Virtual data vectorization management method, device, electronic device and storage medium

Abstract

The embodiment of the present invention provides a virtual data vector management method, including: obtaining virtual data of a target user, the virtual data including asset data and equity data; performing feature vector processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector; performing similarity calculation on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector; performing conversion calculation on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain a conversion value between the target asset feature vector and the target equity feature vector; performing value conversion on the target asset data and the target equity data based on the conversion value. The present invention can simplify the conversion process and conversion relationship between virtual data, improve the conversion efficiency between assets and equity, and thus improve the accuracy of virtual data vector management.

Description

Technical Field

[0001] The present invention relates to the field of data management, and in particular to a virtual data vectorization management method, device, electronic equipment and storage medium. Background Art

[0002] With the development of digital technology, especially with the application of big data and artificial intelligence technology, the acquisition and processing of virtual data has become an indispensable part of many industries. In the fields of finance, insurance, asset management, etc., the generation and application of virtual data is particularly important, which can help enterprises accurately carry out multiple business operations such as customer profiling, asset evaluation and equity management.

[0003] The existing virtual data vectorization management method has the problem that the conversion between different virtual data requires a fixed mapping table, the conversion relationship is complex, and the conversion between different virtual data needs to be compared one by one according to the mapping table, which results in low virtual data conversion efficiency and limits user selection and flexibility. Summary of the invention

[0004] An embodiment of the present invention provides a virtual data vectorization management method, which aims to solve the problem that the conversion between different virtual data in the existing virtual data vectorization management method requires a fixed mapping table, the conversion relationship is complex, and the conversion between different virtual data needs to be compared one by one according to the mapping table, thereby resulting in low virtual data conversion efficiency and limiting user selection and flexibility.

[0005] In a first aspect, an embodiment of the present invention provides a virtual data vectorization management method, the method comprising:

[0006] Acquire virtual data of a target user, wherein the virtual data includes asset data and equity data;

[0007] Performing feature vector quantization processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector;

[0008] Calculate the similarity of the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector;

[0009] Based on a preset conversion strategy, the target asset feature vector and the target equity feature vector are converted and calculated to obtain a conversion value between the target asset feature vector and the target equity feature vector;

[0010] Based on the conversion value, the target asset data and the target equity data are converted into values.

[0011] Optionally, obtaining virtual data of the target user includes:

[0012] Extracting metadata and description information from the asset data to obtain first metadata and first description information;

[0013] Extracting metadata and description information from the equity data to obtain second metadata and second description information;

[0014] Storing the first metadata and the second metadata in a relational database, where the relational database is used to store basic information of virtual data;

[0015] The first description information and the second description information are stored in a vector database, where the vector database is used to store feature information of virtual data.

[0016] Optionally, the performing feature vectorization processing on the asset data and the equity data to obtain an asset feature vector and an equity feature vector includes:

[0017] Normalizing at least one dimension data in the first metadata and the second metadata to obtain a metadata feature vector in at least one asset feature vector and a rights and interests feature vector;

[0018] The first description information and the second description information are processed by a preset large language model to obtain a description information feature vector in at least one asset feature vector and a rights and interests feature vector.

[0019] Optionally, the performing similarity calculation on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector includes:

[0020] Based on user needs, determine the target feature vector combination for conversion, wherein the target feature vector combination includes three combinations: asset feature vector and equity feature vector, asset feature vector and asset feature vector, and equity feature vector and equity feature vector;

[0021] By using a cosine similarity algorithm, similarity is calculated for the first metadata and the second metadata corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion, to obtain corresponding metadata similarity;

[0022] By using a preset similarity algorithm, similarity calculation is performed on the first description information and the second description information corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion, so as to obtain the corresponding description information similarity;

[0023] Based on the metadata similarity and the description information similarity, a target asset feature vector and a corresponding target equity feature vector are determined.

[0024] Optionally, before performing conversion calculation on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain a conversion value between the target asset feature vector and the target equity feature vector, the method further includes:

[0025] Determining the similarity between the target asset feature vector and the target equity feature vector;

[0026] Based on the comparison between the similarity and a preset similarity threshold, determining whether there is a conversion relationship between the target asset feature vector and the target equity feature vector;

[0027] If the similarity between the target asset feature vector and the target equity feature vector is greater than the preset similarity threshold, determining that the target asset feature vector and the target equity feature vector are allowed to be converted, and calculating the corresponding conversion value;

[0028] If the similarity between the target asset feature vector and the target equity feature vector is less than or equal to the preset similarity threshold, it is determined that the conversion between the target asset feature vector and the target equity feature vector is not allowed, and the conversion value is not calculated accordingly.

