Financial application monitoring method, device, electronic device and storage medium

Abstract

The present application relates to a method, device, electronic device, storage medium and computer program product for monitoring financial applications. The method comprises: inputting a target financial application set into an application classification model to obtain a classification result, and then determining abnormal applications in the target financial application set according to the classification result and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set. Among them, the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold, thereby excluding some target monitoring indicators that lead to a large number of abnormal applications, avoiding a large number of abnormal applications caused by the abnormal conditions corresponding to the monitoring indicators not being updated in time, and improving the monitoring accuracy.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, electronic device and storage medium for monitoring financial applications. Background Art

[0002] With the rapid development of science and technology, the increasingly powerful functions of computers have made them widely used in people's daily study, work and life.

[0003] For enterprises, various business service platforms provided by computers can be used to provide users with a variety of business services and corresponding business management, such as abnormal monitoring. In traditional technology, people usually use the abnormal recognition model installed in the computer to realize abnormal monitoring of different businesses.

[0004] However, the above-mentioned anomaly identification model needs to be determined based on the abnormal conditions corresponding to different monitoring indicators. Faced with monitoring indicators with changing abnormal conditions, especially monitoring indicators for financial applications, users are often unable to update the abnormal conditions corresponding to each monitoring indicator in the model in a timely manner, resulting in a lag in the model. Continuing to use it will generate a large number of abnormal applications, reducing the accuracy of monitoring. Summary of the invention

[0005] Based on this, it is necessary to provide a financial application monitoring method, device, electronic device, storage medium and computer program product to address the above technical issues.

[0006] In a first aspect, the present application provides a method for monitoring a financial application, comprising:

[0007] Input the target financial application set into the application classification model to obtain the classification result; wherein the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set;

[0008] The abnormal application in the target financial application set is determined according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set.

[0009] In one embodiment, the method further comprises:

[0010] Input the historical financial application set into the original application classification model to obtain the original classification result;

[0011] The abnormal proportion of abnormal applications in the historical financial application set is determined according to the original classification results and the abnormal conditions corresponding to each original monitoring indicator in the original monitoring indicator set.

[0012] In one embodiment, the process of determining the target monitoring indicator set using the monitoring indicator optimization model includes:

[0013] Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set;

[0014] Determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set;

[0015] The candidate monitoring indicator set corresponding to the minimum value of the objective function is determined as the target monitoring indicator set.

[0016] In one embodiment, the objective function of the monitoring indicator optimization model is determined according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set, including:

[0017] Obtain the sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set as the objective function of the monitoring indicator optimization model; or,

[0018] The weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model.

[0019] In one embodiment, the process of determining the application classification model includes:

[0020] Determine the target monitoring indicator included in each historical financial application in the historical financial application set, and construct an indicator vector group according to the target monitoring indicator included in each historical financial application;

[0021] Construct an indicator directed graph based on the indicator vector group corresponding to all historical financial applications;

[0022] The application classification model is determined using a general clustering algorithm and an indicator directed graph.

[0023] In one embodiment, a general clustering algorithm and an index directed graph are used to determine an application classification model, including:

[0024] The initial application classification model is determined by using a general clustering algorithm and an indicator directed graph; wherein the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement;

[0025] The model parameters in the initial application classification model are solved based on the simulated annealing algorithm, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein the model parameters include the target number.

[0026] In a second aspect, the present application also provides a monitoring device for financial applications, including:

[0027] An application classification module is used to input a target financial application set into an application classification model to obtain a classification result; wherein the application classification model is a classification model determined based on a target monitoring indicator set and a historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set based on the training time cost and the result deviation cost using a monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined based on the original monitoring indicator set and the historical financial application set;

[0028] The abnormal monitoring module is used to determine the abnormal applications in the target financial application set according to the classification results and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set.

[0029] In a third aspect, the present application further provides an electronic device. The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0030] Input the target financial application set into the application classification model to obtain the classification result; wherein the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set;

[0031] The abnormal application in the target financial application set is determined according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set.

[0032] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0033] Input the target financial application set into the application classification model to obtain the classification result; wherein the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set;

[0034] The abnormal application in the target financial application set is determined according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set.

[0035] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0036] Input the target financial application set into the application classification model to obtain the classification result; wherein the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set;

[0037] The abnormal application in the target financial application set is determined according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set.

