CN116362346A

CN116362346A - Digital wallet recognition model training, digital wallet recognition method, device and equipment

Info

Publication number: CN116362346A
Application number: CN202310195274.XA
Authority: CN
Inventors: 靳晓松; 王阳; 卢厚祥; 杨叶平
Original assignee: Ccb Trust Co ltd
Current assignee: Ccb Trust Co ltd
Priority date: 2023-02-24
Filing date: 2023-02-24
Publication date: 2023-06-30

Abstract

The embodiment of the invention discloses digital wallet recognition model training and digital wallet recognition methods, devices and equipment. The digital wallet identifying model training method comprises the following steps: acquiring each digital wallet sample data in the digital wallet sample data set; determining digital wallet sample characteristics corresponding to the digital wallet sample data, and performing characteristic derivative processing on the digital wallet sample characteristics to obtain a data characteristic set corresponding to the digital wallet sample data; determining sample data labels corresponding to the sample data of each digital wallet, and determining a digital wallet training sample data set according to each sample data label; and determining digital wallet training sample data in the digital wallet training sample data set, and performing model training on the digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data. The technical scheme of the embodiment of the invention can accurately identify the digital wallet to be identified, thereby improving the accuracy of the digital wallet identification model.

Description

Digital wallet recognition model training, digital wallet recognition method, device and equipment

Technical Field

The embodiment of the invention relates to the technical field of computers, in particular to a digital wallet recognition model training method, a digital wallet recognition device and digital wallet recognition equipment.

Background

With the promotion of the digital rmb, the act of achieving a specific purpose with the digital rmb is accompanied, for example, fraud with the digital rmb, etc., so that the digital wallet needs to be identified. The traditional identification method usually adopts a single learner or an integrated learner to carry out identification, but the generalization capability of a model is weaker due to disturbance of sample data or disturbance of characteristic attributes, and high requirements are imposed on samples and characteristic variables.

However, in the identification scenario of digital wallets, the data samples are extremely unbalanced samples and there are no fixed distinct features, and the double random perturbation of sample data and feature attributes is particularly large. Therefore, if the digital wallet is identified by adopting the traditional algorithm, the model training result often has good performance in the training set and the verification set, but the prediction effect in the actual application scene is poor, and the digital wallet cannot be identified accurately.

Disclosure of Invention

The embodiment of the invention provides a digital wallet recognition model training method, a digital wallet recognition device, electronic equipment and a storage medium, which can accurately recognize a digital wallet to be recognized, thereby improving the accuracy of the digital wallet recognition model.

According to an aspect of the present invention, there is provided a digital wallet identifying model training method, including:

acquiring each digital wallet sample data in the digital wallet sample data set;

determining digital wallet sample characteristics corresponding to the digital wallet sample data, and performing characteristic derivation processing on the digital wallet sample characteristics to obtain a data characteristic set corresponding to the digital wallet sample data;

determining a sample data label corresponding to each digital wallet sample data, and determining a digital wallet training sample data set according to each sample data label;

and determining digital wallet training sample data in the digital wallet training sample data set, and carrying out model training on the digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data.

According to another aspect of the present invention, there is provided a digital wallet identifying method, comprising:

Acquiring a digital wallet to be identified, and determining the characteristics of the digital wallet to be identified corresponding to the digital wallet to be identified;

performing feature derivation processing on the feature of the digital wallet to be identified to obtain a feature set of the data to be identified corresponding to the digital wallet to be identified;

inputting the feature set of the data to be identified into a pre-trained digital wallet identifying model to generate a digital wallet identifying result through the digital wallet identifying model;

the digital wallet identifying model is obtained through training by the digital wallet identifying model training method.

According to another aspect of the present invention, there is provided a digital wallet identifying model training apparatus comprising:

the sample data acquisition module is used for acquiring each digital wallet sample data in the digital wallet sample data set;

the data feature set determining module is used for determining digital wallet sample features corresponding to the digital wallet sample data, and performing feature derivation processing on the digital wallet sample features to obtain a data feature set corresponding to the digital wallet sample data;

the training sample data set determining module is used for determining sample data labels corresponding to the digital wallet sample data and determining a digital wallet training sample data set according to the sample data labels;

And the model training module is used for determining digital wallet training sample data in the digital wallet training sample data set and carrying out model training on the digital wallet recognition model according to the data characteristic set corresponding to the digital wallet training sample data.

According to another aspect of the present invention, there is provided a digital wallet identifying apparatus comprising:

the digital wallet feature determining module is used for acquiring the digital wallet to be identified and determining the digital wallet feature to be identified corresponding to the digital wallet to be identified;

the to-be-identified data feature set determining module is used for carrying out feature derivation processing on the to-be-identified digital wallet features to obtain to-be-identified data feature sets corresponding to the to-be-identified digital wallets;

the digital wallet identifying result generating module is used for inputting the data characteristic set to be identified into a pre-trained digital wallet identifying model so as to generate a digital wallet identifying result through the digital wallet identifying model;

According to another aspect of the present invention, there is provided an electronic apparatus including:

At least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the digital wallet identifying model training method of any of the embodiments of the invention or to perform the digital wallet identifying method of any of the embodiments of the invention.

