CN116187581A - Financial institution traffic prediction method and device, processor and electronic equipment - Google Patents
Financial institution traffic prediction method and device, processor and electronic equipment Download PDFInfo
- Publication number
- CN116187581A CN116187581A CN202310272900.0A CN202310272900A CN116187581A CN 116187581 A CN116187581 A CN 116187581A CN 202310272900 A CN202310272900 A CN 202310272900A CN 116187581 A CN116187581 A CN 116187581A
- Authority
- CN
- China
- Prior art keywords
- target
- feature vector
- historical data
- data information
- predicted
- 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.)
- Pending
Links
Images
Classifications
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Abstract
The application discloses a financial institution's business volume prediction method and device, a processor and electronic equipment, relates to artificial intelligence technical field, and the method includes: acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: processing historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; and predicting the traffic corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result. According to the method and the device, the problem that in the related art, the service volume of a financial institution is predicted according to single variable information, so that the accuracy of the service volume prediction is low is solved.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a method and apparatus for predicting a financial institution's traffic, a processor, and an electronic device.
Background
The time sequence predictive analysis is a quantitative analysis method, which is to build a predictive model by using a certain mathematical method on the basis of time sequence variable analysis, so that the time trend extends outwards, thereby predicting the development change trend of the future market and determining a predicted value. The traditional time sequence prediction method obtains variable values to be predicted through a series of time sequence analysis and operation, ignores dynamic correlation among time sequence data, but has the problems that different variables have different influences on a prediction result and local dependency relations among different variables exist for a multi-variable time sequence, so that the dependency relations and the time dynamic correlation among multi-variable factors cannot be accurately captured by the traditional time sequence prediction method, and the problem is that the accuracy is low because the traditional time sequence prediction method is adopted for predicting the traffic of a financial institution.
Aiming at the problem that the accuracy of the traffic prediction is lower because the traffic of a financial institution is predicted according to single variable information in the related art, no effective solution is proposed at present.
Disclosure of Invention
The main purpose of the present application is to provide a method and apparatus for predicting the traffic of a financial institution, a processor and an electronic device, so as to solve the problem that in the related art, the accuracy of traffic prediction is low due to the prediction of the traffic of the financial institution according to single variable information.
In order to achieve the above object, according to one aspect of the present application, there is provided a traffic prediction method for a financial institution. The method comprises the following steps: acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result.
Further, the target neural network model at least includes: the convolution layer, the multi-head self-attention layer and the two-way long short-time memory layer process the historical data information through a target neural network model, and the obtaining of the target feature vector corresponding to the historical data information comprises the following steps: extracting features of the historical data information through the convolution layer to obtain a first feature vector matrix; calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and extracting features of the second feature vector matrix through the two-way long and short-time memory layer to obtain the target feature vector.
Further, the target neural network model further includes a full connection layer, and the predicting, according to the target feature vector, the traffic of the target financial institution corresponding to the date to be predicted, and obtaining a prediction result includes: fusing the target feature vectors through the full connection layer to obtain fused target feature values; and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the fused target characteristic value to obtain the prediction result.
Further, extracting features of the historical data information to obtain an initial feature vector matrix; normalizing the initial feature vector matrix to obtain a normalized initial feature vector matrix; and carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain the first eigenvector matrix.
Further, calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix; and obtaining the second eigenvector matrix with weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
Further, the second eigenvector matrix is processed through the two-way long and short-time memory layer in a forward propagation calculation mode, so that a first hidden eigenvector is obtained; processing the second eigenvector matrix by the bidirectional long short-time memory layer in a back propagation calculation mode to obtain a second hidden eigenvector; and performing splicing processing on the first hidden feature vector and the second hidden feature vector to obtain the target feature vector.
Further, acquiring the actual traffic corresponding to the date to be predicted; calculating the difference value between the actual traffic and the predicted traffic in the predicted result to obtain target difference value data; and optimizing the target neural network model according to the target difference data.
Further, a training sample set is obtained, wherein the training sample set comprises a plurality of training sample data sets, the statistical time period corresponding to each training sample data set is different, and each training sample data set at least comprises: a plurality of dates and a plurality of second factor value amounts corresponding to each date; and training the initial neural network model according to the training sample set to obtain the target neural network model and the preset time period.
In order to achieve the above object, according to another aspect of the present application, there is provided a traffic predicting apparatus for a financial institution. The device comprises: the first obtaining unit is used for obtaining historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; the processing unit is used for processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; and the prediction unit is used for predicting the traffic corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result.
