Method, device, equipment and medium for determining entity triples in asset custody

Abstract

The present application provides a method, apparatus, device, and medium for determining entity triples in asset custody. The method relates to the field of financial technology or other related fields, including: a candidate triple includes a candidate entity identifier and a candidate entity relationship; at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy is sequentially used to determine the candidate entity identifier as a target entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time; in response to the candidate entity identifier being the target entity identifier, the candidate entity relationship is sequentially determined as the target entity relationship using at least one of a preset entity relationship library and a preset entity relationship classification strategy; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time; in response to the candidate entity relationship being the target entity relationship, the candidate triple is determined to be the target triple.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and in particular to a method, apparatus, device and medium for determining entity triples in asset custody. Background Art

[0002] In the field of FinTech, users can entrust their assets to certain financial institutions for asset management. For example, the financial institutions can use the assets for investment purposes, thereby achieving asset custody for the users. In asset custody scenarios, it is necessary to extract the entity identifiers and entity relationships involved in asset custody to obtain triples.

[0003] In the prior art, after a financial institution device obtains text related to asset custody, it uses a pattern matching entity relationship extraction method to extract entity identifiers and entity relationships included in the text to obtain triples corresponding to the text.

[0004] However, the pattern matching method can only identify the subject, predicate and object in the text, and cannot make substantive judgments, so the accuracy of the obtained triples is low; at the same time, the obtained triples have a low degree of fit with the actual asset custody situation. Summary of the Invention

[0005] The present application provides a method, apparatus, device and medium for determining entity triples in asset custody, which is used to solve the problem in the prior art that substantive judgment cannot be made, resulting in low accuracy of the triples obtained; at the same time, the obtained triples have a low degree of fit with the actual asset custody situation.

[0006] In a first aspect, the present application provides a method for determining entity triples in asset custody, the method comprising:

[0007] Get target text about asset custody;

[0008] Obtaining a candidate triplet according to the target text; the candidate triplet includes a candidate entity identifier and a candidate entity relationship between entities; the candidate triplet is related to asset custody;

[0009] A preset entity identifier discrimination strategy is adopted to determine the candidate entity identifier as a target entity identifier; the preset entity identifier discrimination strategy is to sequentially adopt at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the candidate entity identifier as a target entity identifier; the preset entity identifier library includes at least one preset entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time;

[0010] In response to the candidate entity identifier being a target entity identifier, a preset entity relationship identification strategy is used to determine the candidate entity relationship as the target entity relationship; the preset entity relationship identification strategy is to sequentially use at least one of a preset entity relationship library and a preset entity relationship classification strategy to determine the candidate entity relationship as the target entity relationship; the preset entity relationship library includes at least one preset entity relationship; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time;

[0011] In response to the candidate entity relationship being a target entity relationship, the candidate triple is determined to be a target triple.

[0012] In a second aspect, the present application provides a device for determining entity triples in asset custody, the device comprising:

[0013] The acquisition module is used to obtain target text about asset custody;

[0014] An acquisition module, configured to obtain a candidate triple according to the target text; the candidate triple includes a candidate entity identifier and a candidate entity relationship between entities; and the candidate triple is related to asset custody;

[0015] A first determination module is configured to determine the candidate entity identifier as a target entity identifier using a preset entity identifier discrimination strategy; the preset entity identifier discrimination strategy is to sequentially use at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the candidate entity identifier as the target entity identifier; the preset entity identifier library includes at least one preset entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time;

[0016] a second determination module configured to, in response to the candidate entity identifier being a target entity identifier, determine the candidate entity relationship as a target entity relationship using a preset entity relationship identification strategy; the preset entity relationship identification strategy is to sequentially use at least one of a preset entity relationship library and a preset entity relationship classification strategy to determine the candidate entity relationship as a target entity relationship; the preset entity relationship library includes at least one preset entity relationship; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time;

[0017] The third determining module is configured to determine the candidate triple as a target triple in response to the candidate entity relationship being a target entity relationship.

[0018] In a third aspect, the present application provides an electronic device, comprising: a processor, and a memory communicatively connected to the processor;

[0019] The memory stores computer-executable instructions;

[0020] The processor executes the computer-executable instructions stored in the memory to implement the method described in the first aspect or any one of the above methods.

[0021] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method described in the first aspect or any one of the above methods.

[0022] In a fifth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method described in the first aspect or any one of the above methods.

[0023] The present application provides a method, device, equipment and medium for determining entity triples in asset custody. The entity triple determination device in asset custody of the present application (hereinafter referred to as the determination device) first obtains the candidate triples based on the target text. Since the existing technology has limitations on the target text extraction method and does not consider substantive judgment, the candidate entity identifiers and the candidate entity relationships in the candidate triples will be further determined in the present application. The determination device in the present application first uses at least one of the preset pronoun library, the preset entity identifier library and the preset entity identifier classification strategy to determine the candidate entity identifier as the target entity identifier. Among them, the preset pronoun library can effectively determine that the candidate entity identifier is a clear and substantial entity. For entities that are not clear and substantial, the present application can also determine them based on the preset pronoun library. Therefore, the present application can improve the accuracy of the target triples based on the preset pronoun library. The preset pronoun library used in the present application can be used as the first screening; in addition, the present application can also be combined with After the first screening, the preset entity identifier library is used to determine the target entity identifier, wherein the preset entity identifier library includes at least one preset entity identifier. At the same time, the preset entity identifier library in this application is an identifier library that receives new preset entity identifiers related to asset custody in real time, so the preset entity identifier library in this application will continue to be enriched, so that it will be more accurate to determine the candidate entity identifier as the target entity identifier based on the preset entity identifier library. In addition, the preset entity identifier received by this application is still related to asset custody, so the target entity identifier determined by this application is also related to asset custody, so the actual fit is improved. The preset entity identifier library in this application can be used as a second screening; in addition, this application can also use the preset entity identifier strategy to determine the candidate entity identifier as the target entity identifier after the second screening. The preset entity identifier strategy in this application can be set to an application that complies with asset custody, so this application can improve the accuracy of determining the target entity identifier and the fit with the actual situation. Then, the present application sequentially adopts at least one of the preset entity relationship library and the preset entity relationship classification strategy to determine that the selected entity relationship is the target entity relationship. Similarly, the preset entity relationship library is a real-time updated relationship library, so it will be continuously enriched, thereby improving the accuracy of determining the target entity relationship. The present application can also adopt the preset entity relationship classification strategy after adopting the preset entity relationship library. Since the preset entity classification strategy can be pre-set to comply with asset custody applications, the present application can improve the accuracy of determining the target entity relationship and its fit with reality, thereby improving the accuracy and fit of determining the target triples. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0025] Figure 1 This is an application scenario diagram of a method for determining entity triples in asset custody provided by this application;

[0026] Figure 2 A schematic flow chart of a method for determining entity triples in asset custody provided in Example 1;

[0027] Figure 3 A flow chart of a method for determining entity triples in asset custody provided in Example 2;

[0028] Figure 4 A flowchart of a method for determining entity triples in asset custody provided in Example 3;

[0029] Figure 5 A flowchart of a method for determining entity triples in asset custody provided in Example 4;

[0030] Figure 6 A flowchart of a method for determining entity triples in asset custody provided in Example 5;

[0031] Figure 7 A schematic diagram of the structure of an entity triplet determination device in asset custody provided in Example 9;

[0032] Figure 8 A schematic diagram of the structure of an electronic device provided in Example 10.

