CN113010798A

CN113010798A - Information recommendation method, information recommendation device, electronic equipment and readable storage medium

Info

Publication number: CN113010798A
Application number: CN202110445250.6A
Authority: CN
Inventors: 魏珊珊; 陈惠梅; 戴子纬
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2021-04-23
Filing date: 2021-04-23
Publication date: 2021-06-22

Abstract

The embodiment of the disclosure provides an information recommendation method, an information recommendation device, electronic equipment and a readable storage medium. The method comprises the following steps: acquiring target financial behavior information of a target user, wherein the target financial behavior information of the target user comprises at least one of the following items: user flow information, user risk assessment information and financial product information of the target user; inputting the target financial behavior information of the target user into an information recommendation model, and outputting a recommendation score for each candidate financial information in a plurality of candidate financial information; determining target financial information from the plurality of candidate financial information according to the plurality of recommendation scores; and recommending the target financial information to the target user. The information recommendation method and device provided by the embodiment of the disclosure can be used in the financial field or other fields.

Description

Information recommendation method, information recommendation device, electronic equipment and readable storage medium

Technical Field

The disclosed embodiments relate to the technical field of artificial intelligence, and more particularly, to an information recommendation method, an information recommendation apparatus, an electronic device, a computer-readable storage medium, and a computer program product.

Background

With the rapid development of the financial industry, at present, financial institutions recommend some financial information to be promoted to each user.

In implementing the disclosed concept, the inventors found that there are at least the following problems in the related art, with which it is difficult to implement personalized financial information recommendation.

Disclosure of Invention

In view of the above, the disclosed embodiments provide an information recommendation method, an information recommendation apparatus, an electronic device, a computer-readable storage medium, and a computer program product.

One aspect of the embodiments of the present disclosure provides an information recommendation method, including: acquiring target financial behavior information of a target user, wherein the target financial behavior information of the target user comprises at least one of the following items: the user flow information, the user risk assessment information and the financial product information of the target user; inputting the target financial behavior information of the target user into an information recommendation model, and outputting a recommendation score for each candidate financial information in a plurality of candidate financial information; determining target financial information from the candidate financial information according to the plurality of recommendation scores; and recommending the target financial information to the target user.

According to an embodiment of the present disclosure, the acquiring target financial behavior information of the target user includes: acquiring original financial behavior information of the target user; determining user sensitive information related to user privacy in the original financial behavior information; processing the user sensitive information by using a data desensitization method to obtain processed user sensitive information; and determining the target financial behavior information according to the original financial behavior information and the processed user sensitive information.

According to the embodiment of the disclosure, the information recommendation model comprises a click rate model and a profit model, wherein the click rate model is obtained by joint training of a first long-short term memory neural network model, a first deep neural network model and a first neural network model, and the profit model is obtained by joint training of a second long-short term memory neural network model, a second deep neural network model and a second neural network model.

According to an embodiment of the present disclosure, the inputting the target financial behavior information of the target user into the information recommendation model and outputting the recommendation score for each of the plurality of candidate financial information includes: inputting the target financial behavior information of the target user into the click rate model, and outputting a click rate for each of the plurality of candidate financial information; inputting the target financial behavior information of the target user into the profit model, and outputting a benefit for each of the plurality of candidate financial information; and determining a recommendation score for each of the plurality of candidate financial information according to the click rate and the benefit of the candidate financial information.

According to an embodiment of the present disclosure, the click rate model is obtained by joint training of a first long-short term memory neural network model, a first deep neural network model, and a first neural network model, and includes: obtaining a first training sample, wherein the first training sample comprises a plurality of positive samples and a plurality of negative samples, each positive sample comprises target financial behavior information of a sample user and a real click rate of the sample user for a sample product, each negative sample comprises target training financial behavior information of the sample user and real display information of the sample user for the sample product, and the target training financial behavior information of the sample user comprises user running information, user risk assessment information and financial product information of the sample user; inputting user flow information of a sample user included in each of the plurality of positive samples into the first long-short term memory neural network model, and outputting a first vector; inputting user risk assessment information and financial product information of a sample user included in each of the plurality of positive samples into the first deep neural network model, and outputting a second vector; splicing the second vector and the second vector to obtain a third vector; inputting the third vector into the first neural network model, and outputting a predicted click rate of the sample user included in the positive sample for the sample product; inputting the user flow information of the sample user included in each negative sample in the plurality of negative samples into the first long-short term memory neural network model, and outputting a fourth vector; inputting user risk assessment information and financial product information of the sample user included in each negative sample in the negative samples into the first deep neural network model, and outputting a fifth vector; splicing the fourth vector and the fifth vector to obtain a sixth vector; inputting the sixth vector into the first neural network model, and outputting the predicted display information of the sample user included in the negative sample for the sample product; and training the first long-short term memory neural network model, the first deep neural network model and the first neural network model according to the real click rates, the predicted click rates, the real display information and the predicted display information to obtain the click rate model.

According to an embodiment of the present disclosure, the revenue model is obtained by joint training using a second long-short term memory neural network model, a second deep neural network model, and a second neural network model, and includes: acquiring a second training sample, wherein the second training sample comprises a plurality of training groups, each training group comprises target financial behavior information of the sample user and real benefit of the sample user for a sample product, and the target training financial behavior information of the sample user comprises user running information, user risk assessment information and financial product information of the sample user; inputting the user flow information of the sample users included in each of the plurality of training sets into the second long-short term memory neural network model, and outputting a seventh vector; inputting the user risk assessment information and the financial product information of the sample user included in each of the plurality of training groups into the second deep neural network model, and outputting an eighth vector; splicing the seventh vector and the eighth vector to obtain a ninth vector; inputting the ninth vector into the second neural network model, and outputting the predicted benefit of the sample user included in the training set for the sample product; and training the second long-short term memory neural network model, the second deep neural network model and the second neural network model according to the plurality of real benefits and the plurality of predicted benefits to obtain the profit model.

