Market subject industry determination model training method, device, equipment and program product

Abstract

Embodiments of the present disclosure disclose a market subject industry determination model training method, device, equipment and program product. The market subject industry determination model training method comprises: determining an initial market subject industry determination model; obtaining training market subject industry determination input data and training market subject industry determination result data corresponding to the training market subject industry determination input data; training the initial market subject industry determination model with the training market subject industry determination input data as input and the training market subject industry determination result data corresponding thereto as output to obtain a target market subject industry determination model. The technical solution can automatically identify one or more industries corresponding to a market subject based on the business scope of the market subject, ensuring the accuracy and comprehensiveness of the automatic industry identification and providing effective data support for the matching of subsequent industry policies.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of market subject data processing, in particular to a market subject industry determination model training method, device, equipment and program product. BACKGROUND

[0002] With the development and progress of society, the number of enterprises is increasing, and policies for enterprises in various industries are emerging in an endless stream. In order to enable enterprises to obtain corresponding policies in a timely, comprehensive and accurate manner, and to protect the interests of enterprises, a scheme is urgently needed to automatically identify the industry of an enterprise according to its registration information. However, when an enterprise is registered, it can usually only fill in an industry category, and some enterprises fill in an inaccurate industry category, which brings difficulties to the automatic identification of the industry based on the registration information of the enterprise and the subsequent policy matching. For example, an industrial production enterprise cannot match the manufacturing industry policy because it only registered the retail and wholesale industry when it was registered. SUMMARY

[0003] The present disclosure provides a market subject industry determination model training method, device, equipment and program product.

[0004] In a first aspect, a market subject industry determination model training method is provided in the embodiments of the present disclosure.

[0005] Specifically, the market subject industry determination model training method comprises:

[0006] determining an initial market subject industry determination model;

[0007] obtaining a training market subject industry determination data set, wherein the training market subject industry determination data set comprises training market subject industry determination input data and training market subject industry determination result data corresponding to the training market subject industry determination input data, and the training market subject industry determination input data is generated based on the market subject business scope of the training market subject;

[0008] training the initial market subject industry determination model with the training market subject industry determination input data as input and the training market subject industry determination result data corresponding thereto as output to obtain a target market subject industry determination model.

[0009] In an implementation manner of the present disclosure, the market subject industry determination model is a model for classifying market subjects in industries based on similarity.

[0010] In an implementation manner of the present disclosure, obtaining the training market subject industry determination input data comprises:

[0011] obtain a joint embedded vector of the training market subject business scope item and the preset business scope item as a first embedded vector;

[0012] obtain a second embedded vector corresponding to the name and the registered industry of the training market subject;

[0013] merge the first embedded vector and the second embedded vector to obtain the training market subject industry determination input data.

[0014] In an implementation manner of the present disclosure, the obtaining of the joint embedded vector of the training market subject business scope item and the preset business scope item as the first embedded vector comprises:

[0015] obtaining a first sentence expression of the training market subject business scope item and a second sentence expression of the preset business scope item;

[0016] inputting the first sentence expression and the second sentence expression into a first preset encoding model to obtain the joint embedded vector of the training market subject business scope item and the preset business scope item as the first embedded vector, wherein the first preset encoding model is an encoding model obtained by training based on a basic encoding model and additionally adding a multi-industry classification task.

[0017] In an implementation manner of the present disclosure, before the obtaining of the first sentence expression of the training market subject business scope item, the method further comprises:

[0018] obtaining a training market subject business scope;

[0019] performing noise cleaning on the training market subject business scope;

[0020] extracting a training market subject business scope item based on the training market subject business scope after the noise cleaning.

[0021] In an implementation manner of the present disclosure, the obtaining of the second embedded vector corresponding to the name and the registered industry of the training market subject comprises:

[0022] obtaining the name and the registered industry of the training market subject;

[0023] generating the second embedded vector corresponding to the name and the registered industry of the training market subject by using a second preset encoding model, wherein the second preset encoding model comprises a convolutional neural network module and a max-pooling operation module.

[0024] In a second aspect, an embodiment of the present disclosure provides a market subject industry determination method.

[0025] Specifically, the market subject industry determination method comprises:

[0026] obtain a joint embedded vector of the target market subject's business scope item and the preset business scope item as a third embedded vector;

[0027] obtain a fourth embedded vector corresponding to the target market subject's name and registered industry;

[0028] merge the third embedded vector and the fourth embedded vector to obtain target market subject industry determination input data;

[0029] input the target market subject industry determination input data into the target market subject industry determination model trained by the training method to obtain the industry determination result corresponding to the target market subject.

[0030] In an implementation manner of the present disclosure, the market subject industry determination device further comprises:

[0031] perform a preset operation according to the industry determination result corresponding to the target market subject, wherein the preset operation comprises one or more of the following operations: industry expansion according to the industry determination result, industry policy matching according to the industry determination result, and industry message sending according to the industry determination result.