Further, the target neural network model at least includes: a convolutional layer, a multi-headed self-attention layer, and a two-way long short term memory layer, the processing unit comprising: the first extraction module is used for extracting the characteristics of the historical data information through the convolution layer to obtain a first characteristic vector matrix; the first operation module is used for operating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and the second extraction module is used for extracting the characteristics of the second characteristic vector matrix through the two-way long and short-time memory layer to obtain the target characteristic vector.
Further, the prediction unit includes: the fusion module is used for fusing the target feature vectors through the full connection layer to obtain fused target feature values; and the prediction module is used for predicting the business volume corresponding to the date to be predicted of the target financial institution according to the fused target characteristic value to obtain the prediction result.
Further, the first extraction module includes: the first extraction submodule is used for carrying out feature extraction on the historical data information to obtain an initial feature vector matrix; the processing sub-module is used for carrying out normalization processing on the initial feature vector matrix to obtain a normalized initial feature vector matrix; and the first operation submodule is used for carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain the first eigenvector matrix.
Further, the first operation module includes: the second operation sub-module is used for calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix; and the third operation sub-module is used for obtaining the second eigenvector matrix with the weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
Further, the second extraction module includes: the second extraction submodule is used for processing the second eigenvector matrix through the bidirectional long and short-term memory layer in a forward propagation calculation mode to obtain a first hidden eigenvector; the third extraction submodule is used for processing the second eigenvector matrix through the bidirectional long and short-time memory layer in a back propagation calculation mode to obtain a second hidden eigenvector; and the splicing module is used for carrying out splicing processing on the first hidden characteristic vector and the second hidden characteristic vector to obtain the target characteristic vector.
Further, the apparatus further comprises: and the second obtaining unit is used for obtaining the actual traffic corresponding to the date to be predicted after predicting the traffic corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result.
Further, the second acquisition unit further includes: the second operation module is used for calculating the difference value between the actual traffic volume and the predicted traffic volume in the predicted result to obtain target difference value data; and the optimization module is used for optimizing the target neural network model according to the target difference value data.
Further, the apparatus further comprises: the third obtaining unit is configured to obtain a training sample set before the historical data information is processed through the target neural network model to obtain a target feature vector corresponding to the historical data information, where the training sample set includes a plurality of training sample data sets, a statistical time period corresponding to each training sample data set is different, and each training sample data set at least includes: a plurality of dates and a plurality of second factor values corresponding to each date; the training unit is used for training the initial neural network model according to the training sample set to obtain the target neural network model and the preset time period.
In order to achieve the above object, according to another aspect of the present application, there is further provided a processor, where the processor processes a program, and when the program runs, controls a device where the processor is located to execute the traffic prediction method of any one of the above financial institutions.
To achieve the above object, according to one aspect of the present application, there is provided an electronic device including one or more processors and a memory for storing a traffic prediction method of one or more processors implementing any one of the above financial institutions.
Through the application, the following steps are adopted: acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; the business volume corresponding to the date to be predicted of the target financial institution is predicted according to the target feature vector to obtain a prediction result, and the problem that the business volume prediction accuracy is low due to the fact that the business volume of the financial institution is predicted according to single variable information in the related technology is solved. According to the scheme, the historical data information of the multi-factor value is subjected to feature extraction through the target neural network model to obtain the target feature vector with the context feature information, and the historical data information can be aggregated in space and feature dimensions through the context feature information, so that the effect of improving the service prediction accuracy is achieved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application, illustrate and explain the application and are not to be construed as limiting the application. In the drawings:
FIG. 1 is a flow chart of a method of traffic prediction for a financial institution provided in accordance with an embodiment of the present application;
FIG. 2 is a schematic diagram of a traffic prediction model for a financial institution provided in accordance with an embodiment of the present application;
FIG. 3 is a schematic diagram of a self-attention layer computation process provided in accordance with an embodiment of the present application;
FIG. 4 is a schematic diagram of a self-attention layer model provided in accordance with an embodiment of the present application;
FIG. 5 is a schematic diagram of a method of traffic prediction for a financial institution provided in accordance with an embodiment of the present application;
FIG. 6 is a schematic diagram of a financial institution's traffic prediction device provided in accordance with an embodiment of the present application;
fig. 7 is a schematic diagram of an electronic device provided according to an embodiment of the present application.
Detailed Description
It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
In order to make the present application solution better understood by those skilled in the art, the following description will be made in detail and with reference to the accompanying drawings in the embodiments of the present application, it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, shall fall within the scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of the present application 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 in order to describe the embodiments of the present application 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.
It should be noted that, related information (including, but not limited to, user equipment information, user personal information, etc.) and data (including, but not limited to, data for presentation, analyzed data, etc.) related to the present disclosure are information and data authorized by a user or sufficiently authorized by each party. For example, an interface is provided between the system and the relevant user or institution, before acquiring the relevant information, the system needs to send an acquisition request to the user or institution through the interface, and acquire the relevant information after receiving the consent information fed back by the user or institution.