According to an embodiment of the present disclosure, the determining target financial information from the plurality of candidate financial information according to the plurality of recommendation scores includes: sorting the plurality of recommendation scores to obtain a sorting result; and determining the target financial information from the candidate financial information according to the sorting result.

According to an embodiment of the present disclosure, the user pipelining information includes at least one of: transaction time, transaction amount and transaction property of the transaction within a preset time period; the financial product information includes at least one of: risk rating of financial products, fund manager, asset size, buy cycle and redeem cycle.

According to the embodiment of the disclosure, the information recommendation model is obtained by training through a federal learning framework.

Another aspect of the disclosed embodiments provides an information recommendation apparatus, including an obtaining module, an output module, a determining module, and a recommending module. The acquisition module is configured to acquire target financial behavior information of a target user, where the target financial behavior information of the target user includes at least one of the following: the user flow information, the user risk assessment information and the financial product information of the target user; an output module, configured to input the target financial behavior information of the target user into an information recommendation model, and output a recommendation score for each of a plurality of candidate financial information; a determining module for determining target financial information from the candidate financial information according to the plurality of recommendation scores; and a recommending module for recommending the target financial information to the target user.

Another aspect of an embodiment of the present disclosure provides an electronic device including: one or more processors; 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 as described above.

Another aspect of the embodiments of the present disclosure provides a computer-readable storage medium storing computer-executable instructions for implementing the method as described above when executed.

Another aspect of an embodiment of the present disclosure provides a computer program product comprising computer executable instructions for implementing the method as described above when executed.

According to the embodiment of the disclosure, the recommendation score of each candidate financial information in the plurality of candidate financial information is obtained by processing the target financial behavior of the target user by using the information recommendation model, and the target financial information recommended to the target user is determined according to the plurality of recommendation scores. The target financial behavior information of the target user can reflect the personalized financial information of the user, so the target financial behavior information of the target user is processed by using the information recommendation model to obtain the recommendation score of each candidate financial information in the candidate financial information, the target financial information recommended to the target user is determined according to the recommendation scores, personalized recommendation is realized, and the target financial behavior information is processed by using the information recommendation model, so the recommendation accuracy is improved, and the user requirements are met.

Drawings

The above and other objects, features and advantages of the present disclosure will become more apparent from the following description of embodiments of the present disclosure with reference to the accompanying drawings, in which:

FIG. 1 schematically illustrates an exemplary system architecture 100 to which the information recommendation method may be applied, according to an embodiment of the disclosure;

FIG. 2 schematically illustrates a flow diagram of an information recommendation method 200 according to an embodiment of the disclosure;

FIG. 3 schematically illustrates a schematic diagram of an information recommendation model training process according to an embodiment of the present disclosure;

fig. 4 schematically shows a block diagram of an information recommendation device 400 according to an embodiment of the present disclosure; and

fig. 5 schematically shows a block diagram of an electronic device 500 adapted to implement the above described method according to an embodiment of the present disclosure.

Detailed Description

Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is illustrative only and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present disclosure.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.

Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B and C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). Where a convention analogous to "A, B or at least one of C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B or C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.).

In the process of realizing the concept disclosed by the invention, the inventor finds that the financial information is recommended mostly in a manual mode in the related technology, the same financial information is recommended to different target users, and personalized recommendation is difficult to realize. In addition, the method for realizing financial information recommendation by using the model in the related art is often realized based on static user information, so that the requirement change of the user on the financial information is ignored, the accuracy of user recommendation is influenced, and the requirement of the user is difficult to meet.

In order to solve the problems in the related art, the inventor proposes a scheme for determining target financial information to be recommended by processing the target financial behavior information of a target user by using an information recommendation model, wherein the target financial behavior information of the target user comprises at least one of user flow information, user risk assessment information and financial product information of the target user. The target financial behavior information of the target user can reflect the personalized financial information of the user, so that the target financial behavior information of the target user is processed by using the information recommendation model, the target financial information recommended to the target user is determined, the personalized recommendation is realized, and the target financial behavior information is processed by using the information recommendation model, so that the recommendation accuracy is improved, and the user requirements are further met.

Specifically, the embodiment of the disclosure provides an information recommendation method, an information recommendation device, an electronic device and a readable storage medium. The method comprises an acquisition process, an output process, a determination process and a recommendation process. In the obtaining process, obtaining target financial behavior information of a target user, where the target financial behavior information of the target user includes at least one of: user flow information, user risk assessment information and financial product information of the target user; in the output process, inputting the target financial behavior information of the target user into an information recommendation model, and outputting a recommendation score for each candidate financial information in a plurality of candidate financial information; in the determining process, determining target financial information from a plurality of candidate financial information according to a plurality of recommendation scores; in the recommendation process, target financial information is recommended to the target user.

It should be noted that the information recommendation method and apparatus provided by the embodiments of the present disclosure may be applied to the field of finance, for example, for bank staff to recommend financial products to customers; in addition, the information recommendation method and apparatus provided by the embodiment of the disclosure can also be applied to other fields except the financial field, for example, in the field of internet technology, an information provider can perform appropriate content push according to the behavior of a user, so as to improve the user retention rate of the information provider. The application fields of the information recommendation method and the information recommendation device provided by the embodiment of the disclosure are not limited.