[0033] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0034] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0035] In the prior art, after a financial institution's equipment obtains text related to asset custody, it uses a pattern matching entity relationship extraction method to extract the entity identifiers and entity relationships included in the text, where two entity identifiers are included. The two entity identifiers and entity relationships are then determined as triples to obtain the triples corresponding to the text. Based on the triples, the financial institution can know the corresponding asset custody relationship.

[0036] However, the pattern matching method can only identify the subject, predicate and object in the text, and cannot make substantive judgments, so the accuracy of the triples obtained is low; at the same time, the triples are actually meaningless triples, so the triples have a low degree of fit with the actual asset custody situation.

[0037] In order to solve the defects of the prior art, the inventor of this solution has designed a new solution after creative research. This solution provides a method for determining entity triples in asset custody. In order to solve the problem that the prior art cannot make substantive judgments, and thus the triples obtained have low accuracy and low fit, this solution uses at least one of the preset pronoun library, the preset entity identification library and the preset entity classification strategy to determine the candidate entity identifier as the target entity identifier, wherein the preset pronoun library can filter out unclear and non-substantive identifiers, and then the preset entity identification library can be used to determine it as the target entity identifier. Since the preset entity identification library is an identification library that receives new preset entity identifications related to asset custody in real time, it can be seen that the preset entity identification library will continue to be enriched. Therefore, based on the preset entity identification library, this application can improve the accuracy of determining the target entity identification. In addition, the preset entity identification in the preset entity identification library is related to asset custody, so that the present solution can improve the fit with asset custody. In terms of fit, this solution can also continue to use a preset entity identification classification strategy to determine the candidate entity identification as the target entity identification. Since the preset entity identification strategy can be preset to be related to asset custody, this solution can improve the fit. Similarly, this application sequentially uses at least one of the preset entity relationship library and the preset entity relationship classification strategy to determine the candidate entity relationship as the target entity relationship. Therefore, it is first determined as the target entity relationship through the preset entity relationship library. Since the preset entity relationship library receives new preset entity relationships related to asset custody in real time, it will continue to be enriched, thereby improving the accuracy of determining the target entity relationship. In addition, this solution can also determine the candidate entity relationship as the target entity relationship in combination with the preset entity relationship classification strategy. The preset entity relationship classification strategy can improve the fit. It can be seen that this solution can improve the accuracy and fit of determining the target triple.

[0038] The following introduces the application scenarios of the method, device, equipment and medium for determining entity triples in asset custody provided by this application.

[0039] Figure 1 This is an application scenario diagram of a method for determining entity triples in asset custody provided by this application. Figure 1 As shown, the application scenario diagram includes an electronic device 101.

[0040] The electronic device 101 may be a mobile phone, a computer or other device, which is not limited here.

[0041] The electronic device 101 includes an entity triple determination device (hereinafter referred to as determination device) 102. The determination device 102 includes a storage area, wherein the storage area includes a preset pronoun library, a preset entity identification library, and a preset entity relationship library.

[0042] Specifically, the electronic device 101 starts the entity triplet determination task, and the determination device 102 obtains the target text about asset custody, and then obtains the candidate triplet according to the target text, which includes the candidate entity identifier and the candidate entity relationship between the entities.

[0043] Furthermore, the determining device 102 uses a preset entity identifier identification strategy to determine whether the candidate entity identifier is the target entity identifier. The preset entity identifier identification strategy is to sequentially use at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine whether the candidate entity identifier is the target entity identifier.

[0044] Specifically, if the candidate entity identifier is determined to be the target entity identifier, a preset entity relationship identification strategy is used to determine whether the candidate entity relationship is the target entity relationship. The entity relationship identification strategy is assumed to sequentially use at least one of a preset entity relationship library and a preset entity relationship classification strategy to determine whether the candidate entity relationship is the target entity relationship.

[0045] Specifically, if it is determined that the entity identifier to be selected is not the target entity identifier, then it is determined that the triplet to be selected is not the target triplet.

[0046] Further, the determining means 102 determines that the candidate entity relationship is the target entity relationship and determines that the candidate triple is the target triple. In another case, the determining means 102 determines that the candidate entity relationship is not the target entity relationship and determines that the candidate triple is not the target triple.

[0047] Further, the determining means 102 determines that it is a target triple and displays the target triple; if it is determined not to be a target triple, it displays that the determination failed, indicating that the target text does not have an asset relationship between entities in the asset custody.

[0048] It should be noted that the preset entity identification library in this application includes at least one preset entity identification, and all preset entity identifications are related to asset custody. Similarly, the preset entity relationship library includes at least one preset entity relationship, and all preset entity relationships are related to asset custody. Therefore, this application can improve the fit of determining the target triple.

[0049] It is understandable that if the candidate triple is a syntactically and formally correct triple that is superior to an application that does not comply with asset custody, it will also be determined not to be the target triple.

[0050] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0051] It should be noted that the method, device, equipment and medium for determining entity triples in asset custody provided by this application can be applied to the field of financial technology technology, and can also be applied to any field other than financial technology. The application field of this application is not limited.

[0052] The present application provides a method for determining entity triples in asset custody, which aims to solve the above technical problems in the prior art.

[0053] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0054] Example 1

[0055] The executing entity of this application is the entity triplet determination device in asset custody (hereinafter referred to as the determination device), which is located in the electronic device.

[0056] Figure 2 This is a flow chart of a method for determining entity triples in asset custody provided in Example 1. Figure 2 As shown, the specific steps are as follows.

[0057] S201, obtaining target text regarding asset custody.

[0058] The target text is a string of text related to asset custody. It includes words such as subject, predicate, and object. For example, the target text may be "The asset custodian of fund company A is financial institution A."

[0059] Among them, asset custody means that users entrust their own assets to certain financial institutions or fund companies for asset management in order to obtain more profits.

[0060] S202, obtaining a candidate triple according to the target text; the candidate triple includes a candidate entity identifier and a candidate entity relationship between entities; the candidate triple is related to asset custody.

[0061] The candidate triplet refers to the triplet that is about to be determined as the target triplet. The candidate triplet includes the candidate entity identifiers and the candidate entity relationships between the entities. The candidate triplet is related to asset custody.

[0062] Here, the entity refers to the user related to asset custody. For example, fund company A and financial institution A are both entities.

[0063] The entity identifier refers to the name corresponding to the entity, such as the identifier of fund company A or the identifier of financial institution A. It should be noted that the entity identifier is already reflected in the target text.

[0064] The candidate entity identifier refers to an entity identifier that is about to be determined as a target entity identifier.

[0065] The candidate entity relationship refers to the relationship between two entities, which is the entity relationship to be determined whether it is the target entity relationship.

[0066] In one embodiment, in this step, a pattern matching extraction method can be used to obtain a candidate triple based on the target text. Specifically, the target text is preprocessed according to syntactic analysis, wherein the preprocessing includes operations such as word segmentation, part-of-speech tagging, dependency syntactic analysis, and entity identifier recognition, so that the candidate entity identifier and the candidate entity relationship can be determined according to the subject-verb-object relationship. It should be noted that there are two candidate entity identifiers, and the two candidate entity identifiers have a certain relationship direction. The format of the triple is <entity identifier 1, entity relationship, entity identifier 2>, wherein entity 1 corresponding to entity identifier 1 is the user in asset custody, that is, the asset owner, and entity 2 corresponding to entity identifier 2 is the asset custodian or manager. It can be understood that entity 1 entrusts its assets to entity 2.