According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement the digital wallet identifying model training method of any of the embodiments of the present invention or the digital wallet identifying method of any of the embodiments of the present invention when executed.

According to the technical scheme, the digital wallet sample data in the digital wallet sample data set are obtained, the digital wallet sample characteristics corresponding to the digital wallet sample data are determined, the characteristic derivative processing is carried out on the digital wallet sample characteristics to obtain the data characteristic set, the sample data label corresponding to the digital wallet sample data is determined, the digital wallet training sample data set is determined according to the sample data label, the digital wallet training sample data is determined in the digital wallet training sample data set, the digital wallet identification model is trained according to the data characteristic set corresponding to the digital wallet training sample data, the digital wallet characteristics to be identified are determined after the digital wallet to be identified is obtained, the characteristic derivative processing is carried out on the digital wallet characteristics to be identified to obtain the data characteristic set to be identified, the data characteristic set to be identified is input into the digital wallet identification model, the digital wallet identification result is generated through the digital wallet identification model, the problem that the digital wallet cannot be accurately identified in the prior art is solved, the digital wallet to be identified can be accurately identified, and accordingly the accuracy of the digital wallet to be identified is improved.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flowchart of a digital wallet identifying model training method according to an embodiment of the present invention;

fig. 2 is a flowchart of a digital wallet identifying model training method according to a second embodiment of the present invention;

fig. 3 is a flowchart of a digital wallet identifying method according to a third embodiment of the present invention;

fig. 4 is a schematic diagram of a digital wallet identifying model training apparatus according to a fourth embodiment of the present invention;

fig. 5 is a schematic diagram of a digital wallet identifying device according to a fifth embodiment of the present invention;

Fig. 6 is a schematic structural diagram of an electronic device implementing the digital wallet identifying model training method or the digital wallet identifying method according to the embodiment of the invention.

Detailed Description

In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. The technical scheme of the invention obtains, stores, uses or processes the data and the like, and accords with the relevant regulations of national laws and regulations.

Example 1

Fig. 1 is a flowchart of a digital wallet identifying model training method provided in an embodiment of the present invention, where the method may be applied to accurately identify a digital wallet, and the method may be performed by a digital wallet identifying model training apparatus, where the apparatus may be implemented in software and/or hardware, and may generally be directly integrated in an electronic device that performs the method, where the electronic device may be a terminal device or a server device, and the embodiment of the present invention does not limit the type of electronic device that performs the digital wallet identifying model training method. Specifically, as shown in fig. 1, the digital wallet identifying model training method specifically includes the following steps:

s110, acquiring each digital wallet sample data in the digital wallet sample data set.

Wherein the digital wallet sample data set may be a set of a plurality of digital wallet sample data. The digital wallet sample data may be a digital rmb wallet capable of being used as sample data in training the digital wallet recognition model.

In the embodiment of the invention, after the digital wallet sample data set is acquired, each digital wallet sample data in the digital wallet sample data set can be further acquired. It is to be appreciated that a plurality of digital wallet sample data can be included in the digital wallet sample data set.

S120, determining digital wallet sample characteristics corresponding to the digital wallet sample data, and performing characteristic derivation processing on the digital wallet sample characteristics to obtain a data characteristic set corresponding to the digital wallet sample data.

The digital wallet sample feature may be any feature attribute of the digital wallet sample data, for example, may be a digital wallet transaction amount, a digital wallet account balance, a digital wallet account available balance, or a digital wallet preferential amount, which is not limited by the embodiment of the present invention. The feature derivation process may be deriving digital wallet sample features. The data feature set may be a feature set of digital wallet sample data obtained after feature derivation, and may include, for example, digital wallet sample features, conventional statistics features, wallet traffic features, or the like, which is not limited by the embodiment of the invention.

In the embodiment of the invention, after each digital wallet sample data in the digital wallet sample data set is obtained, the digital wallet sample characteristics corresponding to each digital wallet sample data can be further determined so as to perform characteristic derivative processing on each digital wallet sample characteristic, thereby obtaining the data characteristic set corresponding to each digital wallet sample data. It is understood that different digital wallet sample data corresponds to different sets of data characteristics.

It can be understood that when the digital wallet identifying model is trained, the digital wallet sample data has fewer digital wallet sample characteristics and is not representative enough, and the digital wallet identifying model obtained by training has low accuracy and poor generalization capability, so that the digital wallet sample characteristics are subjected to characteristic derivation processing.

S130, determining sample data labels corresponding to the digital wallet sample data, and determining a digital wallet training sample data set according to the sample data labels.