The invention will be described with reference to preferred embodiments, and FIG. 1 is a flow chart of a method for predicting the traffic of a financial institution according to an embodiment of the present application, as shown in FIG. 1, comprising the steps of:
step S101, acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: and presetting a plurality of first factor values corresponding to each date in a time period before the date to be predicted.
For example, historical data information before the date to be predicted is acquired, for example, the weather condition of one year before the date to be predicted, whether or not it is a workday, the number of self-service terminals of the target financial institution, the number of employees of the target financial institution, the number of registered customers of the target financial institution, the number of transacted business windows of the target financial institution, the historical business volume of the target financial institution, and the total number of financial institutions within a preset range.
Step S102, historical data information is processed through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information.
For example, the obtained historical data information is input into a target neural network model, and the target neural network model is utilized to perform feature extraction on the historical data information, so as to obtain the target feature vector.
It should be noted that, the target feature vector includes context feature information between the historical data information, and the historical data information can be aggregated in space and feature dimensions through the context feature information, so as to achieve the effect of improving the accuracy of service prediction.
And step S103, predicting the business volume corresponding to the date to be predicted of the target financial institution according to the target feature vector to obtain a prediction result.
For example, the target feature vector is fused and predicted through the target neural network model, so that the business volume corresponding to the date to be predicted of the financial institution is obtained.
In summary, in the scheme, the historical data information of the multi-factor value is subjected to feature extraction through the target neural network model to obtain the target feature vector with the context feature information, the historical data information can be aggregated in space and feature dimensions through the context feature information, the problem that in the related art, the service volume of a financial institution is predicted according to single variable information, the accuracy of the service volume prediction is low is solved, and the effect of accurately predicting the service volume of the financial institution is achieved.
Optionally, in the method for predicting the traffic of a financial institution provided in the embodiment of the present application, the target neural network model structure at least includes: the convolution layer, the multi-head self-attention layer and the two-way long short-time memory layer process the historical data information through the target neural network model, and the obtaining of the target feature vector corresponding to the historical data information comprises the following steps: extracting features of the historical data information through a convolution layer to obtain a first feature vector matrix; calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and extracting features of the second feature vector matrix through the bidirectional long and short-time memory layer to obtain a target feature vector.
For example, the target neural network model structure, as shown in fig. 2, at least includes: convolution layer, multi-head self-attention layer and two-way long short-term memory layer. The method comprises the steps of inputting historical data information into a target neural network model, carrying out convolution operation on the data information through a convolution layer in the target neural network model to obtain a first feature vector matrix corresponding to the data information through the history, inputting the first feature vector matrix into a multi-head self-attention layer, fusing self-attention results calculated by all heads in the multi-head self-attention layer to obtain a second feature vector matrix with weight information, inputting the second feature vector matrix into a bidirectional long and short-time memory layer, and carrying out bidirectional propagation extraction through a long and short-time memory network to obtain a target feature vector.
By combining a multi-head self-attention mechanism and a time sequence prediction method of a bidirectional long-short-time memory network, the feature and learning feature information can be automatically extracted from the historical data information, and different weights are distributed for different feature vectors, so that the problem that the accuracy of traffic prediction is low due to the fact that the same weight value is distributed for each feature vector in the prior art is solved, and therefore, the traffic of a financial institution can be accurately predicted by the method provided by the application.
Optionally, in the method for predicting the traffic of the financial institution provided in the embodiment of the present application, the target neural network model further includes a full connection layer, predicting, according to the target feature vector, the traffic of the target financial institution corresponding to the date to be predicted, where obtaining the prediction result includes: fusing the target feature vectors through the full connection layer to obtain fused target feature values; and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the fused target characteristic value to obtain a prediction result.
For example, as shown in fig. 2, the target neural network model further includes a full connection layer, the target feature vector is input to the full connection layer, fusion calculation is performed on the full connection layer to obtain a target feature value, and the traffic of the target financial institution on the date to be predicted is predicted according to the target feature value.
And carrying out fusion calculation on the target feature vector in space and feature dimension through the full connection layer, so as to obtain the business volume of the target financial institution on the date to be predicted.
Optionally, performing feature extraction on the historical data information through the convolution layer, and obtaining a first feature vector matrix includes: extracting features of the historical data information to obtain an initial feature vector matrix; normalizing the initial feature vector matrix to obtain a normalized initial feature vector matrix; and carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain a first eigenvector matrix.