[0029] Optionally, the conversion calculation is performed on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain a conversion value between the target asset feature vector and the target equity feature vector, including:

[0030] Based on the preset conversion strategy and user needs, determine the conversion scenario of the target asset feature vector and the target equity feature vector;

[0031] Based on the conversion scenario, determining the conversion weights of the target asset feature vector and the target equity feature vector;

[0032] The target asset feature vector and the target equity feature vector are converted and calculated according to the conversion weight to obtain a conversion value between the target asset feature vector and the target equity feature vector.

[0033] Optionally, the performing value conversion on the target asset data and the target equity data based on the conversion value includes:

[0034] According to the conversion value, the unit values ​​corresponding to the target asset data and the target equity data are converted to obtain the conversion unit value of the corresponding virtual data;

[0035] Based on the conversion unit value, the unit value of the virtual data is converted to obtain the target value corresponding to the target asset data and the target equity data.

[0036] In a second aspect, an embodiment of the present invention further provides a virtual data vectorization management device, the virtual data vectorization management device comprising:

[0037] A first acquisition module, used to acquire virtual data of a target user, wherein the virtual data includes asset data and equity data;

[0038] A first processing module, configured to perform feature vectorization processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector;

[0039] A first calculation module is used to perform similarity calculation on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector;

[0040] A second calculation module is used to perform conversion calculation on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain a conversion value between the target asset feature vector and the target equity feature vector;

[0041] The first conversion module is used to perform value conversion on the target asset data and the target equity data based on the conversion value.

[0042] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the virtual data vectorization management method provided in the embodiment of the present invention when executing the computer program.

[0043] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the virtual data vectorization management method provided in the embodiment of the invention are implemented.

[0044] In an embodiment of the present invention, virtual data of a target user is obtained, and the virtual data includes asset data and equity data; feature vectorization is performed on the asset data and equity data to obtain an asset feature vector and an equity feature vector; similarity is calculated on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector; based on a preset conversion strategy, conversion calculation is performed on the target asset feature vector and the target equity feature vector to obtain a conversion value between the target asset feature vector and the target equity feature vector; based on the conversion value, the target asset data and the target equity data are converted in value. By simplifying the conversion process and conversion relationship between virtual data, the conversion efficiency between assets and equity is improved, thereby improving the accuracy of virtual data vectorization management. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 is a flow chart of a virtual data vectorization management method provided by an embodiment of the present invention;

[0047] Figure 2 It is a structural schematic diagram of a virtual data vectorization management device provided in an embodiment of the present invention;

[0048] Figure 3 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] like Figure 1 As shown, Figure 1 1 is a method flow chart of a virtual data vectorization management method provided by an embodiment of the present invention. The virtual data vectorization management method comprises the steps of:

[0051] 101. Obtain virtual data of target users.

[0052] In an embodiment of the present invention, the virtual data vectorization management method can be deployed on a user management platform, which can be constructed based on a server or a server cluster and has functions such as data reception, data analysis, data storage, and data transmission. The user management platform can analyze the user's virtual data and user needs according to the virtual data vectorization management method deployed inside the platform, and then feed it back to the user management platform, and generate corresponding analysis results for the user to select and judge.

[0053] The above-mentioned virtual data may include but is not limited to asset data and equity data. The above-mentioned asset data may include but is not limited to asset metadata, asset description information and other data used to confirm the characteristics of the user's financial assets. The above-mentioned asset metadata may include but is not limited to data content such as asset type, asset value, supplier, asset business information, etc. The above-mentioned asset description information may include but is not limited to data that can briefly describe the asset data, such as the characteristics, uses, attributes and other information of the asset data, and may also be the measurement unit, usage unit and basic storage structure of the asset equity.

[0054] The above-mentioned rights and interests data may include but are not limited to rights and interests metadata, rights and interests description information and other data used to confirm the characteristics of user financial rights and interests. The above-mentioned rights and interests metadata may include but are not limited to data content such as rights and interests type, rights and interests value, supplier, rights and interests business information, etc. The above-mentioned rights and interests description information may include but is not limited to data that can briefly describe the rights and interests data, such as the characteristics, uses, attributes and other information of the rights and interests data, and may also be the measurement unit, usage unit and basic storage structure of the asset rights and interests.

[0055] In a possible embodiment, through the above-mentioned user management platform, when a user registers, purchases or injects funds on the platform, the corresponding data is stored in the current user's database, and the virtual data changes of the corresponding operation are obtained, and the virtual data of the corresponding user is obtained according to the changes.

[0056] 102. Perform feature vectorization processing on the asset data and the equity data to obtain an asset feature vector and an equity feature vector.

[0057] In an embodiment of the present invention, the feature vectorization processing may be a process of converting the structured data in the asset data and equity data into a numerical vector. Generally speaking, the process may replace the corresponding structured information with a numerical vector by presetting a large language model or other algorithm. It should be noted that since there are corresponding metadata and description information in the asset data and equity data, where the metadata and description information are different due to feature information, different specific feature vectorization processing methods may be used to convert the metadata and description information into vectors according to different feature information.