[0032] In a third aspect, the present disclosure provides a market subject industry determination model training device.

[0033] Specifically, the market subject industry determination model training device comprises:

[0034] a first determination module configured to determine an initial market subject industry determination model;

[0035] a first obtaining module configured to obtain a training market subject industry determination data set, wherein the training market subject industry determination data set comprises training market subject industry determination input data and training market subject industry determination result data corresponding to the training market subject industry determination input data, and the training market subject industry determination input data is generated based on the training market subject's market subject business scope;

[0036] a training module configured to train the initial market subject industry determination model by taking the training market subject industry determination input data as input and taking the training market subject industry determination result data corresponding thereto as output to obtain a target market subject industry determination model.

[0037] In a fourth aspect, the present disclosure provides a market subject industry determination device.

[0038] Specifically, the market subject industry determination device comprises:

[0039] The second obtaining module is configured to obtain a joint embedded vector of the target market subject business scope item and the preset business scope item as a third embedded vector.

[0040] The third obtaining module is configured to obtain a fourth embedded vector corresponding to the name and the registered industry of the target market subject.

[0041] The merging module is configured to merge the third embedded vector and the fourth embedded vector to obtain the target market subject industry determination input data.

[0042] The second determining module is configured to input the target market subject industry determination input data into the target market subject industry determination model trained by the training device to obtain the industry determination result corresponding to the target market subject.

[0043] In a fifth aspect, an electronic device is provided, including a memory and at least one processor, wherein the memory is configured to store one or more computer instructions, and the one or more computer instructions are executed by the at least one processor to implement the method steps of the market subject industry determination model training method and / or the market subject industry determination method.

[0044] In a sixth aspect, a computer readable storage medium is provided for storing computer instructions for a market subject industry determination device, which includes computer instructions for implementing the market subject industry determination model training method and / or the market subject industry determination method involved in the market subject industry determination model training device and / or the market subject industry determination device.

[0045] In a seventh aspect, a computer program product is provided, including computer programs / instructions, wherein the computer programs / instructions are executed by a processor to implement the method steps of the market subject industry determination model training method and / or the market subject industry determination method.

[0046] The technical solutions provided by the embodiments of the present disclosure can include the following beneficial effects:

[0047] The above technical solutions provide a market subject industry determination model training method, which takes market subject industry determination input data generated based on market subject business scope as input to obtain one or more industries corresponding to the market subject. This technical solution can automatically identify one or more industries corresponding to the market subject based on the market subject business scope, ensuring the accuracy and comprehensiveness of industry automatic identification and providing effective data support for subsequent industry policy matching.

[0048] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0049] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. In the drawings:

[0050] Figure 1 A flowchart is shown illustrating a method for training a market entity industry determination model according to an embodiment of this disclosure;

[0051] Figure 2 A flowchart illustrating the overall process of a market entity industry determination model training method according to an embodiment of this disclosure is provided.

[0052] Figure 3 A flowchart illustrating a method for determining the industry of market entities according to an embodiment of this disclosure is shown;

[0053] Figure 4 A structural block diagram of a market entity industry determination model training device according to an embodiment of the present disclosure is shown.

[0054] Figure 5 A structural block diagram of a market entity industry determination device according to an embodiment of the present disclosure is shown.

[0055] Figure 6 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown;

[0056] Figure 7 This is a schematic diagram of the structure of a computer system suitable for implementing a market entity industry determination model training and / or a market entity industry determination method according to an embodiment of this disclosure. Detailed Implementation

[0057] In the following, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to readily implement them. Furthermore, for clarity, portions unrelated to the description of the exemplary embodiments have been omitted from the drawings.

[0058] In this disclosure, it should be understood that terms such as “comprising” or “having” are intended to indicate the presence of features, figures, steps, behaviors, components, parts or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, figures, steps, behaviors, components, parts or combinations thereof.

[0059] It should also be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0060] The technical solution provided by the embodiments of the present disclosure provides a market subject industry determination model training method. The market subject industry determination model takes market subject industry determination input data generated based on the market subject business scope as input to obtain one or more industries corresponding to the market subject. The technical solution can automatically identify one or more industries corresponding to the market subject based on the market subject business scope, ensuring the accuracy and comprehensiveness of industry automatic identification, and providing effective data support for subsequent industry policy matching.

[0061] Figure 1 A flowchart of a market subject industry determination model training method according to an embodiment of the present disclosure is shown in FIG. 1. As shown in FIG. 1, the market subject industry determination model training method includes the following steps S101-S103: Figure 1

[0062] In step S101, an initial market subject industry determination model is determined.

