated.

[0131] In the disclosed embodiment, it may be necessary to identify whether there is a second-stage evaluation result in each first-stage evaluation result. When it is identified that there is a second-stage evaluation result in each first-stage evaluation result, it can be determined that the stop-loss processing effect of the stop-loss stage corresponding to the second-stage evaluation result does not meet expectations. In this scenario, it is necessary to update and adjust the first-stage strategy of the stop-loss stage.

[0132] Among them, the stop loss stage corresponding to the second stage evaluation result can be marked as an abnormal stop loss stage, and the abnormal information of the abnormal stop loss stage can be obtained based on the second stage evaluation result, and then the strategy update information of the first stage strategy corresponding to the abnormal stop loss stage can be generated according to the acquired abnormal information.

[0133] Furthermore, the first-stage strategy corresponding to the abnormal stop-loss stage is updated and adjusted according to the strategy update information to obtain a new first-stage strategy, and the new first-stage strategy is used in the subsequent processing of the same type of stop-loss stage.

[0134] In the embodiment of the present disclosure, for the first loss event that is being processed for stop loss, the level of each stop loss stage may also be monitored, which may be understood in conjunction with the following contents:

[0135] Optionally, the first stage level of each stop loss stage of the first loss event is obtained, and during the stop loss processing of the first loss event, the level of each stop loss stage is monitored to obtain the second stage level of each stop loss stage.

[0136] In the disclosed embodiment, before the first loss event begins to perform stop-loss processing, the initially set stage-by-stage level information of each stop-loss stage corresponding to the first loss event can be obtained and marked as the first stage level of each stop-loss stage.

[0137] Furthermore, in the process of performing stop-loss processing on the first loss event based on the first stop-loss strategy, the changes in the stage levels of each stop-loss stage during the stop-loss processing process can be monitored, wherein the stage-by-stage level information of each stop-loss stage obtained by monitoring each stop-loss stage in this scenario can be marked as the second stage level of each stop-loss stage.

[0138] Optionally, for any stop loss stage, in response to the second stage level of the stop loss stage being different from the first stage level, a second stage strategy corresponding to the second stage level is obtained, and stop loss processing is performed on the stop loss stage based on the second stage strategy.

[0139] In the disclosed embodiment, for any stop-loss stage, when the second-stage level monitored in the stop-loss stage is different from the first-stage level initially set, it can be determined that the actual loss situation in the stop-loss stage may be different from the estimated loss situation before processing. In this scenario, the stop-loss processing strategy of the stop-loss stage needs to be adjusted.

[0140] Among them, the stage strategy corresponding to the monitored second stage level can be obtained as the second stage strategy, and the first stage strategy adopted in the stop loss stage can be adjusted to the second stage strategy, so as to perform stop loss processing in the stop loss stage according to the new second stage strategy.

[0141] As an example, Figure 6 As shown, it can be Figure 6The bucketing strategy of the target loss event bucket to which the first loss event belongs and the loss parameters of the first loss event are shown. The first-stage level of each stop-loss stage of the first loss event is predicted and determined, and the level of each stop-loss stage is monitored during the stop-loss processing of the first loss event to obtain the second-stage level of each stop-loss stage.

[0142] like Figure 6 As shown, the first-stage level and the second-stage level of any stop-loss stage can be compared to identify whether there is a need to adjust the level of the stop-loss stage.

[0143] Furthermore, when it is identified that the first stage level is different from the second stage level, it can be determined that there is a need to adjust the stage level of the stop loss stage, and then stop loss processing is performed on the stop loss stage according to the second stage strategy corresponding to the second stage level.

[0144] The loss processing evaluation method proposed in the present invention identifies whether there is a need for a policy update for the first stop-loss strategy based on the stop-loss compliance rate of the target loss event bucket, and updates the first stop-loss strategy after identifying the need for a policy update to obtain a new first stop-loss strategy corresponding to the target loss event bucket, thereby achieving timely optimization and adjustment of the stop-loss strategy, optimizing the stop-loss processing effect of the stop-loss strategy on loss events, improving the stop-loss processing efficiency of loss events, shortening the loss duration of loss events, reducing the amount of resource loss caused by loss events, optimizing the stop-loss control capability of loss events, and improving the stability of system operation.