Fig. 1 schematically shows an exemplary system architecture 100 to which the information recommendation method may be applied, according to an embodiment of the present disclosure. It should be noted that fig. 1 is only an example of a system architecture to which the embodiments of the present disclosure may be applied to help those skilled in the art understand the technical content of the present disclosure, and does not mean that the embodiments of the present disclosure may not be applied to other devices, systems, environments or scenarios.

As shown in fig. 1, the system architecture 100 according to this embodiment may include

terminal devices

101, 102, 103, a network 104 and a server 105. The network 104 serves as a medium for providing communication links between the

terminal devices

101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired and/or wireless communication links, and so forth.

The account may use the

terminal devices

101, 102, 103 to interact with the server 105 over the network 104 to receive or send messages or the like. The

terminal devices

101, 102, 103 may have installed thereon various communication client applications, such as a shopping-like application, a web browser application, a search-like application, an instant messaging tool, a mailbox client, and/or social platform software, etc. (by way of example only).

The

terminal devices

101, 102, 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, and the like.

The server 105 may be a server providing various services, such as a background management server (for example only) providing support for websites browsed by the accounts using the

terminal devices

101, 102, 103. The background management server may analyze and perform other processing on the received data such as the account request, and feed back a processing result (e.g., a webpage, information, or data obtained or generated according to the account request) to the terminal device.

It should be noted that the information recommendation method provided by the embodiment of the present disclosure may be generally executed by the server 105. Accordingly, the information recommendation device provided by the embodiment of the present disclosure may be generally disposed in the server 105. The information recommendation method provided by the embodiment of the present disclosure may also be executed by a server or a server cluster different from the server 105 and capable of communicating with the

terminal devices

101, 102, 103 and/or the server 105. Accordingly, the information recommendation device provided by the embodiment of the present disclosure may also be disposed in a server or a server cluster different from the server 105 and capable of communicating with the

terminal devices

101, 102, 103 and/or the server 105. Alternatively, the information recommendation method provided by the embodiment of the present disclosure may be executed by the

terminal device

101, 102, or 103, or may be executed by another terminal device different from the

terminal device

101, 102, or 103. Accordingly, the information recommendation device provided by the embodiment of the present disclosure may also be disposed in the

terminal device

101, 102, or 103, or in another terminal device different from the

terminal device

101, 102, or 103.

For example, the target financial behavior information of the target user may be originally stored in any one of the

terminal devices

101, 102, or 103 (e.g., the terminal device 101, but not limited thereto), or stored on an external storage device and may be imported into the terminal device 101. Then, the terminal device 101 may locally execute the information recommendation method provided by the embodiment of the present disclosure, or transmit the target financial behavior information to another terminal device, a server, or a server cluster, and execute the information recommendation method provided by the embodiment of the present disclosure by another terminal device, a server, or a server cluster that receives the target financial behavior information.

It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

Fig. 2 schematically shows a flow chart of an information recommendation method 200 according to an embodiment of the present disclosure.

As shown in FIG. 2, the method 200 includes operations S210-S240.

In operation S210, target financial behavior information of a target user is acquired.

According to an embodiment of the present disclosure, the target financial behavior information of the target user may refer to information that the target user generates in daily transaction behaviors and is recorded by the database, and may include one or a combination of more of user flow information, user risk assessment information, and financial product information.

According to the embodiment of the disclosure, the target financial behavior information can be extracted from the database by signing a private information use protocol with the target user and after soliciting the consent of the target user.

It should be noted that, in the technical solution of the present disclosure, the acquisition, storage, application, and the like of the personal information of the related user all conform to the regulations of the relevant laws and regulations, and do not violate the customs of the public order.

In operation S220, the target financial behavior information of the target user is input to the information recommendation model, and a recommendation score for each of the plurality of candidate financial information is output.

According to an embodiment of the disclosure, the information recommendation model may be trained by using historical financial behavior information of a plurality of users as training samples and actual financial information selection of the plurality of users as labels based on an existing model architecture. The number of the neurons of the output layer of the information recommendation model is larger than that of the candidate financial information, so that the expansibility of the information recommendation model is realized.

According to embodiments of the present disclosure, the financial information may refer to a specific financial product, such as a fund product issued by a fund issuer entrusted bank; on the other hand, financial information may also refer to financial services, such as investment service recommendations for credit, stocks, etc.; in addition, the financial information may also include news, science popularization articles, product introduction articles, videos, and the like related to finance.

According to an embodiment of the present disclosure, the recommendation score of the candidate financial information may represent a degree of matching between the target user and the candidate financial information based on the financial behavior information analysis, and the higher the recommendation score is, the greater the probability that the target user selects the candidate financial information is.

In operation S230, target financial information is determined from the plurality of candidate financial information according to the plurality of recommendation scores.

In operation S240, target financial information is recommended to a target user.

The method of fig. 2 is further described with reference to fig. 3 in conjunction with specific embodiments.

According to the embodiment of the disclosure, the user flow information includes information generated by a target user performing financial behaviors through various channels within a preset time range. Specifically, the user flow information may include information such as transaction time, transaction amount, transaction nature, and the like. For example, the user flow information may include information on the time and amount of stock trading performed by the user within 180 days.

According to the embodiment of the disclosure, the user risk assessment information may be determined by the product provider based on the historical financial behavior information of the target user, for example, a bank may determine the risk resistance of the target user by inquiring the credit information of the target user, so as to determine the risk assessment information of the target user. In addition, the user risk assessment information may also be obtained by the user through a self-testing means, for example, the target user may determine the risk assessment information of the target user by completing a questionnaire test according to the personality tendency to obtain an investment tendency.

According to an embodiment of the present disclosure, the financial product information may include a risk level of the financial product, a fund manager, an asset size, a buy cycle, a redeem cycle, and the like.