[0067] For example, assuming that "the asset custodian of fund company A is financial institution A", then according to the execution of this step, the candidate triplet can be obtained as <fund company A, asset custodian, financial institution A>.

[0068] S203, using a preset entity identifier discrimination strategy to determine the entity identifier to be selected as the target entity identifier; the preset entity identifier discrimination strategy is to sequentially use at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the entity identifier to be selected as the target entity identifier; the preset entity identifier library includes at least one preset entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time.

[0069] The preset entity identifier identification strategy refers to a preset strategy for determining whether a candidate entity identifier is a target entity identifier. The preset entity identifier identification strategy sequentially uses at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine whether a candidate entity identifier is a target entity identifier.

[0070] It is understandable that in this application, the preset pronoun library can be used to first determine whether the entity identifier to be selected is the target entity identifier. If there is no judgment result, the preset entity identifier library can be used to determine whether the entity identifier to be selected is the target entity identifier; if there is a judgment result, the preset entity identifier library and the preset entity identifier classification strategy can be used. Further, when the preset entity identifier library is used to determine whether the entity identifier to be selected is the target entity identifier, if there is no judgment result, the preset entity identifier classification strategy can be used to determine whether the entity identifier to be selected is the target entity identifier; if there is a judgment result, the preset entity identifier classification strategy can be used to finally determine that the entity identifier to be selected is the target entity identifier. Among them, the judgment result refers to the result that the entity identifier to be selected is the target entity identifier or the entity identifier to be selected is not the target entity identifier.

[0071] The target entity identifier refers to an entity identifier that is clear and consistent with the entity identifier in asset custody. It should be noted that the target entity identifier is a specific and unambiguous entity identifier, not an entity identifier with no substantive representation. For example, the candidate triple determined based on the target text may be <he, asset custodian, financial institution A>, where "he" is a pronoun and does not represent a clear and specific entity, thus being an entity identifier with no substantive representation.

[0072] It should be noted that this step is a step for determining the candidate entity identifier as the target entity identifier, so this step adopts a preset entity identifier strategy for determination.

[0073] In this step, at least one of the preset pronoun library, the preset entity identifier library and the preset entity identifier classification strategy is used in sequence to determine whether the selected entity identifier is the target entity identifier, which can accurately and appropriately determine whether it is the target entity identifier in accordance with asset custody.

[0074] The preset pronoun library includes at least one preset pronoun, wherein the preset pronoun can be a pronoun similar to "you", "I" and "he", or a pronoun without substantive content of the enterprise, which is not limited here.

[0075] The "preset entity identifier library" refers to a pre-set identifier library for storing preset entity identifiers, wherein the preset entity identifier library includes at least one preset entity identifier. It should be noted that the preset entity identifier library receives new preset entity identifiers related to asset custody in real time, thereby continuously enriching and supplementing the preset entity identifier library, and can be used to determine whether a candidate entity identifier is the target entity identifier. It is understood that the preset entity identifier is related to asset custody.

[0076] The preset entity identification classification strategy refers to a preset strategy for entity identification classification. In one embodiment, the preset entity identification classification strategy can be implemented based on a preset entity identification algorithm.

[0077] S204, in response to the candidate entity identifier being the target entity identifier, a preset entity relationship identification strategy is adopted to determine the candidate entity relationship as the target entity relationship; the preset entity relationship identification strategy is to sequentially adopt at least one of a preset entity relationship library and a preset entity relationship classification strategy to determine the candidate entity relationship as the target entity relationship; the preset entity relationship library includes at least one preset entity relationship; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time.

[0078] Furthermore, if it is determined in step S203 that the entity identifier to be selected is the target entity identifier, a preset entity relationship identification strategy is adopted to determine whether the entity relationship to be selected is the target entity relationship.

[0079] The preset entity relationship identification strategy refers to a preset strategy for determining whether a candidate entity relationship is a target entity relationship. The preset entity relationship identification strategy sequentially uses at least one of a preset entity relationship library and a preset entity relationship classification strategy to determine whether a candidate entity relationship is a target entity relationship.

[0080] The "preset entity relationship library" refers to a pre-set relationship library that stores preset entity relationships, and includes at least one preset entity relationship. It should be noted that the preset entity relationship library receives new preset entity relationships related to asset custody in real time, thereby continuously enriching and supplementing the preset entity relationship library, and can be used to determine candidate entity relationships as target entity relationships. It is understood that the preset entity relationships are related to asset custody.

[0081] Specifically, in this step, the determination device may use a preset entity relationship library to determine whether the candidate entity relationship is the target entity relationship. If no determination result is found, the preset entity relationship strategy may be used to determine whether the candidate entity relationship is the target entity relationship. If a determination result is found, the preset entity relationship strategy may not be used to determine whether the candidate entity relationship is the target entity relationship. The determination result refers to whether the candidate entity relationship is the target entity relationship or whether the candidate entity relationship is not the target entity relationship.

[0082] S205 : In response to the entity relationship to be selected being the target entity relationship, determining the triple to be selected as the target triple.

[0083] Further, in S204, it is determined that the entity relationship to be selected is the target entity relationship, then the entity relationship to be selected is determined to be the target entity relationship, and then in this embodiment, the entity identifier to be selected in the triple to be selected is determined to be the target entity identifier, and the entity relationship to be selected is the target entity relationship, and then the triple to be selected is determined to be the target triple.

[0084] Among them, the target triplet refers to a triplet that is clear and meets the requirements of asset custody.

[0085] The present embodiment provides a method for determining entity triples in asset custody. The entity triple determination device in asset custody of the present application (hereinafter referred to as the determination device) first obtains a candidate triple based on the target text. Since the prior art has limitations on the target text extraction method and does not consider substantive judgment, the candidate entity identifier and the candidate entity relationship in the candidate triple will be further determined in the present application. The determination device in the present application first uses at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the candidate entity identifier as a target entity identifier. The preset pronoun library can effectively determine that the candidate entity identifier is a clear and substantial entity. For entities that are not clear and substantial, the present application can also determine them based on the preset pronoun library. Therefore, the present application can improve the accuracy of the target triple based on the preset pronoun library. The preset pronoun library used in the present application can be used as the first screening; in addition, the present application can also be combined with the first After screening, the preset entity identification library is used to determine the target entity identification, wherein the preset entity identification library includes at least one preset entity identification. At the same time, the preset entity identification library in this application is an identification library that receives new preset entity identifications related to asset custody in real time, so the preset entity identification library in this application will continue to be enriched, so that it will be more accurate to determine the selected entity identification as the target entity identification based on the preset entity identification library. In addition, the preset entity identification received by this application is still related to asset custody, so the target entity identification determined by this application is also related to asset custody, so the actual fit is improved. The preset entity identification library used in this application can be used as a second screening; in addition, this application can also use the preset entity identification strategy to determine the selected entity identification as the target entity identification after the second screening. The preset entity identification strategy in this application can be set to an application that complies with asset custody, so this application can improve the accuracy of determining the target entity identification and the fit with the actual situation. Then, the present application sequentially adopts at least one of the preset entity relationship library and the preset entity relationship classification strategy to determine that the selected entity relationship is the target entity relationship. Similarly, the preset entity relationship library is a real-time updated relationship library, so it will be continuously enriched, thereby improving the accuracy of determining the target entity relationship. The present application can also adopt the preset entity relationship classification strategy after adopting the preset entity relationship library. Since the preset entity classification strategy can be pre-set to comply with asset custody applications, the present application can improve the accuracy of determining the target entity relationship and its fit with reality, thereby improving the accuracy and fit of determining the target triples.