[0038] The above-mentioned financial application monitoring method, device, electronic device, storage medium and computer program product obtain the classification result by inputting the target financial application set into the application classification model, and then determine the abnormal applications in the target financial application set according to the classification result and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set. Among them, the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold, excluding some target monitoring indicators that cause a large number of abnormal applications, updating the original application classification model, reducing the abnormal output rate of monitoring, avoiding a large number of abnormal applications caused by the abnormal conditions corresponding to the monitoring indicators not being updated in time, and thus relatively improving the accuracy of monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of a flow chart of a method for monitoring financial applications in one embodiment;

[0040] Figure 2 A schematic diagram of a process for determining an abnormality ratio in one embodiment;

[0041] Figure 3 A schematic diagram of a process for determining a target monitoring indicator set using a monitoring indicator optimization model in one embodiment;

[0042] Figure 4 A schematic diagram of a process for determining an application classification model in one embodiment;

[0043] Figure 5 is a structural block diagram of a monitoring device for financial applications in one embodiment;

[0044] Figure 6 FIG. 4 is a diagram showing the internal structure of an electronic device in one embodiment. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0046] In one embodiment, Figure 1As shown, a method for monitoring financial applications is provided. This embodiment uses the method applied to electronic devices, such as computer devices, as an example for illustration. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0047] S110 , input the target financial application set into the application classification model to obtain a classification result.

[0048] Among them, the application classification model is a classification model determined based on the target monitoring indicator set and the historical financial application set. When the target monitoring indicator set is the monitoring indicator set obtained by the original application classification model and the proportion of abnormal applications is greater than the abnormal proportion threshold, the monitoring indicator optimization model is used to screen the original monitoring indicator set according to the training time cost and the result deviation cost. The original application classification model is a classification model determined based on the original monitoring indicator set and the historical financial application set.

[0049] It should be noted that monitoring indicators can be understood as the content to be monitored for financial applications, such as amount, reason, etc.

[0050] The above training time cost is the time spent on training a large number of financial application samples using a certain number of monitoring indicators to obtain a classification model; the above result deviation cost is the number or proportion of a category of financial applications that are not classified by the classification model when the classification model obtained through the above training is used to classify financial applications. For example, the time spent on training a classification model using target monitoring indicators A, B, C and a large number of training samples of financial applications is the model's training time cost; and the classification model obtained through training can only perform classification based on the abnormal conditions of the above indicators A, B, and C, and the classification results only include abnormalities or normal conditions of indicators A, B, and C. The training samples actually also include abnormal conditions of the target monitoring indicator D. The number of applications with abnormal target monitoring indicator D in the training samples or the proportion of applications are counted as the model's result deviation cost.

[0051] It should be noted that the target financial application set is a set of target financial applications determined in the financial application database based on the monitoring constraints. Optionally, the monitoring constraints may be pre-set or input by a user. For example, the monitoring constraints may be a time range, application type, department, etc., which are used to identify the target financial applications in the financial application database.

[0052] Optionally, the computer device may accept monitoring constraints input by the user, such as a time range, to obtain all target financial applications within the time range at the time of generation in the financial application database to form the target financial application set.

[0053] The application classification model is used to classify the target financial applications in the target financial application set according to the target monitoring indicators included in the target financial application, so as to classify the target financial applications of the same type into one category. That is, the application classification model is used to classify the input target financial application set into multiple subsets according to the target monitoring indicators, each subset represents a type, and the target financial applications in the same subset belong to the same type.

[0054] Optionally, the computer device inputs the target financial application set determined above into the application classification model, identifies the target monitoring indicators included in each target financial application through the application classification model, and determines the data similarity between different target financial applications based on the actual data of the target monitoring indicators in each target financial application, and classifies the target financial applications whose data similarity is greater than the similarity threshold into one category, thereby obtaining the above-mentioned classification result.

[0055] S120: Determine abnormal applications in the target financial application set according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set.

[0056] Among them, abnormal applications are financial applications whose monitoring indicators meet abnormal conditions.

