CN115879608A

CN115879608A - Resource information prediction method, resource information prediction device, computer equipment and storage medium

Info

Publication number: CN115879608A
Application number: CN202211505002.7A
Authority: CN
Inventors: 曾旭峰; 毛培芳; 薛煜峰; 廖杰
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2022-11-29
Filing date: 2022-11-29
Publication date: 2023-03-31

Abstract

The application relates to a resource information prediction method, a resource information prediction device, computer equipment, a storage medium and a computer program product, relates to the technical field of artificial intelligence, and can be used in the field of financial science and technology or other related fields. The method comprises the following steps: responding to a resource information prediction request aiming at an object to be predicted, and acquiring resource characteristic information of the object to be predicted in the current time period; inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period; the pre-trained resource information prediction model is obtained by training according to sample resource characteristic information of a sample object; inputting the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range. By adopting the method, the prediction accuracy of the resource information can be improved.

Description

Resource information prediction method, resource information prediction device, computer equipment and storage medium

Technical Field

The present application relates to the field of artificial intelligence technologies, and in particular, to a resource information prediction method, apparatus, computer device, storage medium, and computer program product.

Background

Along with the continuous development of economy, the flow scale of cross-border resources is obviously increased, and the method is more important for the fluid management of the cross-border resources; in order to manage cross-border resources conveniently, resource information in a future time period, such as resource demand information, needs to be predicted in advance.

In the traditional technology, when resource information of a future time period is predicted, prediction is generally performed on the basis of simple expert rules or artificial historical experience; however, the expert rules or historical experiences have subjective factors, and the resource information in the future time period is influenced by various factors, so that the predicted resource information is not accurate enough, and the prediction accuracy of the resource information is low.

Disclosure of Invention

In view of the above, it is necessary to provide a resource information prediction method, a resource information prediction apparatus, a computer device, a computer readable storage medium, and a computer program product, which can improve the prediction accuracy of resource information, in view of the above technical problems.

In a first aspect, the present application provides a resource information prediction method. The method comprises the following steps:

responding to a resource information prediction request aiming at an object to be predicted, and acquiring resource characteristic information of the object to be predicted in the current time period;

inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period; the pre-trained resource information prediction model is obtained by training according to sample resource characteristic information of a sample object;

inputting the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range.

In one embodiment, the inputting the resource feature information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in a next time period of the current time period includes:

inputting the resource characteristic information into a pre-trained resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period;

screening out a preset resource information range with the maximum prediction probability from all the preset resource information ranges;

and taking the preset resource information range with the maximum prediction probability as a target resource information range of the object to be predicted in the next time period of the current time period.

In one embodiment, the inputting the resource feature information into a pre-trained resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period includes:

determining a resource identifier associated with the resource characteristic information;

screening out a target resource information prediction model matched with the resource identifier from a pre-trained resource information prediction model;

and inputting the resource characteristic information into the target resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period.

In one embodiment, the pre-trained resource information prediction model is trained by:

acquiring sample resource characteristic information of a sample object in a sample time period;

inputting the sample resource characteristic information into a resource information prediction model to be trained to obtain a target resource information range of the sample object in the next time period of the sample time period;

obtaining a loss value according to the difference between the actual resource information range and the target resource information range of the sample object in the next time period of the sample time period;

and training the resource information prediction model to be trained according to the loss value to obtain a trained resource information prediction model serving as the pre-trained resource information prediction model.

In one embodiment, the method further comprises:

acquiring feedback data aiming at the pre-trained resource information prediction model;

and updating the pre-trained resource information prediction model according to the feedback data to obtain an updated resource information prediction model.

In one embodiment, after inputting the resource feature information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period, the method further includes:

acquiring initial resource report information of the object to be predicted in the next time period;

updating the initial resource reporting information of the object to be predicted in the next time period according to the predicted resource information to obtain target resource reporting information;

and executing a corresponding resource submission process according to the target resource submission information.

In a second aspect, the present application further provides a resource information prediction apparatus. The device comprises:

the characteristic acquisition module is used for responding to a resource information prediction request aiming at an object to be predicted and acquiring resource characteristic information of the object to be predicted in the current time period;

the range prediction module is used for inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period; the pre-trained resource information prediction model is obtained by training according to sample resource characteristic information of a sample object;

the information prediction module is used for inputting the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range.