The sample data tag may be a tag corresponding to digital wallet sample data, and may be used to identify digital wallet sample data. The digital wallet training sample data set may be a collection of digital wallet sample data that can be used to train a digital wallet recognition model.

In the embodiment of the invention, after the feature derivation processing is performed on the sample features of each digital wallet to obtain the data feature set corresponding to the sample data of each digital wallet, the sample data label corresponding to the sample data of each digital wallet can be further determined, so that the digital wallet training sample data set is determined according to the sample data labels.

It should be noted that, in the digital wallet sample data set, the distribution of the digital wallet sample data corresponding to the different sample data labels is not balanced, if the data set is not balanced, the model accuracy during training is very high, but the model accuracy in the test set is very low, and the model generalization capability is not strong. Therefore, according to the technical scheme, the digital wallet training sample data set is determined according to the sample data labels, so that sample data during training of the digital wallet recognition model can be distributed uniformly, and the accuracy of the digital wallet recognition model and the generalization capability of the model are improved.

S140, determining digital wallet training sample data in the digital wallet training sample data set, and performing model training on a digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data.

The digital wallet training sample data may be digital wallet sample data that can be used to train a digital wallet recognition model, among other things. The digital wallet identifying model may be used to identify digital wallets. In the embodiment of the invention, after the digital wallet training sample data set is determined according to each sample data label, digital wallet training sample data can be further determined in the digital wallet training sample data set, so that model training is carried out on the digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data.

Alternatively, the digital wallet identifying model may be constructed using a random forest algorithm.

In an application scenario of digital rmb wagering-related fraud, model training is performed on the digital wallet identifying model according to a data feature set corresponding to digital wallet training sample data, and may include randomly selecting a preset number (e.g., 10) of data features in the data feature set corresponding to the digital wallet training sample data, performing decision tree feature attribute splitting on a first tree in the digital wallet identifying model, randomly selecting a preset number (e.g., 10) of data features in the data feature set corresponding to the digital wallet training sample data, performing decision tree feature splitting on a second tree in the digital wallet identifying model, and so on. It will be appreciated that the process of splitting the feature attributes of the decision tree may continue to split leaf nodes until all nodes cannot re-split leaf nodes. Each decision tree feature attribute split can form a tree, and a large number of decision trees can be constructed to form a random forest, thereby constructing a digital wallet recognition model.

According to the technical scheme, through the acquisition of each digital wallet sample data in the digital wallet sample data set, the digital wallet sample characteristics corresponding to each digital wallet sample data are determined, the characteristic derivation processing is carried out on each digital wallet sample characteristic to obtain the data characteristic set, the sample data label corresponding to each digital wallet sample data is determined, the digital wallet training sample data set is determined according to each sample data label, and the digital wallet training sample data is determined in the digital wallet training sample data set, so that model training is carried out on a digital wallet identification model according to the data characteristic set corresponding to the digital wallet training sample data, the problem that a digital wallet cannot be accurately identified in the prior art is solved, the digital wallet to be identified can be accurately identified, and the accuracy of the digital wallet identification model is improved.

Example two

Fig. 2 is a flowchart of a digital wallet identifying model training method according to a second embodiment of the present invention, where the foregoing technical solutions are further refined, and feature derivation processing is performed on each digital wallet sample feature to obtain a data feature set corresponding to each digital wallet sample data, a digital wallet training sample data set is determined according to each sample data tag, and a digital wallet identifying model is trained according to the data feature set corresponding to the digital wallet training sample data. The technical solution in this embodiment may be combined with each of the alternatives in one or more embodiments described above. As shown in fig. 2, the method may include the steps of:

S210, acquiring each digital wallet sample data in the digital wallet sample data set.

S220, determining digital wallet sample characteristics corresponding to the digital wallet sample data, and carrying out statistic derivation processing on the digital wallet sample characteristics to obtain statistic characteristics corresponding to the digital wallet sample data.

The statistic derivation process may be any statistic derivation of the digital wallet sample feature, for example, may be any statistic derivation of the digital wallet sample feature such as average, variance, standard deviation, or median, which is not limited by the embodiment of the present invention. The statistic feature may be a statistic-related feature derived from the statistic-derived processing of the digital wallet sample data feature.

In the embodiment of the invention, after each digital wallet sample data in the digital wallet sample data set is acquired, the digital wallet sample characteristics corresponding to each digital wallet sample data can be further determined, so that statistic derived processing is carried out on each digital wallet sample characteristic, and thus the statistic characteristics corresponding to each digital wallet sample data are obtained.

And S230, carrying out traffic volume derivation processing on the digital wallet sample characteristics to obtain traffic volume characteristics corresponding to the digital wallet sample data.

The traffic derivation process may be any derivation of digital wallet sample features, for example, may be a derivation of digital wallet sample features for digital wallet transaction services, and the embodiment of the present invention is not limited in this respect. The traffic characteristics may be characteristics related to wallet traffic obtained by performing a deriving process on the digital wallet sample data characteristics, for example, may be transaction times in a preset time period, transaction amounts in a preset time period, or opponent times in a preset time period, which is not limited in the embodiment of the present invention.