According to the embodiment of the disclosure, the acquired financial behavior information of the target user may include sensitive information related to privacy or confidentiality of the target user, for example, transaction amount information in the user flow information. Since the direct use of sensitive information for financial information recommendation or training of an information recommendation model may bring information security risks, in the embodiment of the present disclosure, a data desensitization method is adopted in the acquisition process of financial behavior information.

According to an embodiment of the present disclosure, acquiring target financial behavior information of a target user may include the following operations.

Acquiring original financial behavior information of a target user; determining user sensitive information related to user privacy in the original financial behavior information; processing the user sensitive information by using a data desensitization method to obtain the processed user sensitive information; and determining target financial behavior information according to the original financial behavior information and the processed user sensitive information.

For example, for the transaction amount in the original financial behavior information, the operation of taking the logarithm and rounding the transaction amount can be performed, as shown in formula (1):

Y′＝[log Y] (1)

in the equation, Y represents the original transaction amount, Y' represents the desensitized transaction amount, and [. cndot. ] represents rounding down.

It should be noted that the desensitization method shown in formula (1) is only an exemplary embodiment, and on the basis of ensuring that the transaction amount after desensitization still maintains the partial ordering relationship consistent with the original transaction amount, any mathematical method can be used for data desensitization, for example, before the original transaction amount is subjected to the logarithm operation, the original transaction amount can be subjected to linear transformation.

According to the embodiment of the disclosure, by using the data desensitization method, on one hand, the transaction amount after desensitization cannot be directly and accurately reduced to the original transaction amount, so that the information security of the user is effectively ensured, and the problem of information leakage possibly occurring in the use process of the information recommendation model is effectively avoided. On the other hand, because the distribution between the original transaction amount and the transaction amount is usually power exponential distribution, which is not beneficial to model learning, the distribution between the desensitized transaction amount and the transaction amount can be close to normal distribution by using a data desensitization method for the original transaction amount, which is beneficial to improving the accuracy of the information recommendation model.

FIG. 3 schematically shows a schematic diagram of an information recommendation model training process according to an embodiment of the disclosure.

As shown in fig. 3, the information recommendation model may be obtained by writing model training codes based on an open source framework such as tensorflow, and running a training task on a training cluster deployed with the open source framework by using desensitized target financial behavior information 310 and tag information 320 of a sample user, where the training cluster may be deployed on a private cloud or a secure computer to avoid leakage of user information.

According to an embodiment of the present disclosure, the desensitized target financial behavior information 310 of the sample user includes user pipelining information 311, user risk assessment information 312, and financial product information 313 of the sample user; the tag information 320 includes a real click through rate 321, real presentation information 322, and real benefit 323.

In addition, the information recommendation model of the embodiment of the disclosure can be jointly modeled with other organizations through a federal learning framework, and by a federal transfer learning method, the distributed data of the user in each organization can be fully utilized under the condition of not revealing privacy, so that more accurate information recommendation service is improved for the user.

According to an embodiment of the present disclosure, the information recommendation model includes a click-through rate model 330 and a revenue model 340.

According to an embodiment of the present disclosure, the click rate model 330 may be obtained by joint training using the first long-short term memory neural network model 331, the first deep neural network model 332, and the first neural network model 333.

According to an embodiment of the present disclosure, the first training samples used in training click-through rate model 330 may include a plurality of positive samples and a plurality of negative samples. Wherein, the target financial behavior information 310 can be included in each positive sample, and the real click rate 321 of the sample user for the sample product is used as a label; target financial behavior information 310 may be included in each negative example and tagged with the actual display information 322 of the example user for the example product.

According to an embodiment of the disclosure, each iteration of training click-through rate model 330 is as follows:

first, a plurality of positive samples in the first training sample are input into the click rate model 330, so as to obtain a predicted click rate.

According to an embodiment of the present disclosure, the user running information 311 of the sample user included in each of the plurality of positive samples may be input into the first long-short term memory neural network model 331, and the user risk assessment information 312 and the financial product information 313 of the sample user included in each of the plurality of positive samples may be input into the first deep neural network model 332.

According to the embodiment of the present disclosure, a third vector may be obtained by stitching the first vector output by the first long-short term memory neural network model 331 and the second vector output by the first deep neural network model 332.

According to an embodiment of the present disclosure, inputting the third vector into the first neural network model 333 may result in a predicted click rate of the sample user for the sample product included in the output positive sample.

Then, a plurality of negative samples in the first training sample are input into the click rate model 330, so as to obtain the predicted display information.

According to the embodiment of the present disclosure, the user flow information 311 of the sample user included in each negative sample of the plurality of negative samples may be input into the first long-short term memory neural network model 331, and the user risk assessment information 312 and the financial product information 313 of the sample user included in each negative sample of the plurality of negative samples may be input into the first deep neural network model 332.

According to the embodiment of the disclosure, a sixth vector may be obtained by stitching the fourth vector output by the first long-short term memory neural network model 331 and the fifth vector output by the first deep neural network model 332.

According to the embodiment of the present disclosure, inputting the sixth vector into the first neural network model 333 may result in the predicted display information of the sample user for the sample product, which is included in the output negative sample.

Finally, the alternation of the model parameters of the click rate model 330 is accomplished using a loss function.

According to an embodiment of the disclosure, the plurality of real click-through rates 321 and the plurality of predicted click-through rates, and the plurality of real presentation information 322 and the plurality of predicted presentation information may be sequentially processed by using the first loss function 334 to obtain a loss value for completing the alternation of the model parameters.

According to the embodiment of the present disclosure, the network model may be trained by alternately inputting the positive sample and the negative sample into the first long-short term memory neural network model, the first deep neural network model and the first neural network model, so as to obtain the click rate model 330.