[0145] Regarding the target loss parameter prediction model in the above embodiment, the present disclosure also proposes a loss parameter prediction model training method, which can be combined with Figure 7 understand, Figure 7 FIG. 1 is a flow chart of a method for training a loss parameter prediction model according to an embodiment of the present disclosure. Figure 7 As shown, the method includes:

[0146] S701, obtaining a candidate loss parameter prediction model to be trained.

[0147] In the disclosed embodiment, a model that needs to be trained for predicting loss parameters of loss events may be marked as a candidate loss parameter prediction model to be trained.

[0148] As an example, the construction of the candidate loss parameter prediction model can be understood in combination with the following:

[0149]

[0150] In the above formula, C i It represents the loss amount of the loss event at the i-th unit time granularity, Mi represents the duration of the i-th unit time granularity, D[P i ] indicates P i Minimum value of income loss interval of the level, P i represents the estimated loss level of the i-th unit time granularity, where P i The maximum value is P0.

[0151] In this example, the above formula can be simplified to:

[0152] C*M <D[max(P0,P c )]

[0153] In the above formula, C represents the average loss speed, M represents the total loss duration, and P c Indicates the upper limit of the expected loss level when the loss rate is C.

[0154] S702, obtaining a sample loss event, a sample loss event bucket to which the sample loss event belongs, a sample bucketing strategy corresponding to the sample loss event bucket, and a sample resource loss threshold value corresponding to the sample loss event bucket.

[0155] In the disclosed embodiment, some historical loss events that meet preset sample screening conditions can be obtained from the system's historical loss event database, and these historical loss events can be used as sample loss events for the candidate loss parameter prediction model.

[0156] Optionally, the event bucket to which the sample loss event belongs may be obtained, and the event bucket may be marked as the sample loss event bucket to which the sample loss event belongs, and the bucketing strategy corresponding to the sample loss event bucket may be marked as the sample bucketing strategy.

[0157] Among them, each sample loss event included in the sample loss event bucket has a corresponding resource loss threshold value, which can be marked as the sample resource loss threshold value corresponding to the sample loss event bucket.

[0158] S703: Generate training samples of candidate loss parameter prediction models according to sample loss events, sample bucketing strategies, and sample resource loss thresholds.

[0159] In the embodiment of the present disclosure, a data set consisting of sample loss events, sample bucketing strategies, and sample resource loss threshold values ​​may be obtained, and redundant abnormal data may be filtered to obtain a filtered data set.

[0160] Optionally, obtain a sample construction algorithm in related technology, perform algorithmic processing on the filtered data set, and then obtain samples consisting of sample loss events, sample bucketing strategies, and sample resource loss threshold values ​​based on the results of the algorithmic processing as training samples for the candidate loss parameter prediction model.

[0161] S704, performing model training on the candidate loss parameter prediction model based on the training samples until the training is completed, thereby obtaining a trained target loss parameter prediction model.

[0162] Among them, the target loss parameter prediction model is used to achieve the above Figures 1 to 6 Embodiments of the proposed method.

[0163] In the disclosed embodiment, the training sample can be input into the candidate loss parameter prediction model to obtain the output result of the candidate loss parameter prediction model, and the training loss of the candidate loss parameter prediction model can be obtained based on the output result and the label information of the training sample.

[0164] Furthermore, the parameters of the candidate loss parameter prediction model are adjusted according to the training loss to obtain the candidate loss parameter prediction model after the parameters are adjusted, and the next training sample is returned to continue training the candidate loss parameter prediction model after the parameters are adjusted until the training end condition is met. The model training of the candidate loss parameter prediction model can be ended, and the model obtained after the training is determined as the trained target loss parameter prediction model.