For example, the convolution layer in the target neural network model is composed of a plurality of filters, the height of the filters is h, the width of the filters is n (the width is set to be the same as the variable number of the first factor value), and the step length is s. The kth filter scans the acquired historical data information X to obtain output, and the calculation process is shown in the following formula 1:
h k ＝f(W k *X+b k ) (equation 1)
Where, represents convolution operation, h k Representing the output vector corresponding to the filter, b k Representing the bias term, the function f is a nonlinear activation function, here using a ReLU, as shown in equation 2 below:
and then nonlinear transformation is carried out on the normalized initial feature vector matrix by using a nonlinear activation function f, random inactivation (dropout) operation is added after the nonlinear transformation, complex cooperative adaptation among neurons is reduced, and each neuron learns more robust features, so that the over-fitting of a calculation result is avoided, and finally a first feature vector matrix is obtained.
And the characteristic extraction is carried out on the historical data information through the convolution layer, so that the characteristic information related to the date to be predicted is obtained, and the accuracy of traffic prediction is improved.
Optionally, the calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information includes: calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix; and obtaining a second eigenvector matrix with weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
For example, the first eigenvector matrix is calculated by the multi-head self-attention layer, the calculation process is shown in fig. 3, and the self-attention layer mainly adopts a scaled dot product attention formula to calculate, specifically as shown in formula 3:
wherein Q, K, V respectively represent a query, a key, a value matrix, d k Representing the dimension of the key. In order to make the model learn information from different representing subspaces, repeating the Attention operation for a plurality of times through a plurality of heads, and making each head process different information, so as to calculate a weight value corresponding to each feature vector in the first vector matrix, which is specifically shown in formula 4:
Wherein Q is i 、K i 、V i A query, key, value matrix representing the ith subspace,
representing a trainable parameter matrix. After the weight value corresponding to each feature vector is obtained, multi-space fusion is carried out, and a second feature vector matrix with weight information is obtained.
The multi-head self-attention layer model, as shown in fig. 4, assigns different weights to different feature vectors through self-attention layer operation, namely, large weights are assigned to key feature vectors, small weights are assigned to most of irrelevant feature vectors, so that the problem of lower accuracy of traffic prediction due to the fact that the same weight value is assigned to each feature vector in the prior art is solved.
Optionally, extracting features of the second feature vector matrix by using the bidirectional long and short time memory layer, and obtaining the target feature vector includes: processing the second eigenvector matrix by adopting a forward propagation calculation mode through the bidirectional long and short-time memory layer to obtain a first hidden eigenvector; processing the second eigenvector matrix by adopting a back propagation calculation mode through the bidirectional long and short-time memory layer to obtain a second hidden eigenvector; and performing splicing processing on the first hidden feature vector and the second hidden feature vector to obtain a target feature vector.
For example, the second eigenvector matrix is input to the bidirectional long short-time memory layer, the first hidden eigenvector and the second hidden eigenvector are obtained through forward propagation of the long short-time memory network and backward propagation calculation of the long short-time memory network, and then the two eigenvectors are spliced to obtain the target eigenvector.
In summary, the common long-short-term memory network can only process data in a one-way transmission mode, and the bidirectional long-short-term memory network provided by the application processes historical data information through forward transmission and backward transmission, so that corresponding context characteristic information is accurately obtained, and the accuracy of traffic prediction of a financial institution is further improved.
Optionally, after predicting the traffic corresponding to the date to be predicted by the target financial institution according to the target feature vector, a prediction result is obtained, the method further includes: acquiring actual traffic corresponding to a date to be predicted; calculating the difference between the actual traffic and the predicted traffic in the predicted result to obtain target difference data; and optimizing the target neural network model according to the target difference data.
For example, by calculating the difference between the real traffic volume and the predicted traffic volume of the date to be predicted, the calculated data is input into the target neural network model to optimize the target neural network model, and the accuracy of the target neural network model in predicting the traffic volume is improved.
The accuracy of the target neural network model for predicting the traffic of the financial institutions is further improved through optimization processing of the target neural network model, and the target neural network model is optimized after each prediction, so that the prediction result of the target neural network model approaches to the real traffic infinitely.
Optionally, before the historical data information is processed through the target neural network model to obtain the target feature vector corresponding to the historical data information, the method further includes: the method comprises the steps of obtaining a training sample set, wherein the training sample set comprises a plurality of training sample data sets, the corresponding statistical time period of each training sample data set is different, and each training sample data set at least comprises: a plurality of dates and a plurality of second factor values corresponding to each date; and training the initial neural network model according to the training sample set to obtain a target neural network model and a preset time period.
For example, a training sample set is obtained that includes a plurality of training sample data sets that include a plurality of dates and weather conditions corresponding to each date, whether workdays, a number of self-service terminals for the target financial institution, a number of employees for the target financial institution, a number of registered customers for the target financial institution, a number of transacted business windows for the target financial institution, a historical business volume for the target financial institution, and a total number of financial institutions within a preset range.