According to the embodiment of the disclosure, the network model is trained by inputting the positive sample and the negative sample into the first long-short term memory neural network model, the first deep neural network model and the first neural network model, so that the network model can learn the characteristics of the positive sample and the negative sample in a contrast manner. Because the characteristics of the positive sample and the negative sample can be compared and learned, the effective identification of the positive sample is realized, and the prediction precision of the click rate model is further ensured. In addition, the positive samples and the negative samples are alternately input into the network model to train the network model, so that the model training speed is accelerated.

According to an embodiment of the present disclosure, the profit model 340 may be jointly trained using the second long-short term memory neural network model 341, the second deep neural network model 342, and the second neural network model 343.

According to an embodiment of the present disclosure, the second training samples used in training the revenue model 340 include a plurality of training sets. Wherein each training set includes target financial behavior information 310, using the actual benefit 323 of the sample user for the sample product as a label.

In accordance with an embodiment of the present disclosure, each iteration of training the revenue model 340 is as follows:

first, a plurality of training sets in the second training sample are input into the profit model 340 to obtain the predicted profit.

According to an embodiment of the present disclosure, the user flow information 311 of the sample users included in each training set in the second training sample may be input into the second long-short term memory neural network model 341, resulting in an output seventh vector.

According to an embodiment of the present disclosure, the user risk assessment information 312 and the financial product information 313 of the sample user included in each training set in the second training sample may be input into the first deep neural network model 342, resulting in an output eighth vector.

According to the embodiment of the present disclosure, the ninth vector obtained by splicing the seventh vector and the eighth vector is input into the second neural network model 343, and the prediction benefit of the sample user included in the output training set for the sample product can be obtained.

Thereafter, the alternation of the model parameters of the revenue model 340 is accomplished using a loss function.

According to an embodiment of the present disclosure, the plurality of real benefits 323 and the plurality of predicted benefits may be processed using the second loss function 344, resulting in a loss value for completing the alternation of the model parameters.

The first and

second loss functions

334, 344 may be, for example, log loss functions, square loss functions, hubor loss functions, or the like, in accordance with embodiments of the present disclosure. The same loss function may be used for the first and

second loss functions

334 and 344, or different loss functions may be used.

According to an embodiment of the present disclosure, during the training process of the click rate model 330 and the profit model 340, a gradient descent method, a random gradient descent method, or the like may be used for the alternation of the model parameters.

According to an embodiment of the present disclosure, an end condition of the model training may be preset in the process of training the click rate model 330 and the profit model 340. For example, the training may be stopped when the loss value of the loss function is smaller than a preset value in each iteration, or the training may be stopped when the gradient of the loss function is smaller than a preset value in each iteration, or the training may be stopped after the number of times of training reaches a preset value.

According to an embodiment of the disclosure, after the training of the click rate model 330 and the profit model 340 is completed, the target financial behavior information 310 of the target user may be input into the click rate model 330 and the profit model 340, respectively, resulting in the output click rate 351 and benefit 352 for each candidate financial information of the plurality of candidate financial information. From the click through rate 351 and the benefit 352 of each candidate financial information, a recommendation score 353 for that candidate financial information may be determined, for example, the calculation of the recommendation score may be performed according to the method of equation (2):

score＝ctr×value (2)

in the formula, score represents the recommendation score of the candidate financial information, ctr represents the click rate of the candidate financial information, and value represents the benefit of the candidate financial information.

According to the embodiment of the disclosure, the click rate of the candidate financial information represents the use probability of the user after receiving the candidate financial information, and the benefit of the candidate financial information represents the benefit recommended by the information provider based on the financial information. The form of the click rate and benefit of the financial information varies according to the specific form of the financial information. For example, if the candidate financial information represents a specific financial product, the click rate of the candidate financial information represents the probability that the user selects the financial product after receiving the recommendation of the financial product; the benefit of the candidate financial information represents an amount of the deposit of the user after receiving the recommendation for the financial product.

According to the embodiments of the present disclosure, according to the recommendation scores 353 of the plurality of candidate financial information, the target financial information may be determined and recommended to the target user. For example, the plurality of recommendation scores may be ranked, and the candidate financial information having the highest score may be selected as the target financial information from the ranking result. Furthermore, the target financial information may be determined using a classifier, for example, a softmax classifier.

According to the financial information recommendation method for processing financial behavior information based on deep learning, because the financial behavior information comprises user flow information, user risk evaluation information and financial product information of a target user, the information can reflect changes of user risk bearing capacity, cash flow state and investment financing habits, the method can automatically learn the changes of the user risk bearing capacity, the cash flow state and the investment financing habits, and therefore, a more personalized, higher-order and more suitable financial information recommendation service can be provided for the user, the cost for different users to obtain suitable financial information is reduced, the requirements of the user are met, and the use experience of the user is improved.

Fig. 4 schematically shows a block diagram of an information recommendation apparatus 400 according to an embodiment of the present disclosure.

As shown in fig. 4, the information recommendation apparatus 400 includes an acquisition module 410, an output module 420, a determination module 430, and a recommendation module 440.

An obtaining module 410, configured to obtain target financial behavior information of a target user, where the target financial behavior information of the target user includes at least one of: user flow information, user risk assessment information, and financial product information for the target user.

An output module 420, configured to input the target financial behavior information of the target user into the information recommendation model, and output a recommendation score for each candidate financial information of the plurality of candidate financial information.

A determining module 430, configured to determine the target financial information from the plurality of candidate financial information according to the plurality of recommendation scores.

And a recommending module 440 for recommending the target financial information to the target user.