[0057] Specifically, the computer device verifies whether each type of target financial application in the classification result is an abnormal application based on the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set. Among them, if any target monitoring indicator in the target financial application reaches the corresponding abnormal condition, the target financial application is determined to be an abnormal application. Optionally, the computer device can also determine the abnormal type of the abnormal application based on the abnormal indicators of the abnormal application. For example, the target monitoring indicators include indicators A and B, and the target financial application 1 includes both indicators A and B. It is verified that indicator A reaches the corresponding abnormal condition, that is, indicator A in the target financial application 1 is abnormal, and indicator B does not reach the corresponding abnormal condition, that is, indicator B in the target financial application 1 is normal. It is comprehensively determined that the target financial application 1 is an abnormal application, and the abnormal type is that indicator A is abnormal.

[0058] Optionally, there is a high data similarity between target financial applications of the same type in the above classification results. Therefore, the computer device only needs to verify whether one target financial application in each type in the classification results is abnormal to determine whether all target financial applications of the same type are abnormal. For example, the target monitoring indicators include A and B, and the classification results for the target financial application set include 3 categories, among which type 1 includes target financial applications 1 to 4, type 2 includes target financial applications 5 to 8, and type 3 includes target financial applications 9 to 10. Verifying that target financial application 1 is an abnormal application, and the abnormal type is that indicator A is abnormal, the computer device can determine that all target financial applications (1 to 4) in type 1 are abnormal applications, and the abnormal type is that indicator A is abnormal; verifying that target financial application 8 is an abnormal application, and the abnormal type is that indicator A is abnormal and indicator B is abnormal, the computer device can determine that all target financial applications (5 to 8) in type 2 are abnormal applications, and the abnormal type is that indicators AB are abnormal; verifying that indicators A and B of target financial application 10 are normal, the computer device can determine that all target financial applications (9 to 10) in type 3 are normal applications.

[0059] In this embodiment, the computer device inputs the target financial application set into the application classification model to obtain the classification result, and then determines the abnormal applications in the target financial application set according to the classification result and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set. Among them, the application classification model is determined by the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold, excluding some target monitoring indicators that cause a large number of abnormal applications, updating the original application classification model, reducing the abnormal output rate of monitoring, avoiding a large number of abnormal applications caused by the abnormal conditions corresponding to the monitoring indicators not being updated in time, and thus relatively improving the accuracy of monitoring.

[0060] In one of the embodiments, when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold, the original classification model needs to be updated to obtain the above application classification model.

[0061] like Figure 2 As shown, the above method also includes a process of determining the abnormality ratio, which specifically includes:

[0062] S210: Input the historical financial application set into the original application classification model to obtain the original classification result.

[0063] The historical financial application set is composed of financial applications generated before the target financial application. The method for determining the historical financial application set can refer to the method for determining the target financial application set, which will not be described in detail here.

[0064] Similar to the aforementioned application classification model, the original application classification model is used to classify the historical financial applications in the historical financial application set according to the original monitoring indicators included in the historical financial applications, so as to classify the historical financial applications belonging to the same type into one category. That is, the original application classification model is used to divide the input historical financial application set into multiple subsets according to the original monitoring indicators, each subset represents a type, and the historical financial applications in the same subset belong to the same type.

[0065] Optionally, the computer device inputs the set of historical financial applications into the application classification model, obtains the original monitoring indicators included in each historical financial application through the original application classification model, and determines the data similarity between different historical financial applications based on the actual data of the original monitoring indicators in each historical financial application, and classifies the historical financial applications whose data similarity is greater than the similarity threshold into one category, thereby obtaining the above-mentioned original classification result.

[0066] S220: Determine the abnormal proportion of abnormal applications in the historical financial application set according to the original classification result and the abnormal condition corresponding to each original monitoring indicator in the original monitoring indicator set.

[0067] Optionally, the process of determining abnormal applications in the historical financial application set may refer to the aforementioned process of determining abnormal applications in the target financial application set, which will not be described in detail here.

[0068] Optionally, after determining the abnormal applications in the historical financial application set, the computer device may further count the number of abnormal applications and calculate the ratio of the number of abnormal applications to the total number of historical financial applications included in the historical financial application set, which is used as the above-mentioned abnormal ratio. For example, the historical financial application set includes 500 historical financial applications, of which there are 180 abnormal applications, and the abnormal ratio of abnormal applications in the historical financial application set is 180 / 500=0.36.