In the embodiment of the invention, after the digital wallet sample characteristics corresponding to the digital wallet sample data are determined, the service volume derivation processing can be further carried out on the digital wallet sample characteristics, so that the service volume characteristics corresponding to the digital wallet sample data are obtained.

Optionally, performing the traffic derivatization on each digital wallet sample feature may include configuring a time dimension parameter, a packet dimension parameter, or a polynomial derivatization parameter, and implementing the traffic derivatization by using a polynominal features function (polynomial regression function). The time dimension parameters may include, among other things, a day, a week, a month, or a quarter, etc. The group dimension parameters may include in-group, out-group, or trade-hand group, etc. The polynomial derived parameters may include a number of transaction times derived or a transaction amount derived, etc.

Illustratively, in the application scenario of digital renminbi to the fraud, the criminal fraud features of the fraudulent wallet may include: frequent transactions in a short time, the wallet of a fraudulent client is dispersed and supervised, money is remitted to a plurality of wallets of the same client in a short time by a plurality of wallets, the money is large, frequent transfer in and transfer out among the wallets are closely related, the operation time is extremely short, and the like. Performing traffic derivatization processing on the digital wallet sample characteristics to obtain traffic characteristics corresponding to the digital wallet sample data may include: the total number of the transfer-in/transfer-out transaction in one day, the total amount of the transfer-in/transfer-out same transaction opponent in one day, the number of the transfer-in/transfer-out transaction opponents in one day, the total number of the transfer-in/transfer-out transaction in one week, the total number of the transfer-in/transfer-out transaction opponents in one week, the total number of the transfer-in/transfer-out transaction in one month, the total number of the transfer-in/transfer-out transaction in 3 months, the total number of the transfer-in/transfer-out transaction opponents in 3 months, and the like.

According to the technical scheme, the service meaning of the derivative variable can be increased by carrying out service quantity derivative processing on the digital wallet sample characteristics, so that the derivative variable with the service meaning is obtained, and more effective characteristic variables are obtained.

S240, determining a data feature set corresponding to each digital wallet sample data according to each digital wallet sample feature, each statistic feature and each traffic volume feature.

In the embodiment of the invention, after obtaining the statistic characteristics corresponding to the digital wallet sample data and obtaining the traffic characteristics corresponding to the digital wallet sample data, the data characteristic set corresponding to the digital wallet sample data can be further determined according to the digital wallet sample characteristics, the statistic characteristics and the traffic characteristics.

Optionally, the digital wallet sample feature, the statistic feature and the traffic feature may be combined to form a large-width table to obtain the data feature set.

S250, determining a sample data label corresponding to each digital wallet sample data, and determining the digital wallet sample data as first target sample data under the condition that the sample data label is determined to be the target data label.

Wherein the target data tag may be one of various sample data tags. The first target sample data may be one of a plurality of digital wallet sample data.

In the embodiment of the invention, after feature derivation processing is performed on the sample features of each digital wallet to obtain the data feature set corresponding to each digital wallet sample data, the sample data tag corresponding to each digital wallet sample data can be further determined, so that the digital wallet sample data is determined to be the first target sample data when the sample data tag is the target data tag.

And S260, determining the digital wallet sample data as second target sample data under the condition that the sample data tag is determined to be a non-target data tag.

Wherein the second target sample data may be another target data of the plurality of digital wallet sample data.

In the embodiment of the invention, after feature derivation processing is performed on the sample features of each digital wallet to obtain the data feature set corresponding to each digital wallet sample data, the sample data tag corresponding to each digital wallet sample data can be further determined, so that the digital wallet sample data is determined to be the second target sample data when the sample data tag is not the target data tag.

S270, determining the digital wallet training sample data set according to a first preset number of first target sample data and a second preset number of second target sample data.

The first preset number may be a preset number. The second preset number may be another preset number. It is to be understood that the first preset number and the second preset number may be the same or different, and the embodiment of the present invention is not limited thereto.

In the embodiment of the invention, after the first target sample data and the second target sample data are determined, a first preset number of first target sample data can be further acquired, and a second preset number of second target sample data can be acquired, so that a digital wallet training sample data set is determined according to the first preset number of first target sample data and the second preset number of second target sample data.

Optionally, determining the digital wallet training sample data set according to the first target sample data of the first preset number and the second target sample data of the second preset number may include merging the first target sample data of the first preset number and the second target sample data of the second preset number to obtain merged sample data, so as to divide the merged sample data into the digital wallet training sample data set and the digital wallet verification sample data set according to a preset proportion.