According to an embodiment of the present disclosure, the obtaining module 410 includes a first obtaining unit, a second obtaining unit, a third obtaining unit, and a fourth obtaining unit. Wherein:

the system comprises a first acquisition unit, a second acquisition unit and a processing unit, wherein the first acquisition unit is used for acquiring original financial behavior information of a user;

the second acquisition unit is used for determining user sensitive information related to user privacy in the original financial behavior information;

the third acquisition unit is used for processing the user sensitive information by using a data desensitization method to obtain the processed user sensitive information; and

and the fourth acquisition unit is used for determining the target financial behavior information according to the original financial behavior information and the processed user sensitive information.

According to the embodiment of the disclosure, the information recommendation model comprises a click rate model and a profit model, and the information recommendation device further comprises a first training module and a second training module, wherein:

the first training module is used for obtaining a click rate model by utilizing the joint training of the first long-short term memory neural network model, the first deep neural network model and the first neural network model;

and the first training module is used for obtaining a profit model by utilizing the combined training of the second long-short term memory neural network model, the second deep neural network model and the second neural network model.

According to an embodiment of the present disclosure, the output module 420 includes a first output unit, a second output unit, and a third output unit. Wherein:

a first output unit for inputting target financial behavior information of a target user into the click rate model, and outputting a click rate for each of a plurality of candidate financial information;

a second output unit for inputting the target financial behavior information of the target user into the profit model, and outputting a profit for each of the plurality of candidate financial information; and

and a third output unit, configured to determine a recommendation score for each of the plurality of candidate financial information according to the click rate and the benefit of the candidate financial information.

According to an embodiment of the present disclosure, the first training module includes a first training unit, a second training unit, a third training unit, a fourth training unit, a fifth training unit, a sixth training unit, a seventh training unit, an eighth training unit, a ninth training unit, and a tenth training unit.

Wherein:

the system comprises a first training unit, a second training unit and a third training unit, wherein the first training sample comprises a plurality of positive samples and a plurality of negative samples, each positive sample comprises target financial behavior information of a sample user and a real click rate of the sample user for a sample product, each negative sample comprises target training financial behavior information of the sample user and real display information of the sample user for the sample product, and the target training financial behavior information of the sample user comprises user flow information, user risk assessment information and financial product information of the sample user;

the second training unit is used for inputting user flow information of sample users included in each positive sample in the plurality of positive samples into the first long-short term memory neural network model and outputting a first vector;

the third training unit is used for inputting the user risk assessment information and the financial product information of the sample user included in each positive sample in the plurality of positive samples into the first deep neural network model and outputting a second vector;

the fourth training unit is used for splicing the second vector and the second vector to obtain a third vector;

the fifth training unit is used for inputting the third vector into the first neural network model and outputting the predicted click rate of the sample user included in the positive sample for the sample product;

the sixth training unit is used for inputting the user flow information of the sample user included by each negative sample in the negative samples into the first long-short term memory neural network model and outputting a fourth vector;

the seventh training unit is used for inputting the user risk assessment information and the financial product information of the sample user, which are included in each negative sample in the negative samples, into the first deep neural network model and outputting a fifth vector;

the eighth training unit splices the fourth vector and the fifth vector to obtain a sixth vector;

the ninth training unit is used for inputting the sixth vector into the first neural network model and outputting the prediction display information of the sample user aiming at the sample product, wherein the prediction display information comprises the negative sample; and

and the tenth training unit is used for training the first long-short term memory neural network model, the first deep neural network model and the first neural network model according to the plurality of real click rates, the plurality of predicted click rates, the plurality of real display information and the plurality of predicted display information to obtain a click rate model.

According to an embodiment of the present disclosure, the second training module includes an eleventh training unit, a twelfth training unit, a thirteenth training unit, a fourteenth training unit, a fifteenth training unit, and a sixteenth training unit. Wherein:

the eleventh training unit is used for acquiring a second training sample, wherein the second training sample comprises a plurality of training groups, each training group comprises target financial behavior information of a sample user and real benefit of the sample user for a sample product, and the target training financial behavior information of the sample user comprises user running information, user risk assessment information and financial product information of the sample user;

a twelfth training unit, configured to input user flow information of the sample users included in each of the multiple training sets into the second long-short term memory neural network model, and output a seventh vector;

a thirteenth training unit, configured to input, into the second deep neural network model, user risk assessment information and financial product information of the sample user included in each of the plurality of training sets, and output an eighth vector;

the fourteenth training unit is used for splicing the seventh vector and the eighth vector to obtain a ninth vector;

a fifteenth training unit, configured to input the ninth vector into the second neural network model, and output a prediction benefit of the sample user included in the training set for the sample product; and

and the sixteenth training unit is used for training the second long-short term memory neural network model, the second deep neural network model and the second neural network model according to the plurality of real benefits and the plurality of predicted benefits to obtain a profit model.

According to an embodiment of the present disclosure, the determination module 430 includes a first determination unit and a second determination unit. Wherein:

the first determining unit is used for sequencing the recommendation scores to obtain a sequencing result; and

and a second determining unit for determining the target financial information from the plurality of candidate financial information according to the sorting result.

According to an embodiment of the present disclosure, the first training module and the second training module utilize a federated learning framework in training the information recommendation model.

Any number of modules, sub-modules, units, sub-units, or at least part of the functionality of any number thereof according to embodiments of the present disclosure may be implemented in one module. Any one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure may be implemented by being split into a plurality of modules. Any one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure may be implemented at least partially as a hardware Circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented by hardware or firmware in any other reasonable manner of integrating or packaging a Circuit, or implemented by any one of three implementations of software, hardware, and firmware, or any suitable combination of any of them. Alternatively, one or more of the modules, sub-modules, units, sub-units according to embodiments of the disclosure may be at least partially implemented as a computer program module, which when executed may perform the corresponding functions.

