CN114565041A - Payment big data analysis system based on internet finance and analysis method thereof - Google Patents
Payment big data analysis system based on internet finance and analysis method thereof Download PDFInfo
- Publication number
- CN114565041A CN114565041A CN202210190418.8A CN202210190418A CN114565041A CN 114565041 A CN114565041 A CN 114565041A CN 202210190418 A CN202210190418 A CN 202210190418A CN 114565041 A CN114565041 A CN 114565041A
- Authority
- CN
- China
- Prior art keywords
- feature
- classification
- feature map
- semantic
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0203—Market surveys; Market polls
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
Abstract
The application relates to the field of internet financial payment, and particularly discloses a payment big data analysis system and an analysis method thereof based on internet finance. Therefore, the user can be matched with the service to be paid more appropriately, and the credit risk of internet financial platform payment is effectively reduced.
Description
Technical Field
The present application relates to the field of internet financial payment, and more particularly, to a payment big data analysis system based on internet finance and an analysis method thereof.
Background
With the rapid development of internet technology, the amount of data related to payment on internet financial platforms is rapidly increasing in exponential order, and accordingly, providers providing various internet financial services are emerging as in spring. As such, now and in the near future, the data accumulated in the internet financial network will be huge, so it is important how to mine and analyze the huge amount of financial payment data to improve the quality of financial services provided by each provider to the user.
In addition, the time spent on the rapid development of internet finance from the initial online shopping to the online bank and third-party payment to the online financing is only as short as a few years, but a series of credit risk problems of internet finance payment also occur. Therefore, in order to effectively reduce the credit risk of internet financial platform payment, a payment big data analysis method based on internet finance is desired.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides a payment big data analysis system and an analysis method thereof based on internet finance, which respectively extract semantic feature information of user portrait and description of service to be paid through a semantic understanding model based on context, and process the obtained feature matrix by using the same convolutional neural network model so that the extracted associated features have correlation as much as possible in a high-dimensional feature space relative to the user label features and the service semantic features, and further update parameters of the convolutional neural network by using a gradient transfer method so as to improve the classification accuracy. Therefore, the user can be matched with the service to be paid more appropriately, and the credit risk of internet financial platform payment is effectively reduced.
According to an aspect of the present application, there is provided an internet finance-based payment big data analysis system, including:
a training module comprising:
a first training data acquisition unit for acquiring a user profile, the user profile being a series of (labels, weights) key-value pairs;
training a user data encoding unit for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model comprising an embedding layer to obtain a plurality of label feature vectors;
the second training data acquisition unit is used for acquiring the description of the service to be paid;
the training service data coding unit is used for performing word segmentation processing on the description of the service to be paid and then obtaining a plurality of semantic feature vectors through the context-based semantic understanding model containing the embedded layer;
the first neural network unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through a convolutional neural network to obtain a first feature map;
the second neural network unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network to obtain a second feature map;
a feature fusion unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map;
the loss function value calculation unit is used for enabling the classification characteristic map to pass through a classifier so as to obtain a classification loss function value;
a training unit, configured to train the convolutional neural network based on the classification loss function values, wherein in each iteration of training the convolutional neural network, gradient information of the classification loss function values relative to the classification feature map is first transferred to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transferred back to parameters of the convolutional neural network by using branches of the first feature matrix; and
an inference module comprising:
a to-be-analyzed user data unit for acquiring a user representation, wherein the user representation is a series of (labels, weights) key value pairs;
a user portrait semantic encoding unit for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model containing an embedding layer to obtain a plurality of label feature vectors;
the payment service data unit is used for acquiring the description of the payment service;
the semantic coding unit of the service to be paid is used for performing word segmentation processing on the description of the service to be paid and then obtaining a plurality of semantic feature vectors through the context-based semantic understanding model comprising the embedded layer;
the first associated feature coding unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through the convolutional neural network trained by the training module to obtain a first feature map;
the second associated feature coding unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network trained by the training module to obtain a second feature map;
a classification feature map generation unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map; and
and the result output unit is used for enabling the classification characteristic graph to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the user is adaptive to the service to be paid or not.
Compared with the prior art, the internet finance-based payment big data analysis system and the internet finance-based payment big data analysis method respectively extract semantic feature information of user portrait and description of a service to be paid through a context-based semantic understanding model, process the obtained feature matrix by using the same convolutional neural network model, enable the extracted associated features to have correlation as much as possible in a high-dimensional feature space relative to the user tag features and the service semantic features, and further update parameters of the convolutional neural network by using a gradient transfer method so as to improve classification accuracy. Therefore, the user can be matched with the service to be paid more appropriately, and the credit risk of internet financial platform payment is effectively reduced.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally indicate like parts or steps.
Fig. 1 is a schematic training diagram of a convolutional neural network of an internet finance-based payment big data analysis system according to an embodiment of the application.
Fig. 2 is a block diagram of an internet finance-based payment big data analysis system according to an embodiment of the present application.
Fig. 3A is a flow chart of a training phase in an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the application.
Fig. 3B is a flowchart of an inference phase in an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the application.
Fig. 4 is a schematic diagram of an architecture of a training phase in an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the application.
Fig. 5 is a schematic diagram of an architecture of an inference stage in an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only a few embodiments of the present application, and not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.