[0069] After obtaining the above-mentioned abnormal proportion, the computer device determines whether the original application classification model needs to be updated based on the abnormal proportion. Specifically, it can be determined based on the size relationship between the abnormal proportion and the abnormal proportion threshold. Among them, if the abnormal proportion is greater than the abnormal proportion threshold, the computer device uses the above-mentioned monitoring indicator optimization model to screen the original monitoring indicator set to obtain the target monitoring indicator set, and then obtains the application classification model from the target monitoring indicator set and the historical financial application set to update the original application classification model for subsequent financial application monitoring; conversely, if the abnormal proportion is not greater than the abnormal proportion threshold, the computer device can continue to use the above-mentioned original application classification model to monitor subsequent financial applications based on the original monitoring indicators.

[0070] In this embodiment, the computer device inputs the historical financial application set into the original application classification model to obtain the original classification result, and then determines the abnormal proportion of abnormal applications in the historical financial application set based on the original classification result and the abnormal conditions corresponding to each original monitoring indicator in the original monitoring indicator set, so as to subsequently determine whether the original application classification model needs to be updated based on the abnormal proportion, thereby achieving timely updating of the original application classification model and continuous and stable monitoring of financial applications.

[0071] In one embodiment, if Figure 3 As shown in FIG. 1 , the process of using the monitoring indicator optimization model to determine the target monitoring indicator set includes:

[0072] S310. Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set.

[0073] Among them, n increases from 1 to N-1 in sequence, and N is the total number of original monitoring indicators in the original monitoring indicator set.

[0074] In order to determine the appropriate target monitoring indicator, the computer device needs to extract all possible candidate monitoring indicator combinations (including at least one monitoring indicator in the combination) from the original monitoring indicator set except the device itself. The original monitoring indicators in the original monitoring indicator set have been extracted as candidate monitoring indicators to form a candidate monitoring indicator combination. Among them, each candidate monitoring indicator combination constitutes a set of candidate monitoring indicator sets. For example, the original monitoring indicator set includes three original monitoring indicators A, B, and C, that is, N = 3. The computer device uses the above method to first obtain n = 1 candidate monitoring indicators from the original monitoring indicator set, and then form 3 candidate monitoring indicator sets, namely {A}, {B}, {C}, and then obtain n = 2 candidate monitoring indicators from the original monitoring indicator set, and then form 3 candidate monitoring indicator sets, namely {A, B}, {B, C}, {A, C}, and a total of 6 candidate monitoring indicator sets are obtained.

[0075] S320. Determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set.

[0076] Optionally, the computer device may obtain the sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set as the objective function of the monitoring indicator optimization model; it may also obtain the weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set as the objective function of the monitoring indicator optimization model.

[0077] Among them, the training time cost is used to measure the time efficiency of the model, and the result deviation cost is used to measure the recognition quality of the model.

[0078] Alternatively, for the efficiency measurement, assume that a model has A links, which is represented by M, and M = {m a |a=1,2,3...,A}. Each link involves B data categories, which are represented by N, and N={n b |f=1,2,3,...B}. For this model, the time required to iterate whether a data category is suitable for inclusion in the model is Therefore, for this time efficiency metric, the cost is expressed as Z e ,have:

[0079]

[0080] Optionally, for the measurement of recognition quality, for each link, there are C recognized anomalies, which are expressed as E, and E = {e c |c=1,2,3,...C}. Therefore, for a model, its recognition quality cost is Z q ,have:

[0081]

[0082] Therefore, for the entire project, the total cost model is:

[0083] Z=αZ e +(1-α)Z q +β0

[0084] Substituting into the above formula, we have:

[0085]

[0086] in:

[0087] α represents the equilibrium point coefficient, ω κ ,ω λrepresents the inspection time coefficient and error weight. β0 represents a constant.

[0088] The corresponding constraints are:

[0089] 0≤α≤1

[0090] M={m a |a=1,2,3...,A}

[0091] N={n b |f=1,2,3,...B}

[0092] E={e c |c=1,2,3,...C}

[0093] ω κ ,ω λ >0

[0094] β0>0

[0095] S330. Determine the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set.

[0096] The minimum value of the objective function means the minimum comprehensive model cost. The comprehensive model cost is the sum of the training time cost and the result deviation cost, or the weighted sum.

[0097] Specifically, the computer device solves the objective function and determines the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set.