For example, any plurality of the obtaining module 410, the outputting module 420, the determining module 430, and the recommending module 440 may be combined and implemented in one module/unit/sub-unit, or any one of the modules/units/sub-units may be split into a plurality of modules/units/sub-units. Alternatively, at least part of the functionality of one or more of these modules/units/sub-units may be combined with at least part of the functionality of other modules/units/sub-units and implemented in one module/unit/sub-unit. According to an embodiment of the disclosure, at least one of the obtaining module 410, the outputting module 420, the determining module 430, and the recommending module 440 may be implemented at least partially as a hardware circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented in hardware or firmware in any other reasonable manner of integrating or packaging a circuit, or in any one of three implementations of software, hardware, and firmware, or in any suitable combination of any of them. Alternatively, at least one of the obtaining module 410, the outputting module 420, the determining module 430, and the recommending module 440 may be at least partially implemented as a computer program module, which when executed, may perform a corresponding function.

According to the embodiments of the present disclosure, the information recommendation device part in the embodiments of the present disclosure corresponds to the information recommendation method part in the embodiments of the present disclosure, and the description of the information recommendation device part specifically refers to the information recommendation method part, which is not described herein again.

Fig. 5 schematically shows a block diagram of an electronic device 500 adapted to implement the above described method according to an embodiment of the present disclosure. The electronic device shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 5, an electronic device 500 according to an embodiment of the present disclosure includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a Read-Only Memory (ROM) 502 or a program loaded from a storage section 508 into a Random Access Memory (RAM) 503. The processor 501 may comprise, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or associated chipset, and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), among others. The processor 501 may also include onboard memory for caching purposes. Processor 501 may include a single processing unit or multiple processing units for performing different actions of a method flow according to embodiments of the disclosure.

In the RAM 503, various programs and data necessary for the operation of the electronic apparatus 500 are stored. The processor 501, the ROM502, and the RAM 503 are connected to each other by a bus 504. The processor 501 performs various operations of the method flows according to the embodiments of the present disclosure by executing programs in the ROM502 and/or the RAM 503. Note that the programs may also be stored in one or more memories other than the ROM502 and the RAM 503. The processor 501 may also perform various operations of method flows according to embodiments of the present disclosure by executing programs stored in the one or more memories.

According to an embodiment of the present disclosure, electronic device 500 may also include an input/output (I/O) interface 505, input/output (I/O) interface 505 also being connected to bus 504. The electronic device 500 may also include one or more of the following components connected to the I/O interface 505: an input portion 506 including a keyboard, a mouse, and the like; an output portion 507 including a Display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and a speaker; a storage portion 508 including a hard disk and the like; and a communication section 509 including a network interface card such as a LAN card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. The driver 510 is also connected to the I/O interface 505 as necessary. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as necessary, so that a computer program read out therefrom is mounted into the storage section 508 as necessary.

According to embodiments of the present disclosure, method flows according to embodiments of the present disclosure may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable storage medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 509, and/or installed from the removable medium 511. The computer program, when executed by the processor 501, performs the above-described functions defined in the system of the embodiments of the present disclosure. The systems, devices, apparatuses, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the present disclosure.

The present disclosure also provides a computer-readable storage medium, which may be contained in the apparatus/device/system described in the above embodiments; or may exist separately and not be assembled into the device/apparatus/system. The computer-readable storage medium carries one or more programs which, when executed, implement the method according to an embodiment of the disclosure.

According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. Examples may include, but are not limited to: a portable Computer diskette, a hard disk, a Random Access Memory (RAM), a Read-Only Memory (ROM), an Erasable Programmable Read-Only Memory (EPROM) or flash Memory), a portable compact Disc Read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the preceding. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

For example, according to embodiments of the present disclosure, a computer-readable storage medium may include ROM502 and/or RAM 503 and/or one or more memories other than ROM502 and RAM 503 described above.

Embodiments of the present disclosure also include a computer program product comprising a computer program containing program code for performing the method provided by the embodiments of the present disclosure, when the computer program product is run on an electronic device, the program code being adapted to cause the electronic device to carry out the information recommendation method provided by the embodiments of the present disclosure.

The computer program, when executed by the processor 501, performs the above-described functions defined in the system/apparatus of the embodiments of the present disclosure. The systems, apparatuses, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the present disclosure.

In one embodiment, the computer program may be hosted on a tangible storage medium such as an optical storage device, a magnetic storage device, or the like. In another embodiment, the computer program may also be transmitted, distributed in the form of a signal on a network medium, downloaded and installed through the communication section 509, and/or installed from the removable medium 511. The computer program containing program code may be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.

In accordance with embodiments of the present disclosure, program code for executing computer programs provided by embodiments of the present disclosure may be written in any combination of one or more programming languages, and in particular, these computer programs may be implemented using high level procedural and/or object oriented programming languages, and/or assembly/machine languages. The programming language includes, but is not limited to, programming languages such as Java, C + +, python, the "C" language, or the like. The program code may execute entirely on the account computing device, partly on the account device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the account computing device through any kind of Network, including a Local Area Network (LAN) or Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not expressly recited in the present disclosure. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present disclosure may be made without departing from the spirit or teaching of the present disclosure. All such combinations and/or associations are within the scope of the present disclosure.

The embodiments of the present disclosure have been described above. However, these examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described separately above, this does not mean that the measures in the embodiments cannot be used in advantageous combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be devised by those skilled in the art without departing from the scope of the present disclosure, and such alternatives and modifications are intended to be within the scope of the present disclosure.