For example, in an application scenario of digital rmb betting fraud, the sample data tag may include a betting fraud tag and a non-betting fraud tag, the target data tag may be a betting fraud tag, the first target sample data may be betting fraud digital wallet sample data, and the second target sample data may be non-betting fraud digital wallet sample data. Assuming that the first preset number is ten thousand and the second preset number is ninety thousand, combining ten thousand of the fraud digital wallet sample data with ninety thousand of the non-fraud digital wallet sample data to obtain combined sample data to divide the combined sample data into a digital wallet training sample data set and a digital wallet verification sample data set according to a ratio of 8:2. It can be understood that in the application scenario of the digital rmb to the fraud, the ratio of the number of wallets of the sample data without the fraud to the fraud is 100:1, so that the sample data needs to be balanced.

According to the technical scheme, the digital wallet training sample data set is determined according to the first target sample data with the first preset number and the second target sample data with the second preset number, so that the digital wallet sample data can be subjected to balanced processing, the duty ratio of the sample data is improved, the calculated amount is reduced, and the accuracy of model training is improved.

S280, determining digital wallet training sample data in the digital wallet training sample data set, and performing model training on a digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data.

Optionally, performing model training on the digital wallet identifying model according to the data feature set corresponding to the digital wallet training sample data may include: according to the data feature set corresponding to the digital wallet training sample data, performing first model training to obtain a first digital wallet identification model; determining the contribution degree of each data feature in the data feature set to the model of the first digital wallet identifying model; sequencing the data features according to the model contribution degree, and acquiring a third preset number of target data features according to the sequencing result; training a second model according to the characteristics of each target data to obtain a second digital wallet recognition model; the digital wallet identifying model is determined based on the first digital wallet identifying model and/or the second digital wallet identifying model.

Wherein the first model training may be a model training process. The first digital wallet identifying model may be a digital wallet identifying model. The data feature may be any data feature in the set of data features. The model contribution degree may be a contribution of the first digital wallet identifying model in identifying the digital wallet. The third predetermined number may be another predetermined number. The target data feature may be one of the data features. The second model training may be another model training process. The second digital wallet identifying model may be another digital wallet identifying model. It is understood that the model parameters of the first digital wallet identifying model and the model parameters of the second digital wallet identifying model may be different.

Specifically, after the digital wallet training sample data is determined in the digital wallet training sample data set, a first model training can be further performed according to a data feature set corresponding to the digital wallet training sample data to obtain a first digital wallet recognition model, the model contribution degree of each data feature in the data feature set to the first digital wallet recognition model is determined, so that each data feature is ordered according to the model contribution degree, a third preset number of target data features are obtained according to the ordering result, a second model training is performed according to each target data feature to obtain a second digital wallet recognition model, and then the digital wallet recognition model is determined according to the first digital wallet recognition model, the second digital wallet recognition model, or the first digital wallet recognition model and the second digital wallet recognition model.

It should be noted that, in the application scenario of digital rmb wagering, it is assumed that the data feature set corresponding to the digital wallet training sample data may include 130 data features, while some data features (such as gender, wallet available balance or wallet account balance) have a low contribution to the digital wallet recognition model, if all the data features are used for model training, disturbance of the sample data may affect feature extraction, and on the other hand, a great amount of calculation performance may be consumed, so that model training speed is slow. Therefore, in the embodiment of the invention, the model contribution degree of each data feature in the data feature set to the first digital wallet recognition model is determined, and each data feature is sequenced according to the model contribution degree, so that a third preset number (such as 8) of target data features (such as total wallet number, customer age, preferential amount, total amount of transactions transferred in the last month, total number of times transferred out of the same transaction opponent in the last month, total number of transactions transferred in the last 3 months and total number of times transferred out of the same transaction opponent in the last 3 months) are obtained according to the sequencing result, and thus, a second model training is performed according to each target data feature, and a second digital wallet recognition model is obtained.

Optionally, determining the digital wallet identifying model according to the first digital wallet identifying model and/or the second digital wallet identifying model may include: and determining the first digital wallet identifying model as a digital wallet identifying model, or determining the second digital wallet identifying model as a digital wallet identifying model, or determining the model accuracy of the first digital wallet identifying model and the model accuracy of the second digital wallet identifying model to determine the digital wallet identifying model with the high model accuracy.

Optionally, after model training is performed on the digital wallet identifying model according to the data feature set corresponding to the digital wallet training sample data, the method may further include: in the case that the model accuracy of the digital wallet identifying model is determined to have not met the model training stop condition, the operation of determining the digital wallet training sample data set from the respective sample data tags is performed back.

The model training stopping condition may be a condition for stopping model training, for example, the model accuracy rate may reach a preset value, or the model accuracy rate may not be greatly improved, which is not limited in the embodiment of the present invention.

Specifically, after model training is performed on the digital wallet identifying model according to the data feature set corresponding to the digital wallet training sample data, whether the model accuracy of the digital wallet identifying model meets the model training stopping condition can be further determined, and when the model accuracy of the digital wallet identifying model does not meet the model training stopping condition, the digital wallet training sample data set is determined according to each sample data label in a returning mode to obtain a new digital wallet training sample data set, so that model training is performed on the digital wallet identifying model according to the data feature set corresponding to the digital wallet training sample data in the new digital wallet training sample data set.