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the following steps when executing the computer program:

In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:

In a fifth aspect, the present application further provides a computer program product. The computer program product comprising a computer program which when executed by a processor performs the steps of:

The resource information prediction method, the resource information prediction device, the computer equipment, the storage medium and the computer program product respond to the resource information prediction request aiming at the object to be predicted and acquire the resource characteristic information of the object to be predicted in the current time period; then inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period; the pre-trained resource information prediction model is obtained by training according to sample resource characteristic information of a sample object; finally, inputting the resource characteristic information and the target resource information range of the current time period into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range. Therefore, when the resource information is predicted, a target resource information range of the object to be predicted in the next time period of the current time period is predicted through a pre-trained resource information prediction model, and then the predicted resource information of the object to be predicted in the next time period is predicted from the target resource information range through a pre-trained regression prediction model; in the whole process, the pre-trained resource information prediction model and the pre-trained regression prediction model are fully utilized, and the expert rules and the historical experience are not required to be relied on, so that the predicted resource information is more accurate, the improvement of the prediction accuracy of the resource information is facilitated, and the defect that the predicted resource information is not accurate enough due to the expert rules and the historical experience is avoided.

Drawings

FIG. 1 is a flow diagram illustrating a method for resource information prediction in one embodiment;

FIG. 2 is a flowchart illustrating the steps of determining a target resource information range in one embodiment;

FIG. 3 is a flowchart illustrating the training steps of the resource information prediction model in one embodiment;

FIG. 4 is a flowchart illustrating a resource information prediction method according to another embodiment;

FIG. 5 is an overall block diagram of the funds planning apparatus in one embodiment;

FIG. 6 is a block diagram of a funding position presentation device in accordance with an embodiment;

FIG. 7 is a block diagram showing the structure of a position predicting device according to an embodiment;

FIG. 8 is a block diagram of an apparatus for resource information prediction in one embodiment;

FIG. 9 is a diagram illustrating an internal structure of a computer device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

In an embodiment, as shown in fig. 1, a resource information prediction method is provided, and this embodiment is illustrated by applying the method to a terminal, and it is to be understood that the method may also be applied to a server, and may also be applied to a system including the terminal and the server, and is implemented through interaction between the terminal and the server. The terminal can be but not limited to various personal computers, notebook computers, smart phones, tablet computers, internet of things equipment and portable wearable equipment, and the internet of things equipment can be smart sound boxes, smart televisions, smart air conditioners, smart vehicle-mounted equipment and the like. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, and the like. The server may be implemented as a stand-alone server or as a server cluster consisting of a plurality of servers. In this embodiment, the method includes the steps of:

step S101, responding to the resource information prediction request aiming at the object to be predicted, and acquiring the resource characteristic information of the object to be predicted in the current time period.

The resource information related to the application refers to resource demand information, specifically refers to actual demand amount of position fund, actual usage amount of position fund and the like.

The object to be predicted related to the present application refers to an object that needs to predict resource information, and specifically refers to a bank, such as an overseas bank. In an actual scene, the object to be predicted refers to an overseas bank position account.

The resource information prediction request refers to a request for obtaining predicted resource information of an object to be predicted in a next time period of the current time period.

The current time period refers to the current day, and the next time period refers to the next day, the next week and the like.

The resource characteristic information refers to characteristic information related to future resource information change of an object to be predicted, such as a business type, a Fund status, bookkeeping details, a delivery amount, currency and the like, and other influence factors are introduced, such as an Exchange rate, an international Exchange rate, an external trade amount, a national policy weighting factor, a gold ETF (Exchange Traded Fund) fluctuation index, a dollar index, a Daojones industrial index, a policy uncertainty index, a Chinese evidence index and the like.

Specifically, the terminal responds to a resource information prediction request aiming at an object to be predicted, and confirms the object to be predicted; the method comprises the steps of obtaining object data of an object to be predicted, inputting the object data of the object to be predicted into a resource characteristic information extraction model, and performing characteristic extraction processing on the object data of the object to be predicted through the resource characteristic information extraction model to obtain characteristic information related to future resource information change of the object to be predicted, wherein the characteristic information is used as resource characteristic information of the object to be predicted in the current time period.

For example, in response to a resource information prediction request for an object to be predicted, a terminal performs time period sliding window processing on a position account report quantity, a report supplement quantity and a use quantity of the object to be predicted in a current time period to generate sequence characteristic values such as a maximum value, a minimum value, a mean value, a range, a difference and the like of the report quantity and the use quantity in an observation period; then, according to the date dimension of the month and the day, integrating and calculating the average value, the maximum value and the minimum value of the historical synchronization, and taking the average value, the maximum value and the minimum value as new characteristics; meanwhile, other characteristics such as external data such as exchange rate, gold ETF fluctuation index, dollar index, doujones industrial index, policy uncertainty index, and middle-school index are supplemented; and finally, taking the various characteristics together as the resource characteristic information of the object to be predicted in the current time period.

Step S102, inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period; and the pre-trained resource information prediction model is obtained by training according to the sample resource characteristic information of the sample object.