[0098] Optionally, the process of solving the above objective function is as follows:

[0099] Based on the genetic algorithm, it is combined with the iterative strategy model to solve the problem. Among them, the decision variable is the parameter category in the model. If it is selected, the value is 1, otherwise it is 0, so the variable Ψ is defined c To refer to the set, we have Ψ u ={φ b ={0,1}|b=1,2,3,...B}, where φ b Indicates that file b is a monitoring model consisting of a set of certain data categories. Further, define the population symbol Ω T , used to represent the population of the Tth generation, for each population, there is Ω T ={Ψ1,Ψ2,Ψ3,...,Ψ d |d=1,2,3,...,D}, where D represents the population size.

[0100] (1) Crossover process

[0101] Assume that the crossover probability of each iteration is pcross To ensure the reasonable conduct of crossover, the following assumptions are made: 1) Gene fragments are inseparable, and a group of genes (individuals) is the basic unit of crossover; 2) When the two gene loci are the same, crossover does not occur and is directly passed on to the offspring, otherwise crossover is implemented.

[0102] (2) Mutation process

[0103] Assume that each chromosome has p mutation The probability of mutation occurs, the mutation process is 0,1 interchange, and the mutation position is random.

[0104] (3) Selection process

[0105] The selection process is to allow good individuals to be inherited. Here, the conventional roulette method is used, and the individual fitness is the objective function. Assume that the fitness of an individual is Z(Ψ), and the total fitness is Z total , then the probability of selecting an individual is:

[0106]

[0107] So far, the entire solution steps are given:

[0108] Step 1: Initialize relevant parameters and arbitrarily generate a dynamic monitoring model, i.e., a population, based on random numbers.

[0109] Step 2: Implement iterations based on different models and run relevant data.

[0110] Step 3 calculates the fitness of individuals and groups based on the data results.

[0111] Step 4: Perform mutation, crossover, and selection operations on individuals in the population to generate a new population collection;

[0112] Step 5 determines the number of iterations. If it exceeds the limit, go to Step 6; otherwise, go to Step 2.

[0113] Step 6 outputs the optimal iteration strategy.

[0114] Through the above algorithms, users can intelligently evaluate different models while defining their own balance points, and seek a balance between recognition quality and time efficiency to achieve the lowest operating cost.

[0115] Among them, the lowest operating cost means the minimum of the above objective function.

[0116] In this embodiment, the computer device obtains n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set; then, according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set, the objective function of the monitoring indicator optimization model is determined, so as to determine the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set, and then realize the determination of the target monitoring indicator set by the monitoring indicator optimization model, so as to prepare a data foundation for determining the subsequent update of the original application classification model. The above process is easy to implement by computer equipment, which can improve the efficiency of determining the target monitoring indicator set and help to update the application classification model in a timely manner.

[0117] In one embodiment, the determination of the above application classification can be achieved by introducing a mathematical programming model to supervise the clustering algorithm, such as Figure 4 As shown, the process of determining the application classification model includes:

[0118] S410: Determine the target monitoring indicator included in each historical financial application in the historical financial application set, and construct an indicator vector group according to the target monitoring indicator included in each historical financial application.

[0119] S420. Construct an indicator directed graph according to the indicator vector groups corresponding to all historical financial applications.

[0120] S430. Use a general clustering algorithm and an indicator directed graph to determine an application classification model.

[0121] Specifically, the computer device can use a general clustering algorithm and an indicator directed graph to determine an initial application classification model; wherein the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement. Then, the model parameters in the initial application classification model are solved based on the simulated annealing algorithm, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein the model parameters include the target number.

[0122] Optionally, the overall process of determining the application classification model is as follows:

[0123] Index vector group As an indicator to measure the similarity of applications. Assume that a candidate set of V applications can be regarded as a vector group set consisting of V vectors, which can be constructed as a directed graph G = (V, E, w). Among them, E represents the arc between single application vectors, and E = {e|e = (v w ,v v ),v w ,v v ∈V,v w≠v v}; w represents the weight of the arc, w={w e |e∈E}.

[0124] For graph G, define its weighted adjacency matrix A j , whose dimension is o×o. j The elements in are:

[0125]

[0126] Where w and v represent indices. Further, define the degree matrix D j , whose diagonal elements are:

[0127]

[0128] The remaining elements are all 0. At this time, L j It is Figure G j The Laplace matrix of j =D j -A j .