[0086] Example 2

[0087] This embodiment is a further refinement of the above embodiment. In this embodiment, the preset pronoun library includes at least one preset pronoun, and there are two entity identifiers to be selected.

[0088] The fact that there are two entity identifiers to be selected is determined based on the triple format, which is <entity identifier 1, entity relationship, entity identifier 2>. Therefore, there are also two entity identifiers to be selected. In this application, both entity identifiers must be determined to be target entity identifiers. However, if one of them is determined not to be the target entity identifier, then the candidate triple is not the target triple.

[0089] This embodiment is an optional method of using a preset entity identifier identification strategy to determine the candidate entity identifier as the target entity identifier.

[0090] Figure 3 This is a flow chart of a method for determining entity triples in asset custody provided in Example 2. Figure 3 As shown, the specific steps are as follows.

[0091] S301, obtaining a preset pronoun library and a preset entity identifier library.

[0092] Specifically, the preset pronoun library and the preset entity identifier library are obtained from the own storage area.

[0093] S302: When executing the preset entity identifier identification strategy, it is determined from the preset pronoun library whether there is a preset pronoun that is consistent with the entity identifier to be selected.

[0094] The preset pronoun library includes at least one preset pronoun.

[0095] For example, assuming that the triple to be selected is <he, asset custodian, financial institution A>, and assuming that one of the preset pronouns included in the preset pronoun library is "he", further, it is determined in the preset pronoun library whether there is a preset pronoun that is consistent with the entity identifier to be selected (i.e., "he" and "financial institution A"). It can be seen that "he" in the entity identifier to be selected is consistent with a preset pronoun (i.e., "he") in the preset pronoun library, so the preset pronoun is determined to be a consistent preset pronoun.

[0096] This embodiment also includes: if it is determined that there is a consistent preset pronoun, determining that the candidate entity identifier is not the target entity identifier.

[0097] Further, if the determining device determines that there is a consistent preset pronoun, it is determined that there is an ambiguous entity identifier in the candidate triple, and thus the determining device determines that the candidate entity identifier is not the target entity identifier.

[0098] It should be noted that in this step, both of the two entity identifiers to be selected can be pronouns. In fact, it is determined that neither of the two entity identifiers to be selected is the target entity identifier.

[0099] It should be noted that, in this application, as long as there is a candidate entity identifier that is not the target entity identifier, the steps in Example 6 can be performed.

[0100] S303: If it is determined that there is no consistent preset pronoun, it is determined whether there is a preset entity identifier that is consistent with the entity identifier to be selected in the preset entity identifier library.

[0101] Specifically, if it is determined that there is no consistent preset pronoun, it means that neither of the two candidate entity identifiers is a pronoun, thereby screening out that both candidate entity identifiers are actually clear entities, so the determination device then determines whether there is a preset entity identifier consistent with the candidate entity identifier in the preset entity identifier library.

[0102] Table 1: Schematic table of the preset entity identification library.

[0103] Preset entity identifier First preset entity identifier: Fund Company A Second preset entity identifier: Fund Enterprise B Third preset entity identifier: Financial institution A

[0104] As shown in Table 1, the preset entity identifiers include a first preset entity identifier (fund enterprise A), a second preset entity identifier (fund enterprise B), and a third preset entity identifier (financial institution A).

[0105] For example, assuming that the triplet to be selected is <fund company A, asset custodian, financial institution A>, it is determined in the preset entity identifier library (ie, Table 1) whether there is a preset entity identifier that is consistent with the two entity identifiers to be selected.

[0106] S304: If the two entity identifiers to be selected have the same preset entity identifier, the entity identifier to be selected is determined as the target entity identifier.

[0107] Furthermore, in Table 1, a preset entity identifier (i.e., the first preset entity identifier) ​​that is consistent with the entity identifier to be selected 1 (i.e., fund enterprise A) is queried, and in Table 1, a preset entity identifier (i.e., the third preset entity identifier) ​​that is consistent with the entity identifier to be selected 2 (i.e., financial institution A) is queried.

[0108] Furthermore, in response to the existence of a consistent preset entity identifier in both the two entity identifiers to be selected, the determining device determines that the entity identifier to be selected is the target entity identifier, and then the content of the fourth embodiment can be executed.

[0109] S305: If one or two entity identifiers to be selected do not have a consistent preset entity identifier, a preset entity identifier classification strategy is used to determine the entity identifier to be selected as the target entity identifier.

[0110] Here, if one or two entity identifiers to be selected do not have a consistent preset entity identifier, it means that one entity identifier to be selected does not have a consistent preset entity identifier, or two entity identifiers to be selected do not have a consistent preset entity identifier.

[0111] For example, if the triplet to be selected is <fund enterprise A, asset custodian, financial institution B>, in Table 1, it is found that the entity identifier to be selected 1 (i.e., the fund enterprise) has a consistent preset entity identifier in the preset entity identifier library, i.e., the first preset entity identifier. However, no preset entity identifier consistent with the entity identifier to be selected 2 is found in Table 1, so the determination device responds to the fact that there is no consistent preset entity identifier for one of the entity identifiers to be selected, and thus adopts the preset entity identifier classification strategy to determine whether the entity identifier to be selected is the target entity identifier.

[0112] This embodiment provides a method for determining entity triples in asset custody. In this embodiment, a preset pronoun library is sequentially used to determine whether two candidate entity identifiers are target entity identifiers. Because the pronouns are ambiguous, no specific entity can be obtained. Therefore, if the embodiment determines that a consistent preset pronoun exists, the candidate entity identifier is determined not to be the target entity identifier. Therefore, this embodiment improves accuracy. If no consistent preset pronoun exists, the preset entity identifier library is used to determine the target entity identifier. If both candidate entity identifiers have a consistent preset entity identifier, it indicates that both candidate entity identifiers are currently eligible for asset custody and are therefore determined as target entity identifiers. If one or both candidate entity identifiers do not have a consistent preset entity identifier, it indicates that one of the candidate entity identifiers does not have a consistent preset entity identifier, or that both candidate entity identifiers do not have a consistent preset entity identifier. Therefore, at most two candidate entity identifiers do not meet the requirements for asset custody. Therefore, a preset entity identifier classification strategy is used to accurately determine the candidate entity identifiers as target entity identifiers.

[0113] Example 3

[0114] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional way of using a preset entity identifier classification strategy to determine a candidate entity identifier as a target entity identifier.

[0115] Figure 4 This is a flow chart of a method for determining entity triples in asset custody provided in Example 3. Figure 4 As shown, the specific steps are as follows.

[0116] S401 , when executing a preset entity identifier classification strategy, performing feature calculation on the content included in the entity identifier to be selected, and obtaining a target entity identifier feature vector corresponding to the entity identifier to be selected.