The pre-trained resource information prediction model is a model for predicting the resource information range of the object in the future time period, such as a machine learning model, a deep learning model, a neural network model, and the like. In an actual scene, the resource information prediction model refers to a LightGBM model; the LightGBM model is a lightweight gradient elevator, has the characteristics of high training speed and low memory occupation, and has strong generalization capability; under general conditions, the LightGBM model can achieve better effect than other models, and better model training effect is achieved.

The next time period refers to a time period of the next unit of the current time period, the next day, the next week, the next month, and the like.

The resource information range refers to a resource demand information range, specifically refers to an actual demand range of position funds, such as 2130-2690 million, 2690-3220 million, and the like.

The target resource information range of the object to be predicted in the next time period of the current time period refers to the actual demand amount range of position funds of the oversea bank in the next time period (for example, the next day).

Wherein, the sample object refers to a training object of the resource information prediction model and also refers to an overseas bank; meanwhile, the sample resource feature information also refers to feature information related to future resource information changes of the sample object.

Specifically, the terminal inputs resource characteristic information of an object to be predicted in the current time period into a pre-trained resource information prediction model, and performs prediction processing on the resource characteristic information through each decision tree model in the resource information prediction model to obtain a resource information range output by each decision tree model; and selecting the resource information range with the largest occupation ratio from the resource information ranges output by the decision tree models as a target resource information range of the object to be predicted in the next time period of the current time period.

For example, the resource information ranges output by the decision tree models in the resource information prediction model are a resource information range a, a resource information range B, a resource information range C, a resource information range B, a resource information range D, a resource information range E, a resource information range B, and a resource information range C, respectively; and if the occupation proportion of the resource information range B is 0.5 and the occupation proportion is the maximum, taking the resource information range B as a target resource information range of the object to be predicted in the next time period of the current time period.

Step S103, inputting the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the prediction resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range.

The pre-trained regression prediction model is a model for predicting the prediction resource information of the object to be predicted in the next time period, such as a machine learning model, a deep learning model, a neural network model, and the like.

The resource information of the object to be predicted in the next time period is a specific numerical value of the position fund actual demand of the oversea bank in the next time period (for example, the next day), for example, 2650 ten thousand, 3000 ten thousand, and the like.

The predicted resource information refers to certain resource information in a target resource information range, such as the actual demand of funds of a certain position.

Specifically, the terminal inputs the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model, and performs regression prediction on the target resource information range through the pre-trained regression prediction model based on the resource characteristic information so as to predict the predicted resource information of the object to be predicted in the next time period from the target resource information range.

For example, assuming that the target resource information range is 2690-3220 million, performing regression prediction on the target resource information range through a regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period of 3000 million.

In the resource information prediction method, in response to a resource information prediction request for an object to be predicted, resource characteristic information of the object to be predicted in a current time period is acquired; then inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period; the pre-trained resource information prediction model is obtained by training according to sample resource characteristic information of a sample object; finally, inputting the resource characteristic information and the target resource information range of the current time period into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range. Therefore, when resource information is predicted, a target resource information range of the object to be predicted in the next time period of the current time period is predicted through a pre-trained resource information prediction model, and then predicted resource information of the object to be predicted in the next time period is predicted from the target resource information range through a pre-trained regression prediction model; in the whole process, the pre-trained resource information prediction model and the pre-trained regression prediction model are fully utilized, and the expert rules and the historical experience are not required to be relied on, so that the predicted resource information is more accurate, the improvement of the prediction accuracy of the resource information is facilitated, and the defect that the predicted resource information is not accurate enough due to the expert rules and the historical experience is avoided.

In an embodiment, as shown in fig. 2, in the step S102, the resource feature information is input into a resource information prediction model trained in advance, so as to obtain a target resource information range of the object to be predicted in a time period next to the current time period, which specifically includes the following steps:

step S201, inputting the resource characteristic information into a resource information prediction model trained in advance, and obtaining the prediction probability of each preset resource information range corresponding to the next time period of the current time period.

Step S202, screening out the preset resource information range with the maximum prediction probability from all the preset resource information ranges.

Step S203, using the preset resource information range with the maximum prediction probability as the target resource information range of the object to be predicted in the next time period of the current time period.

Each preset resource information range refers to a preset resource information range, such as 0-2130 million, 2130-2690 million, 2690-3220 million, 3220-4000 million, more than 4000 million, and the like.

The prediction probability is used for measuring the probability that the resource information of the object to be predicted in the next time period of the current time period belongs to a certain preset resource information range.