[0063] In step S102, a training market subject industry determination data set is obtained, wherein the training market subject industry determination data set includes training market subject industry determination input data and training market subject industry determination result data corresponding to the training market subject industry determination input data, and the training market subject industry determination input data is generated based on the market subject business scope of the training market subject.

[0064] In step S103, the initial market subject industry determination model is trained with the training market subject industry determination input data as input and the training market subject industry determination result data corresponding thereto as output, to obtain a target market subject industry determination model.

[0065] As mentioned above, with the development and progress of society, the number of enterprises is increasing, and policies for enterprises in various industries are emerging in an endless stream. In order to enable enterprises to obtain corresponding policies in a timely, comprehensive and accurate manner and protect the interests of enterprises, a scheme is needed to automatically identify the industry of an enterprise based on its registration information. However, when registering, an enterprise can usually only fill in one industry category, and some enterprises fill in an inaccurate industry category, which brings difficulties to the automatic identification of the industry based on the registration information of the enterprise and the subsequent policy matching. For example, an industrial production enterprise cannot match the manufacturing industry policy because it only registered the retail and wholesale industry when registering.

[0066] ​In view of the above defects, in this embodiment, a market subject industry determination model training method is provided, and the market subject industry determination model trained by the method is used to obtain one or more industries corresponding to a market subject based on market subject business scope generated market subject industry determination input data. The technical solution can automatically identify one or more industries corresponding to the market subject based on the market subject business scope, ensuring the accuracy and comprehensiveness of the automatic industry identification, and providing effective data support for the matching of subsequent industry policies.

[0067] In an embodiment of the present disclosure, the market subject industry determination model training method can be applied to a computer, a computing device, an electronic device, a server, a service cluster, or the like for training a market subject industry determination model.

[0068] In an embodiment of the present disclosure, the market subject refers to a subject with a certain industry and capable of performing certain market behavior, such as an enterprise, an individual business, and other non-enterprise units.

[0069] In an embodiment of the present disclosure, the initial market subject industry determination model can be selected according to the actual application needs. In an embodiment of the present disclosure, the initial market subject industry determination model is selected as a model for classifying market subjects based on similarity.

[0070] In an embodiment of the present disclosure, the training market subject industry determination input data refers to data used as input data for training the market subject industry determination model. The training market subject industry determination input data can be generated based on the market subject business scope of the training market subject.

[0071] In an embodiment of the present disclosure, the training market subject industry determination input data corresponds to training market subject industry determination result data. The training market subject industry determination result data may, for example, be one or more industries corresponding to a training market subject determined based on certain training market subject industry determination input data, and the like.

[0072] In the above embodiment, when training the market subject industry determination model, first, an initial market subject industry determination model is determined; then, training market subject industry determination input data and training market subject industry determination result data corresponding to the training market subject industry determination input data are obtained; then, the initial market subject industry determination model is trained by taking the training market subject industry determination input data as input and taking the training market subject industry determination result data corresponding to the training market subject industry determination input data as output, and when the training result converges, a target market subject industry determination model is obtained. The learning and training of the market subject industry determination model can be realized by using the learning and training method mastered by those skilled in the art, and the specific learning and training implementation method of the market subject industry determination is not particularly limited in the present disclosure.

[0073] In an embodiment of the present disclosure, the step of obtaining the training market subject industry determination input data can include the following steps:

[0074] Obtaining a joint embedded vector of the training market subject business scope item and the preset business scope item as a first embedded vector;

[0075] Obtaining a second embedded vector corresponding to the name and registered industry of the training market subject;

[0076] Combining the first embedded vector and the second embedded vector to obtain the training market subject industry determination input data.

[0077] In this embodiment, when obtaining the training market subject industry determination input data, first, a joint embedded vector of the training market subject business scope item and the preset business scope item is obtained as a first embedded vector by means of a first preset encoding model, wherein the first preset encoding model is an encoding model obtained by training based on an underlying encoding model with additional multi-industry classification tasks, which will be described in detail below; then, a second embedded vector corresponding to the name and registered industry of the training market subject is obtained by means of a second preset encoding model, wherein the second preset encoding model includes a convolutional neural network module and a max-pooling operation module, which will be described in detail below; finally, the first embedded vector and the second embedded vector are combined to obtain the training market subject industry determination input data.

[0078] In an embodiment of the present disclosure, the step of obtaining a joint embedded vector of the training market subject business scope item and the preset business scope item as a first embedded vector can include the following steps:

[0079] obtaining a first sentence expression of a training market subject business scope item and a second sentence expression of a preset business scope item;

[0080] inputting the first sentence expression and the second sentence expression into a first preset encoding model to obtain a joint embedded vector of the training market subject business scope item and the preset business scope item as a first embedded vector, wherein the first preset encoding model is an encoding model obtained by training based on a basic encoding model and adding a multi-industry classification task.