[0058] It can be understood that the above-mentioned asset feature vector and equity feature vector both contain metadata vectors and description information feature vectors after feature vectorization processing of the corresponding asset data and equity data.

[0059] In a possible embodiment, the user management platform extracts metadata and description information from the asset data and equity data, and uses a preset large language model or other vector conversion algorithms to perform feature vectorization on the metadata and description information to obtain corresponding asset feature vectors and equity feature vectors.

[0060] 103. Perform similarity calculation on the asset feature vector and the equity feature vector to obtain the target asset feature vector and the corresponding target equity feature vector.

[0061] In an embodiment of the present invention, the similarity calculation can be determined by calculating the asset feature vector and the equity feature vector. The similarity can be used to describe the degree of matching between the asset data and the equity data. Generally speaking, when the similarity between the asset feature vector and the equity feature vector is greater than a preset similarity threshold, it means that the asset data and the equity data corresponding to the asset feature vector and the equity feature vector can be exchanged, that is, they can be replaced or purchased in proportion. It should be noted that in addition to calculating the similarity between the asset feature vector and the equity feature vector, the asset feature vector and the equity feature vector can also be combined to form three combinations of asset feature vector and asset feature vector, equity feature vector and equity feature vector, and asset feature vector and equity feature vector, that is, asset data and asset data, and equity data and equity data can also be converted into each other, that is, if the similarity between the two is greater than the similarity threshold, it is determined that there is a conversion relationship between the two, and they are targets with the user's conversion intention.

[0062] The above-mentioned target asset feature vector and the corresponding target equity feature vector can be the above-mentioned user performing operations in the above-mentioned user management platform, such as storing funds and purchasing equity, and then analyzing the stored funds and purchased equity, that is, after the above-mentioned similarity calculation of the feature vectors, the asset data and corresponding equity data that the user may need to convert or exchange are obtained through analysis.

[0063] 104. Based on a preset conversion strategy, a conversion calculation is performed on the target asset feature vector and the target equity feature vector to obtain a conversion value between the target asset feature vector and the target equity feature vector.

[0064] In an embodiment of the present invention, the preset conversion strategy may be a process of performing conversion calculation according to the proportion of the corresponding object combination conversion based on the object combination that the user may want to convert. Specifically, due to different object combinations selected for conversion, the conversion weights between objects may also be different. For example, when the user needs to convert between assets and equity, the weight of assets can be reduced and the weight of equity can be increased. The target asset feature vector and the equity feature vector are calculated according to the adjusted weights to obtain the conversion value between the target asset feature vector and the target equity feature vector.

[0065] It should be noted that the conversion value can be the final exchange ratio of the target asset feature vector and the target equity feature vector. The unit values ​​corresponding to the target asset feature vector and the target equity feature vector can be converted according to the conversion value. For example, the unit value of asset a is 10, and the unit value of equity b is 1. The similarity of asset a is calculated to be 0.451, and the similarity of equity b is 0.567. The exchange feasibility is passed, and the final similarity is calculated to be 0.457. It can be obtained by checking the similarity exchange table that when asset a needs to be exchanged for equity b, the unit value of a needs to be processed at 70% off, that is, 1 unit of asset a can be exchanged for 7 units of equity b. The above-mentioned similarity exchange table can be a table for setting the conversion value according to the similarity interval, such as shown in Table 1.

[0066]

[0067] Table 1

[0068] It should be noted that the interval of the above-mentioned final similarity range can be determined based on the analysis of the user's daily operating behavior. When the user's daily operating behavior is analyzed as a multi-conversion transaction type, a relatively large number of similarity range intervals are set to meet the customer's needs for multi-conversion transactions.

[0069] 105. Based on the conversion value, the target asset data and the target equity data are converted into value.

[0070] In an embodiment of the present invention, in a possible embodiment, the above-mentioned user management platform standardizes the target asset feature vector and the target equity feature vector to unify all features to the same scale, uses a weighted combination method to weighted sum the standardized features to obtain the weight values ​​of the target assets and target equity, calculates the weighted target assets and target equity according to the similarity algorithm, and after determining the corresponding conversion value, converts the unit value of the target asset and target equity by the conversion value to obtain the corresponding target asset data and target equity data.

[0071] In an embodiment of the present invention, virtual data of a target user is obtained, and the virtual data includes asset data and equity data; feature vectorization processing is performed on the asset data and equity data to obtain an asset feature vector and an equity feature vector; similarity calculation is performed on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector; based on a preset conversion strategy, conversion calculation is performed on the target asset feature vector and the target equity feature vector to obtain a conversion value between the target asset feature vector and the target equity feature vector; based on the conversion value, value conversion is performed on the target asset data and the target equity data. By simplifying the conversion process and conversion relationship between virtual data, the conversion efficiency between assets and equity is improved, thereby improving the accuracy of virtual data vectorization management.