Specifically, the resource characteristic information of the object to be predicted of the terminal in the current time period is input into a pre-trained resource information prediction model, and the resource characteristic information is subjected to convolution processing and full connection processing through the resource information prediction model to obtain a one-dimensional prediction probability vector; determining the prediction probability that the resource information of the object to be predicted in the next time period of the current time period belongs to each preset resource information range from the one-dimensional prediction probability vector; screening out a preset resource information range with the maximum prediction probability from all preset resource information ranges; and taking the preset resource information range with the maximum prediction probability as a target resource information range of the object to be predicted in the next time period of the current time period.

For example, it is assumed that the prediction probabilities of the resource information range a, the resource information range B, the resource information range C, the resource information range D, and the resource information range E are 0.1, 0.2, 0.4, and 0.1, respectively, and the prediction probability of the resource information range D is the maximum, which indicates that the target resource information range of the object to be predicted in the next time period of the current time period is the resource information range D.

In the embodiment, the resource characteristic information is input into a pre-trained resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period, and the target resource information range of the object to be predicted in the next time period of the current time period is screened out from each preset resource information range according to the prediction probability; therefore, the resource information prediction model trained in advance is comprehensively utilized, the resource information range is favorably and accurately determined, and the prediction accuracy of the resource information is further improved; meanwhile, the whole process does not need manual participation, so that the manpower is reduced, and the prediction efficiency of the resource information is improved.

In an embodiment, in step S201, the resource feature information is input into a pre-trained resource information prediction model, so as to obtain a target resource information range of the object to be predicted in a next time period of the current time period, which specifically includes the following contents: determining a resource identifier associated with the resource characteristic information; screening out a target resource information prediction model matched with the resource identifier from the pre-trained resource information prediction model; and inputting the resource characteristic information into a target resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period.

Wherein, the resource identification refers to currency, such as RMB, USD, euro, etc. The resource information prediction models corresponding to different resource identifications are different.

Specifically, the terminal determines a resource identifier associated with the resource feature information according to the resource identifier identification instruction; screening out a resource information prediction model with the corresponding resource identification matched with the resource identification from pre-trained resource information prediction models to serve as a target resource information prediction model; inputting the resource characteristic information into a target resource information prediction model to obtain a one-dimensional prediction probability vector; determining the prediction probability that the resource information of the object to be predicted in the next time period of the current time period belongs to each preset resource information range from the one-dimensional prediction probability vector; and screening out a preset resource information range with the maximum prediction probability from all the preset resource information ranges, wherein the preset resource information range is used as a target resource information range of the object to be predicted in the next time period of the current time period, such as the actual demand range of position funds of different currencies in the next time period.

In the embodiment, according to the resource identifier associated with the resource characteristic information, a target resource information prediction model meeting the requirement is screened out from pre-trained resource information prediction models, and according to the target resource information prediction model, the prediction probability of each preset resource information range corresponding to the next time period of the current time period is obtained; therefore, by distinguishing the resource information prediction models corresponding to different resource identifications, the accuracy of the predicted resource information range is further improved, the subsequently determined predicted resource information is more accurate, and the prediction accuracy of the resource information is improved.

In one embodiment, as shown in fig. 3, the pre-trained resource information prediction model is obtained by training the following steps:

step S301, sample resource characteristic information of the sample object in the sample time period is obtained.

Step S302, inputting the sample resource characteristic information into a resource information prediction model to be trained, and obtaining a target resource information range of the sample object in the next time period of the sample time period.

Step S303, a loss value is obtained according to a difference between the actual resource information range and the target resource information range of the sample object in the next time period of the sample time period.

And step S304, training the resource information prediction model to be trained according to the loss value to obtain the trained resource information prediction model as a pre-trained resource information prediction model.

Wherein, the sample time period refers to the time period participating in training; the next time period of the sample time period refers to the time period of the next unit of the sample time period, the next day, the next week, the next month, and the like.

The loss value is used for measuring the prediction accuracy of the resource information prediction model; for example, the smaller the loss value, the higher the prediction accuracy of the resource information prediction model.

Specifically, the terminal acquires sample resource characteristic information of a sample object in a sample time period from a local database; inputting the sample resource characteristic information into a resource information prediction model to be trained, and performing a series of processing, such as convolution processing, full-connection processing and the like, on the sample resource characteristic information through the resource information prediction model to be trained to obtain the prediction probability that the resource information of the sample object in the next time period of the sample time period belongs to each preset resource information range; screening out a preset resource information range with the maximum prediction probability from all preset resource information ranges, and taking the preset resource information range as a target resource information range of the sample object in the next time period of the sample time period; acquiring an actual resource information range of the sample object in a next time period of the sample time period, and calculating to obtain a loss value according to the difference between the actual resource information range and the target resource information range and in combination with a loss function; and training the resource information prediction model to be trained according to the loss value until a training end condition is reached, and taking the trained resource information prediction model reaching the training end condition as a pre-trained resource information prediction model.