Optionally, in an application scenario of digital rmb wagering and fraud, the model parameters in the digital wallet identifying model may include: the optimators represent the trees in the random forest, i.e. how many wagering-related intelligent recognition base learners (decision trees) to generate; boost indicates whether self-service sampling is adopted to generate a sampling set; obb _score indicates whether or not the out-of-bag data is used to estimate the validity of the intelligent identification model of the wagering involved in the fraud; criterion represents a criterion for selecting an optimal partitioning of the characteristic attributes of the wagering-related fraud, default "gini", optionally "entopy"; max_depth represents the maximum depth of the decision tree; max_features represents the maximum feature number (attribute sampling) of the randomly extracted wagering feature candidate partition attribute set; min_samples_split represents the minimum number of samples required for internal node subdivision; min_samples_leaf represents the minimum number of feature samples that the leaf node is involved in betting on; max_leaf_nodes represents the maximum leaf node number; min_weight_fraction_leaf represents the minimum sample weight sum of leaf nodes; min_input_split represents the node partition minimum unrepeace.

It can be appreciated that adjusting model parameters in the digital wallet identifying model is a key step in model training, and the training process of the digital wallet identifying model can be controlled by parameters, so that the digital wallet identifying model has stronger generalization capability. In the prior art, parameters are usually adjusted empirically, however, for different data distributions and different data characteristics, a large amount of computing resources are often required to be consumed, and a long time is required to obtain the model parameters of the optimal digital wallet identifying model.

In the embodiment of the invention, the optimized parameter configuration can be found by adopting a method of combining random zdSearchCV (sample random search) and random forest regression, so that a large amount of computing resources can be saved, and the optimal configuration can be found rapidly.

According to the technical scheme, model training of the digital wallet recognition model can be highly parallelized, and the digital wallet recognition model can be effectively operated on a large data set; the feature derivation processing is carried out on the wallet sample data, and the digital wallet training sample data set is determined according to each sample data label, so that a model can be efficiently trained when the feature dimension of the sample is high, and the model generalization capability obtained by training is high; sequencing the data features according to the model contribution degree, and acquiring a third preset number of target data features according to the sequencing result, so that feature screening can be performed on the arrangement weights of the predicted targets by counting the features; the tolerance to the deletion of partial features is high; the out-of-bag data can be used as a validation set to verify the validity of the model without additional partitioning of the data set; deviations and variances in the statistics features can affect the performance of the model through accuracy; the model parameters can be adjusted to achieve great harmony between the deviation and the variance of the overall model.

According to the technical scheme, through the acquisition of each digital wallet sample data in the digital wallet sample data set, the digital wallet sample characteristics corresponding to each digital wallet sample data are determined, statistics deriving processing is carried out on each digital wallet sample characteristic to obtain statistics characteristics, traffic deriving processing is carried out on each digital wallet sample characteristic to obtain traffic characteristics, so that the data characteristic set corresponding to each digital wallet sample data is determined according to each digital wallet sample characteristic, each statistics characteristic and each traffic characteristic, then the sample data label corresponding to each digital wallet sample data is determined, the digital wallet sample data is determined to be first target sample data when the sample data label is the target data label, the digital wallet sample data is determined to be second target sample data when the sample data label is the non-target data label, and the digital wallet training sample data set is determined according to the first target sample data with the first preset number and the second target sample data with the second preset number, so that the digital wallet sample data is determined in the digital wallet training sample data set, the digital wallet sample data is treated according to the data characteristic set corresponding to the digital wallet training data, the digital wallet training model is treated, the problem that the digital wallet sample data can not be accurately identified in the prior art is solved, and the digital wallet sample recognition model can not be accurately identified is accurately identified.

Example III

Fig. 3 is a flowchart of a digital wallet identifying method according to a third embodiment of the present invention, where the method may be applied to accurately identify a digital wallet, and the method may be performed by a digital wallet identifying apparatus, where the apparatus may be implemented in software and/or hardware, and may be generally directly integrated in an electronic device that performs the method, where the electronic device may be a terminal device or a server device, and the embodiment of the present invention does not limit the type of electronic device that performs the digital wallet identifying method. Specifically, as shown in fig. 3, the digital wallet identifying method specifically includes the following steps:

s310, acquiring the digital wallet to be identified, and determining the characteristics of the digital wallet to be identified corresponding to the digital wallet to be identified.

The digital wallet to be identified can be a digital RMB wallet to be identified. The digital wallet feature to be identified may be any feature attribute of the digital rmb wallet to be identified.

In the embodiment of the invention, the digital wallet to be identified is obtained to determine the characteristics of the digital wallet to be identified corresponding to the digital wallet to be identified.

And S320, performing feature derivation processing on the feature of the digital wallet to be identified to obtain a feature set of the data to be identified corresponding to the digital wallet to be identified.