Overview of a scene
As described above, with the rapid development of internet technology, the amount of data related to payment on internet financial platforms is rapidly increasing in exponential order, and accordingly, providers providing various internet financial services are emerging as if they were coming spring shoots after rain. As such, now and in the near future, the data accumulated in the internet financial network will be huge, so it is important how to mine and analyze the huge amount of financial payment data to improve the quality of financial services provided by each provider to the user.
In addition, the time spent on the rapid development of internet finance from the initial online shopping to the online bank and third-party payment to the online financing is only as short as a few years, but a series of credit risk problems of internet finance payment also occur. Therefore, in order to effectively reduce the credit risk of internet financial platform payment, a payment big data analysis method based on internet finance is desired.
Therefore, in the solution of the present application, a user portrait, which is represented as a series of (labels, weights) key-value pairs, is first obtained and passed through a context-based semantic understanding model including an embedding layer, such as a Bert model, where weights are used to weight label embedding vectors obtained after the embedding layer to obtain a label feature vector corresponding to each label, so that the user portrait is converted into a plurality of label feature vectors.
Also, the description of the transaction to be paid is used with a context-based semantic understanding model to obtain a plurality of semantic feature vectors.
In order to further extract the associated features for matching, the plurality of label feature vectors are subjected to two-dimensional splicing to obtain a first feature matrix and input to a convolutional neural network to obtain a first feature map, and meanwhile, the plurality of semantic feature vectors are subjected to two-dimensional splicing to obtain a second feature matrix and input to the convolutional neural network to obtain a second feature map. Here, in order to make the extracted associated features have as much correlation as possible with respect to the user tag features and the business semantic features in the high-dimensional feature space, the same convolutional neural network is used for feature extraction, that is, the convolutional neural network can embody the associated mapping of the user tag features and the business semantic features.
However, since the same convolutional neural network is used, the parameter update of the convolutional neural network depends on the parameter distribution of the first feature matrix and the second feature matrix at the same time, and compared with the second feature matrix which is a business semantic feature, the parameter distribution of the first feature matrix based on the mature user imaging technology is obviously more standard, so the application further uses the gradient transfer method to update the parameter of the convolutional neural network.
Specifically, first, a weighted sum of the first feature map and the second feature map is calculated to obtain a classification feature map, and then a classification loss function value L is obtained from the classification feature map (F1+ F2), so that the parameter update of the convolutional neural network is generally expressed as:
however, in the present application, the gradient is first passed to the classification signature graph, and the branches of the first signature matrix are then used to pass the gradient back to the parameters of the convolutional neural network, expressed as:
as shown in fig. 1, wherein the solid line represents forward direction transfer and the dashed line represents reverse direction transfer.
Based on this, the application provides an internet finance-based payment big data analysis system, which comprises a training module and an inference module. Wherein, the training module includes: a first training data acquisition unit for acquiring a user profile, the user profile being a series of (labels, weights) key-value pairs; training a user data encoding unit for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model comprising an embedding layer to obtain a plurality of label feature vectors; the second training data acquisition unit is used for acquiring the description of the service to be paid; the training service data coding unit is used for performing word segmentation processing on the description of the service to be paid and then obtaining a plurality of semantic feature vectors through the context-based semantic understanding model containing the embedded layer; the first neural network unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through a convolutional neural network to obtain a first feature map; the second neural network unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network to obtain a second feature map; a feature fusion unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map; the loss function value calculation unit is used for enabling the classification characteristic map to pass through a classifier so as to obtain a classification loss function value; a training unit, configured to train the convolutional neural network based on the classification loss function value, where in each iteration of training the convolutional neural network, gradient information of the classification loss function value relative to the classification feature map is first transferred to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transferred back to a parameter of the convolutional neural network using a branch of the first feature matrix. Wherein, the inference module comprises: a to-be-analyzed user data unit for acquiring a user representation, wherein the user representation is a series of (labels, weights) key value pairs; a user portrait semantic encoding unit for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model containing an embedding layer to obtain a plurality of label feature vectors; the payment service data unit is used for acquiring the description of the payment service; the semantic coding unit of the service to be paid is used for performing word segmentation processing on the description of the service to be paid and then obtaining a plurality of semantic feature vectors through the context-based semantic understanding model comprising the embedded layer; the first associated feature coding unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through the convolutional neural network trained by the training module to obtain a first feature map; the second associated feature coding unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network trained by the training module to obtain a second feature map; a classification feature map generation unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map; and the result output unit is used for enabling the classification characteristic graph to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the user is adaptive to the service to be paid or not.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary System