[0081] In this embodiment, when determining the first embedded vector, first, a first sentence expression of a training market subject business scope item and a second sentence expression of a preset business scope item are obtained, wherein the sentence expression can be generated based on the content of the literal description by means of the sentence expression conversion means in the prior art, and the present disclosure does not make a specific description thereof, wherein the preset business scope item refers to a business scope item in a standard business scope system that is constructed in advance, the standard business scope system refers to a standard business scope system that is constructed based on a national standard, is manually audited and corrected, and includes a business scope code, a business scope name corresponding to the business scope code, one or more industry codes and industry names to which the business scope belongs, and an operating activity corresponding to the business scope, and the like. As mentioned above, in the prior art, an enterprise can usually only fill in one industry category when registering, and some enterprises fill in an inaccurate industry category. In the standard business scope system, however, almost all industry information related to each business scope exists, which can provide data support for subsequent industry determination based on the business scope. Then, the obtained first sentence expression and second sentence expression are input into a first preset encoding model to obtain a joint embedded vector of the training market subject business scope item and the preset business scope item, which can be used as the first embedded vector, wherein the first preset encoding model is an encoding model obtained by training based on a basic encoding model and adding a multi-industry classification task. In an embodiment of the present disclosure, the basic encoding model can be selected according to the actual application needs, for example, a Bert (Bidirectional Encoder Representations from Transformers) language model can be selected. In this embodiment, the first preset encoding model is an encoding model obtained by training based on the Bert model and adding a multi-industry classification task, which can realize the adaptation from a business scope to one or more industries.

[0082] In an embodiment of the present disclosure, before obtaining the first sentence expression of the training market subject business scope item, the method further comprises:

[0083] obtain training market subject business scope;

[0084] noise cleaning is performed on the training market subject business scope;

[0085] Based on the training market subject business scope after noise cleaning, training market subject business scope items are extracted.

[0086] In view of the fact that there may be content unrelated to industry classification in the market subject business scope, in this embodiment, before generating the first sentence expression of the training market subject business scope item, data cleaning is also required to be performed on the market subject business scope of the training market subject, to remove the content unrelated to industry classification in the market subject business scope, so as to obtain the content in the market subject business scope that is conducive to industry classification. Specifically, first, the market subject business scope of the training market subject is obtained; then noise cleaning is performed on the training market subject business scope, wherein the noise refers to the content unrelated to industry classification, such as the template text that usually appears in the market subject business scope, such as “except for projects prohibited and regulated by laws and regulations and projects that must be approved before registration”; then based on the training market subject business scope after noise cleaning, training market subject business scope items are extracted, such as “permitted business items: artificial intelligence public service platform technical consulting services”, “general business items: production and sale of western medicine (including chemical medicine raw materials, chemical medicine preparations, biological preparations, biochemical preparations, and other unlisted western medicines)” and the like. The content in the business items can also be split and filled in the minimum expression unit, for example, if the business item content is “production and sale: food”, it can be split and filled to obtain “food production and processing” and “food sales” two business items.

[0087] In an embodiment of the present disclosure, the step of obtaining the second embedded vector corresponding to the name and registered industry of the training market subject can include the following steps:

[0088] obtain the name and registered industry of the training market subject;

[0089] generate the second embedded vector corresponding to the name and registered industry of the training market subject by using a second preset encoding model, wherein the second preset encoding model includes a convolutional neural network module and a max-pooling operation module.

[0090] In this embodiment, in determining the second embedded vector, first, the name of the training market subject and the industry information registered when the training market subject is registered for business are obtained; then the name of the training market subject and the registered industry are input into a second preset encoding model, and the second embedded vector corresponding to the name and the registered industry of the training market subject is generated by using the second preset encoding model, wherein the second preset encoding model comprises a convolutional neural network module and a max-pooling operation module, and the application of the convolutional neural network module and the max-pooling operation module belongs to the technical means that should be mastered by those skilled in the art, and the specific application details thereof will not be described herein.

[0091] Figure 2 The overall flowchart of the market subject industry determination model training method according to an embodiment of the present disclosure is shown in FIG. 1. Figure 2 As shown in FIG. 1, in training the market subject industry determination model, first, an initial market subject industry determination model is determined; then, training market subject industry determination input data is obtained, including: noise cleaning is performed on the business scope of the training market subject, based on the business scope of the training market subject after noise cleaning, the training market subject business scope item is extracted, the first sentence expression of the training market subject business scope item and the second sentence expression of the preset business scope item are obtained, and then the first sentence expression and the second sentence expression are input into the Bert model after adding a multi-industry classification task, i.e., a first preset encoding model, to obtain the joint embedded vector of the training market subject business scope item and the preset business scope item, i.e., a first embedded vector; then, the name of the training market subject and the registered industry are input into a second preset encoding model comprising a convolutional neural network module and a max-pooling operation module, to generate a second embedded vector corresponding to the name and the registered industry of the training market subject; then, the first embedded vector and the second embedded vector are combined as the input of the initial market subject industry determination model; the training market subject industry determination result corresponding to the training market subject industry determination input data is taken as the output of the initial market subject industry determination model, and the initial market subject industry determination model is trained, i.e., a target market subject industry determination model is obtained.