[0165] Among them, the corresponding training end conditions can be set according to the model training rounds. For the current round of model training, when the total number of model training rounds after the end of this round of training meets the training end conditions, the model training can be ended, and the model obtained after the end of the current round of training is determined as the trained target loss parameter prediction model.

[0166] In addition, corresponding training end conditions can be set according to the training output of the model. For the current round of model training, when the output result of this round of training meets the training end conditions, the model training can be ended, and the model obtained after the current round of training is determined as the trained target loss parameter prediction model.

[0167] The training method of the loss parameter prediction model proposed in the present invention obtains training samples of candidate loss parameter prediction models according to sample loss events, sample bucketing strategies and sample resource loss threshold values, so that the trained target loss parameter prediction model learns the relationship between sample loss events, sample bucketing strategies and sample resource loss threshold values, thereby improving the prediction accuracy and prediction efficiency of the target loss parameter prediction model.

[0168] An embodiment of the present disclosure also proposes a loss handling assessment device. Since the loss handling assessment device proposed in the embodiment of the present disclosure corresponds to the loss handling assessment methods proposed in the above-mentioned embodiments, the implementation methods of the above-mentioned loss handling assessment methods are also applicable to the loss handling assessment device proposed in the embodiment of the present disclosure, and will not be described in detail in the following embodiments.

[0169] Figure 8 FIG. 1 is a schematic diagram of a loss handling and evaluation device according to an embodiment of the present disclosure. Figure 8 As shown, the loss processing and evaluation device 800 includes a first acquisition module 81, a second acquisition module 82, a prediction module 83 and an evaluation module 84, wherein:

[0170] A first acquisition module 81 is used to acquire a first loss event to be evaluated and a first stop-loss strategy for the first loss event;

[0171] A second acquisition module 82 is used to acquire a first loss parameter of a first loss event under a first stop-loss strategy;

[0172] A prediction module 83, used to obtain a trained target loss parameter prediction model, and obtain a second loss parameter of the first loss event through the target loss parameter prediction model;

[0173] The evaluation module 84 is used to obtain a first target stop loss evaluation result of the first loss event according to the first loss parameter and the second loss parameter.

[0174] In the disclosed embodiment, the prediction module 83 also includes: obtaining the target loss event bucket to which the first loss event belongs, the bucketing strategy corresponding to the target loss event bucket, and the resource loss threshold value corresponding to the target loss event bucket; generating a target input data set of a target loss parameter prediction model based on the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter; and obtaining the second loss parameter of the first loss event based on the target loss parameter prediction model and the target input data set.

[0175] In the disclosed embodiment, the prediction module 83 also includes: obtaining a preset set of candidate loss event buckets and the first loss dimension and the first loss speed of each of the candidate loss event buckets; obtaining the second loss dimension and the second loss speed of the first loss event; for any candidate loss event bucket, in response to the first loss dimension corresponding to the candidate loss event bucket matching the second loss dimension, and the first loss speed matching the second loss speed, determining that the candidate loss event bucket is the target loss event bucket for the first loss event.

[0176] In the disclosed embodiment, the prediction module 83 also includes: generating an initial input data set according to the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter; performing redundant abnormal data filtering processing on the initial input data set to obtain a processed candidate input data set; obtaining the data input format of the target loss parameter prediction model, and integrating the candidate input data sets according to the data input format to obtain a processed target input data set.

[0177] In the disclosed embodiment, the assessment module 84 also includes: in response to the first loss parameter matching the second loss parameter, determining that the first target stop-loss assessment result of the first loss event is that the stop-loss standard is met; in response to the first loss parameter not matching the second loss parameter, determining that the first target stop-loss assessment result of the first loss event is that the stop-loss standard is not met.

[0178] In the disclosed embodiment, the first acquisition module 81 also includes: acquiring the third loss rate of each second stop-loss strategy in a preset set of second stop-loss strategies; acquiring the fourth loss rate of the target loss event bucket to which the first loss event belongs; for any second stop-loss strategy, in response to the fourth loss rate matching the third loss rate of the second stop-loss strategy, determining that the second stop-loss strategy is the first stop-loss strategy adopted for the first loss event.