[0129] The goal of spectral clustering is to cut a graph into k subgraphs, and to ensure that the sum of the interval weights connecting the subgraphs is minimized and the sum of the weights within the subgraphs is large.

[0130] Suppose the graph is cut into two subgraphs, for subgraphs I and Assume that the set of connection paths between the two is E c ,have:

[0131]

[0132] Then the weight of the path set is W c , whose value is:

[0133]

[0134] The partitioning objective function based on spectral clustering is as follows:

[0135]

[0136] Here, the present application uses the multi-path canonical cut set criterion as an implementation method, which has higher execution efficiency than other criteria, and its recognition model is as follows:

[0137]

[0138] Based on this model, the goal of this step is to divide the graph composed of V application vectors into a number of κ subgraphs, and after the division, the similarity between the application vectors in the subgraph reaches the highest, and the similarity between the application vectors in the subgraph reaches the lowest. This process needs to be driven by an intelligent algorithm, and the solution process is introduced below.

[0139] Model solving process based on simulated annealing:

[0140] After the model is established, there are two parameters to be optimized, one is the Gaussian kernel function σ, and the other is the number of separated clusters κ. In order to determine the optimal value, an optimization algorithm based on simulated annealing is proposed to optimize the values ​​of the two parameters.

[0141] Therefore, the constraints are:

[0142]

[0143] Where n is the number of iterations, Te is the temperature threshold, and η is the temperature decay rate.

[0144] Here is the specific solution process of the model:

[0145] Step 1 Initialize σ and κ, calculate Aj, D j , L j ;

[0146] Step2 Normalize the Laplacian matrix Calculate L S The eigenvectors corresponding to the first K small eigenvalues ​​of Composition vector space

[0147] Step 3 Convert all row vectors in Ψ into unit vectors to obtain matrix X;

[0148] Step 4: Use the k-means algorithm on each row of the matrix to obtain k clusters;

[0149] Step 5 When the e-th row of matrix X belongs to category f, the data point x e Classify into category f;

[0150] Step 6 calculates the objective function value F, random σ and κ, repeats Step 1-Step 5 to get the new function value;

[0151] Step 7: Calculate the increment ΔF and choose whether to accept the new solution;

[0152] If the number of iterations is reached in Step 8, go to Step 9, otherwise go to Step 6;

[0153] Check whether Step 9 meets the termination condition. If so, go to Step 10. Otherwise, cool down and go to Step 6.

[0154] Step 10 outputs the optimal solution.

[0155] In this embodiment, the computer device determines the target monitoring indicators included in each historical financial application in the historical financial application set, constructs an indicator vector group according to the target monitoring indicators included in each historical financial application, and constructs an indicator directed graph according to the indicator vector groups corresponding to all historical financial applications; and then uses the general clustering algorithm and the indicator directed graph to determine the application classification model. By vectorizing each financial application in the above manner, and determining the above application classification model based on a mathematical programming model, it is easy to implement the computer device, and improves the updating efficiency and accuracy of the application classification model.

[0156] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0157] In one embodiment, Figure 5 As shown, a financial application monitoring device is provided, including: an application classification module 501 and an abnormality monitoring module 502, wherein:

[0158] The application classification module 501 is used to input the target financial application set into the application classification model to obtain a classification result; wherein the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by using the monitoring indicator optimization model to screen the original monitoring indicator set according to the training time cost and the result deviation cost when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set;

[0159] The abnormal monitoring module 502 is used to determine abnormal applications in the target financial application set according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set.

[0160] In one embodiment, the application classification module 501 is further configured to:

[0161] The historical financial application set is input into the original application classification model to obtain the original classification result; the abnormal proportion of abnormal applications in the historical financial application set is determined according to the original classification result and the abnormal conditions corresponding to each original monitoring indicator in the original monitoring indicator set.

[0162] In one embodiment, the application classification module 501 is specifically used to:

[0163] Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set; determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set; determine the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set.

[0164] In one embodiment, the application classification module 501 is specifically used to:

[0165] The sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model; or, the weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model.