[0092] Figure 3 The flowchart of the market subject industry determination method according to an embodiment of the present disclosure is shown in FIG. 2. Figure 3 As shown in FIG. 2, the market subject industry determination method comprises the following steps S301-S304:

[0093] In step S301, the joint embedded vector of the target market subject business scope item and the preset business scope item is obtained as a third embedded vector;

[0094] In step S302, the name of the target market subject and the fourth embedded vector corresponding to the registered industry are obtained.

[0095] In step S303, the third embedded vector and the fourth embedded vector are merged to obtain the target market subject industry determination input data.

[0096] In step S304, the target market subject industry determination input data is input into the target market subject industry determination model trained by the above training method to obtain the industry determination result corresponding to the target market subject.

[0097] As mentioned above, with the development and progress of society, the number of enterprises is increasing, and policies for enterprises in various industries are also emerging. In order to enable enterprises to obtain corresponding policies in a timely, comprehensive and accurate manner, and to protect the interests of enterprises, there is an urgent need for a scheme that can automatically identify the industry of an enterprise based on its registration information. However, when registering, an enterprise can usually only fill in one industry category, and some enterprises fill in an inaccurate industry category, which brings difficulties to the automatic identification of the industry based on the registration information of the enterprise and the subsequent policy matching. For example, an industrial production enterprise cannot match the manufacturing industry policy because it only registered the retail and wholesale industry when registering.

[0098] In view of the above defects, in this embodiment, a market subject industry determination method is proposed, which uses a pre-trained market subject industry determination model to determine one or more industries corresponding to a market subject based on the business scope of the market subject. This technical solution can automatically identify one or more industries corresponding to a market subject based on the business scope of the market subject, ensuring the accuracy and comprehensiveness of the automatic identification of the industry and providing effective data support for the subsequent matching of industry policies.

[0099] In an embodiment of the present disclosure, the market subject industry determination method can be applied to a computer, a computing device, an electronic device, a server, a service cluster, etc. for determining the industry of a market subject.

[0100] In an embodiment of the present disclosure, the third embedded vector is obtained in a similar manner to the first embedded vector, i.e., first, the target market subject business scope is cleaned of noise, based on the target market subject business scope after noise cleaning, the target market subject business scope items are extracted, the third sentence expression of the target market subject business scope item and the fourth sentence expression of the preset business scope item are obtained, and then the third sentence expression and the fourth sentence expression are input into the Bert model after adding a multi-industry classification task, i.e., the first preset encoding model, to obtain the joint embedded vector of the target market subject business scope item and the preset business scope item, i.e., the third embedded vector.

[0101] That is, in an embodiment of the present disclosure, the step of obtaining the joint embedded vector of the target market subject business scope item and the preset business scope item as the third embedded vector can include the following steps:

[0102] Obtaining the third sentence expression of the target market subject business scope item and the fourth sentence expression of the preset business scope item;

[0103] Inputting the third sentence expression and the fourth sentence expression into a first preset encoding model to obtain the joint embedded vector of the target market subject business scope item and the preset business scope item as the third embedded vector, wherein the first preset encoding model is an encoding model obtained by training based on a basic encoding model and adding a multi-industry classification task.

[0104] Further, in an embodiment of the present disclosure, before obtaining the third sentence expression of the target market subject business scope item, it further includes:

[0105] Obtaining the target market subject business scope;

[0106] Noise cleaning for the target market subject business scope;

[0107] Based on the target market subject business scope after noise cleaning, the target market subject business scope item is extracted.

[0108] In an embodiment of the present disclosure, the fourth embedded vector is obtained in a similar process to the second embedded vector, that is, the name and registered industry of the target market subject are input into a second preset encoding model including a convolutional neural network module and a max-pooling operation module to generate the fourth embedded vector corresponding to the name and registered industry of the target market subject.

[0109] That is, in an embodiment of the present disclosure, the step of obtaining the fourth embedded vector corresponding to the name and registered industry of the target market subject can include the following steps:

[0110] Obtaining the name and registered industry of the target market subject;

[0111] Generating the fourth embedded vector corresponding to the name and registered industry of the target market subject by using a second preset encoding model, wherein the second preset encoding model includes a convolutional neural network module and a max-pooling operation module.