[0117] The target entity identification feature vector refers to a feature vector obtained by encoding and calculating the content included in the target entity identification.

[0118] In one method, through the above-mentioned embodiment 2, it is determined that one of the candidate entity identifiers is the target entity identifier, and there is no consistent preset entity identifier for the other candidate entity identifier. Then, in this embodiment, the preset entity identifier classification strategy is used to determine whether the other candidate entity identifier is the target entity identifier.

[0119] Alternatively, after the above-mentioned second embodiment determines that there is no consistent preset entity identifier for the two candidate entity identifiers, the preset entity identifier classification strategy is used to determine whether the two candidate entity identifiers are target entity identifiers in this embodiment.

[0120] Specifically, in this step, it is assumed that there is no consistent preset entity identifier for the two candidate entity identifiers, and the contents included in the two candidate entity identifiers are respectively feature calculated using a preset encoding algorithm to obtain the target entity identifier feature vectors corresponding to the two candidate entity identifiers.

[0121] It should be noted that the two candidate entity identifiers correspond to their respective target entity identifier feature vectors.

[0122] S402: Input the target entity identification feature vector into a preset entity identification classification algorithm to obtain a target entity identification feature value.

[0123] The preset entity identification classification algorithm refers to a preset algorithm for calculating entity identification classification. It should be noted that the preset entity identification classification algorithm can be a support vector machine (SVM) classifier.

[0124] Furthermore, the target entity identification feature vectors are respectively input into a preset entity identification classification algorithm, thereby obtaining two target entity identification feature values ​​respectively.

[0125] S403 : In response to the target entity identifier characteristic value being greater than or equal to the preset entity identifier characteristic value, determining the category of the entity identifier to be selected as the preset entity category, and determining the entity identifier to be selected as the target entity identifier.

[0126] The preset entity identification characteristic value refers to the characteristic value of the entity identification set in advance for asset custody, which is a threshold used to determine the size of the target entity identification characteristic value.

[0127] The term "preset entity category" refers to an entity category pre-set based on a preset entity identification classification algorithm. In this application, the preset entity category may be an entity category in asset custody. It should be noted that the preset entity category is related to the preset entity identification classification algorithm.

[0128] Furthermore, the determination device responds that the target entity identifier characteristic value is greater than or equal to the preset entity identifier characteristic value, indicating that the category of the entity identifier to be selected is the entity category in asset custody, thereby determining that the category of the entity identifier to be selected is the preset entity category, and determining that the entity identifier to be selected is the target entity identifier.

[0129] This embodiment also includes: in response to the target entity identifier characteristic value being less than the preset entity identifier characteristic value, determining that the category of the entity identifier to be selected is not the preset entity category, and determining that the entity identifier to be selected is not the target entity identifier.

[0130] It should be noted that one of the two candidate entity identifiers may be determined to be the target entity identifier, while the other is determined not to be the target entity identifier, and ultimately the candidate triple is determined not to be the target triple.

[0131] This embodiment provides a method for determining entity triples in asset custody. When a preset entity identifier classification strategy is adopted in this embodiment, feature calculation is first performed on the entity identifier to be selected, thereby obtaining a target entity identifier feature vector, which is further input into the preset entity identifier classification algorithm to obtain a target entity identifier feature value. Since the preset entity identifier classification algorithm can be used to classify entity identifiers, it can be determined whether the entity identifier to be selected is the target entity identifier based on the size of the target entity identifier feature value and the preset entity identifier feature value. Therefore, in this embodiment, the preset entity identifier classification strategy can also be used for the final screening to accurately determine whether it is the target entity identifier.

[0132] Example 4

[0133] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional way of using a preset entity relationship identification strategy to determine that a candidate entity relationship is a target entity relationship.

[0134] Figure 5 This is a flow chart of a method for determining entity triples in asset custody provided in Example 4. Figure 5 As shown, the specific steps are as follows.

[0135] S501, obtaining a preset entity relationship library.

[0136] Specifically, the preset entity relationship library is obtained from its own storage area.

[0137] S502 : When executing the preset entity relationship identification strategy, it is determined from the preset entity relationship library whether there is a preset entity relationship that is consistent with the entity relationship to be selected.

[0138] The preset entity relationship library includes at least one preset entity relationship.

[0139] Among them, the preset entity relationship is the entity relationship related to asset custody.

[0140] As shown in Table 2, the preset entity relationships are respectively a first preset entity relationship and a second preset entity relationship, wherein the first preset entity relationship is the asset manager and the second preset entity relationship is the asset custodian.

[0141] Table 2: Schematic diagram of the preset entity relationship library.

[0142] Preset entity relationships First default entity relationship: asset manager Second default entity relationship: asset custodian

[0143] For example, assuming the candidate triple is <fund company A, asset custodian, financial institution A>, the above embodiment has determined that two candidate entity identifiers (i.e., fund company A and financial institution A) are target entity identifiers. Furthermore, the candidate entity relationship is determined to be "asset custodian." Specifically, the determination device queries the preset entity relationship library (i.e., Table 2) to determine whether there is a preset entity relationship that is consistent with the candidate entity relationship (i.e., asset custodian).

[0144] Furthermore, the determining device determines that the preset entity relationship that is consistent with the entity relationship to be selected is the second preset entity relationship (ie, the asset custodian).

[0145] S503 : In response to the existence of a consistent preset entity relationship, determining the candidate entity relationship as the target entity relationship.

[0146] Furthermore, the determining device determines the candidate entity relationship as the target entity relationship in response to the existence of a consistent preset entity relationship.

[0147] S504 : In response to the absence of a consistent preset entity relationship, a preset entity relationship classification strategy is used to determine whether the candidate entity relationship is a target entity relationship.

[0148] For example, assuming that the triplet to be selected is <fund enterprise A, asset storage party, financial institution A>, where the entity relationship to be selected is "asset storage party", the device then queries in Table 2 whether there is a preset entity relationship consistent with the entity relationship to be selected. If it is found that there is no consistent preset entity relationship, the device then determines whether the preset entity relationship classification strategy is used to determine whether the entity relationship to be selected is the target entity relationship.

[0149] This embodiment provides a method for determining entity triples in asset custody. In this embodiment, the determination device first uses a preset entity relationship library to determine whether there is a preset entity relationship that is consistent with the entity relationship to be selected. Since the preset entity relationships in the preset entity relationship library are related to asset custody, it is possible to accurately determine whether the entity relationship to be selected is the target entity relationship. If there is no consistent preset entity relationship, the preset entity relationship classification strategy is then used to continue determining. Therefore, in this embodiment, whether the entity relationship to be selected is the target entity relationship can be determined through multi-layer screening, thereby making the result more accurate.

[0150] Example 5

[0151] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional way of using a preset entity relationship classification strategy to determine a candidate entity relationship as a target entity relationship.

[0152] Figure 6 This is a flow chart of a method for determining entity triples in asset custody provided in Example 5. Figure 6 As shown, the specific steps are as follows.

[0153] S601 , when executing a preset entity relationship classification strategy, performing feature calculation on the content included in the candidate entity relationship to obtain a target entity relationship feature vector corresponding to the candidate entity relationship.

[0154] The execution method of this embodiment is the same as that of the third embodiment, and will not be described in detail here.

[0155] The target entity relationship feature vector refers to a feature vector obtained by encoding and calculating the content included in the candidate entity relationship.