[0072] Optionally, in the step of obtaining the virtual data of the target user, it also includes extracting metadata and description information from the asset data to obtain first metadata and first description information; extracting metadata and description information from the equity data to obtain second metadata and second description information; storing the first metadata and the second metadata in a relational database; and storing the first description information and the second description information in a vector database.

[0073] In the embodiment of the present invention, the above-mentioned process of extracting the asset data and the equity data can be obtained by detecting the data changes in the user database through the above-mentioned user management platform.

[0074] It is understandable that multidimensional extraction processing can be performed on metadata items to obtain corresponding multidimensional metadata. For example, if the metadata item is a supplier, it can be decomposed into four dimensions: supplier source, supplier industry, supplier name, and credit rating.

[0075] The descriptive information is collected and converted into vectors using the text2vec method through a preset large prediction model.

[0076] The above-mentioned relational database can be used to store basic information of virtual data, including but not limited to basic information used to describe asset data and equity data, such as name, etc.

[0077] The above-mentioned vector database can be used to store the converted feature vectors corresponding to the feature information of virtual data, including but not limited to the feature information content represented by the feature vectors used to describe asset data and equity data, such as purpose, attributes, etc.

[0078] In a possible embodiment, the user management platform analyzes the asset data and equity data in the user database to obtain first metadata and first description information corresponding to the asset data, and stores them in a relational database and a vector database, respectively. Similarly, the user management platform analyzes the second metadata and second description information corresponding to the equity data, and stores them in a relational database and a vector database, respectively.

[0079] Optionally, in the step of performing feature vectorization processing on the asset data and the equity data to obtain the asset feature vector and the equity feature vector, it also includes normalizing at least one dimension data in the first metadata and the second metadata to obtain the metadata feature vector in at least one asset feature vector and the equity feature vector; and processing the first description information and the second description information by means of a preset large language model to obtain the description information feature vector in at least one asset feature vector and the equity feature vector.

[0080] In an embodiment of the present invention, the normalization process may be a process of scaling the vectors of at least one dimension data in the first metadata and the second metadata to the same range or unit. Specifically, for each dimension data vector, there may be different data ranges and units. In order to perform effective feature vector representation, the feature vectors may be standardized or normalized and scaled to the same range or unit (usually defined as 0 to 1.0).

[0081] The above-mentioned preset large language model can be any deep learning model that can vectorize the ownership, usage rights, and description information of privatized virtual data obtained through subscription, redemption, collection, and gift in the user database, including but not limited to Zhipu's text_embedding model.

[0082] Specifically, we can select vectors of dimensional data that need to be normalized, such as the value vector of assets and the shareholder equity ratio vector of equity, and use the Min-Max normalization method to scale the vector data to between 0 and 1. We can also use a preset large language model (such as GPT, BERT, etc.) to perform text processing on the first description information (such as product description, market positioning, etc. of assets) and the second description information (such as company vision, market performance, etc. of equity) to obtain feature vectors of descriptive information of assets and equity.

[0083] By normalizing the metadata, dimensional differences are eliminated, and by using a preset large language model to perform vector processing on the descriptive information, the semantic information of the text is fully extracted, making the feature vectors of the target asset data and target equity data more accurate and comprehensive.

[0084] Optionally, in the step of calculating the similarity of the asset feature vector and the equity feature vector to obtain the target asset feature vector and the corresponding target equity feature vector, it also includes determining the target feature vector combination to be converted based on user needs; calculating the similarity of the first metadata and the second metadata corresponding to the user's asset feature vector and the equity feature vector in the target feature vector combination selected for conversion by using a cosine similarity algorithm to obtain the corresponding metadata similarity; calculating the similarity of the first description information and the second description information corresponding to the user's asset feature vector and the equity feature vector in the target feature vector combination selected for conversion by using a preset similarity algorithm to obtain the corresponding description information similarity; and determining the target asset feature vector and the corresponding target equity feature vector based on the metadata similarity and the description information similarity.

[0085] In an embodiment of the present invention, the above-mentioned user demand may refer to the selection of a corresponding target feature vector combination according to a specific business background or goal before the user converts assets and equity. For example, when the user needs to use the balance to purchase equity, the conversion between assets and equity is considered, that is, the target feature vector combination selects an asset feature vector and an equity feature vector. At this time, only the conversion ratio between assets and equity needs to be paid attention to.

[0086] The above target feature vector combinations include three combinations of asset feature vector and equity feature vector, asset feature vector and asset feature vector, and equity feature vector and equity feature vector.