For example, when the loss value is greater than or equal to the preset threshold, the terminal adjusts the model parameter of the resource information prediction model to be trained according to the loss value, and repeats steps S302 to S304 to train the resource information prediction model after model parameter adjustment again until the loss value obtained through the trained resource information prediction model is less than the preset threshold, and then the trained resource information prediction model is used as the pre-trained resource information prediction model.

Further, after obtaining the target resource information range of the sample object in the next time period of the sample time period, the terminal can predict the target resource information range through a regression prediction model to be trained to obtain the predicted resource information of the sample object in the next time period; and performing joint training on the resource information prediction model to be trained and the regression prediction model to be trained according to the difference between the predicted resource information and the actual resource information of the sample object in the next time period to obtain the trained resource information prediction model and the trained regression prediction model.

In this embodiment, the resource information prediction model to be trained is repeatedly trained according to the sample resource feature information of the sample object in the sample time period, which is beneficial to improving the accuracy of the resource information range predicted by the trained resource information prediction model, thereby improving the prediction accuracy of the resource information.

In an embodiment, the resource information prediction method provided by the present application further includes an update step of the resource information prediction model, which specifically includes the following steps: acquiring feedback data of a pre-trained resource information prediction model; and updating the pre-trained resource information prediction model according to the feedback data to obtain an updated resource information prediction model.

The feedback data refers to newly generated or newly mined data, such as resource feature information of an actual object, an actual resource information range, a target resource information range of the actual object in the next time period, which is output by using a pre-trained resource information prediction model, and the like.

Specifically, the terminal acquires feedback data aiming at a pre-trained resource information prediction model from a local database; and then, according to the feedback data, repeatedly executing the steps S302 to S304 to update the network parameters of the pre-trained resource information prediction model to obtain a resource information prediction model with updated network parameters, and taking the resource information prediction model as the updated resource information prediction model.

Further, in step S102, the resource feature information is input into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in a next time period of the current time period, which specifically includes the following contents: and the terminal inputs the resource characteristic information into the updated resource information prediction model, and performs prediction processing on the resource characteristic information through the updated resource information prediction model to obtain a target resource information range of the object to be predicted in the next time period of the current time period.

In this embodiment, the resource information prediction model trained in advance is updated according to the feedback data for the resource information prediction model trained in advance, so that the purpose of optimizing the resource information prediction model is achieved, and the accuracy of the resource information range predicted by the resource information prediction model is further improved.

In an embodiment, in step S103, after inputting the resource feature information and the target resource information range of the current time period into a regression prediction model trained in advance to obtain the predicted resource information of the object to be predicted in the next time period, the method further includes: acquiring initial resource report information of an object to be predicted in the next time period; updating the initial resource reporting information of the object to be predicted in the next time slot according to the predicted resource information to obtain target resource reporting information; and executing the corresponding resource submission flow according to the target resource submission information.

Wherein, the initial resource reporting information refers to the original position fund reporting amount.

Wherein, the resource reporting process is a position fund reporting process.

Specifically, the terminal acquires initial resource reporting information of an object to be predicted in the next time period from a database; updating the initial resource reporting information of the object to be predicted in the next time slot according to the predicted resource information, so that the updated resource reporting information is closer to the predicted resource information; taking the updated resource submission information as target resource submission information; triggering the resource submission process on the line according to the target resource submission information, such as automatically entering the next submission node.

For example, if the predicted resource information is 3000 ten thousand, the initial resource submission information is 1800 ten thousand, and the difference from 3000 ten thousand is large, the initial resource submission information is updated to a value close to 3000 ten thousand, for example 2900 ten thousand, so that the difference from 3000 ten thousand is avoided; and then, automatically triggering an online resource submission process according to the updated resource submission information.

In this embodiment, according to the obtained predicted resource information, the initial resource reporting information of the object to be predicted in the next time period is updated, which is beneficial to improving the accuracy of resource reporting and avoiding resource idleness or insufficient preparation.

In an embodiment, as shown in fig. 4, another resource information prediction method is provided, which is described by taking an example that the method is applied to a terminal, and includes the following steps:

step S401, in response to the resource information prediction request for the object to be predicted, acquiring resource feature information of the object to be predicted in the current time period.

Step S402, determining a resource identifier associated with the resource characteristic information.

Step S403, screening out a target resource information prediction model matched with the resource identifier from the pre-trained resource information prediction models.

Step S404, inputting the resource characteristic information into the target resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period.

Step S405, screening out a preset resource information range with the maximum prediction probability from all preset resource information ranges.