[0156] S602: Input the target entity relationship feature vector into a preset entity relationship classification algorithm to obtain a target entity relationship feature value.

[0157] The preset entity relationship classification algorithm refers to a preset algorithm for calculating entity relationship classification. It should be noted that the preset entity relationship classification algorithm can be a support vector machine (SVM) classifier.

[0158] S603 : In response to the target entity relationship feature value being greater than or equal to the preset entity relationship feature value, determining the category of the entity relationship to be selected as the preset entity relationship category, and determining the entity relationship to be selected as the target entity relationship.

[0159] The preset entity relationship characteristic value refers to the characteristic value of the entity relationship set in advance for asset custody, which is a threshold used to determine the size of the target entity relationship characteristic value.

[0160] The preset entity relationship category refers to an entity relationship category pre-set based on a preset entity relationship classification algorithm. In this application, the preset entity relationship category may be an entity relationship category in asset custody. It should be noted that the preset entity relationship category is related to the preset entity relationship classification algorithm.

[0161] This embodiment also includes: in response to the target entity relationship feature value being less than the preset entity relationship feature value, determining that the category of the entity relationship to be selected is not the preset entity relationship category, and determining that the entity relationship to be selected is not the target entity relationship.

[0162] This embodiment provides a method for determining entity triples in asset custody. In this embodiment, when executing a preset entity relationship classification strategy, the determination device first calculates the target entity relationship feature value corresponding to the entity relationship to be selected, and then inputs the target entity relationship feature value into the preset entity relationship classification algorithm to obtain the target entity relationship feature value. Based on the size of the target entity relationship feature value and the preset entity relationship feature value, it is accurately determined whether the category of the entity relationship to be selected is the preset entity relationship category.

[0163] Example 6

[0164] This embodiment is a further refinement of any of the above embodiments, and includes the following contents.

[0165] In response to satisfying the non-triplet condition, it is determined that the candidate triple corresponding to the target text is not the target triple; the non-triplet condition is that the candidate entity identifier is not the target entity identifier, the category of the candidate entity identifier is not the preset entity category, and the category of the candidate entity relationship is not the preset entity relationship category.

[0166] The non-triplet condition is that the entity identifier to be selected is not the target entity identifier, the category of the entity identifier to be selected is not the preset entity category, and the category of the entity relationship to be selected is not the preset entity relationship category.

[0167] The entity identifier to be selected is not the target entity identifier, which means that one entity identifier to be selected is not the target entity identifier or both entity identifiers to be selected are not the target entity identifier.

[0168] The category of the entity identifier to be selected is not the preset entity category, which means that the entity identifier to be selected is determined not to be the preset entity category after the preset entity identifier classification strategy is applied.

[0169] The category of the entity relationship to be selected is not the preset entity relationship category, which means that the entity relationship to be selected is determined not to be the preset entity relationship category after the preset entity relationship classification strategy is applied.

[0170] It should be noted that, as long as any one of the non-triplet conditions is satisfied, it can be determined that the candidate triple corresponding to the target text is not the target triple.

[0171] This embodiment provides a method for determining entity triples in asset custody. In this embodiment, when a non-triplet condition is met, it is determined that the to-be-selected triple is not the target triple. In this embodiment, the non-triplet condition is that the to-be-selected entity identifier is not the target entity identifier, the category of the to-be-selected entity identifier is not the preset entity category, and the category of the to-be-selected entity relationship is not the preset entity relationship category. Therefore, the non-triplet condition in this embodiment is comprehensive, so it can accurately determine that the to-be-selected triple is not the target triple.

[0172] Example 7

[0173] This embodiment is a further refinement of any of the above embodiments, and includes the following contents.

[0174] In response to the fact that no consistent preset entity identifier exists for any of the candidate entity identifiers and the candidate entity identifier is determined to be the target entity identifier, the target entity identifier is updated as a new preset entity identifier into the preset entity identifier library.

[0175] It should be noted that in this embodiment, there is no consistent preset entity identifier for any of the candidate entity identifiers, but the preset entity identifier classification strategy determines that the candidate entity identifier is the target entity identifier, and thus the target entity identifier is updated as a new preset entity identifier to the preset entity identifier library, thereby continuously enriching the preset entity identifier library.

[0176] This embodiment provides a method for determining entity triples in asset custody. In this embodiment, for an entity that is not in the preset entity identifier library and the entity identifier to be selected is determined to be the target entity identifier, the target entity identifier is determined as a new preset entity identifier, and then updated to the preset entity identifier library, thereby enriching and supplementing the preset entity identifier library and facilitating subsequent use.

[0177] Example 8

[0178] This embodiment is a further refinement of any of the above embodiments, and includes the following contents.

[0179] In response to the absence of a consistent preset entity relationship and determining that the candidate entity relationship is the target entity relationship, the target entity relationship is updated as a new preset entity relationship into the preset entity relationship library.

[0180] It should be noted that there is no consistent preset entity relationship in this embodiment, but the preset entity relationship classification strategy is used to determine that the candidate entity relationship is the target entity relationship, so that the target entity relationship is used as a new preset entity relationship and is updated to the preset entity relationship library, thereby continuously enriching and supplementing the preset entity relationship library.

[0181] This embodiment provides a method for determining entity triples in asset custody. In this embodiment, for an entity relationship that is not in the preset entity relationship library and is determined to be a target entity relationship, the target entity relationship is determined as a new preset entity relationship, and then updated to the preset entity relationship library, thereby enriching and supplementing the preset entity relationship library and facilitating subsequent use.

[0182] In one embodiment, the determination result can also be displayed, where the determination result is a positive result or a negative result. A positive result is the target triple and the accuracy of the target triple, and a negative result is that the target triple was not determined. The accuracy of the target triple can be preset or calculated. The accuracy of the target triple is divided into entity identification accuracy and entity relationship accuracy. The entity identification accuracy includes two types: the first entity identification accuracy and the second entity identification accuracy.

[0183] In one approach, the accuracy of the target triples can be obtained by obtaining the entity identification accuracy and the entity relationship accuracy respectively.

[0184] In one approach, for a target entity identifier determined using a preset entity identifier library, the preset entity identifier accuracy rate can be used as the entity identifier accuracy rate. For a target entity identifier determined using a preset entity identifier classification strategy, an entity identifier accuracy rate can be calculated based on the difference between the target entity identifier feature value and the preset entity identifier feature value. Based on the two different scenarios described above, a first entity identifier accuracy rate and a second entity identifier accuracy rate are obtained separately.

[0185] In one approach, for a target entity relationship determined using a preset entity relationship library, the preset entity relationship accuracy rate can be used as the entity relationship accuracy rate. However, if a target entity relationship is determined using a preset entity relationship classification strategy, an entity relationship accuracy rate can be calculated based on the difference between the target entity relationship feature value and the preset entity relationship feature value.