In an alternative embodiment, the statistical time period corresponding to each training sample data set may be years, half years, three months, one month, etc., the training sample data sets corresponding to each statistical time period are respectively constructed, the training sample data sets are used to train the initial neural network model, the trained initial neural network model corresponding to each training sample data set is respectively obtained, and finally the target neural network model is determined from the plurality of trained initial neural network models. It should be noted that, the final target neural network model is one of the plurality of trained initial neural network models with highest prediction accuracy, and meanwhile, the statistical time period corresponding to the final target neural network model is determined as the preset time period.
The training sample set is input into the initial neural network model to train the initial neural network model, the training sample data sets with multiple time periods are adopted, and the prediction results of the neural network models with all the time periods are compared, so that the preset time period is set more reasonably, the training effect on the initial neural network model is improved, and further the target neural network model can predict the traffic more accurately.
In an alternative embodiment, the process of traffic prediction for a financial institution may be implemented using a schematic diagram as shown in fig. 5, with the following steps: the method comprises the steps of obtaining historical data information before a date to be predicted, and processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target neural network model at least comprises: the convolution layer, the multi-head self-attention layer and the two-way long short-time memory layer process the historical data information through the target neural network model, and the obtaining of the target feature vector corresponding to the historical data information comprises the following steps: extracting features of the historical data information through a convolution layer to obtain a first feature vector matrix; calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and extracting features of the second feature vector matrix through the bidirectional long and short time memory layer to obtain a target feature vector, and finally predicting the business volume corresponding to the target financial institution on the date to be predicted according to the target feature vector to obtain a prediction result.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowcharts, in some cases the steps illustrated or described may be performed in an order other than that illustrated herein.
According to the business volume prediction method for the financial institution, historical data information before a date to be predicted is obtained, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; the business volume corresponding to the date to be predicted of the target financial institution is predicted according to the target feature vector to obtain a prediction result, and the problem that the business volume prediction accuracy is low due to the fact that the business volume of the financial institution is predicted according to single variable information in the related technology is solved. According to the scheme, the historical data information of the multi-factor value is subjected to feature extraction through the target neural network model to obtain the target feature vector with the context feature information, and the historical data information can be aggregated in space and feature dimensions through the context feature information, so that the effect of improving the service prediction accuracy is achieved.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowcharts, in some cases the steps illustrated or described may be performed in an order other than that illustrated herein.
The embodiment of the application also provides a service volume predicting device for a financial institution, and it should be noted that the service volume predicting device for a financial institution of the embodiment of the application can be used for executing the service volume predicting method for a financial institution provided by the embodiment of the application. The following describes a traffic prediction apparatus for a financial institution according to an embodiment of the present application.
Fig. 6 is a schematic diagram of a traffic prediction device of a financial institution according to an embodiment of the present application. As shown in fig. 6, the apparatus includes: a first acquisition unit 601, a processing unit 602 and a prediction unit 603.
The first obtaining unit 601 is configured to obtain historical data information before a date to be predicted, where the historical data information at least includes: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted;
the processing unit 602 is configured to process the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, where the target feature vector at least includes context feature information between the historical data information;
And the prediction unit 603 is configured to predict, according to the target feature vector, a traffic volume corresponding to the target financial institution on the date to be predicted, so as to obtain a prediction result.
Optionally, in the traffic prediction device of a financial institution provided in the embodiment of the present application, the first obtaining unit 601 obtains historical data information before a date to be predicted, where the historical data information includes at least: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; the processing unit 602 processes the historical data information through the target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; the prediction unit 603 is configured to predict, according to the target feature vector, a traffic volume corresponding to a target financial institution on a date to be predicted, so as to obtain a prediction result, thereby solving a problem in the related art that the accuracy of traffic volume prediction is low due to the prediction of the traffic volume of the financial institution according to single variable information. According to the scheme, the historical data information of the multi-factor value is subjected to feature extraction through the target neural network model to obtain the target feature vector with the context feature information, and the historical data information can be aggregated in space and feature dimensions through the context feature information, so that the effect of improving the service prediction accuracy is achieved.
Optionally, the target neural network model includes at least: a convolutional layer, a multi-headed self-attention layer, and a two-way long short term memory layer, the processing unit 602 includes: the first extraction module is used for extracting the characteristics of the historical data information through the convolution layer to obtain a first characteristic vector matrix; the first operation module is used for operating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and the second extraction module is used for extracting the characteristics of the second characteristic vector matrix through the two-way long and short-time memory layer to obtain the target characteristic vector.
Optionally, the prediction unit 603 includes: the fusion module is used for fusing the target feature vectors through the full connection layer to obtain fused target feature values; and the prediction module is used for predicting the business volume corresponding to the date to be predicted of the target financial institution according to the fused target characteristic value to obtain a prediction result.