[0087] Specifically, the process of calculating the similarity between the first metadata and the second metadata corresponding to the asset feature vector and the equity feature vector of the user can be obtained through the following specific steps:

[0088] First, define the similarity between the first metadata and the second metadata as S1, the feature vector of the first metadata as ua=[a1,a2,…,an], and the feature vector of the second metadata as ub=[b1,b2,…,bn];

[0089] Through the formula:

[0090]

[0091] Calculate the similarity S1; where ua·ub is the dot product of the feature vectors corresponding to the first metadata and the second metadata, is the square of the modulus of the feature vectors corresponding to the first metadata and the second metadata. The similarity S1 between the first metadata and the second metadata is calculated according to the above formula.

[0092] The above-mentioned process of calculating the similarity of the first description information and the second description information corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion by using the preset similarity algorithm can be obtained through the following specific steps:

[0093] The similarity between the first description information and the second description information is defined as S2, va is the feature vector of the first description information, vb is the feature vector of the second description information, and S2 is calculated by the following formula:

[0094] ;

[0095] The specific calculation method may refer to the vector calculation method of S1, and finally the similarity between the first description information and the second description information is obtained.

[0096] If the calculated S1 and S2 are both greater than 0.3, it means that there is a corresponding conversion value between the asset characteristic vector selected for exchange or conversion and the equity characteristic vector, that is, they can be converted according to the ratio of the conversion value. Therefore, the two can be used as the target asset characteristic vector and the corresponding target equity characteristic vector.

[0097] Optionally, in the step before performing conversion calculation on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain the conversion value between the target asset feature vector and the target equity feature vector, it also includes determining the similarity between the target asset feature vector and the target equity feature vector; comparing the similarity with a preset similarity threshold to determine whether there is a conversion relationship between the target asset feature vector and the target equity feature vector; if the similarity between the target asset feature vector and the target equity feature vector is greater than the preset similarity threshold, determining that conversion is allowed between the target asset feature vector and the target equity feature vector, and calculating the corresponding conversion value; if the similarity between the target asset feature vector and the target equity feature vector is less than or equal to the preset similarity threshold, determining that conversion is not allowed between the target asset feature vector and the target equity feature vector, and the corresponding conversion value is not calculated.

[0098] In an embodiment of the present invention, the preset similarity threshold can be used to confirm whether there is a conversion relationship between the target asset feature vector and the target equity feature vector. Specifically, since the target asset feature vector and the target equity feature vector have different distinguishing features in terms of dimension and content, which are specifically manifested in the different corresponding unit values, for example, the target asset feature vector represents the possession of 1 unit amount, while the target equity feature vector represents "can be exchanged for ten minutes of talk time", and at this time, it is not possible to directly perform a reciprocal exchange, that is, at this time, the similarity calculation of the feature vector is introduced. If the similarity is greater than the preset similarity threshold, it means that there is a connection between the two feature vectors and they can be converted. For example, the target asset feature vector of 1 unit amount can be exchanged for ten minutes of talk time according to the exchange relationship in Table 1 above.

[0099] More specifically, the preset similarity threshold can be adjusted according to user needs or the feature vector combination selected for conversion. Preferably, 0.3 is generally taken as the preset similarity threshold. For example, when the user's goals and preferences are more inclined to risk and usage needs, the corresponding preset similarity threshold is adjusted up or down according to the risk level or usage needs, so that the correlation and influencing factors between different assets are reduced, meeting the user's conversion needs.

[0100] In a possible embodiment, the user management platform obtains user needs by analyzing the user's operating behavior, and calculates the similarity between the target asset feature vector and the target equity feature vector based on the user needs and the selected conversion object combination, and determines the conversion relationship between the selected target asset feature vector and the target equity feature vector by comparing the similarity with a preset similarity threshold.

[0101] Optionally, in the step of performing conversion calculations on the target asset characteristic vector and the target equity characteristic vector based on a preset conversion strategy to obtain a conversion value between the target asset characteristic vector and the target equity characteristic vector, it also includes determining a conversion scenario of the target asset characteristic vector and the target equity characteristic vector based on the preset conversion strategy and user needs; determining conversion weights of the target asset characteristic vector and the target equity characteristic vector based on the conversion scenario; and performing conversion calculations on the target asset characteristic vector and the target equity characteristic vector according to the conversion weights to obtain a conversion value between the target asset characteristic vector and the target equity characteristic vector.

[0102] In an embodiment of the present invention, the preset conversion strategy may be a process of performing conversion calculation according to the proportion of the corresponding object combination conversion based on the object combination that the user may want to convert. Specifically, due to different object combinations selected for conversion, the conversion weights between objects may also be different. For example, when the user needs to convert between assets and equity, the weight of assets can be reduced and the weight of equity can be increased. The target asset feature vector and the equity feature vector are calculated according to the adjusted weights to obtain the conversion value between the target asset feature vector and the target equity feature vector.