[0179] In the disclosed embodiment, the second acquisition module 82 also includes: acquiring the event perception time and event positioning time of the first loss event; performing stop-loss processing on the loss event according to the first stop-loss strategy to obtain the stop-loss processing duration of the first loss event under the first stop-loss strategy; acquiring the loss duration of the first loss event according to the event perception time, the event positioning time and the stop-loss processing duration; and obtaining the first loss parameter of the first loss event under the first stop-loss strategy according to the loss duration and the second loss speed of the first loss event.

[0180] In the disclosed embodiment, the device also includes an updating module, which is used to: obtain a set of second loss events in the target loss event bucket to which the first loss event belongs, and obtain a second target stop-loss assessment result for each second loss event in the second loss event set; identify whether there is a need to update the first stop-loss strategy corresponding to the target loss event bucket based on each second target stop-loss assessment result; in response to identifying that there is a need to update the first stop-loss strategy, update the first stop-loss strategy to obtain an updated third stop-loss strategy, and use the third stop-loss strategy as the new first stop-loss strategy for the target loss event bucket.

[0181] In the disclosed embodiment, the update module is also used to: obtain a third target stop loss assessment result set in which the stop loss standard is met, and a fourth target stop loss assessment result set in which the stop loss standard is not met, from each second target stop loss assessment result; obtain the stop loss compliance rate for the target loss event bucket based on the third target stop loss assessment result set and the fourth target stop loss assessment result set; and identify whether there is a need to update the first stop loss strategy based on the stop loss compliance rate.

[0182] In the disclosed embodiment, the update module is also used to: obtain each stop loss stage of the first loss event under the first stop loss strategy, and obtain the first stage strategy of the first stop loss strategy in each stop loss stage; obtain the first stage evaluation result of each stop loss stage, and identify whether there is a second stage evaluation result that does not meet the stop loss standard in each first stage evaluation result; in response to identifying the existence of the second stage evaluation result in each first stage evaluation result, obtain the abnormal stop loss stage corresponding to the second stage evaluation result, and update the strategy of the first stage strategy corresponding to the abnormal stop loss stage.

[0183] In the disclosed embodiment, the update module is also used to: obtain the first stage level of each stop loss stage of the first loss event; during the stop loss processing of the first loss event, monitor the level of each stop loss stage to obtain the second stage level of each stop loss stage; for any stop loss stage, in response to the second stage level of the stop loss stage being different from the first stage level, obtain the second stage strategy corresponding to the second stage level, and perform stop loss processing on the stop loss stage based on the second stage strategy.

[0184] The loss processing evaluation device proposed in the present disclosure obtains a first loss event and a first stop-loss strategy for the first loss event to obtain a first loss parameter of the first loss event under the first stop-loss strategy, obtains a trained target loss parameter prediction model to obtain a second loss parameter corresponding to the first loss event, and then obtains a first target stop-loss evaluation result of the first loss event based on the first loss parameter and the second loss parameter. In the present disclosure, the predicted second loss parameter of the first loss event is obtained according to the trained target loss parameter prediction model, which improves the efficiency and accuracy of obtaining the predicted loss parameter. In the scenario where the stop-loss effect is evaluated based on the predicted second loss parameter and the first loss parameter, the evaluation accuracy and efficiency of the first target stop-loss evaluation result are improved, which provides accurate data support for downstream tasks and optimizes the evaluation method of the stop-loss effect of the loss event.

[0185] An embodiment of the present disclosure further proposes a training device for a loss parameter prediction model. Since the training device for a loss parameter prediction model proposed in the embodiment of the present disclosure corresponds to the training methods for a loss parameter prediction model proposed in the above-mentioned embodiments, the implementation methods of the above-mentioned training methods for loss parameter prediction models are also applicable to the training device for a loss parameter prediction model proposed in the embodiment of the present disclosure and will not be described in detail in the following embodiments.