[0186] Embodiment 9

[0187] The following are device embodiments of the present application. Figure 7 This is a schematic diagram of the structure of an entity triplet determination device in asset custody provided by Example 9. Figure 7 As shown, the determining device 700 includes the following modules:

[0188] An acquisition module 701 is used to acquire a target text regarding asset custody;

[0189] An acquisition module 702 is configured to obtain a candidate triple based on the target text; the candidate triple includes a candidate entity identifier and a candidate entity relationship between entities; and the candidate triple is related to asset custody;

[0190] The first determination module 703 is configured to determine the candidate entity identifier as the target entity identifier using a preset entity identifier discrimination strategy; the preset entity identifier discrimination strategy is to sequentially use at least one of a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the candidate entity identifier as the target entity identifier; the preset entity identifier library includes at least one preset entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time;

[0191] The second determining module 704 is configured to, in response to the candidate entity identifier being the target entity identifier, determine the candidate entity relationship as the target entity relationship using a preset entity relationship identification strategy; the preset entity relationship identification strategy is to sequentially use at least one of a preset entity relationship library and a preset entity relationship classification strategy to determine the candidate entity relationship as the target entity relationship; the preset entity relationship library includes at least one preset entity relationship; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time;

[0192] The third determining module 705 is configured to determine the candidate triple as the target triple in response to the candidate entity relationship being the target entity relationship.

[0193] In one embodiment, the preset pronoun library includes at least one preset pronoun; there are two entity identifiers to be selected;

[0194] The first determination module 703 is specifically configured to:

[0195] Obtain a preset pronoun library and a preset entity identifier library;

[0196] When executing the preset entity identification discrimination strategy, it is determined from the preset pronoun library whether there is a preset pronoun that is consistent with the entity identification to be selected;

[0197] If it is determined that there is no consistent preset pronoun, then determine whether there is a preset entity identifier that is consistent with the entity identifier to be selected in the preset entity identifier library;

[0198] If the two candidate entity identifiers both have a consistent preset entity identifier, the candidate entity identifier is determined to be the target entity identifier;

[0199] If one or two candidate entity identifiers do not have a consistent preset entity identifier, the preset entity identifier classification strategy is used to determine the candidate entity identifier as the target entity identifier.

[0200] The first determining module 703 is further configured to determine that the candidate entity identifier is not the target entity identifier if it is determined that a consistent preset pronoun exists.

[0201] In one embodiment, when the preset entity identifier classification strategy is adopted to determine the candidate entity identifier as the target entity identifier, the first determining module 703 is specifically configured to:

[0202] When executing the preset entity identification classification strategy, the content included in the candidate entity identification is subjected to feature calculation to obtain the target entity identification feature vector corresponding to the candidate entity identification;

[0203] Inputting the target entity identification feature vector into a preset entity identification classification algorithm to obtain the target entity identification feature value;

[0204] In response to the target entity identifier characteristic value being greater than or equal to the preset entity identifier characteristic value, determining the category of the entity identifier to be selected to be the preset entity category, and determining the entity identifier to be selected to be the target entity identifier;

[0205] The first determining module 703 is further configured to: in response to the target entity identifier characteristic value being less than the preset entity identifier characteristic value, determine that the category of the candidate entity identifier is not the preset entity category, and determine that the candidate entity identifier is not the target entity identifier.

[0206] In one embodiment, when the preset entity relationship identification strategy is adopted to determine that the candidate entity relationship is the target entity relationship, the second determination module 704 is specifically configured to:

[0207] Get the preset entity relationship library;

[0208] When executing the preset entity relationship identification strategy, it is determined from the preset entity relationship library whether there is a preset entity relationship that is consistent with the entity relationship to be selected;

[0209] In response to the existence of a consistent preset entity relationship, determining the candidate entity relationship as the target entity relationship;

[0210] In response to the absence of a consistent preset entity relationship, a preset entity relationship classification strategy is adopted to determine the candidate entity relationship as the target entity relationship.

[0211] In one embodiment, when the preset entity relationship classification strategy is adopted to determine that the candidate entity relationship is the target entity relationship, the second determining module 704 is specifically configured to:

[0212] When executing the preset entity relationship classification strategy, the content included in the candidate entity relationship is subjected to feature calculation to obtain the target entity relationship feature vector corresponding to the candidate entity relationship;

[0213] Input the target entity relationship feature vector into a preset entity relationship classification algorithm to obtain the target entity relationship feature value;

[0214] In response to the target entity relationship feature value being greater than or equal to the preset entity relationship feature value, determining the category of the entity relationship to be selected to be the preset entity relationship category, and determining the entity relationship to be selected to be the target entity relationship;

[0215] The second determining module 704 is further configured to:

[0216] In response to the target entity relationship feature value being less than the preset entity relationship feature value, it is determined that the category of the entity relationship to be selected is not the preset entity relationship category, and it is determined that the entity relationship to be selected is not the target entity relationship.

[0217] In one embodiment, the third determining module 705 is further configured to:

[0218] In response to satisfying the non-triplet condition, it is determined that the candidate triple corresponding to the target text is not the target triple; the non-triplet condition is that the candidate entity identifier is not the target entity identifier, the category of the candidate entity identifier is not the preset entity category, and the category of the candidate entity relationship is not the preset entity relationship category.

[0219] In one embodiment, the first determining module 703 is further configured to:

[0220] In response to the fact that no consistent preset entity identifier exists for any of the candidate entity identifiers and the candidate entity identifier is determined to be the target entity identifier, the target entity identifier is updated as a new preset entity identifier into the preset entity identifier library.

[0221] In one embodiment, the second determining module 704 is further configured to:

[0222] In response to the absence of a consistent preset entity relationship and determining that the candidate entity relationship is the target entity relationship, the target entity relationship is updated as a new preset entity relationship into the preset entity relationship library.

[0223] Example 10

[0224] Figure 8 This is a schematic diagram of the structure of an electronic device provided in Example 10. Figure 8 As shown, the electronic device 800 may include: a processor 801, and a memory 802 in communication with the processor 801. The memory 802 stores computer-executable instructions; the processor 801 executes the computer-executable instructions stored in the memory 802 to implement any of the method embodiments in the first to eighth embodiments described above. The specific implementation methods and technical effects are similar and will not be repeated here.

[0225] In this embodiment, the memory 802 and the processor 801 are connected via a bus. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0226] Example 11

[0227] The present application provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are executed by a processor, they are used to implement any one of the method embodiments of Examples 1 to 8 above. The specific implementation methods and technical effects are similar and will not be repeated here.

[0228] Example 12

[0229] The present application provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements any one of the method embodiments of the above-mentioned embodiments 1 to 8. The specific implementation methods and technical effects are similar and will not be repeated here.

[0230] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0231] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected to achieve the purpose of this embodiment based on actual needs.

[0232] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing module, each module may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or hardware plus software functional modules.

[0233] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. Such program code can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable interrupt handling device, so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0234] In the context of the present invention, machine-readable medium can be a tangible medium that can contain or store a program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0235] In addition, although adopting specific order to describe each operation, this should be understood as requiring such operation to be carried out in the specific order shown or in sequential order, or requiring all illustrated operations to be carried out to obtain desired result.Under certain environment, multitasking and parallel processing may be advantageous.Similarly, although comprising some specific implementation details in the above discussion, these should not be construed as limiting the scope of the present invention.Some features described in the context of independent embodiment can also be realized in single realization in combination.On the contrary, the various features described in the context of independent realization also can be realized in multiple realizations individually or in the mode of any suitable subcombination.