[0103] The above conversion scenarios may include but are not limited to purchase behavior scenarios, redemption behavior scenarios, and equity replacement scenarios. Specifically, the corresponding conversion scenarios may correspond to different object combinations. For example, the purchase behavior scenario and the redemption behavior scenario select the target asset feature vector and the target equity feature vector for conversion. The difference is that when the conversion is performed, the corresponding weights will change. Simply put, the sold apple has a value of 2, and you may need to pay a value of 3 to buy it back. The equity replacement behavior scenario selects the target equity feature vector and the target equity feature vector for conversion.

[0104] According to the specific conversion scenario, the conversion weight of the target asset feature vector and the target equity feature vector can be set, that is, the conversion ratio value. For example, in the purchase behavior scenario, the target asset feature vector and the target equity feature vector can be set with a weight of 0.5 when calculating the above final similarity, that is, the final similarity is: 0.5×target asset feature vector+0.5×target equity feature vector.

[0105] After the final similarity is determined, the final exchange ratio is calculated according to the similarity exchange table shown in Table 1, and the exchange ratio is used as the conversion value between the target asset feature vector and the target equity feature vector.

[0106] It should be noted that the exchange ratio, i.e., the above-mentioned similarity exchange table, can be set according to user needs and preferences. For example, the exchange ratio between different assets can be adjusted according to the user's needs, goals and preferences to meet the user's risk preferences and usage needs. The exchange ratio can also be adjusted according to the differences between suppliers, so that settlement can be better performed between different suppliers.

[0107] Optionally, in the step of converting the value of the target asset data and the target equity data based on the conversion value, the step also includes converting the unit values ​​corresponding to the target asset data and the target equity data according to the conversion value to obtain the conversion unit value of the corresponding virtual data; based on the conversion unit value, converting the unit value of the virtual data to obtain the target value corresponding to the target asset data and the target equity data.

[0108] In the embodiment of the present invention, the unit values ​​corresponding to the target asset data and the target equity data are converted according to the preset conversion value. Specifically, the unit value of the target asset data (such as the value per share, the asset value per square meter, etc.) and the unit value of the target equity data (such as the shareholder's equity per share, the dividend per share, etc.) can be converted between the unit values ​​through the above conversion value.

[0109] For example, if the unit value of the target asset data is 100 yuan and the conversion value is 0.85, the corresponding target equity unit value can be converted to 85 yuan by multiplication.

[0110] By calculating the conversion unit value, the data of the target assets and target equity are unified, and a conversion basis is provided for the target value of the virtual data. This process ensures that the unit values ​​between assets and equity can be compared and converted under the same standard by standardizing the conversion value.

[0111] like Figure 2 As shown, an embodiment of the present invention provides a virtual data vectorization management device 200, the virtual data vectorization management device comprising:

[0112] A first acquisition module 201 is used to acquire virtual data of a target user, wherein the virtual data includes asset data and equity data;

[0113] The first processing module 202 is used to perform feature vectorization processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector;

[0114] The first calculation module 203 is used to perform similarity calculation on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector;

[0115] The second calculation module 204 is used to perform conversion calculation on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain a conversion value between the target asset feature vector and the target equity feature vector;

[0116] The first conversion module 205 is used to perform value conversion on the target asset data and the target equity data based on the conversion value.

[0117] Optionally, the first acquisition module 201 includes:

[0118] A first acquisition submodule, used to extract metadata and description information from the asset data to obtain first metadata and first description information;

[0119] A second acquisition submodule is used to extract metadata and description information from the equity data to obtain second metadata and second description information;

[0120] A first storage submodule, used to store the first metadata and the second metadata in a relational database, where the relational database is used to store basic information of virtual data;

[0121] The second storage submodule is used to store the first description information and the second description information in a vector database, and the vector database is used to store feature information of virtual data.

[0122] Optionally, the first processing module 202 includes:

[0123] A first processing submodule is used to normalize at least one dimension data in the first metadata and the second metadata to obtain a metadata feature vector in at least one asset feature vector and a rights and interests feature vector;

[0124] The second processing submodule is used to process the first description information and the second description information through a preset large language model to obtain a description information feature vector in at least one asset feature vector and a rights and interests feature vector.

[0125] Optionally, the first calculation module 203 includes:

[0126] A first determination submodule is used to determine a target feature vector combination for conversion based on user needs, wherein the target feature vector combination includes three combinations of an asset feature vector and an equity feature vector, an asset feature vector and an asset feature vector, and an equity feature vector and an equity feature vector;

[0127] The second determination submodule is used to calculate the similarity of the first metadata and the second metadata corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion by using a cosine similarity algorithm to obtain the corresponding metadata similarity;

[0128] A third determination submodule is used to calculate the similarity of the first description information and the second description information corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion by using a preset similarity algorithm to obtain the corresponding description information similarity;

[0129] The fourth determination submodule is used to determine the target asset feature vector and the corresponding target equity feature vector based on the metadata similarity and the description information similarity.