[0186] Fig. 9 FIG. 1 is a schematic diagram of the structure of a training device for a loss parameter prediction model according to an embodiment of the present disclosure. Fig. 9 As shown, the training device 900 of the loss parameter prediction model includes a third acquisition module 91, a fourth acquisition module 92, a generation module 93 and a training module 94, wherein:

[0187] A third acquisition module 91 is used to acquire a candidate loss parameter prediction model to be trained;

[0188] A fourth acquisition module 92 is used to acquire a sample loss event, a sample loss event bucket to which the sample loss event belongs, a sample bucketing strategy corresponding to the sample loss event bucket, and a sample resource loss threshold value corresponding to the sample loss event bucket;

[0189] A generating module 93, for generating training samples of a candidate loss parameter prediction model according to a sample loss event, a sample bucketing strategy, and a sample resource loss threshold value;

[0190] The training module 94 is used to perform model training on the candidate loss parameter prediction model based on the training samples until the training is completed to obtain a trained target loss parameter prediction model, wherein the target loss parameter prediction model is used to achieve the above Figure 8 The device proposed in the embodiment.

[0191] The training device for the loss parameter prediction model proposed in the present invention obtains training samples of candidate loss parameter prediction models according to sample loss events, sample bucketing strategies and sample resource loss threshold values, so that the trained target loss parameter prediction model learns the relationship between sample loss events, sample bucketing strategies and sample resource loss threshold values, thereby improving the prediction accuracy and prediction efficiency of the target loss parameter prediction model.

[0192] According to an embodiment of the present disclosure, the present disclosure also proposes an electronic device, a readable storage medium, and a computer program product.

[0193] Fig.10 A schematic block diagram of an example electronic device 100 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0194] like Fig.10As shown, the device 100 includes a computing unit 101, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 102 or a computer program loaded from a storage unit 109 into a random access memory (RAM) 103. In the RAM 103, various programs and data required for the operation of the device 100 can also be stored. The computing unit 101, the ROM 102, and the RAM 103 are connected to each other via a bus 104. An input / output (I / O) interface 105 is also connected to the bus 104.

[0195] A number of components in the device 100 are connected to the I / O interface 105, including: an input unit 106, such as a keyboard, a mouse, etc.; an output unit 106, such as various types of displays, speakers, etc.; a storage unit 109, such as a disk, an optical disk, etc.; and a communication unit 109, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 109 allows the device 100 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0196] The computing unit 101 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 101 performs the various methods and processes described above, such as the loss processing assessment method and / or the training method of the loss parameter prediction model. For example, in some embodiments, the loss processing assessment method and / or the training method of the loss parameter prediction model may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 109. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 100 via the ROM 102 and / or the communication unit 109. When the computer program is loaded into the RAM 103 and executed by the computing unit 101, one or more steps of the loss processing assessment method and / or the training method of the loss parameter prediction model described above may be performed. Alternatively, in other embodiments, the computing unit 101 may be configured to execute the loss processing assessment method and / or the loss parameter prediction model training method in any other suitable manner (eg, by means of firmware).

[0197] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0198] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be presented to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, implements the functions / operations specified in the flow chart and / or block diagram. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0199] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, 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 foregoing.

[0200] To propose interactions with a user account, the systems and techniques described herein may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user account; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user account can propose input to the computer. Other types of devices may also be used to propose interactions with the user account; for example, feedback proposed to the user account may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user account may be received in any form (including acoustic input, voice input, or tactile input).

[0201] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user account computer with a graphical user account interface or a web browser through which a user account can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0202] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0203] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0204] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for assessing loss handling, wherein: The method comprises: Obtaining a first loss event to be evaluated and a first stop-loss strategy for the first loss event; Obtaining a first loss parameter of the first loss event under the first stop-loss strategy; Obtaining a trained target loss parameter prediction model, and obtaining a second loss parameter of the first loss event through the target loss parameter prediction model; A first target stop-loss assessment result of the first loss event is obtained according to the first loss parameter and the second loss parameter.