Step S406, using the preset resource information range with the maximum prediction probability as the target resource information range of the object to be predicted in the next time period of the current time period.

Step S407, inputting the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the prediction resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range.

Step S408, acquiring initial resource reporting information of the object to be predicted in the next time period.

Step S409, updating the initial resource reporting information of the object to be predicted in the next time slot according to the predicted resource information to obtain target resource reporting information; and executing the corresponding resource submission flow according to the target resource submission information.

When the resource information is predicted, a target resource information range of an object to be predicted in the next time period of the current time period is predicted through a pre-trained resource information prediction model, and then the predicted resource information of the object to be predicted in the next time period is predicted from the target resource information range through a pre-trained regression prediction model; in the whole process, the pre-trained resource information prediction model and the pre-trained regression prediction model are fully utilized, and the expert rules and the historical experience are not required to be relied on, so that the predicted resource information is more accurate, the improvement of the prediction accuracy of the resource information is facilitated, and the defect that the predicted resource information is not accurate enough due to the expert rules and the historical experience is avoided. In addition, according to the obtained predicted resource information, the initial resource reporting information of the object to be predicted in the next time period is updated, so that the accuracy of resource reporting is improved, and the resource is prevented from being idle or insufficient in preparation.

In an embodiment, in order to more clearly illustrate the resource information prediction method provided in the embodiments of the present application, the resource information prediction method is specifically described below with a specific embodiment. In one embodiment, the method for planning and predicting the capital position of the overseas bank is characterized in that an artificial intelligent model based on machine learning is established by improving the capital position reporting process, so that the whole process management of position capital reporting, calculation, summarization, limitation, early warning and prediction is realized, manual processing is reduced, the capital position risk of the bank is reduced, the capital use benefit is improved, advance prediction, in-process monitoring and after-event analysis are realized, and decision support is provided for the capital application of the bank. Wherein, fig. 5 provides an overall frame diagram of a fund planning device, which comprises a position account balance summarizing device, a bookkeeping detail calculating device, a fund position reporting device, a position planning summarizing device and a position predicting device. Fig. 6 is a block diagram showing a structure of a fund position submitting apparatus, which includes a UI (User Interface) presentation unit, a workflow unit, a logic calculation unit, and a data storage unit. Fig. 7 provides a block diagram of a position prediction apparatus including a sample selection unit, a target value definition unit, a model design unit, and a result presentation unit. The method specifically comprises the following steps:

step one, linking a bank clearing system according to the currency type related to overseas banks, and acquiring the balance condition of the account in real time. For example, in the morning of each day, the position account balance summarizing device shown in fig. 5 traverses all the related currencies, traverses the balances inquired by all position users in the currencies aiming at each currency, and is linked with the bank clearing system to acquire the balance of the position user in the bank of the T-1 day; firstly, collecting according to single currency, and combining T-1 day exchange rate, converting into general currency such as dollar, port currency, etc. to collect to obtain the use condition of different currency funds and the total fund condition after converting into general currency.

Linking a banking system to acquire bookkeeping information every day; estimating daily fund use condition according to the service classification and the term factors of bookkeeping service types by day; generally, banks have business systems related to financial markets for registering bookkeeping information, and interaction among systems is performed in a file or interface mode to complete acquisition of the bookkeeping information.

And step three, classifying the book details according to the actual operation condition of the bank, such as financial markets, foreign currency derivatives and the like, and carrying out fund exposure summary according to the book-keeping basic elements (interest date, expiration date, current day exchange rate, amount and currency) and according to currency, type maintenance and date. For example, the financial market comprises the types of putting on and leaving the same industry, borrowing the same industry, fund falling, foreign exchange on-demand, foreign exchange forward, financing tools, bond repurchase and the like, each type of element comprises fund, currency, interest date, expiration date and the like, the effective time range of bookkeeping is determined according to the interest date and the expiration date, and then fund exposure required by all bookkeeping can be summarized according to the currency and the bookkeeping types so as to replace partial manual reporting and reimbursement and simplify the reporting process.

And step four, establishing an on-line manual forecasting channel, supporting cross-department information transfer reporting and estimation and alarm of automatic pre-occupation of each reporting fund, and mainly used for reporting funds before bookkeeping information. For example, referring to fig. 6, the ui display unit is used for providing the display of the functions of forecasting, inquiring and examining and approving funds, so as to facilitate the on-line processing of banking staff; the elements of forecasting include beginning of capital, ending time, currency, forecasting business category, etc. The work flow unit is used for carrying out authority division according to the role of the user in a work flow mode, and provides a function of reporting and flowing between cross-department banks. The logic calculation unit is used for automatically judging whether the liquidity risk of the fund exists after the forecast takes effect, and the main idea is to forecast the position fund balance under the current currency by combining the position fund balance summarizing result under the forecast currency obtained in the step one, the bookkeeping detail obtained in the step three and the forecast success record of the current day, namely: clearing account balance, effective bookkeeping details on the current day, forecast funds on the current day, and newly-increased forecast funds = current-day position fund balance; and allowing to enter forecast circulation only when the position fund balance of the day exceeds a preset threshold, and automatically rejecting the report if the position fund balance of the day exceeds the preset threshold. And the data storage and storage unit is used for storing the reporting element information and the circulation information by relying on the database middleware and is used for next-time forecast frame calculation and post analysis.