[0166] In one embodiment, the application classification module 501 is specifically used to:

[0167] Determine the target monitoring indicators included in each historical financial application in the historical financial application set, and construct an indicator vector group based on the target monitoring indicators included in each historical financial application; construct an indicator directed graph based on the indicator vector groups corresponding to all historical financial applications; use the general clustering algorithm and the indicator directed graph to determine the application classification model.

[0168] In one embodiment, the application classification module 501 is specifically used to:

[0169] The general clustering algorithm and the indicator directed graph are used to determine the initial application classification model; wherein, the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement; based on the simulated annealing algorithm, the model parameters in the initial application classification model are solved, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein, the model parameters include the target number.

[0170] Each module in the above-mentioned financial application monitoring device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0171] In one embodiment, an electronic device is provided. The electronic device may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown. The electronic device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for monitoring financial applications is implemented. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the housing of the electronic device, or an external keyboard, touchpad or mouse, etc.

[0172] Those skilled in the art will understand that Figure 6 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the electronic device to which the scheme of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0173] In one embodiment, an electronic device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0174] The target financial application set is input into the application classification model to obtain the classification result; wherein, the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set; the abnormal applications in the target financial application set are determined according to the classification results and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set.

[0175] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0176] The historical financial application set is input into the original application classification model to obtain the original classification result; the abnormal proportion of abnormal applications in the historical financial application set is determined according to the original classification result and the abnormal conditions corresponding to each original monitoring indicator in the original monitoring indicator set.

[0177] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0178] Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set; determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set; determine the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set.

[0179] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0180] The sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model; or, the weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model.

[0181] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0182] Determine the target monitoring indicators included in each historical financial application in the historical financial application set, and construct an indicator vector group based on the target monitoring indicators included in each historical financial application; construct an indicator directed graph based on the indicator vector groups corresponding to all historical financial applications; use the general clustering algorithm and the indicator directed graph to determine the application classification model.

[0183] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0184] The general clustering algorithm and the indicator directed graph are used to determine the initial application classification model; wherein, the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement; based on the simulated annealing algorithm, the model parameters in the initial application classification model are solved, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein, the model parameters include the target number.

[0185] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0186] The target financial application set is input into the application classification model to obtain the classification result; wherein, the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set; the abnormal applications in the target financial application set are determined according to the classification results and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set.

[0187] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0188] The historical financial application set is input into the original application classification model to obtain the original classification result; the abnormal proportion of abnormal applications in the historical financial application set is determined according to the original classification result and the abnormal conditions corresponding to each original monitoring indicator in the original monitoring indicator set.

[0189] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0190] Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set; determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set; determine the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set.

[0191] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0192] The sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model; or, the weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model.

[0193] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0194] Determine the target monitoring indicators included in each historical financial application in the historical financial application set, and construct an indicator vector group based on the target monitoring indicators included in each historical financial application; construct an indicator directed graph based on the indicator vector groups corresponding to all historical financial applications; use the general clustering algorithm and the indicator directed graph to determine the application classification model.

[0195] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0196] The general clustering algorithm and the indicator directed graph are used to determine the initial application classification model; wherein, the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement; based on the simulated annealing algorithm, the model parameters in the initial application classification model are solved, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein, the model parameters include the target number.

[0197] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0198] The target financial application set is input into the application classification model to obtain the classification result; wherein, the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set; the abnormal applications in the target financial application set are determined according to the classification results and the abnormal conditions corresponding to each target monitoring indicator in the target monitoring indicator set.

[0199] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0200] The historical financial application set is input into the original application classification model to obtain the original classification result; the abnormal proportion of abnormal applications in the historical financial application set is determined according to the original classification result and the abnormal conditions corresponding to each original monitoring indicator in the original monitoring indicator set.

[0201] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0202] Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set; determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set; determine the candidate monitoring indicator set corresponding to the minimum value of the objective function as the target monitoring indicator set.

[0203] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0204] The sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model; or, the weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set is obtained as the objective function of the monitoring indicator optimization model.

[0205] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0206] Determine the target monitoring indicators included in each historical financial application in the historical financial application set, and construct an indicator vector group based on the target monitoring indicators included in each historical financial application; construct an indicator directed graph based on the indicator vector groups corresponding to all historical financial applications; use the general clustering algorithm and the indicator directed graph to determine the application classification model.