[0130] Optionally, the device further comprises:

[0131] A first determination module, used to determine the similarity between the target asset feature vector and the target equity feature vector;

[0132] A second determination module is used to determine whether there is a conversion relationship between the target asset feature vector and the target equity feature vector based on the comparison between the similarity and a preset similarity threshold;

[0133] A third determination module is used to determine that the target asset feature vector and the target equity feature vector are allowed to be converted if the similarity between the target asset feature vector and the target equity feature vector is greater than the preset similarity threshold, and calculate the corresponding conversion value;

[0134] The fourth determination module is used to determine that conversion between the target asset feature vector and the target equity feature vector is not allowed, and correspondingly does not calculate the conversion value if the similarity between the target asset feature vector and the target equity feature vector is less than or equal to the preset similarity threshold.

[0135] Optionally, the second calculation module 204 includes:

[0136] A fifth determination submodule, configured to determine a conversion scenario of the target asset feature vector and the target equity feature vector based on a preset conversion strategy and user needs;

[0137] A sixth determination submodule, configured to determine the conversion weights of the target asset feature vector and the target equity feature vector based on the conversion scenario;

[0138] The seventh determination submodule is used to perform conversion calculation on the target asset feature vector and the target equity feature vector according to the conversion weight to obtain a conversion value between the target asset feature vector and the target equity feature vector.

[0139] Optionally, the first conversion module 205 includes:

[0140] A conversion submodule, used to convert the unit values ​​corresponding to the target asset data and the target equity data according to the conversion value to obtain the conversion unit value of the corresponding virtual data;

[0141] The conversion submodule is used to convert the unit value of the virtual data based on the conversion unit value to obtain the target value corresponding to the target asset data and the target equity data.

[0142] like Figure 3As shown, a schematic diagram of the structure of an electronic device 300 provided by an embodiment of the present invention includes: a memory 302, a processor 301, and a computer program of a virtual data vectorization management method stored in the memory 302 and executable on the processor 301, wherein:

[0143] The processor 301 is used to call the computer program stored in the memory 302 and execute the following steps:

[0144] Acquire virtual data of a target user, wherein the virtual data includes asset data and equity data;

[0145] Performing feature vector quantization processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector;

[0146] Calculate the similarity of the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector;

[0147] Based on a preset conversion strategy, the target asset feature vector and the target equity feature vector are converted and calculated to obtain a conversion value between the target asset feature vector and the target equity feature vector;

[0148] Based on the conversion value, the target asset data and the target equity data are converted into values.

[0149] Optionally, the processor 301 executes the step of obtaining the virtual data of the target user, including:

[0150] Extracting metadata and description information from the asset data to obtain first metadata and first description information;

[0151] Extracting metadata and description information from the equity data to obtain second metadata and second description information;

[0152] Storing the first metadata and the second metadata in a relational database, where the relational database is used to store basic information of virtual data;

[0153] The first description information and the second description information are stored in a vector database, where the vector database is used to store feature information of virtual data.

[0154] Optionally, the processor 301 performs feature vectorization processing on the asset data and the equity data to obtain an asset feature vector and an equity feature vector, including:

[0155] Normalizing at least one dimension data in the first metadata and the second metadata to obtain a metadata feature vector in at least one asset feature vector and a rights and interests feature vector;

[0156] The first description information and the second description information are processed by a preset large language model to obtain a description information feature vector in at least one asset feature vector and a rights and interests feature vector.

[0157] Optionally, the processor 301 performs the similarity calculation on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector, including:

[0158] Based on user needs, determine the target feature vector combination for conversion, wherein the target feature vector combination includes three combinations: asset feature vector and equity feature vector, asset feature vector and asset feature vector, and equity feature vector and equity feature vector;

[0159] By using a cosine similarity algorithm, similarity is calculated for the first metadata and the second metadata corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion, to obtain corresponding metadata similarity;

[0160] By using a preset similarity algorithm, similarity calculation is performed on the first description information and the second description information corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion, so as to obtain the corresponding description information similarity;

[0161] Based on the metadata similarity and the description information similarity, a target asset feature vector and a corresponding target equity feature vector are determined.

[0162] Optionally, before the processor 301 executes the conversion calculation based on the preset conversion strategy to the target asset feature vector and the target equity feature vector to obtain the conversion value between the target asset feature vector and the target equity feature vector, the method further includes:

[0163] Determining the similarity between the target asset feature vector and the target equity feature vector;

[0164] Based on the comparison between the similarity and a preset similarity threshold, determining whether there is a conversion relationship between the target asset feature vector and the target equity feature vector;

[0165] If the similarity between the target asset feature vector and the target equity feature vector is greater than the preset similarity threshold, determining that the target asset feature vector and the target equity feature vector are allowed to be converted, and calculating the corresponding conversion value;

[0166] If the similarity between the target asset feature vector and the target equity feature vector is less than or equal to the preset similarity threshold, it is determined that the conversion between the target asset feature vector and the target equity feature vector is not allowed, and the conversion value is not calculated accordingly.