Optionally, the first extraction module includes: the first extraction submodule is used for carrying out feature extraction on the historical data information to obtain an initial feature vector matrix; the processing sub-module is used for carrying out normalization processing on the initial feature vector matrix to obtain a normalized initial feature vector matrix; and the first operation submodule is used for carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain a first eigenvector matrix.
Optionally, the first operation module includes: the second operation sub-module is used for calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix; and the third operation sub-module is used for obtaining a second eigenvector matrix with weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
Optionally, the second extraction module includes: the second extraction submodule is used for processing the second feature vector matrix through the bidirectional long and short-time memory layer in a forward propagation calculation mode to obtain a first hidden feature vector; the third extraction submodule is used for processing the second eigenvector matrix through the bidirectional long and short-time memory layer in a back propagation calculation mode to obtain a second hidden eigenvector; and the splicing module is used for carrying out splicing processing on the first hidden feature vector and the second hidden feature vector to obtain a target feature vector.
Optionally, the apparatus further comprises: the second obtaining unit is used for obtaining the actual traffic corresponding to the date to be predicted after predicting the traffic corresponding to the date to be predicted of the target financial institution according to the target feature vector to obtain a prediction result.
Optionally, the second acquisition unit further includes: the second operation module is used for calculating the difference value between the actual traffic and the predicted traffic in the predicted result to obtain target difference value data; and the optimization module is used for optimizing the target neural network model according to the target difference value data.
Optionally, the apparatus further comprises: the third obtaining unit is configured to obtain a training sample set before obtaining a target feature vector corresponding to the historical data information by processing the historical data information through the target neural network model, where the training sample set includes a plurality of training sample data sets, a statistical time period corresponding to each training sample data set is different, and each training sample data set at least includes: a plurality of dates and a plurality of second factor values corresponding to each date.
Optionally, the apparatus further comprises: the training unit is used for training the initial neural network model according to the training sample set to obtain a target neural network model and a preset time period.
The traffic predicting device for a financial institution includes a processor and a memory, and the first acquiring unit 601, the processing unit 602, the predicting unit 603, and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.
The processor includes a kernel, and the kernel fetches the corresponding program unit from the memory. The kernel can be provided with one or more, and the prediction of the business volume of the financial institution is realized by adjusting the kernel parameters.
The memory may include volatile memory, random Access Memory (RAM), and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM), among other forms in computer readable media, the memory including at least one memory chip.
Embodiments of the present invention provide a computer-readable storage medium having stored thereon a program that, when executed by a processor, implements a method of traffic prediction for a financial institution.
The embodiment of the invention provides a processor which is used for running a program, wherein the program runs to execute a traffic prediction method of a financial institution.
As shown in fig. 7, an embodiment of the present invention provides an electronic device, where the device includes a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, the following steps are implemented: acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; and predicting the traffic corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result.
Optionally, the target neural network model includes at least: the convolution layer, the multi-head self-attention layer and the two-way long short-time memory layer process the historical data information through the target neural network model, and the obtaining of the target feature vector corresponding to the historical data information comprises the following steps: extracting features of the historical data information through a convolution layer to obtain a first feature vector matrix; calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and extracting features of the second feature vector matrix through the bidirectional long and short-time memory layer to obtain a target feature vector.
Optionally, the target neural network model further includes a full connection layer, and predicting, according to the target feature vector, a traffic volume corresponding to the target financial institution on a date to be predicted, where obtaining a prediction result includes: fusing the target feature vectors through the full connection layer to obtain fused target feature values; and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the fused target characteristic value to obtain a prediction result.
Optionally, extracting features of the historical data information to obtain an initial feature vector matrix; normalizing the initial feature vector matrix to obtain a normalized initial feature vector matrix; and carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain a first eigenvector matrix.
Optionally, calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix; and obtaining a second eigenvector matrix with weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
Optionally, processing the second eigenvector matrix by adopting a forward propagation calculation mode through the two-way long and short time memory layer to obtain a first hidden eigenvector; processing the second eigenvector matrix by adopting a back propagation calculation mode through the bidirectional long and short-time memory layer to obtain a second hidden eigenvector; and performing splicing processing on the first hidden feature vector and the second hidden feature vector to obtain a target feature vector.
Optionally, acquiring the actual traffic corresponding to the date to be predicted; calculating the difference between the actual traffic and the predicted traffic in the predicted result to obtain target difference data; and optimizing the target neural network model according to the target difference data.