2. The method according to claim 1, wherein: The step of obtaining a trained target loss parameter prediction model and obtaining a second loss parameter of the first loss event through the target loss parameter prediction model includes: Obtaining a target loss event bucket to which the first loss event belongs, a bucketing strategy corresponding to the target loss event bucket, and a resource loss threshold value corresponding to the target loss event bucket; Generate a target input data set for the target loss parameter prediction model according to the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter; The second loss parameter of the first loss event is obtained according to the target loss parameter prediction model and the target input data set.

3. The method according to claim 2, wherein: The obtaining a target loss event bucket to which the first loss event belongs includes: Obtaining a preset candidate loss event bucket set and a first loss dimension and a first loss speed of each of the candidate loss event bucket sets; Obtain a second loss dimension and a second loss speed of the first loss event; For any candidate loss event bucket, in response to a first loss dimension corresponding to the candidate loss event bucket matching the second loss dimension and a first loss speed matching the second loss speed, the candidate loss event bucket is determined to be the target loss event bucket of the first loss event.

4. The method according to claim 2, wherein: The step of generating a target input data set for the target loss parameter prediction model according to the first loss event, the bucketing strategy, the resource loss threshold value, and the first loss parameter includes: Generate an initial input data set according to the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter; Performing redundant and abnormal data filtering processing on the initial input data set to obtain a processed candidate input data set; The data input format of the target loss parameter prediction model is obtained, and the candidate input data sets are integrated and processed according to the data input format to obtain the processed target input data set.

5. The method according to claim 1, wherein: The obtaining, according to the first loss parameter and the second loss parameter, a first target stop loss assessment result of the first loss event includes: In response to the first loss parameter matching the second loss parameter, determining that the first target stop-loss assessment result of the first loss event is that the stop-loss target is met; In response to the first loss parameter not matching the second loss parameter, determining the first target stop loss assessment result of the first loss event as failure to meet stop loss criteria.

6. The method according to claim 1, wherein: The obtaining of a first loss event to be evaluated and a first stop-loss strategy for the first loss event includes: Obtaining the third loss speed of each second stop-loss strategy in the preset second stop-loss strategy set; Obtaining a fourth loss rate of a target loss event bucket to which the first loss event belongs; For any second stop-loss strategy, in response to the fourth loss rate matching the third loss rate of the second stop-loss strategy, the second stop-loss strategy is determined to be the first stop-loss strategy adopted for the first loss event.

7. The method according to claim 1, wherein: The obtaining of a first loss parameter of the first loss event under the first stop-loss strategy includes: Obtaining an event perception time and an event location time of the first loss event; Performing stop-loss processing on the loss event according to the first stop-loss strategy to obtain the stop-loss processing duration of the first loss event under the first stop-loss strategy; According to the event perception time, the event location time and the stop loss processing time, the loss duration of the first loss event is obtained; The first loss parameter of the first loss event under the first stop-loss strategy is obtained according to the loss duration and the second loss speed of the first loss event.

8. The method according to claim 1, wherein: The method further comprises: Obtaining a second loss event set in the target loss event bucket to which the first loss event belongs, and obtaining a second target stop-loss assessment result of each second loss event in the second loss event set; According to each second target stop loss evaluation result, identifying whether the first stop loss strategy corresponding to the target loss event bucket has a strategy update requirement; In response to identifying that the first stop-loss strategy has the policy update requirement, the first stop-loss strategy is updated to obtain an updated third stop-loss strategy, and the third stop-loss strategy is used as a new first stop-loss strategy for the target loss event bucket.

9. The method according to claim 8, wherein: The step of identifying, according to each second target stop-loss evaluation result, whether there is a need to update the first stop-loss strategy corresponding to the target loss event bucket comprises: From each second target stop loss assessment result, obtain a third target stop loss assessment result set in which the stop loss meets the target, and a fourth target stop loss assessment result set in which the stop loss does not meet the target; Obtaining the stop-loss compliance rate of the target loss event bucket according to the third target stop-loss assessment result set and the fourth target stop-loss assessment result set; It is identified whether the first stop-loss strategy has the strategy update requirement according to the stop-loss compliance rate.