FIG. 2 illustrates a block diagram of an Internet finance-based payment big data analysis system according to an embodiment of the application. As shown in fig. 2, the internet finance-based payment big data analysis system 200 according to the embodiment of the present application includes: a training module 210 and an inference module 220. Wherein, the training module 210 includes: a first training data acquisition unit 211, configured to acquire a user profile, where the user profile is a series of (labels, weights) key-value pairs; training a user data encoding unit 212 for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model comprising an embedding layer to obtain a plurality of label feature vectors; a second training data obtaining unit 213, configured to obtain a description of a service to be paid; a training service data encoding unit 214, configured to perform word segmentation on the description of the service to be paid, and then obtain a plurality of semantic feature vectors through the context-based semantic understanding model including the embedding layer; a first neural network unit 215, configured to perform two-dimensional stitching on the plurality of label feature vectors to obtain a first feature matrix, and then pass the first feature matrix through a convolutional neural network to obtain a first feature map; a second neural network unit 216, configured to perform two-dimensional stitching on the plurality of semantic feature vectors to obtain a second feature matrix, and then pass the second feature matrix through the convolutional neural network to obtain a second feature map; a feature fusion unit 217, configured to fuse the first feature map and the second feature map to obtain a classification feature map; a loss function value calculation unit 218, configured to pass the classification feature map through a classifier to obtain a classification loss function value; a training unit 219, configured to train the convolutional neural network based on the classification loss function value, where in each iteration of training the convolutional neural network, gradient information of the classification loss function value relative to the classification feature map is first transferred to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transferred back to a parameter of the convolutional neural network by using a branch of the first feature matrix. The inference module 220 includes: a to-be-analyzed user data unit 221, configured to obtain a user representation, where the user representation is a series of (labels, weights) key-value pairs; a user portrait semantic encoding unit 222, configured to pass the series of (label, weight) key-value pairs through a context-based semantic understanding model including an embedding layer to obtain a plurality of label feature vectors; a to-be-paid service data unit 223, configured to obtain a description of a to-be-paid service; a to-be-paid service semantic coding unit 224, configured to perform word segmentation on the description of the to-be-paid service, and then obtain a plurality of semantic feature vectors through the context-based semantic understanding model including the embedded layer; a first associated feature encoding unit 225, configured to perform two-dimensional stitching on the plurality of label feature vectors to obtain a first feature matrix, and then pass the first feature matrix through the convolutional neural network trained by the training module to obtain a first feature map; a second associated feature encoding unit 226, configured to perform two-dimensional splicing on the multiple semantic feature vectors to obtain a second feature matrix, and then pass the second feature matrix through the convolutional neural network trained by the training module to obtain a second feature map; a classification feature map generation unit 227, configured to fuse the first feature map and the second feature map to obtain a classification feature map; and a result output unit 228, configured to pass the classification feature map through a classifier to obtain a classification result, where the classification result is used to indicate whether the user and the service to be paid are adapted.
Specifically, in the embodiment of the present application, in the training module 210, the first training data obtaining unit 211 and the training user data encoding unit 212 are configured to obtain a user portrait, where the user portrait is a series of (label, weight) key-value pairs, and pass the series of (label, weight) key-value pairs through a context-based semantic understanding model including an embedding layer to obtain a plurality of label feature vectors. As previously mentioned, it should be appreciated that some services to be paid for are less expensive or less demanding to purchase, the payment capacity for the user may be less standardized, while the credit rating for the user may be more demanding. Some services to be paid have higher price or higher purchase requirements, for example, requirements for academic calendar, social contribution and the like, so that the payment capacity of the user needs to be higher, and the credit degree of the user can be lowered in order to make the services to be paid reach a full-space state as much as possible. Therefore, in the technical solution of the present application, it is desirable to determine whether the user and the service to be paid are adapted based on the representation of the user, the matching degree between a series of (labels, weights) key value pairs and the description of the service to be paid.
Specifically, in the technical solution of the present application, portrait data of a user, which is a series of (labels, weights) key value pairs, is first acquired from a cloud storage. The series of (label, weight) key-value pairs is then processed through a context-based semantic understanding model, such as a Bert model, that includes an embedding layer to extract semantic information for the series of (label, weight) key-value pairs, where the weights are used to weight label embedding vectors obtained after embedding the layer to obtain a label feature vector corresponding to each of the labels, thereby converting the user representation into a plurality of label feature vectors.
More specifically, in an embodiment of the present application, the training user data encoding unit includes: an embedding subunit, configured to convert, respectively, each tag in the user portrait into a tag embedding vector using an embedding layer of the semantic understanding model; a weighting subunit, configured to weight the tag embedding vector corresponding to each tag in the user representation using the weight corresponding to the tag to obtain a sequence of weighted tag embedding vectors; and a semantic coding subunit, configured to perform semantic coding on the sequence of the weighted tag embedding vectors using a Bert model of the semantic understanding model to obtain the plurality of tag feature vectors.
Specifically, in this embodiment of the application, in the training module 210, the second training data obtaining unit 213 and the training service data encoding unit 214 are configured to obtain a description of a service to be paid, perform word segmentation processing on the description of the service to be paid, and obtain a plurality of semantic feature vectors through the context-based semantic understanding model including the embedding layer. That is, in the technical solution of the present application, in order to accurately obtain the degree of adaptation between the user and the service to be paid, a description of the service to be paid also needs to be obtained. Then, the description of the service to be paid needs to be subjected to word segmentation processing to avoid semantic confusion. Then, the semantic information is processed in the context-based semantic understanding model containing the embedded layer to extract the semantic information of the service description to be paid, so that a plurality of semantic feature vectors are obtained.
More specifically, in the technical solution of the present application, the training service data encoding unit includes: the vector transformation module is used for transforming each word in the description of the service to be paid into an embedded vector by using the embedded layer of the semantic understanding model so as to obtain a sequence of the embedded vector; and a context encoding unit, configured to perform context-based semantic encoding on the sequence of embedded vectors using a Bert model of the semantic understanding model to obtain the plurality of semantic feature vectors.