The data feature set to be identified can be a set formed by data features corresponding to the digital wallet to be identified.

In the embodiment of the invention, after the feature of the digital wallet to be identified corresponding to the digital wallet to be identified is determined, feature derivation processing can be further performed on the feature of the digital wallet to be identified, so as to obtain the feature set of the data to be identified corresponding to the digital wallet to be identified.

Optionally, performing feature derivation processing on the feature of the digital wallet to be identified to obtain a feature set of the data to be identified corresponding to the digital wallet to be identified, which may include: carrying out statistic derivatization processing on the digital wallet characteristics to be identified to obtain statistic characteristics to be identified corresponding to the digital wallet to be identified; carrying out traffic derivatization processing on the characteristics of the digital wallet to be identified to obtain the characteristics of the traffic to be identified, which correspond to the digital wallet to be identified; and determining a feature set of the data to be recognized corresponding to the digital wallet to be recognized according to the feature of the digital wallet to be recognized, the statistic feature to be recognized and the traffic feature to be recognized.

S330, inputting the feature set of the data to be identified into a pre-trained digital wallet identifying model to generate a digital wallet identifying result through the digital wallet identifying model.

The digital wallet identifying model can be obtained through training by the digital wallet identifying model training method. The digital wallet identifying result may be a result obtained by identifying the digital wallet to be identified. For example, in an application scenario of digital rmb wagering and fraud, the digital wallet identification result may be that the digital wallet to be identified is a wagering and fraud digital wallet, or that the digital wallet to be identified is a non-wagering and fraud digital wallet.

In the embodiment of the invention, after feature derivation processing is performed on the feature of the digital wallet to be identified to obtain the feature set of the data to be identified corresponding to the digital wallet to be identified, the feature set of the data to be identified can be further input into a pre-trained digital wallet identification model so as to generate a digital wallet identification result through the digital wallet identification model.

According to the technical scheme, the digital wallet to be identified is obtained, the feature of the digital wallet to be identified corresponding to the digital wallet to be identified is determined, and the feature of the digital wallet to be identified is subjected to feature derivation processing to obtain the feature set of the data to be identified, so that the feature set of the data to be identified is input into the digital wallet identification model, a digital wallet identification result is generated through the digital wallet identification model, the problem that the digital wallet cannot be accurately identified in the prior art is solved, the digital wallet to be identified can be accurately, and the accuracy of the digital wallet identification model is improved.

Example IV

Fig. 4 is a schematic diagram of a digital wallet identifying model training apparatus according to a fourth embodiment of the present invention, as shown in fig. 4, the apparatus includes: a sample data acquisition module 410, a data feature set determination module 420, a training sample data set determination module 430, and a model training module 440, wherein:

a sample data acquisition module 410 for acquiring each digital wallet sample data in the digital wallet sample data set;

the data feature set determining module 420 is configured to determine digital wallet sample features corresponding to the digital wallet sample data, and perform feature derivation processing on the digital wallet sample features to obtain a data feature set corresponding to the digital wallet sample data;

a training sample data set determining module 430, configured to determine a sample data tag corresponding to each digital wallet sample data, and determine a digital wallet training sample data set according to each sample data tag;

the model training module 440 is configured to determine digital wallet training sample data in the digital wallet training sample data set, and perform model training on the digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data.

Optionally, the data feature set determining module 420 may be specifically configured to: carrying out statistic derivation processing on the sample characteristics of each digital wallet to obtain statistic characteristics corresponding to sample data of each digital wallet; carrying out traffic derivatization processing on the sample characteristics of each digital wallet to obtain the corresponding traffic characteristics of the sample data of each digital wallet; and determining a data feature set corresponding to each digital wallet sample data according to each digital wallet sample feature, each statistic feature and each traffic feature.

Optionally, the training sample data set determining module 430 may be specifically configured to: in the case that the sample data tag is determined to be the target data tag, determining the digital wallet sample data as first target sample data; in the case that the sample data tag is determined to be a non-target data tag, determining the digital wallet sample data as second target sample data; and determining a digital wallet training sample data set according to the first target sample data of the first preset quantity and the second target sample data of the second preset quantity.

Optionally, the model training module 440 may be specifically configured to: according to the data feature set corresponding to the digital wallet training sample data, performing first model training to obtain a first digital wallet identification model; determining the contribution degree of each data feature in the data feature set to the model of the first digital wallet identifying model; sequencing the data features according to the model contribution degree, and acquiring a third preset number of target data features according to the sequencing result; training a second model according to the characteristics of each target data to obtain a second digital wallet recognition model; the digital wallet identifying model is determined based on the first digital wallet identifying model and/or the second digital wallet identifying model.

Optionally, the model training module 440 may be further specifically configured to: after model training is performed on the digital wallet identifying model according to the data feature set corresponding to the digital wallet training sample data, the operation of determining the digital wallet training sample data set according to each sample data label is performed back in the case that the model accuracy of the digital wallet identifying model is determined to not meet the model training stop condition.