[0112] In the above embodiment, firstly, a joint embedded vector of the target market subject business scope item and the preset business scope item is obtained as a third embedded vector; then, a fourth embedded vector corresponding to the name and the registered industry of the target market subject is obtained; the third embedded vector and the fourth embedded vector are combined to obtain the target market subject industry determination input data; then, the target market subject industry determination input data is input into the target market subject industry determination model trained according to the training method to obtain the industry determination result corresponding to the target market subject.

[0113] In an embodiment of the present disclosure, the method can further include the following steps:

[0114] According to the industry determination result corresponding to the target market subject, a preset operation is performed, wherein the preset operation includes one or more of the following operations: industry expansion on the industry determination result, industry policy matching according to the industry determination result, and industry message sending according to the industry determination result.

[0115] In this embodiment, after obtaining the industry determination result corresponding to the target market subject, a preset operation can be performed according to the industry determination result, wherein the preset operation can include one or more of the following operations: industry expansion on the industry determination result, such as expanding the industry small class to obtain the industry large class corresponding to the industry small class according to the corresponding relationship between the industry small class and the industry large class if the result of determining the industry corresponding to the target market subject is the industry small class corresponding to the market subject; industry policy matching according to the industry determination result, that is, finding the industry policy matching the industry according to the industry determination result and pushing the hit industry policy to the market subject; and industry message sending according to the industry determination result, that is, determining the notification message of the corresponding industry according to the industry determination result and sending the notification message to the market subject.

[0116] For example, for a food limited company, the registered industry of the company is wholesale and retail when it is registered for business, but the "bread processing" recorded in the registered business scope of the company can determine that the company also belongs to the "cake and bread manufacturing" industry small class, and the further expansion of the industry small class determines that the company also belongs to the "manufacturing" industry large class.

[0117] Figure 3 The technical terms and technical features involved in the embodiments shown and related embodiments are the same as or similar to those in the embodiments shown and related embodiments Figures 1-2 The technical terms and technical features mentioned in the embodiments shown and related embodiments are the same as or similar to those in the embodiments shown and related embodiments Figure 3 The explanations and descriptions of the technical terms and technical features involved in the embodiments shown and related embodiments can refer to the above explanations and descriptions of the technical terms and technical features involved in the embodiments shown and related embodiments Figures 1-2The above description of the illustrated embodiments and related implementations is not intended to be exhaustive or to be necessarily limited to the precise embodiments disclosed. Numerous variations and modifications will become apparent to those skilled in the art in light of this disclosure, and are intended to be within the scope of this disclosure.

[0118] The following is an embodiment of the device of the present disclosure, which can be used to execute the method embodiment of the present disclosure.

[0119] Figure 4 A structural block diagram of a market subject industry determination model training device according to an embodiment of the present disclosure is shown, which can be realized by software, hardware or a combination of both as part of or all of an electronic device. As shown in the figure, Figure 4 The market subject industry determination model training device includes:

[0120] The first determination module 401 is configured to determine an initial market subject industry determination model.

[0121] The first acquisition module 402 is configured to acquire a training market subject industry determination data set, wherein the training market subject industry determination data set includes training market subject industry determination input data and training market subject industry determination result data corresponding to the training market subject industry determination input data, and the training market subject industry determination input data is generated based on the market subject business scope of the training market subject.

[0122] The training module 403 is configured to train the initial market subject industry determination model with the training market subject industry determination input data as input and the training market subject industry determination result data corresponding thereto as output, to obtain a target market subject industry determination model.

[0123] As mentioned above, with the development and progress of society, the number of enterprises is increasing, and policies for enterprises in various industries are emerging in an endless stream. In order to enable enterprises to obtain corresponding policies in a timely, comprehensive and accurate manner and protect the interests of enterprises, a scheme is needed to automatically identify the industry of an enterprise according to its registration information. However, when registering, an enterprise can usually only fill in an industry category, and some enterprises fill in an inaccurate industry category, which brings difficulties to the automatic identification of the industry based on the registration information of the enterprise and the subsequent policy matching. For example, an industrial production enterprise cannot match the manufacturing industry policy because it only registered the retail and wholesale industry when registering.

[0124] In view of the above defects, in this embodiment, a market subject industry determination model training device is provided, and the market subject industry determination model trained by the device takes the market subject industry determination input data generated based on the market subject business scope as input to obtain one or more industries corresponding to the market subject. This technical solution can automatically identify one or more industries corresponding to the market subject based on the market subject business scope, ensuring the accuracy and comprehensiveness of the automatic industry identification, and providing effective data support for the subsequent matching of industry policies.

[0125] In an embodiment of the present disclosure, the market subject industry determination model training device can be implemented as a computer, a computing device, an electronic device, a server, a service cluster, or the like for training the market subject industry determination model.