Specifically, in this embodiment of the present application, in the training module 210, the first neural network unit 215 and the second neural network unit 216 are configured to perform two-dimensional stitching on the plurality of label feature vectors to obtain a first feature matrix, and then pass the first feature matrix through a convolutional neural network to obtain a first feature map, and perform two-dimensional stitching on the plurality of semantic feature vectors to obtain a second feature matrix, and then pass the second feature matrix through the convolutional neural network to obtain a second feature map. It should be understood that, in order to further extract the associated features for matching, in the technical solution of the present application, the plurality of label feature vectors are further two-dimensionally spliced to obtain a first feature matrix, and the first feature matrix is input into a convolutional neural network to be processed, so as to extract the associated implicit features of each position in the first feature matrix, thereby obtaining a first feature map. Meanwhile, the semantic feature vectors are subjected to two-dimensional splicing to obtain a second feature matrix, and the second feature matrix is input into a convolutional neural network to obtain a second feature map. In particular, it is worth mentioning that, in order to make the extracted associated features have as much correlation as possible in the high-dimensional feature space with respect to the user tag features and the business semantic features, in the technical solution of the present application, the same convolutional neural network is used for feature extraction, that is, the convolutional neural network can embody the associated mapping of the user tag features and the business semantic features.
Accordingly, in one specific example, each layer of the convolutional neural network performs convolution processing, pooling processing and activation processing on input data during forward pass of the layer to output the first feature map or the second feature map by the last layer of the convolutional neural network, wherein the input of the first layer of the convolutional neural network is the first feature matrix or the second feature matrix.
Specifically, in the embodiment of the present application, in the training module 210, the feature fusion unit 217 and the loss function value calculation unit 218 are configured to fuse the first feature map and the second feature map to obtain a classification feature map, and pass the classification feature map through a classifier to obtain a classification loss function value. It should be understood that, since the same convolutional neural network is used, the parameter updating of the convolutional neural network depends on the parameter distribution of the first feature matrix and the second feature matrix at the same time, and it is obvious that the parameter distribution of the first feature matrix based on the mature user imaging technology is more standard than the second feature matrix as the business semantic feature, in the technical solution of the present application, the parameter of the convolutional neural network is further updated using a gradient transfer method.
That is, specifically, in the technical solution of the present application, a weighted sum by position between the first feature map and the second feature map is first calculated to fuse the first feature map and the second feature map, thereby obtaining a classification feature map. The classification signature is then passed through a classifier to obtain a classification loss function value L (F1+ F2).
More specifically, in an embodiment of the present application, the loss function value calculating unit is further configured to: processing the classification feature map using the classifier with a formula of softmax { (W) to obtain a classification resultn,Bn):…:(W1,B1) L project (F), where project (F) represents the projection of the classification feature map as a vector, W1To WnAs a weight matrix for all connected layers of each layer, B1To BnA bias matrix representing the layers of the fully-connected layer; and calculating a cross entropy value between the classification result and the real value as the classification loss function value.
Specifically, in the embodiment of the present application, in the training module 210, the training unit 219 is configured to train the convolutional neural network based on the classification loss function values, wherein in each iteration of training the convolutional neural network, gradient information of the classification loss function values relative to the classification feature map is first transferred to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transferred back to parameters of the convolutional neural network by using branches of the first feature matrix. It will be appreciated that the parameter update for a convolutional neural network is typically expressed as:
however, in the technical solution of the present application, firstly, the gradient is transferred to the classification feature map, and then the branch of the first feature matrix is used to transfer the gradient back to the parameter of the convolutional neural network, which is expressed as:
as shown in fig. 1, wherein the solid line represents forward direction transfer and the dashed line represents reverse direction transfer.
After the training is completed, the classification feature map can be obtained according to the method in the inference module. And then, the classification characteristic graph is used for obtaining a classification result which is used for indicating whether the user is adaptive to the service to be paid or not through a classifier.
Specifically, in the embodiment of the present application, first, a user profile is obtained, where the user profile is a series of (labels, weights) key-value pairs. Then, the series of (label, weight) key-value pairs is passed through a context-based semantic understanding model including an embedding layer to obtain a plurality of label feature vectors. Then, a description of the service to be paid is obtained. And then, after word segmentation processing is carried out on the description of the service to be paid, a plurality of semantic feature vectors are obtained through the context-based semantic understanding model comprising the embedding layer. Then, after the label feature vectors are subjected to two-dimensional splicing to obtain a first feature matrix, the first feature matrix passes through the convolutional neural network trained by the training module to obtain a first feature map. And then, after the semantic feature vectors are subjected to two-dimensional splicing to obtain a second feature matrix, the second feature matrix passes through the convolutional neural network trained by the training module to obtain a second feature map. Then, the first feature map and the second feature map are fused to obtain a classification feature map. And finally, the classification characteristic graph is used for obtaining a classification result through a classifier, and the classification result is used for indicating whether the user is adaptive to the service to be paid or not.
In summary, the internet finance-based payment big data analysis system 200 according to the embodiment of the present application is clarified, which extracts semantic feature information from a user portrait and a description of a service to be paid through a context-based semantic understanding model, processes the obtained feature matrix using the same convolutional neural network model again, so that the extracted associated features have a correlation as much as possible in a high-dimensional feature space with respect to the user tag feature and the service semantic feature, and further updates parameters of the convolutional neural network using a gradient transfer method, so as to improve classification accuracy. Therefore, the user can be matched with the service to be paid more appropriately, and the credit risk of internet financial platform payment is effectively reduced.