The digital wallet identifying model training device provided by the embodiment of the invention can execute the digital wallet identifying model training method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the executing method.

Example five

Fig. 5 is a schematic diagram of a digital wallet identifying apparatus according to a fifth embodiment of the present invention, as shown in fig. 5, the apparatus includes: a digital wallet characteristics determination module 510, a data characteristics set to be identified determination module 520, and a digital wallet identification result generation module 530, wherein:

the digital wallet feature determining module 510 is configured to obtain a digital wallet to be identified, and determine a digital wallet feature to be identified corresponding to the digital wallet to be identified;

the to-be-identified data feature set determining module 520 is configured to perform feature derivation processing on the to-be-identified digital wallet features to obtain a to-be-identified data feature set corresponding to the to-be-identified digital wallet;

A digital wallet identifying result generating module 530, configured to input the feature set of data to be identified into a pre-trained digital wallet identifying model, so as to generate a digital wallet identifying result through the digital wallet identifying model;

Optionally, the to-be-identified data feature set determining module 520 may be specifically configured to: carrying out statistic derivatization processing on the digital wallet characteristics to be identified to obtain statistic characteristics to be identified corresponding to the digital wallet to be identified; carrying out traffic derivatization processing on the characteristics of the digital wallet to be identified to obtain the characteristics of the traffic to be identified, which correspond to the digital wallet to be identified; and determining a feature set of the data to be recognized corresponding to the digital wallet to be recognized according to the feature of the digital wallet to be recognized, the statistic feature to be recognized and the traffic feature to be recognized.

The digital wallet identifying device provided by the embodiment of the invention can execute the digital wallet identifying method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the executing method.

Example six

Fig. 6 shows a schematic diagram of the structure of an electronic device 10 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.

As shown in fig. 6, the electronic device 10 includes at least one processor 11, and a memory, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, etc., communicatively connected to the at least one processor 11, in which the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 may also be stored. The processor 11, the ROM 12 and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.

Various components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, etc.; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.

The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the digital wallet recognition model training method or the digital wallet recognition method.

In some embodiments, the digital wallet identifying model training method or digital wallet identifying method may be implemented as a computer program tangibly embodied on a computer readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the digital wallet identifying model training method or digital wallet identifying method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the digital wallet identifying model training method or the digital wallet identifying method in any other suitable manner (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.

The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.

The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.

Claims

1. A digital wallet identifying model training method, comprising:

2. The method according to claim 1, wherein the performing feature derivation processing on each of the digital wallet sample features to obtain a data feature set corresponding to each of the digital wallet sample data includes:

carrying out statistic derivation processing on the digital wallet sample characteristics to obtain statistic characteristics corresponding to the digital wallet sample data;

performing traffic derivatization processing on the digital wallet sample characteristics to obtain traffic characteristics corresponding to the digital wallet sample data;

and determining a data feature set corresponding to each digital wallet sample data according to each digital wallet sample feature, each statistic feature and each traffic feature.

3. The method of claim 1, wherein said determining a digital wallet training sample dataset from each of said sample data tags comprises:

determining the digital wallet sample data as first target sample data if the sample data tag is determined to be a target data tag;

Determining the digital wallet sample data as second target sample data if the sample data tag is determined to be a non-target data tag;

and determining the digital wallet training sample data set according to the first target sample data of the first preset quantity and the second target sample data of the second preset quantity.

4. The method of claim 1, wherein the model training the digital wallet recognition model according to the data feature set corresponding to the digital wallet training sample data comprises:

according to the data feature set corresponding to the digital wallet training sample data, performing a first model training to obtain a first digital wallet identification model;

determining a model contribution degree of each data feature in the data feature set to the first digital wallet identifying model;

sorting the data features according to the model contribution degree, and acquiring a third preset number of target data features according to sorting results;

training a second model according to each target data characteristic to obtain a second digital wallet identification model;

and determining the digital wallet identifying model according to the first digital wallet identifying model and/or the second digital wallet identifying model.

5. The method of claim 1, further comprising, after the model training of the digital wallet identifying model from the set of data characteristics corresponding to the digital wallet training sample data:

and returning to execute the operation of determining the digital wallet training sample data set according to each sample data label under the condition that the model accuracy of the digital wallet identifying model is determined to not meet the model training stopping condition.

6. A digital wallet identifying method, comprising:

the digital wallet identifying model is trained by the digital wallet identifying model training method of any one of claims 1-5.

7. A digital wallet identifying model training device, comprising:

8. A digital wallet identifying device, comprising:

9. An electronic device, the electronic device comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the digital wallet identifying model training method of any of claims 1-5 or to perform the digital wallet identifying method of claim 6.

10. A computer readable storage medium storing computer instructions for causing a processor to implement the digital wallet identifying model training method of any one of claims 1-5 or the digital wallet identifying method of claim 6 when executed.