[0126] Figure 5 A structural block diagram of a market subject industry determination device according to an embodiment of the present disclosure is shown, which can be implemented as part or all of an electronic device by software, hardware, or a combination of both. As shown in the figure, the market subject industry determination device includes: Figure 5

[0127] The second acquisition module 501 is configured to acquire the joint embedded vector of the target market subject business scope item and the preset business scope item as the third embedded vector;

[0128] The third acquisition module 502 is configured to acquire the fourth embedded vector corresponding to the name and registered industry of the target market subject;

[0129] The merging module 503 is configured to merge the third embedded vector and the fourth embedded vector to obtain the target market subject industry determination input data;

[0130] The second determination module 504 is configured to input the target market subject industry determination input data into the target market subject industry determination model trained by the above training device to obtain the industry determination result corresponding to the target market subject.

[0131] As mentioned above, with the development and progress of society, the number of enterprises is increasing, and policies for enterprises in various industries are emerging in an endless stream. In order to enable enterprises to obtain corresponding policies in a timely, comprehensive, and accurate manner, and to protect the interests of enterprises, a scheme is urgently needed to automatically identify the industry of an enterprise based on its registration information. However, when registering, an enterprise can usually only fill in one industry category, and some enterprises fill in an inaccurate industry category, which brings difficulties to the automatic identification of the industry based on the registration information of the enterprise and the subsequent policy matching. For example, an industrial production enterprise can not match the manufacturing industry policy because it only registered the retail and wholesale industry when registering.​

[0132] To address the aforementioned shortcomings, this embodiment proposes a market entity industry determination device. This device utilizes a pre-trained market entity industry determination model to identify one or more industries corresponding to the market entity based on its business scope. This technical solution can automatically identify one or more industries corresponding to a market entity based on its business scope, ensuring the accuracy and comprehensiveness of automatic industry identification and providing effective data support for subsequent industry policy matching.

[0133] In one embodiment of this disclosure, the market entity industry determination device can be implemented as a computer, computing device, electronic device, server, service cluster, etc., for determining the industry of market entities.

[0134] The technical terms and features involved in the above-mentioned device embodiments are the same as or similar to those mentioned in the above-mentioned method embodiments. For explanations and descriptions of the technical terms and features involved in the above-mentioned device embodiments, please refer to the explanations of the above-mentioned method embodiments. They will not be repeated here.

[0135] This disclosure also discloses an electronic device, Figure 6 This diagram illustrates a structural block diagram of an electronic device according to an embodiment of the present disclosure, such as... Figure 6 As shown, the electronic device 600 includes a memory 601 and a processor 602; wherein,

[0136] The memory 601 is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor 602 to implement the above method steps.

[0137] Figure 7 This is a schematic diagram of the structure of a computer system suitable for implementing a market entity industry determination model training and / or a market entity industry determination method according to an embodiment of this disclosure.

[0138] like Figure 7 As shown, the computer system 700 includes a processing unit 701, which can execute various processes described above based on a program stored in a read-only memory (ROM) 702 or a program loaded from a storage section 708 into a random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the computer system 700. The processing unit 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0139] The following components are connected to the I / O interface 705: an input part 706 including a keyboard, a mouse, and the like; an output part 707 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), and the like, and a speaker, and the like; a storage part 708 including a hard disk, and the like; and a communication part 709 including a network interface card such as a LAN card, a modem, and the like. The communication part 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as necessary. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is attached to the drive 710 as necessary, so that a computer program read out therefrom is installed in the storage part 708 as necessary. Among them, the processing unit 701 can be implemented as a CPU, a GPU, a TPU, an FPGA, an NPU, and the like processing unit.

[0140] In particular, according to embodiments of the present disclosure, the method described above can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program tangibly embodied on a non-transitory computer readable medium, the computer program containing program code for executing the method. In such embodiments, the computer program can be downloaded and installed from a network via the communication part 709, and / or installed from the removable medium 711.

[0141] The flow and block diagrams in the drawings show possible architectures, functional and operational, of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flow and block diagrams can represent a module, a segment, or a portion of code, which comprises one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks can 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 and / or flowchart illustrations, and combinations thereof, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0142] The units or modules described in the embodiments of the present disclosure can be implemented by software, or can be implemented by hardware. The described units or modules can also be provided in a processor, and the names of these units or modules do not constitute a limitation on the units or modules themselves in some cases.

[0143] As another aspect, the disclosure also provides a computer readable storage medium, which can be the computer readable storage medium included in the apparatus described in the above embodiments; or can exist separately and not be assembled into the apparatus. The computer readable storage medium stores one or more programs for being executed by one or more processors to perform the method described in the disclosure.