[0236] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0237] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for determining entity triples in asset custody, characterized in that: The method comprises: Get target text about asset custody; Obtaining a candidate triplet according to the target text; the candidate triplet includes a candidate entity identifier and a candidate entity relationship between entities; the candidate triplet is related to asset custody; A preset entity identifier discrimination strategy is adopted to determine the candidate entity identifier as a target entity identifier; the preset entity identifier discrimination strategy is to sequentially adopt a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the candidate entity identifier as a target entity identifier; the preset entity identifier library includes at least one preset entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time; In response to the candidate entity identifier being a target entity identifier, a preset entity relationship identification strategy is used to determine the candidate entity relationship as a target entity relationship; the preset entity relationship identification strategy is to sequentially use a preset entity relationship library and a preset entity relationship classification strategy to determine the candidate entity relationship as a target entity relationship; the preset entity relationship library includes at least one preset entity relationship; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time; In response to the candidate entity relationship being a target entity relationship, determining the candidate triple as a target triple; The preset pronoun library includes at least one preset pronoun; the number of entity identifiers to be selected is two; The adopting of a preset entity identifier discrimination strategy to determine the candidate entity identifier as the target entity identifier includes: Obtaining the preset pronoun library and the preset entity identifier library; When executing the preset entity identifier discrimination strategy, determining whether there is a preset pronoun consistent with the entity identifier to be selected from the preset pronoun library; If it is determined that there is no consistent preset pronoun, determining whether there is a preset entity identifier consistent with the entity identifier to be selected in the preset entity identifier library; If the two entity identifiers to be selected both have the same preset entity identifier, determining the entity identifier to be selected as the target entity identifier; If one or two of the candidate entity identifiers do not have the consistent preset entity identifier, a preset entity identifier classification strategy is adopted to determine the candidate entity identifier as the target entity identifier; The method further comprises: If it is determined that there is a consistent preset pronoun, then it is determined that the candidate entity identifier is not the target entity identifier; The method further comprises: In response to the fact that the consistent preset entity identifier does not exist in any of the candidate entity identifiers and the candidate entity identifier is determined to be the target entity identifier, the target entity identifier is updated as a new preset entity identifier into the preset entity identifier library.

2. The method according to claim 1, characterized in that The adopting of a preset entity identifier classification strategy to determine the candidate entity identifier as the target entity identifier includes: When executing the preset entity identification classification strategy, performing feature calculation on the content included in the candidate entity identification to obtain a target entity identification feature vector corresponding to the candidate entity identification; Inputting the target entity identification feature vector into a preset entity identification classification algorithm to obtain a target entity identification feature value; In response to the target entity identifier characteristic value being greater than or equal to a preset entity identifier characteristic value, determining that the category of the entity identifier to be selected is a preset entity category, and determining the entity identifier to be selected as the target entity identifier; The method further comprises: In response to the target entity identifier characteristic value being less than the preset entity identifier characteristic value, it is determined that the category of the entity identifier to be selected is not a preset entity category, and it is determined that the entity identifier to be selected is not a target entity identifier.

3. The method according to claim 1, characterized in that The method of using a preset entity relationship identification strategy to determine the candidate entity relationship as the target entity relationship includes: Obtaining the preset entity relationship library; When executing the preset entity relationship identification strategy, determining whether there is a preset entity relationship consistent with the entity relationship to be selected from the preset entity relationship library; In response to the existence of the consistent preset entity relationship, determining the candidate entity relationship as a target entity relationship; In response to the absence of the consistent preset entity relationship, a preset entity relationship classification strategy is adopted to determine the candidate entity relationship as a target entity relationship.

4. The method according to claim 3, characterized in that The adopting of a preset entity relationship classification strategy to determine the candidate entity relationship as a target entity relationship includes: When executing the preset entity relationship classification strategy, performing feature calculation on the content included in the candidate entity relationship to obtain a target entity relationship feature vector corresponding to the candidate entity relationship; Inputting the target entity relationship feature vector into a preset entity relationship classification algorithm to obtain a target entity relationship feature value; In response to the target entity relationship feature value being greater than or equal to a preset entity relationship feature value, determining the category of the candidate entity relationship to be a preset entity relationship category, and determining the candidate entity relationship to be a target entity relationship; The method further comprises: In response to the target entity relationship feature value being less than the preset entity relationship feature value, it is determined that the category of the candidate entity relationship is not a preset entity relationship category, and it is determined that the candidate entity relationship is not a target entity relationship.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: In response to satisfying the non-triplet condition, it is determined that the candidate triple corresponding to the target text is not the target triple; the non-triplet condition is that the candidate entity identifier is not the target entity identifier, the category of the candidate entity identifier is not the preset entity category, and the category of the candidate entity relationship is not the preset entity relationship category.

6. The method according to claim 3, characterized in that The method further comprises: In response to the absence of the consistent preset entity relationship and determining that the candidate entity relationship is a target entity relationship, the target entity relationship is updated as a new preset entity relationship into the preset entity relationship library.

7. A device for determining entity triples in asset custody, characterized in that: The device comprises: The acquisition module is used to obtain target text about asset custody; An acquisition module, configured to obtain a candidate triple according to the target text; the candidate triple includes a candidate entity identifier and a candidate entity relationship between entities; and the candidate triple is related to asset custody; A first determination module is configured to determine the candidate entity identifier as a target entity identifier using a preset entity identifier discrimination strategy; the preset entity identifier discrimination strategy is to sequentially use a preset pronoun library, a preset entity identifier library, and a preset entity identifier classification strategy to determine the candidate entity identifier as the target entity identifier; the preset entity identifier library includes at least one preset entity identifier; the preset entity identifier library is an identifier library that receives new preset entity identifiers related to asset custody in real time; a second determination module configured to, in response to the candidate entity identifier being a target entity identifier, determine the candidate entity relationship as a target entity relationship using a preset entity relationship identification strategy; the preset entity relationship identification strategy is to sequentially use a preset entity relationship library and a preset entity relationship classification strategy to determine the candidate entity relationship as a target entity relationship; the preset entity relationship library includes at least one preset entity relationship; the preset entity relationship library is a relationship library that receives new preset entity relationships related to asset custody in real time; a third determining module, configured to determine the candidate triple as a target triple in response to the candidate entity relationship being a target entity relationship; The preset pronoun library includes at least one preset pronoun; the number of entity identifiers to be selected is two; The first determining module specifically performs the following steps when the preset entity identifier discrimination strategy is used to determine that the candidate entity identifier is the target entity identifier, including: Obtaining the preset pronoun library and the preset entity identifier library; When executing the preset entity identifier discrimination strategy, determining whether there is a preset pronoun consistent with the entity identifier to be selected from the preset pronoun library; If it is determined that there is no consistent preset pronoun, determining whether there is a preset entity identifier consistent with the entity identifier to be selected in the preset entity identifier library; If the two entity identifiers to be selected both have the same preset entity identifier, determining the entity identifier to be selected as the target entity identifier; If one or two of the candidate entity identifiers do not have the consistent preset entity identifier, a preset entity identifier classification strategy is adopted to determine the candidate entity identifier as the target entity identifier; The first determination module is further configured to determine that the candidate entity identifier is not the target entity identifier if it is determined that a consistent preset pronoun exists; The first determining module is further configured to, in response to any candidate entity identifier not having the consistent preset entity identifier and determining that the candidate entity identifier is the target entity identifier, update the target entity identifier as a new preset entity identifier into the preset entity identifier library.

8. An electronic device comprising: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

10. A computer program product comprising a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images