And step five, calculating the position fund use condition in a future period of time according to the result of the step four and the transverse frame of the date, thereby facilitating the fund reservation or allocation in advance.

And step six, according to the result of the step four, the result of the position balance of the current date is longitudinally calculated according to the service types, the bank service development condition is comprehensively mastered, and support is provided for the operation decision.

Finally, forecasting the position fund, calculating the using condition of the fund through the previous steps, mainly aiming at the problems that the capacity of resisting unexpected risks is low on the premise that banks normally operate, particularly, the manual forecasting part is relatively dependent on the experience of people, and the overseas fund environment is complex, the manual judgment is often greatly deteriorated, and factors such as season rate change, history, politics, period, natural disasters and the like need to be considered, so that the method based on LightGBM modeling is also provided by referring to fig. 7, and the position fund is forecasted according to time-related business data. The method specifically comprises the following steps:

a sample selection unit: besides the time, currency and position fund usage amount, the model supplements external data such as exchange rate, gold ETF fluctuation index, dollar index, dojones industry index, policy uncertainty index and Chinese certificate index, meanwhile, the external data is subjected to time period sliding window derivation processing, and finally 70 characteristic fields are selected, wherein the date is used as a sample unique identification field.

A target value definition unit: for the prediction target of the cross-border fund position amount range, five quantile intervals are divided, wherein the five quantile intervals are respectively more than 0-2130 (million), 2130-2690 (million), 2690-3220 (million), 3220-4110 (million) and 4110 (million).

A model design unit: the LightGBM model can obtain better effect than other models, and is favorable for playing better model training effect. In order to guarantee the maximum possible data volume for model training under limited data, a leave-one-out method is used for modeling during training, namely a training data set (assuming that n samples are total) is divided into a training set (n-1 samples) and a testing set (1 sample), and the training is carried out for n times. The model design firstly carries out range prediction on the position historical data, combines the model effect, and carries out regression prediction again on the categories with higher classification prediction accuracy.

A result display unit: the results are shown in Table 1. Referring to table 1, it can be seen that the prediction accuracy exceeds 80% for the fund of which the model prediction exceeds 4000 ten thousand, and the effect is good.

TABLE 1 prediction accuracy for different gears

The method for planning and predicting the fund position of the oversea bank can achieve the following technical effects: (1) The capital position checking process is optimized, the clearing account and bookkeeping information are introduced, the manual accounting process is reduced, and the accuracy and efficiency of the capital position checking are improved; (2) A universal on-line manual forecast circulation mechanism is established, and the unified planning and management of the reported funds are performed, so that the circulation efficiency among bank departments is improved; (3) The key elements of bookkeeping and manual delivery are extracted, and the fund flow is classified, so that the full-dimensional grasp of the position fund use condition is facilitated; (4) Based on machine learning, the fund change condition is predicted in advance, the fund transfer is prepared in advance by overseas institutions, and meanwhile, the prediction result can also be used for the advanced calculation of manual position forecast, so that unnecessary circulation is avoided, and the position reporting efficiency is improved.

It should be understood that, although the steps in the flowcharts related to the embodiments as described above are sequentially displayed as indicated by arrows, the steps are not necessarily performed sequentially as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the embodiments described above may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the execution order of the steps or stages is not necessarily sequential, but may be rotated or alternated with other steps or at least a part of the steps or stages in other steps.

Based on the same inventive concept, the embodiment of the application also provides a resource information prediction device for realizing the resource information prediction method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so specific limitations in one or more embodiments of the resource information prediction device provided below can be referred to the limitations of the resource information prediction method in the foregoing, and details are not described here.

In one embodiment, as shown in fig. 8, there is provided a resource information prediction apparatus, including: a feature acquisition module 801, a range prediction module 802, and an information prediction module 803, wherein:

a feature obtaining module 801, configured to, in response to a resource information prediction request for an object to be predicted, obtain resource feature information of the object to be predicted in a current time period.

The range prediction module 802 is configured to input the resource feature information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in a next time period of the current time period; and the pre-trained resource information prediction model is obtained by training according to the sample resource characteristic information of the sample object.