As described above, the internet-finance-based payment big data analysis system 200 according to the embodiment of the present application may be implemented in various terminal devices, such as a server of an internet-finance-based payment big data analysis algorithm, and the like. In one example, the internet finance-based payment big data analysis system 200 according to the embodiment of the present application may be integrated into a terminal device as one software module and/or hardware module. For example, the internet finance-based payment big data analysis system 200 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the internet finance-based payment big data analysis system 200 may also be one of many hardware modules of the terminal device.
Alternatively, in another example, the internet financial-based payment big data analysis system 200 and the terminal device may be separate devices, and the internet financial-based payment big data analysis system 200 may be connected to the terminal device through a wired and/or wireless network and transmit the interaction information in an agreed data format.
Exemplary method
FIG. 3A illustrates a flow diagram of a training phase in an analysis method of an Internet finance-based payment big data analysis system according to an embodiment of the application. As shown in fig. 3A, an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the present application includes: a training phase comprising the steps of: s110, acquiring a user portrait, wherein the user portrait is a series of (labels, weights) key value pairs; s120, passing the series of (label, weight) key value pairs through a context-based semantic understanding model containing an embedding layer to obtain a plurality of label feature vectors; s130, obtaining the description of the service to be paid; s140, after the description of the service to be paid is subjected to word segmentation processing, a plurality of semantic feature vectors are obtained through the context-based semantic understanding model containing the embedded layer; s150, performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then passing the first feature matrix through a convolutional neural network to obtain a first feature map; s160, performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then passing the second feature matrix through the convolutional neural network to obtain a second feature map; s170, fusing the first feature map and the second feature map to obtain a classification feature map; s180, enabling the classification characteristic graph to pass through a classifier to obtain a classification loss function value; and S190, training the convolutional neural network based on the classification loss function values, wherein in each iteration of training the convolutional neural network, gradient information of the classification loss function values relative to the classification feature map is firstly transmitted to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transmitted back to parameters of the convolutional neural network by using branches of the first feature matrix.
FIG. 3B illustrates a flow diagram of an inference phase in an analytics method of an Internet finance-based payment big data analytics system, according to an embodiment of the present application. Fig. 3B illustrates an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the present application, including: an inference phase comprising the steps of: s210, acquiring a user portrait, wherein the user portrait is a series of (labels and weights) key value pairs; s220, passing the series of (label, weight) key-value pairs through a context-based semantic understanding model comprising an embedding layer to obtain a plurality of label feature vectors; s230, obtaining the description of the service to be paid; s240, after the description of the service to be paid is subjected to word segmentation processing, a plurality of semantic feature vectors are obtained through the context-based semantic understanding model containing the embedded layer; s250, performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then passing the first feature matrix through the convolutional neural network trained by the training module to obtain a first feature map; s260, performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network trained by the training module to obtain a second feature map; s270, fusing the first feature map and the second feature map to obtain a classification feature map; and S280, the classification characteristic graph is processed by a classifier to obtain a classification result, and the classification result is used for indicating whether the user is adaptive to the service to be paid or not.
Fig. 4 illustrates an architecture diagram of a training phase in an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the present application. As shown in fig. 4, in the training phase, first, the obtained series of (label, weight) key-value pairs (e.g., P1 as illustrated in fig. 4) are passed through a context-based semantic understanding model (e.g., SUM as illustrated in fig. 4) containing an embedding layer to obtain a plurality of label feature vectors (e.g., VF1 as illustrated in fig. 4) in the network architecture; then, performing word segmentation processing on the obtained description of the service to be paid (e.g., P2 as illustrated in fig. 4) to obtain a plurality of semantic feature vectors (e.g., VF2 as illustrated in fig. 4) through the context-based semantic understanding model (e.g., SUM as illustrated in fig. 4) including an embedding layer; then, after two-dimensionally stitching the plurality of tagged feature vectors to obtain a first feature matrix (e.g., MF1 as illustrated in fig. 4), passing the first feature matrix through a convolutional neural network (e.g., CNN as illustrated in fig. 4) to obtain a first feature map (e.g., F1 as illustrated in fig. 4); then, after two-dimensionally stitching the plurality of semantic feature vectors to obtain a second feature matrix (e.g., MF2 as illustrated in fig. 4), passing the second feature matrix through the convolutional neural network to obtain a second feature map (e.g., F2 as illustrated in fig. 4); then, fusing the first feature map and the second feature map to obtain a classification feature map (e.g., F as illustrated in fig. 4); then, passing the classification signature through a classifier (e.g., a classifier as illustrated in fig. 4) to obtain a classification loss function value (e.g., a CLV as illustrated in fig. 4); finally, the convolutional neural network is trained based on the classification loss function values, wherein in each iteration of training the convolutional neural network, gradient information of the classification loss function values relative to the classification feature map is firstly transmitted to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transmitted back to parameters of the convolutional neural network by using branches of the first feature matrix.