[0207] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0208] The general clustering algorithm and the indicator directed graph are used to determine the initial application classification model; wherein, the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement; based on the simulated annealing algorithm, the model parameters in the initial application classification model are solved, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein, the model parameters include the target number.

[0209] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0210] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0211] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for monitoring financial applications, characterized in that: The method comprises: Input the target financial application set into the application classification model to obtain a classification result; wherein the application classification model is a classification model determined according to the target monitoring indicator set and the historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using the monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than the abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set; Determine an abnormal application in the target financial application set according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set; The process of determining the application classification model includes: Determine the target monitoring indicator included in each historical financial application in the historical financial application set, and construct an indicator vector group according to the target monitoring indicator included in each historical financial application; Constructing an indicator directed graph according to the indicator vector group corresponding to all the historical financial applications; Determining the application classification model using a spectral clustering algorithm and the index directed graph; Wherein, the use of the spectral clustering algorithm and the indicator directed graph to determine the application classification model includes: The initial application classification model is determined by using a spectral clustering algorithm and the indicator directed graph; wherein the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement; The model parameters in the initial application classification model are solved based on a simulated annealing algorithm, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein the model parameters include the target quantity.

2. The method according to claim 1, characterized in that The method further comprises: Inputting the historical financial application set into the original application classification model to obtain an original classification result; The abnormal proportion of abnormal applications in the historical financial application set is determined according to the original classification result and the abnormal condition corresponding to each original monitoring indicator in the original monitoring indicator set.

3. The method according to claim 1, characterized in that The process of using the monitoring indicator optimization model to determine the target monitoring indicator set includes: Obtain n candidate monitoring indicators from the original monitoring indicator set to form a candidate monitoring indicator set; wherein n is sequentially taken from 1 to N-1, and N is the total number of original monitoring indicators in the original monitoring indicator set; Determine the objective function of the monitoring indicator optimization model according to the training time cost and result deviation cost of the candidate application classification model determined by each of the candidate monitoring indicator sets; The candidate monitoring indicator set corresponding to the minimum value of the objective function is determined as the target monitoring indicator set.

4. The method according to claim 3, characterized in that The objective function of the monitoring indicator optimization model is determined based on the training time cost and result deviation cost of the candidate application classification model determined by each candidate monitoring indicator set, including: Obtaining the sum of the training time cost and the result deviation cost of the candidate application classification model determined by each of the candidate monitoring indicator sets as the objective function of the monitoring indicator optimization model; or, The weighted sum of the training time cost and the result deviation cost of the candidate application classification model determined by each of the candidate monitoring indicator sets is obtained as the objective function of the monitoring indicator optimization model.

5. A monitoring device for financial applications, characterized in that: The device comprises: An application classification module, used for inputting a target financial application set into an application classification model to obtain a classification result; wherein the application classification model is a classification model determined according to a target monitoring indicator set and a historical financial application set, and the target monitoring indicator set is a monitoring indicator set obtained by screening the original monitoring indicator set according to the training time cost and the result deviation cost using a monitoring indicator optimization model when the proportion of abnormal applications obtained by the original application classification model is greater than an abnormal proportion threshold; the original application classification model is a classification model determined according to the original monitoring indicator set and the historical financial application set; An abnormality monitoring module, used for determining abnormal applications in the target financial application set according to the classification result and the abnormal condition corresponding to each target monitoring indicator in the target monitoring indicator set; The process of determining the application classification model includes: Determine the target monitoring indicator included in each historical financial application in the historical financial application set, and construct an indicator vector group according to the target monitoring indicator included in each historical financial application; Constructing an indicator directed graph according to the indicator vector group corresponding to all the historical financial applications; Determining the application classification model using a spectral clustering algorithm and the index directed graph; Wherein, the use of the spectral clustering algorithm and the indicator directed graph to determine the application classification model includes: The initial application classification model is determined by using a spectral clustering algorithm and the indicator directed graph; wherein the initial application classification model is used to divide the indicator directed graph into a target number of subgraphs, and the similarity between the indicator vector groups in each subgraph meets the similarity lower limit requirement; The model parameters in the initial application classification model are solved based on a simulated annealing algorithm, and the model parameters are brought into the initial application classification model to obtain the application classification model; wherein the model parameters include the target quantity.

6. An electronic device, comprising a terminal memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images