The information prediction module 803 is configured to input the resource feature information of the current time period and the target resource information range into a pre-trained regression prediction model, so as to obtain predicted resource information of the object to be predicted in the next time period; the predicted resource information is within the target resource information range.

In an embodiment, the information predicting module 803 is further configured to input the resource feature information into a pre-trained resource information predicting model, so as to obtain a prediction probability of each preset resource information range corresponding to a next time period of the current time period; screening out a preset resource information range with the maximum prediction probability from all the preset resource information ranges; and taking the preset resource information range with the maximum prediction probability as a target resource information range of the object to be predicted in the next time period of the current time period.

In one embodiment, the information prediction module 803 is further configured to determine a resource identifier associated with the resource feature information; screening out a target resource information prediction model matched with the resource identifier from a pre-trained resource information prediction model; and inputting the resource characteristic information into the target resource information prediction model to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period.

In one embodiment, the resource information prediction apparatus further includes a model training module, configured to obtain sample resource feature information of the sample object in the sample time period; inputting the sample resource characteristic information into a resource information prediction model to be trained to obtain a target resource information range of the sample object in the next time period of the sample time period; obtaining a loss value according to the difference between the actual resource information range and the target resource information range of the sample object in the next time period of the sample time period; and training the resource information prediction model to be trained according to the loss value to obtain a trained resource information prediction model serving as the pre-trained resource information prediction model.

In one embodiment, the resource information prediction apparatus further includes a model updating unit for obtaining feedback data for the pre-trained resource information prediction model; and updating the pre-trained resource information prediction model according to the feedback data to obtain an updated resource information prediction model.

In one embodiment, the resource information prediction apparatus further includes an information updating module, configured to obtain initial resource reporting information of the object to be predicted in the next time period; updating the initial resource reporting information of the object to be predicted in the next time period according to the predicted resource information to obtain target resource reporting information; and executing a corresponding resource submission process according to the target resource submission information.

The modules in the resource information prediction apparatus may be wholly or partially implemented by software, hardware, or a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 9. The computer apparatus includes a processor, a memory, an input/output interface, a communication interface, a display unit, and an input device. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The input/output interface of the computer device is used for exchanging information between the processor and an external device. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a resource information prediction method. The display unit of the computer device is used for forming a visual picture and can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

Those skilled in the art will appreciate that the architecture shown in fig. 9 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is further provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.

In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

In an embodiment, a computer program product is provided, comprising a computer program which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data need to comply with the relevant laws and regulations and standards of the relevant country and region.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high-density embedded nonvolatile Memory, resistive Random Access Memory (ReRAM), magnetic Random Access Memory (MRAM), ferroelectric Random Access Memory (FRAM), phase Change Memory (PCM), graphene Memory, and the like. Volatile Memory can include Random Access Memory (RAM), external cache Memory, and the like. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others. The databases referred to in various embodiments provided herein may include at least one of relational and non-relational databases. The non-relational database may include, but is not limited to, a block chain based distributed database, and the like. The processors referred to in the various embodiments provided herein may be, without limitation, general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, or the like.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, and these are all within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for predicting resource information, the method comprising:

responding to a resource information prediction request aiming at an object to be predicted, and acquiring resource characteristic information of the object to be predicted in a current time period;

inputting the resource characteristic information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period; the predicted resource information is located within the target resource information range.

2. The method according to claim 1, wherein the inputting the resource feature information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in a time period next to the current time period comprises:

3. The method according to claim 2, wherein the inputting the resource feature information into a resource information prediction model trained in advance to obtain the prediction probability of each preset resource information range corresponding to the next time period of the current time period comprises:

4. The method of claim 1, wherein the pre-trained resource information prediction model is trained by:

5. The method of claim 4, further comprising:

acquiring feedback data of the pre-trained resource information prediction model;

6. The method according to any one of claims 1 to 5, wherein after inputting the resource feature information of the current time period and the target resource information range into a pre-trained regression prediction model to obtain the predicted resource information of the object to be predicted in the next time period, the method further comprises:

and executing a corresponding resource submission flow according to the target resource submission information.

7. An apparatus for predicting resource information, the apparatus comprising:

the characteristic obtaining module is used for responding to a resource information prediction request aiming at an object to be predicted and obtaining resource characteristic information of the object to be predicted in the current time period;

the range prediction module is used for inputting the resource characteristic information into a pre-trained resource information prediction model to obtain a target resource information range of the object to be predicted in a time period next to the current time period; the pre-trained resource information prediction model is obtained by training according to sample resource characteristic information of a sample object;

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 6.

9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that the computer program realizes the steps of the method of any one of claims 1 to 6 when executed by a processor.