Fig. 5 illustrates an architecture diagram of an inference stage in an analysis method of an internet finance-based payment big data analysis system according to an embodiment of the application. As shown in fig. 5, in the inference phase, first, the obtained series of (label, weight) key-value pairs (e.g., P1 as illustrated in fig. 5) are passed through a context-based semantic understanding model (e.g., SUM as illustrated in fig. 5) containing an embedding layer to obtain a plurality of label feature vectors (e.g., VF1 as illustrated in fig. 5) in the network architecture; then, performing word segmentation processing on the obtained description of the service to be paid (e.g., P2 as illustrated in fig. 5) to obtain a plurality of semantic feature vectors (e.g., VF2 as illustrated in fig. 5) through the context-based semantic understanding model (e.g., SUM as illustrated in fig. 5) including an embedding layer; then, after two-dimensionally splicing the plurality of tagged feature vectors to obtain a first feature matrix (e.g., MF1 as illustrated in fig. 5), passing the first feature matrix through the convolutional neural network (e.g., CN as illustrated in fig. 5) completed by training of the training module to obtain a first feature map (e.g., F1 as illustrated in fig. 5); then, after the plurality of semantic feature vectors are two-dimensionally spliced to obtain a second feature matrix (e.g., MF2 as illustrated in fig. 5), passing the second feature matrix through the convolutional neural network completed by training of the training module to obtain a second feature map (e.g., F2 as illustrated in fig. 5); then, fusing the first feature map and the second feature map to obtain a classification feature map (e.g., F as illustrated in fig. 5); and finally, passing the classification feature map through a classifier (e.g., a classifier as illustrated in fig. 5) to obtain a classification result, wherein the classification result is used for indicating whether the user is adapted to the service to be paid.
In summary, the analysis method of the internet finance-based payment big data analysis system based on the embodiment of the application is clarified, the semantic feature information is extracted through the context-based semantic understanding model respectively for the description of the user portrait and the service to be paid, the obtained feature matrix is processed by using the same convolutional neural network model, so that the extracted associated features have the correlation as much as possible in the high-dimensional feature space relative to the user label features and the service semantic features, and the parameters of the convolutional neural network are further updated by using the gradient transfer method, so as to improve the classification accuracy. Therefore, the user can be matched with the service to be paid more appropriately, and the credit risk of internet financial platform payment is effectively reduced.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, devices, systems referred to in this application are only used as illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. The words "or" and "as used herein mean, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations should be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. An internet finance-based payment big data analysis system, comprising:
a training module comprising:
a first training data acquisition unit for acquiring a user profile, the user profile being a series of (labels, weights) key-value pairs;
training a user data encoding unit for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model comprising an embedding layer to obtain a plurality of label feature vectors;
the second training data acquisition unit is used for acquiring the description of the service to be paid;
the training service data coding unit is used for performing word segmentation processing on the description of the service to be paid and then obtaining a plurality of semantic feature vectors through the context-based semantic understanding model containing the embedded layer;
the first neural network unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through a convolutional neural network to obtain a first feature map;
the second neural network unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network to obtain a second feature map;
a feature fusion unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map;
the loss function value calculation unit is used for enabling the classification characteristic map to pass through a classifier so as to obtain a classification loss function value;
a training unit, configured to train the convolutional neural network based on the classification loss function values, wherein in each iteration of training the convolutional neural network, gradient information of the classification loss function values relative to the classification feature map is first transferred to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transferred back to parameters of the convolutional neural network by using branches of the first feature matrix; and
an inference module comprising:
a to-be-analyzed user data unit for acquiring a user representation, wherein the user representation is a series of (labels, weights) key value pairs;
a user portrait semantic encoding unit for passing the series of (label, weight) key-value pairs through a context-based semantic understanding model containing an embedding layer to obtain a plurality of label feature vectors;
the payment service data unit is used for acquiring the description of the payment service;
the semantic coding unit of the service to be paid is used for performing word segmentation processing on the description of the service to be paid and then obtaining a plurality of semantic feature vectors through the context-based semantic understanding model comprising the embedded layer;
the first associated feature coding unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through the convolutional neural network trained by the training module to obtain a first feature map;
the second associated feature coding unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network trained by the training module to obtain a second feature map;
a classification feature map generation unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map; and
and the result output unit is used for enabling the classification characteristic graph to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the user is adaptive to the service to be paid or not.
2. The internet financial based payment big data analysis system of claim 1, wherein the training user data encoding unit comprises:
an embedding subunit, configured to convert, respectively, each tag in the user portrait into a tag embedding vector using an embedding layer of the semantic understanding model;
a weighting subunit, configured to weight the tag embedding vector corresponding to each tag in the user representation using the weight corresponding to the tag to obtain a sequence of weighted tag embedding vectors; and
a semantic coding subunit, configured to perform semantic coding on the sequence of the weighted tag embedding vectors using a Bert model of the semantic understanding model to obtain the plurality of tag feature vectors.
3. The internet financial-based payment big data analysis system of claim 2, wherein the training service data encoding unit comprises:
the vector transformation module is used for transforming each word in the description of the service to be paid into an embedded vector by using the embedding layer of the semantic understanding model so as to obtain a sequence of the embedded vector; and
a context encoding unit, configured to perform context-based semantic encoding on the sequence of embedded vectors using a Bert model of the semantic understanding model to obtain the plurality of semantic feature vectors.
4. The internet financial-based payment big data analysis system of claim 3, wherein each layer of the convolutional neural network performs convolutional processing, pooling processing, and activation processing on input data during forward pass of a layer to output the first feature map or the second feature map by a last layer of the convolutional neural network, wherein an input of a first layer of the convolutional neural network is the first feature matrix or the second feature matrix.
5. The internet financial-based payment big data analysis system of claim 4, wherein the feature fusion unit is further configured to compute a weighted sum by location between the first feature map and the second feature map to obtain the classification feature map.