[0144] The above description is merely the preferred embodiments of the disclosure and the explanation of the principles of the applied technologies. It should be understood by those skilled in the art that the inventive scope of the disclosure is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by the combinations of the above technical features or equivalent features without departing from the inventive concept. For example, the technical solutions formed by the mutual replacement of the above features and the technical features disclosed in the disclosure (but not limited to) with similar functions.

Claims

1. A method for training a market entity industry determination model, comprising: Determine the initial market entity industry identification model; Obtain a training market entity industry determination data set, wherein the training market entity industry determination data set includes training market entity industry determination input data and training market entity industry determination result data corresponding to the training market entity industry determination input data, and the training market entity industry determination input data is generated based on the market entity business scope of the training market entity; Using the training market entity industry determination input data as input and the corresponding training market entity industry determination result data as output, train the initial market entity industry determination model to obtain the target market entity industry determination model. Obtain input data to identify the industry sectors of the training market participants, including: The system acquires the first statement expression of the business scope items of the training market entity and the second statement expression of the preset business scope items; the preset business scope items are the business scope items in the pre-constructed standard business scope system. The first statement expression and the second statement expression are input into the first preset encoding model to obtain the joint embedded vector of the business scope items of the training market entity and the preset business scope items, which is used as the first embedded vector; wherein, the first preset encoding model is an encoding model trained by adding a multi-industry classification task to the basic encoding model; Obtain the names and registered industries of training market entities; The second preset encoding model is used to generate a second embedded vector corresponding to the name and registered industry of the training market entity, wherein the second preset encoding model includes a convolutional neural network module and a max pooling operation module; The first embedded vector and the second embedded vector are merged to obtain the training market entity industry determination input data.

2. The method according to claim 1, wherein, The market entity industry determination model is a model that classifies market entities by industry based on similarity.

3. The method according to claim 1, further comprising, before obtaining the first statement expression of the business scope items of the training market entity: Obtain the business scope of training market entities; Noise cleanup will be conducted on the business scope of the aforementioned training market entities. Based on the noise-cleaned business scope of training market entities, the business scope items of training market entities are extracted.

4. A method for determining the industry of market entities, comprising: Obtain the joint embedded vector of the target market entity's business scope items and the preset business scope items, and use it as the third embedded vector; Obtain the fourth embedded vector corresponding to the name and registered industry of the target market entity; By merging the third and fourth embedded vectors, the target market entity industry determination input data is obtained. The target market entity industry determination input data is input into the target market entity industry determination model trained by the training method of any one of claims 1-3 to obtain the industry determination result corresponding to the target market entity.

5. The method according to claim 4, further comprising: Based on the industry determination result corresponding to the target market entity, a preset operation is executed, wherein the preset operation includes one or more of the following operations: industry expansion based on the industry determination result, industry policy matching based on the industry determination result, and industry message sending based on the industry determination result.

6. A training device for a market entity industry determination model, comprising: The first determination module is configured to determine the initial market entity industry determination model; The first acquisition module is configured to acquire a training market entity industry determination data set, wherein the training market entity industry determination data set includes training market entity industry determination input data and training market entity industry determination result data corresponding to the training market entity industry determination input data, and the training market entity industry determination input data is generated based on the market entity business scope of the training market entity; The training module is configured to use the training market entity industry determination input data as input and the corresponding training market entity industry determination result data as output to train the initial market entity industry determination model, thereby obtaining the target market entity industry determination model. The first acquisition module is configured to acquire a first statement expression of the business scope items of the training market entity and a second statement expression of the preset business scope items; the preset business scope items are business scope items in a pre-constructed standard business scope system; the first statement expression and the second statement expression are input into a first preset encoding model to obtain a joint embedded vector of the business scope items of the training market entity and the preset business scope items, which serves as the first embedded vector; wherein, the first preset encoding model is an encoding model trained by adding a multi-industry classification task to a basic encoding model; a second preset encoding model is used to generate a second embedded vector corresponding to the name and registered industry of the training market entity, wherein, the second preset encoding model includes a convolutional neural network module and a max pooling operation module; the first embedded vector and the second embedded vector are merged to obtain the industry determination input data of the training market entity.

7. A device for determining the industry of a market entity, comprising: The second acquisition module is configured to acquire a joint embedded vector of the target market entity's business scope items and the preset business scope items, as the third embedded vector; The third acquisition module is configured to acquire the name of the target market entity and the fourth embedded vector corresponding to the registered industry; The merging module is configured to merge the third embedded vector and the fourth embedded vector to obtain the target market entity industry determination input data. The second determining module is configured to input the target market entity industry determining input data into the target market entity industry determining model trained by the training device of claim 6, and obtain the industry determining result corresponding to the target market entity.

8. An electronic device comprising a memory and at least one processor; wherein, The memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the at least one processor to implement the steps of the method according to any one of claims 1-5.

9. A computer program product comprising a computer program / instructions, wherein, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-5.

Images