10. The method according to any one of claims 1 to 9, wherein: The method further comprises: Obtaining each stop-loss stage of the first loss event under the first stop-loss strategy, and obtaining the first-stage strategy of the first stop-loss strategy in each stop-loss stage; Obtain the first-stage evaluation results of each stop-loss stage, and identify whether there is a second-stage evaluation result that does not meet the stop-loss standard in each first-stage evaluation result; In response to identifying that the second-stage evaluation result exists in each first-stage evaluation result, the abnormal stop-loss stage corresponding to the second-stage evaluation result is obtained, and the first-stage strategy corresponding to the abnormal stop-loss stage is updated.

11. The method according to claim 10, wherein: The method further comprises: Obtaining the first stage level of each stop loss stage of the first loss event; During the stop loss processing of the first loss event, each stop loss stage is monitored to obtain a second stage level of each stop loss stage; For any stop loss stage, in response to the second stage level of the stop loss stage being different from the first stage level, a second stage strategy corresponding to the second stage level is obtained, and stop loss processing is performed on the stop loss stage based on the second stage strategy.

12. A method for training a loss parameter prediction model, wherein: The method comprises: Obtain the candidate loss parameter prediction model to be trained; Obtain a sample loss event, a sample loss event bucket to which the sample loss event belongs, a sample bucketing strategy corresponding to the sample loss event bucket, and a sample resource loss threshold value corresponding to the sample loss event bucket; Generate training samples of the candidate loss parameter prediction model according to the sample loss event, the sample bucketing strategy and the sample resource loss threshold value; The candidate loss parameter prediction model is trained based on the training samples until the training is completed to obtain a trained target loss parameter prediction model, wherein the target loss parameter prediction model is used to implement the method described in any one of claims 1 to 11 above.

13. A loss handling assessment device, wherein: The device comprises: A first acquisition module, used to acquire a first loss event to be evaluated and a first stop-loss strategy for the first loss event; A second acquisition module, configured to acquire a first loss parameter of the first loss event under the first stop-loss strategy; A prediction module, used to obtain a trained target loss parameter prediction model, and obtain a second loss parameter of the first loss event through the target loss parameter prediction model; An evaluation module is used to obtain a first target stop loss evaluation result of the first loss event according to the first loss parameter and the second loss parameter.

14. The device according to claim 13, wherein: The prediction module further includes: Obtaining a target loss event bucket to which the first loss event belongs, a bucketing strategy corresponding to the target loss event bucket, and a resource loss threshold value corresponding to the target loss event bucket; Generate a target input data set for the target loss parameter prediction model according to the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter; The second loss parameter of the first loss event is obtained according to the target loss parameter prediction model and the target input data set.

15. The device according to claim 14, wherein: The prediction module further includes: Obtaining a preset candidate loss event bucket set and a first loss dimension and a first loss speed of each of the candidate loss event bucket sets; Obtain a second loss dimension and a second loss speed of the first loss event; For any candidate loss event bucket, in response to a first loss dimension corresponding to the candidate loss event bucket matching the second loss dimension and a first loss speed matching the second loss speed, the candidate loss event bucket is determined to be the target loss event bucket of the first loss event.

16. The device according to claim 14, wherein: The prediction module further includes: Generate an initial input data set according to the first loss event, the bucketing strategy, the resource loss threshold value and the first loss parameter; Performing redundant and abnormal data filtering processing on the initial input data set to obtain a processed candidate input data set; The data input format of the target loss parameter prediction model is obtained, and the candidate input data sets are integrated and processed according to the data input format to obtain the processed target input data set.

17. The device according to claim 13, wherein: The evaluation module further includes: In response to the first loss parameter matching the second loss parameter, determining that the first target stop-loss assessment result of the first loss event is that the stop-loss target is met; In response to the first loss parameter not matching the second loss parameter, determining the first target stop loss assessment result of the first loss event as failure to meet stop loss criteria.