6. The internet financial-based payment big data analysis system of claim 5, wherein the loss function values calculation unit is further to process the classification feature map using the classifier to obtain a classification result with a formula, wherein the formula is softmax { (W {)n,Bn):…:(W1,B1) L project (F), where project (F) represents the projection of the classification feature map as a vector, W1To WnAs a weight matrix for each fully connected layer, B1To BnA bias matrix representing all layers of the fully connected layer; and calculating a cross entropy value between the classification result and the real value as the classification loss function value.
7. The internet financial based payment big data analysis system of claim 6, wherein the training unit is further configured to perform a transfer of gradient information in each iteration according to the following formula;
wherein the formula is:
8. an analysis method of a payment big data analysis system based on internet finance is characterized by comprising the following steps:
a training phase comprising:
obtaining a user representation, wherein the user representation is a series of (labels, weights) key value pairs;
passing the series of (label, weight) key-value pairs through a context-based semantic understanding model that includes an embedding layer to obtain a plurality of label feature vectors;
obtaining a description of a service to be paid;
after the description of the service to be paid is subjected to word segmentation processing, a plurality of semantic feature vectors are obtained through the context-based semantic understanding model comprising the embedded layer;
performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then passing the first feature matrix through a convolutional neural network to obtain a first feature map;
after the semantic feature vectors are subjected to two-dimensional splicing to obtain a second feature matrix, the second feature matrix passes through the convolutional neural network to obtain a second feature map;
fusing the first feature map and the second feature map to obtain a classification feature map;
passing the classification feature map through a classifier to obtain a classification loss function value;
training the convolutional neural network based on the classification loss function values, wherein in each iteration of training the convolutional neural network, gradient information of the classification loss function values relative to the classification feature map is firstly transmitted to the classification feature map, and then gradient information of the classification feature map relative to the first feature matrix is transmitted back to parameters of the convolutional neural network by using branches of the first feature matrix; and
an inference phase comprising:
obtaining a user representation, wherein the user representation is a series of (labels, weights) key value pairs;
passing the series of (label, weight) key-value pairs through a context-based semantic understanding model that includes an embedding layer to obtain a plurality of label feature vectors;
obtaining a description of a service to be paid;
after the description of the service to be paid is subjected to word segmentation processing, a plurality of semantic feature vectors are obtained through the context-based semantic understanding model comprising the embedded layer;
the first associated feature coding unit is used for performing two-dimensional splicing on the plurality of label feature vectors to obtain a first feature matrix, and then enabling the first feature matrix to pass through the convolutional neural network trained by the training module to obtain a first feature map;
the second associated feature coding unit is used for performing two-dimensional splicing on the semantic feature vectors to obtain a second feature matrix, and then enabling the second feature matrix to pass through the convolutional neural network trained by the training module to obtain a second feature map;
a classification feature map generation unit, configured to fuse the first feature map and the second feature map to obtain a classification feature map; and
and the result output unit is used for enabling the classification characteristic graph to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the user is adaptive to the service to be paid or not.
9. The analysis method of the internet financial based payment big data analysis system of claim 8, wherein passing the series of (label, weight) key-value pairs through a context-based semantic understanding model containing an embedding layer to obtain a plurality of label feature vectors comprises:
respectively converting each label in the user portrait into a label embedding vector by using an embedding layer of the semantic understanding model;
weighting the tag embedding vector corresponding to each tag in the user representation using the weight corresponding to the tag to obtain a sequence of weighted tag embedding vectors; and
semantically encoding the sequence of weighted tag-embedded vectors using a Bert model of the semantic understanding model to obtain the plurality of tag feature vectors.
10. The analysis method of the internet finance-based payment big data analysis system according to claim 8, wherein the obtaining of the plurality of semantic feature vectors through the context-based semantic understanding model including the embedding layer after the word segmentation processing is performed on the description of the service to be paid comprises:
converting each word in the description of the service to be paid into an embedded vector by using an embedded layer of the semantic understanding model to obtain a sequence of embedded vectors; and
a context encoding unit, configured to perform context-based semantic encoding on the sequence of embedded vectors using a Bert model of the semantic understanding model to obtain the plurality of semantic feature vectors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210190418.8A CN114565041A (en) | 2022-02-28 | 2022-02-28 | Payment big data analysis system based on internet finance and analysis method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210190418.8A CN114565041A (en) | 2022-02-28 | 2022-02-28 | Payment big data analysis system based on internet finance and analysis method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114565041A true CN114565041A (en) | 2022-05-31 |
Family
ID=81715621