Optionally, a training sample set is obtained, where the training sample set includes a plurality of training sample data sets, and a statistical time period corresponding to each training sample data set is different, and each training sample data set includes at least: a plurality of dates and a plurality of second factor value amounts corresponding to each date; and training the initial neural network model according to the training sample set to obtain a target neural network model and a preset time period.
The device herein may be a server, PC, PAD, cell phone, etc.
The present application also provides a computer program product adapted to perform, when executed on a data processing device, a program initialized with the method steps of: acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted; processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information; and predicting the traffic corresponding to the date to be predicted by the target financial institution by the target feature vector to obtain a prediction result.
Optionally, the target neural network model includes at least: the convolution layer, the multi-head self-attention layer and the two-way long short-time memory layer process the historical data information through the target neural network model, and the obtaining of the target feature vector corresponding to the historical data information comprises the following steps: extracting features of the historical data information through a convolution layer to obtain a first feature vector matrix; calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information; and extracting features of the second feature vector matrix through the bidirectional long and short-time memory layer to obtain a target feature vector.
Optionally, the target neural network model further includes a full connection layer, and predicting, according to the target feature vector, a traffic volume corresponding to the target financial institution on a date to be predicted, where obtaining a prediction result includes: fusing the target feature vectors through the full connection layer to obtain fused target feature values; and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the fused target characteristic value to obtain a prediction result.
Optionally, extracting features of the historical data information to obtain an initial feature vector matrix; normalizing the initial feature vector matrix to obtain a normalized initial feature vector matrix; and carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain a first eigenvector matrix.
Optionally, calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix; and obtaining a second eigenvector matrix with weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
Optionally, processing the second eigenvector matrix by adopting a forward propagation calculation mode through the two-way long and short time memory layer to obtain a first hidden eigenvector; processing the second eigenvector matrix by adopting a back propagation calculation mode through the bidirectional long and short-time memory layer to obtain a second hidden eigenvector; and performing splicing processing on the first hidden feature vector and the second hidden feature vector to obtain a target feature vector.
Optionally, acquiring the actual traffic corresponding to the date to be predicted; calculating the difference between the actual traffic and the predicted traffic in the predicted result to obtain target difference data; and optimizing the target neural network model according to the target difference data.
Optionally, a training sample set is obtained, where the training sample set includes a plurality of training sample data sets, and a statistical time period corresponding to each training sample data set is different, and each training sample data set includes at least: a plurality of dates and a plurality of second factor value amounts corresponding to each date; and training the initial neural network model according to the training sample set to obtain a target neural network model and a preset time period.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.
Claims (11)
1. A method for traffic prediction in a financial institution, comprising:
acquiring historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted;
processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information;
and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result.
2. The method according to claim 1, wherein the target neural network model comprises at least: the convolution layer, the multi-head self-attention layer and the two-way long short-time memory layer process the historical data information through a target neural network model, and the obtaining of the target feature vector corresponding to the historical data information comprises the following steps:
extracting features of the historical data information through the convolution layer to obtain a first feature vector matrix;
Calculating the first eigenvector matrix through the multi-head self-attention layer to obtain a second eigenvector matrix with weight information;
and extracting features of the second feature vector matrix through the two-way long and short-time memory layer to obtain the target feature vector.
3. The method of claim 1, wherein the target neural network model further comprises a full connection layer, and predicting traffic corresponding to the target financial institution on the date to be predicted according to the target feature vector, and obtaining a prediction result comprises:
fusing the target feature vectors through the full connection layer to obtain fused target feature values;
and predicting the business volume corresponding to the date to be predicted by the target financial institution according to the fused target characteristic value to obtain the prediction result.
4. The method of claim 2, wherein performing feature extraction on the historical data information by the convolution layer to obtain a first feature vector matrix comprises:
extracting features of the historical data information to obtain an initial feature vector matrix;
normalizing the initial feature vector matrix to obtain a normalized initial feature vector matrix;
And carrying out nonlinear transformation on the normalized initial eigenvector matrix to obtain the first eigenvector matrix.
5. The method of claim 2, wherein operating the first eigenvector matrix through the multi-headed self-attention layer to obtain a second eigenvector matrix with weight information comprises:
calculating through a scaling dot product attention formula in the multi-head self-attention layer to obtain a weight value corresponding to each feature vector in the first feature vector matrix;
and obtaining the second eigenvector matrix with weight information according to the weight value corresponding to each eigenvector in the first eigenvector matrix.
6. The method of claim 2, wherein extracting features of the second feature vector matrix through the two-way long-short-term memory layer, obtaining a target feature vector comprises:
processing the second eigenvector matrix by the bidirectional long short-time memory layer in a forward propagation calculation mode to obtain a first hidden eigenvector;
processing the second eigenvector matrix by the bidirectional long short-time memory layer in a back propagation calculation mode to obtain a second hidden eigenvector;
And performing splicing processing on the first hidden feature vector and the second hidden feature vector to obtain the target feature vector.