18. The device according to claim 13, wherein: The first acquisition module further includes: Obtaining the third loss speed of each second stop-loss strategy in the preset second stop-loss strategy set; Obtaining a fourth loss rate of a target loss event bucket to which the first loss event belongs; For any second stop-loss strategy, in response to the fourth loss rate matching the third loss rate of the second stop-loss strategy, the second stop-loss strategy is determined to be the first stop-loss strategy adopted for the first loss event.

19. The device according to claim 13, wherein: The second acquisition module further includes: Obtaining an event perception time and an event location time of the first loss event; Performing stop-loss processing on the loss event according to the first stop-loss strategy to obtain the stop-loss processing duration of the first loss event under the first stop-loss strategy; According to the event perception time, the event location time and the stop loss processing time, the loss duration of the first loss event is obtained; The first loss parameter of the first loss event under the first stop-loss strategy is obtained according to the loss duration and the second loss speed of the first loss event.

20. The device according to claim 13, wherein: The device also includes an updating module, which is used to: Obtaining a second loss event set in the target loss event bucket to which the first loss event belongs, and obtaining a second target stop-loss assessment result of each second loss event in the second loss event set; According to each second target stop loss evaluation result, identifying whether the first stop loss strategy corresponding to the target loss event bucket has a strategy update requirement; In response to identifying that the first stop-loss strategy has the policy update requirement, the first stop-loss strategy is updated to obtain an updated third stop-loss strategy, and the third stop-loss strategy is used as a new first stop-loss strategy for the target loss event bucket.

21. The device according to claim 20, wherein: The update module is further used for: From each second target stop loss assessment result, obtain a third target stop loss assessment result set in which the stop loss meets the target, and a fourth target stop loss assessment result set in which the stop loss does not meet the target; Obtaining the stop-loss compliance rate of the target loss event bucket according to the third target stop-loss assessment result set and the fourth target stop-loss assessment result set; It is identified whether the first stop-loss strategy has the strategy update requirement according to the stop-loss compliance rate.

22. The device according to any one of claims 13 to 21, wherein: The update module is further used for: Obtaining each stop-loss stage of the first loss event under the first stop-loss strategy, and obtaining the first-stage strategy of the first stop-loss strategy in each stop-loss stage; Obtain the first-stage evaluation results of each stop-loss stage, and identify whether there is a second-stage evaluation result that does not meet the stop-loss standard in each first-stage evaluation result; In response to identifying that the second-stage evaluation result exists in each first-stage evaluation result, the abnormal stop-loss stage corresponding to the second-stage evaluation result is obtained, and the first-stage strategy corresponding to the abnormal stop-loss stage is updated.

23. The device according to claim 22, wherein: The update module is further used for: Obtaining the first stage level of each stop loss stage of the first loss event; During the stop loss processing of the first loss event, each stop loss stage is monitored to obtain a second stage level of each stop loss stage; For any stop loss stage, in response to the second stage level of the stop loss stage being different from the first stage level, a second stage strategy corresponding to the second stage level is obtained, and stop loss processing is performed on the stop loss stage based on the second stage strategy.

24. A training device for a loss parameter prediction model, wherein: The device comprises: A third acquisition module is used to obtain a candidate loss parameter prediction model to be trained; A fourth acquisition module is used to acquire a sample loss event, a sample loss event bucket to which the sample loss event belongs, a sample bucketing strategy corresponding to the sample loss event bucket, and a sample resource loss threshold value corresponding to the sample loss event bucket; A generation module, configured to generate a training sample of the candidate loss parameter prediction model according to the sample loss event, the sample bucketing strategy and the sample resource loss threshold value; A training module is used to perform model training on the candidate loss parameter prediction model based on the training samples until the training is completed, thereby obtaining a trained target loss parameter prediction model, wherein the target loss parameter prediction model is used to implement the device described in any one of claims 13 to 23.

25. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 11 and / or claim 12.

26. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 11 and / or claim 12.

27. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1-11 and / or claim 12.

Images