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210190418.8A Withdrawn CN114565041A (en) | 2022-02-28 | 2022-02-28 | Payment big data analysis system based on internet finance and analysis method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114565041A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114758424A (en) * | 2022-06-14 | 2022-07-15 | 众旅联(浙江)生态科技有限公司 | Intelligent payment equipment based on multiple verification mechanisms and payment method thereof |
| CN115146488A (en) * | 2022-09-05 | 2022-10-04 | 山东鼹鼠人才知果数据科技有限公司 | Variable business process intelligent modeling system and method based on big data |
| CN115203380A (en) * | 2022-09-19 | 2022-10-18 | 山东鼹鼠人才知果数据科技有限公司 | Text processing system and method based on multi-mode data fusion |
| CN115547466A (en) * | 2022-09-05 | 2022-12-30 | 西安启玥华辰软件咨询开发有限公司 | Medical institution registration and review system and method based on big data |
| CN115602165A (en) * | 2022-09-07 | 2023-01-13 | 杭州优航信息技术有限公司（Cn） | Digital staff intelligent system based on financial system |
| CN115693918A (en) * | 2022-09-07 | 2023-02-03 | 浙江心友机电设备安装有限公司 | Comprehensive intelligent power utilization system and method for building |
| CN115880036A (en) * | 2023-02-23 | 2023-03-31 | 山东金潮交通设施有限公司 | Parking stall level dynamic sharing intelligence management and control transaction platform |
| CN116167833A (en) * | 2023-04-19 | 2023-05-26 | 无锡数信互融科技发展有限公司 | Internet financial risk control system and method based on federal learning |
| CN117114760A (en) * | 2023-05-04 | 2023-11-24 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
| CN117114760B (en) * | 2023-05-04 | 2024-05-28 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
-
2022
- 2022-02-28 CN CN202210190418.8A patent/CN114565041A/en not_active Withdrawn
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114758424A (en) * | 2022-06-14 | 2022-07-15 | 众旅联(浙江)生态科技有限公司 | Intelligent payment equipment based on multiple verification mechanisms and payment method thereof |
| CN115146488A (en) * | 2022-09-05 | 2022-10-04 | 山东鼹鼠人才知果数据科技有限公司 | Variable business process intelligent modeling system and method based on big data |
| CN115547466B (en) * | 2022-09-05 | 2023-06-16 | 西安启玥华辰软件咨询开发有限公司 | Medical institution registration and review system and method based on big data |
| CN115146488B (en) * | 2022-09-05 | 2022-11-22 | 山东鼹鼠人才知果数据科技有限公司 | Variable business process intelligent modeling system and method based on big data |
| CN115547466A (en) * | 2022-09-05 | 2022-12-30 | 西安启玥华辰软件咨询开发有限公司 | Medical institution registration and review system and method based on big data |
| CN115693918A (en) * | 2022-09-07 | 2023-02-03 | 浙江心友机电设备安装有限公司 | Comprehensive intelligent power utilization system and method for building |
| CN115602165A (en) * | 2022-09-07 | 2023-01-13 | 杭州优航信息技术有限公司（Cn） | Digital staff intelligent system based on financial system |
| CN115693918B (en) * | 2022-09-07 | 2023-08-18 | 浙江心友机电设备安装有限公司 | Comprehensive intelligent electricity utilization system and method for building |
| CN115203380A (en) * | 2022-09-19 | 2022-10-18 | 山东鼹鼠人才知果数据科技有限公司 | Text processing system and method based on multi-mode data fusion |
| CN115880036A (en) * | 2023-02-23 | 2023-03-31 | 山东金潮交通设施有限公司 | Parking stall level dynamic sharing intelligence management and control transaction platform |
| CN115880036B (en) * | 2023-02-23 | 2023-06-06 | 山东金潮交通设施有限公司 | Parking space level dynamic sharing intelligent management and control transaction platform |
| CN116167833A (en) * | 2023-04-19 | 2023-05-26 | 无锡数信互融科技发展有限公司 | Internet financial risk control system and method based on federal learning |
| CN116167833B (en) * | 2023-04-19 | 2023-06-30 | 无锡数信互融科技发展有限公司 | Internet financial risk control system and method based on federal learning |
| CN117114760A (en) * | 2023-05-04 | 2023-11-24 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
| CN117114760B (en) * | 2023-05-04 | 2024-05-28 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114565041A (en) | Payment big data analysis system based on internet finance and analysis method thereof | |
| CN111291183B (en) | Method and device for carrying out classification prediction by using text classification model | |
| CN112507995B (en) | Cross-model face feature vector conversion system and method | |
| CN112733768B (en) | Natural scene text recognition method and device based on bidirectional characteristic language model | |
| WO2010068840A1 (en) | Machine optimization devices, methods, and systems | |
| CN111382868A (en) | Neural network structure search method and neural network structure search device | |
| CN111967565B (en) | Neural network system, method and device for risk assessment | |
| CN115512420A (en) | Visitor management system and method thereof | |
| CN114861842B (en) | Few-sample target detection method and device and electronic equipment | |
| CN113628059A (en) | Associated user identification method and device based on multilayer graph attention network | |
| CN116523583A (en) | Electronic commerce data analysis system and method thereof | |
| CN116977001A (en) | Geological disaster prevention and treatment engineering cost management system and method thereof | |
| EP4339831A1 (en) | Image data processing method and apparatus, and device and storage medium | |
| CN114821169A (en) | Method-level non-intrusive call link tracking method under micro-service architecture | |
| CN114529085A (en) | Resident income prediction system based on big data and prediction method thereof | |
| CN114445121A (en) | Advertisement click rate prediction model construction and advertisement click rate prediction method | |
| CN116468885A (en) | Method and device for training semantic segmentation model and image detection method | |
| CN114299304A (en) | Image processing method and related equipment | |
| CN116167833B (en) | Internet financial risk control system and method based on federal learning | |
| CN117036006A (en) | User portrait generation method and device, storage medium and electronic equipment | |
| CN116975347A (en) | Image generation model training method and related device | |
| CN116340635A (en) | Article recommendation method, model training method, device and equipment | |
| CN110717037A (en) | Method and device for classifying users | |
| CN112559640A (en) | Training method and device of atlas characterization system | |
| CN112184431A (en) | User risk determination method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20220531 |
|
| WW01 | Invention patent application withdrawn after publication |