Claims

1. An information recommendation method, comprising:

acquiring target financial behavior information of a target user, wherein the target financial behavior information of the target user comprises at least one of the following items: user flow information, user risk assessment information and financial product information of the target user;

inputting the target financial behavior information of the target user into an information recommendation model, and outputting a recommendation score for each candidate financial information in a plurality of candidate financial information;

determining target financial information from the plurality of candidate financial information based on a plurality of the recommendation scores; and

recommending the target financial information to the target user.

2. The method of claim 1, wherein the obtaining target financial behavior information of the target user comprises:

acquiring original financial behavior information of the target user;

determining user sensitive information related to user privacy in the original financial behavior information;

processing the user sensitive information by using a data desensitization method to obtain the processed user sensitive information; and

and determining the target financial behavior information according to the original financial behavior information and the processed user sensitive information.

3. The method of claim 1, wherein the information recommendation model comprises a click-through rate model and a profit model, wherein the click-through rate model is jointly trained using a first long-short term memory neural network model, a first deep neural network model, and a first neural network model, and the profit model is jointly trained using a second long-short term memory neural network model, a second deep neural network model, and a second neural network model.

4. The method of claim 3, wherein the inputting the target financial behavior information of the target user into an information recommendation model, outputting a recommendation score for each of a plurality of candidate financial information, comprises:

inputting target financial behavior information of the target user into the click-through rate model, and outputting a click-through rate for each of the plurality of candidate financial information;

inputting the target financial behavior information of the target user into the revenue model, outputting a benefit for each of the plurality of candidate financial information; and

determining a recommendation score for each of the plurality of candidate financial information based on the click-through rate and the benefit of the candidate financial information.

5. The method of claim 3, wherein the click-through rate model is jointly trained using a first long-short term memory neural network model, a first deep neural network model, and a first neural network model, comprising:

obtaining a first training sample, wherein the first training sample comprises a plurality of positive samples and a plurality of negative samples, each positive sample comprises target financial behavior information of a sample user and a real click rate of the sample user for a sample product, each negative sample comprises target training financial behavior information of the sample user and real display information of the sample user for the sample product, and the target training financial behavior information of the sample user comprises user running information, user risk assessment information and financial product information of the sample user;

inputting user flow information of sample users included in each positive sample in the plurality of positive samples into the first long-short term memory neural network model, and outputting a first vector;

inputting user risk assessment information and financial product information of a sample user included in each of the plurality of positive samples into the first deep neural network model, and outputting a second vector;

splicing the second vector and the second vector to obtain a third vector;

inputting the third vector into the first neural network model, and outputting a predicted click rate of the sample user for the sample product included in the positive sample;

inputting user flow information of sample users included in each negative sample in the negative samples into the first long-short term memory neural network model, and outputting a fourth vector;

inputting user risk assessment information and financial product information of the sample user included in each negative sample in the plurality of negative samples into the first deep neural network model, and outputting a fifth vector;

splicing the fourth vector and the fifth vector to obtain a sixth vector;

inputting the sixth vector into the first neural network model, and outputting the predicted display information of the sample user for the sample product, wherein the predicted display information is included in the negative sample; and

and training the first long-short term memory neural network model, the first deep neural network model and the first neural network model according to the plurality of real click rates, the plurality of predicted click rates, the plurality of real display information and the plurality of predicted display information to obtain the click rate model.

6. The method of claim 3, wherein the benefit model is jointly trained using a second long-short term memory neural network model, a second deep neural network model, and a second neural network model, comprising:

acquiring a second training sample, wherein the second training sample comprises a plurality of training groups, each training group comprises target financial behavior information of the sample user and real benefit of the sample user for a sample product, and the target training financial behavior information of the sample user comprises user running information, user risk assessment information and financial product information of the sample user;

inputting user flow information of the sample users included in each of the plurality of training sets into the second long-short term memory neural network model, and outputting a seventh vector;

inputting user risk assessment information and financial product information of the sample users included in each of the plurality of training sets into the second deep neural network model, and outputting an eighth vector;

splicing the seventh vector and the eighth vector to obtain a ninth vector;

inputting the ninth vector into the second neural network model, and outputting the predicted benefit of the sample user included in the training set for the sample product; and

and training the second long-short term memory neural network model, the second deep neural network model and the second neural network model according to the plurality of real benefits and the plurality of predicted benefits to obtain the profit model.

7. The method of claim 1, wherein said determining target financial information from the plurality of candidate financial information based on the plurality of recommendation scores comprises:

sequencing the plurality of recommendation scores to obtain a sequencing result; and

determining the target financial information from the plurality of candidate financial information according to the ranking result.

8. The method of any of claims 1-7, wherein the user pipelining information includes at least one of: transaction time, transaction amount and transaction property of the transaction within a preset time period;

the financial product information includes at least one of: risk rating of financial products, fund manager, asset size, buy cycle and redeem cycle.

9. The method of claim 1, wherein the information recommendation model is trained using a federated learning framework.

10. An information recommendation apparatus comprising:

the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring target financial behavior information of a target user, and the target financial behavior information of the target user comprises at least one of the following items: user flow information, user risk assessment information and financial product information of the target user;

an output module, configured to input the target financial behavior information of the target user into an information recommendation model, and output a recommendation score for each candidate financial information of a plurality of candidate financial information;

a determining module for determining target financial information from the plurality of candidate financial information according to a plurality of the recommendation scores; and

and the recommending module is used for recommending the target financial information to the target user.

11. An electronic device, comprising:

one or more processors;

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 any of claims 1-9.

12. A computer readable storage medium having stored thereon executable instructions which, when executed by a processor, cause the processor to carry out the method of any one of claims 1 to 9.

13. A computer program product comprising a computer program which, when executed by a processor, is adapted to carry out the method of any one of claims 1 to 9.