7. The method of claim 1, wherein after predicting traffic corresponding to the date to be predicted by the target financial institution based on the target feature vector, the method further comprises:
acquiring the actual traffic corresponding to the date to be predicted;
calculating the difference value between the actual traffic and the predicted traffic in the predicted result to obtain target difference value data;
and optimizing the target neural network model according to the target difference data.
8. The method of claim 1, wherein before processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, the method further comprises:
obtaining a training sample set, wherein the training sample set comprises a plurality of training sample data sets, the corresponding statistical time period of each training sample data set is different, and each training sample data set at least comprises: a plurality of dates and a plurality of second factor values corresponding to each date;
And training the initial neural network model according to the training sample set to obtain the target neural network model and the preset time period.
9. A financial institution traffic prediction apparatus, comprising:
the first obtaining unit is used for obtaining historical data information before a date to be predicted, wherein the historical data information at least comprises: a plurality of first factor values corresponding to each date in a preset time period before the date to be predicted;
the processing unit is used for processing the historical data information through a target neural network model to obtain a target feature vector corresponding to the historical data information, wherein the target feature vector at least comprises context feature information among the historical data information;
and the prediction unit is used for predicting the business volume corresponding to the date to be predicted by the target financial institution according to the target feature vector to obtain a prediction result.
10. A processor for running a program, wherein the program when run performs the method of traffic prediction for a financial institution of any one of claims 1 to 8.
11. An electronic device comprising one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of traffic prediction for a financial institution of any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310272900.0A CN116187581A (en) | 2023-03-20 | 2023-03-20 | Financial institution traffic prediction method and device, processor and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310272900.0A CN116187581A (en) | 2023-03-20 | 2023-03-20 | Financial institution traffic prediction method and device, processor and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116187581A true CN116187581A (en) | 2023-05-30 |
Family
ID=86446392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310272900.0A Pending CN116187581A (en) | 2023-03-20 | 2023-03-20 | Financial institution traffic prediction method and device, processor and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116187581A (en) |
-
2023
- 2023-03-20 CN CN202310272900.0A patent/CN116187581A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11693917B2 (en) | Computational model optimizations | |
| US11621930B2 (en) | Systems and methods for generating dynamic conversational responses using trained machine learning models | |
| US20220253856A1 (en) | System and method for machine learning based detection of fraud | |
| US11694165B2 (en) | Key-value memory network for predicting time-series metrics of target entities | |
| US11811708B2 (en) | Systems and methods for generating dynamic conversational responses using cluster-level collaborative filtering matrices | |
| US11677693B2 (en) | Systems and methods for generating dynamic conversational responses using ensemble prediction based on a plurality of machine learning models | |
| US20220414343A1 (en) | Systems and methods for generating dynamic conversational responses based on historical and dynamically updated information | |
| Sina Mirabdolbaghi et al. | Model optimization analysis of customer churn prediction using machine learning algorithms with focus on feature reductions | |
| CN112598526A (en) | Asset data processing method and device | |
| US20230351119A1 (en) | Systems and methods for generating dynamic conversational responses through aggregated outputs of machine learning models | |
| CN116187581A (en) | Financial institution traffic prediction method and device, processor and electronic equipment | |
| Raipitam et al. | Comparative Study on Stock Market Prediction using Generic CNN-LSTM and Ensemble Learning | |
| KR102284440B1 (en) | Method to broker deep learning model transactions perfomed by deep learning model transaction brokerage servers | |
| CA3196711A1 (en) | Systems and methods for generating dynamic conversational responses through aggregated outputs of machine learning models | |
| KAMLEY et al. | Multiple regression: A data mining approach for predicting the stock market trends based on open, close and high price of the month | |
| Ouyang et al. | Interpretable Spatial-Temporal Attention Graph Convolution Network for Service Part Hierarchical Demand Forecast | |
| CN115730823A (en) | Risk probability identification method and device, processor and electronic equipment | |
| CN117351297A (en) | Method and device for identifying noise interference image, processor and electronic equipment | |
| CN116800831A (en) | Service data pushing method, device, storage medium and processor | |
| Mulkkalla | Web Traffic Time Series Forecasting | |
| CN117933568A (en) | Operation decision method, apparatus, device, medium and program product | |
| CA3108609A1 (en) | System and method for machine learning based detection of fraud | |
| CN117474603A (en) | Popularization information generation method and device, storage medium and electronic equipment | |
| Chandrasekaran | Prediction of Litecoin Prices using ARIMA and LSTM | |
| CN117611352A (en) | Vehicle insurance claim processing method, device, computer equipment and storage medium |
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 |