[0167] Optionally, the processor 301 executes the conversion calculation based on the preset conversion strategy to the target asset feature vector and the target equity feature vector to obtain the conversion value between the target asset feature vector and the target equity feature vector, including:

[0168] Based on the preset conversion strategy and user needs, determine the conversion scenario of the target asset feature vector and the target equity feature vector;

[0169] Based on the conversion scenario, determining the conversion weights of the target asset feature vector and the target equity feature vector;

[0170] The target asset feature vector and the target equity feature vector are converted and calculated according to the conversion weight to obtain a conversion value between the target asset feature vector and the target equity feature vector.

[0171] Optionally, the processor 301 performs the value conversion on the target asset data and the target equity data based on the conversion value, including:

[0172] According to the conversion value, the unit values ​​corresponding to the target asset data and the target equity data are converted to obtain the conversion unit value of the corresponding virtual data;

[0173] Based on the conversion unit value, the unit value of the virtual data is converted to obtain the target value corresponding to the target asset data and the target equity data.

[0174] The electronic device provided by the embodiment of the present invention can implement each process implemented by the virtual data vectorization management method in the above method embodiment, and can achieve the same beneficial effect. To avoid repetition, it will not be described here.

[0175] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the computer program implements the various processes of the virtual data vectorization management method or the application-side virtual data vectorization management method provided by the embodiment of the present invention, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0176] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0177] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A virtual data vectorization management method, characterized in that: include: Acquire virtual data of a target user, wherein the virtual data includes asset data and equity data; Performing feature vector quantization processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector; Calculate the similarity of the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector; Based on a preset conversion strategy, the target asset feature vector and the target equity feature vector are converted and calculated to obtain a conversion value between the target asset feature vector and the target equity feature vector; Based on the conversion value, value conversion is performed on the target asset data and the target equity data; The step of obtaining the virtual data of the target user includes: Extracting metadata and description information from the asset data to obtain first metadata and first description information; Extracting metadata and description information from the equity data to obtain second metadata and second description information; Storing the first metadata and the second metadata in a relational database, where the relational database is used to store basic information of virtual data; Storing the first description information and the second description information in a vector database, where the vector database is used to store feature information of virtual data; The performing feature vectorization processing on the asset data and the equity data to obtain an asset feature vector and an equity feature vector includes: Normalizing at least one dimension data in the first metadata and the second metadata to obtain a metadata feature vector in at least one asset feature vector and a rights and interests feature vector; Processing the first description information and the second description information by using a preset large language model to obtain a description information feature vector in at least one asset feature vector and a rights and interests feature vector; The similarity calculation of the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector includes: Based on user needs, determine the target feature vector combination for conversion, wherein the target feature vector combination includes three combinations: asset feature vector and equity feature vector, asset feature vector and asset feature vector, and equity feature vector and equity feature vector; By using a cosine similarity algorithm, similarity is calculated for the first metadata and the second metadata corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion, to obtain corresponding metadata similarity; By using a preset similarity algorithm, similarity calculation is performed on the first description information and the second description information corresponding to the asset feature vector and the equity feature vector of the user in the target feature vector combination selected for conversion, so as to obtain the corresponding description information similarity; Based on the metadata similarity and the description information similarity, a target asset feature vector and a corresponding target equity feature vector are determined.

2. The method according to claim 1, characterized in that Before performing conversion calculation on the target asset feature vector and the target equity feature vector based on the preset conversion strategy to obtain the conversion value between the target asset feature vector and the target equity feature vector, the method further includes: Determining the similarity between the target asset feature vector and the target equity feature vector; Based on the comparison between the similarity and a preset similarity threshold, determining whether there is a conversion relationship between the target asset feature vector and the target equity feature vector; If the similarity between the target asset feature vector and the target equity feature vector is greater than the preset similarity threshold, determining that the target asset feature vector and the target equity feature vector are allowed to be converted, and calculating the corresponding conversion value; If the similarity between the target asset feature vector and the target equity feature vector is less than or equal to the preset similarity threshold, it is determined that the conversion between the target asset feature vector and the target equity feature vector is not allowed, and the conversion value is not calculated accordingly.

3. The method according to claim 1, characterized in that The conversion calculation is performed on the target asset feature vector and the target equity feature vector based on the preset conversion strategy to obtain the conversion value between the target asset feature vector and the target equity feature vector, including: Based on the preset conversion strategy and user needs, determine the conversion scenario of the target asset feature vector and the target equity feature vector; Based on the conversion scenario, determining the conversion weights of the target asset feature vector and the target equity feature vector; The target asset feature vector and the target equity feature vector are converted and calculated according to the conversion weight to obtain a conversion value between the target asset feature vector and the target equity feature vector.

4. The method according to claim 1, characterized in that The performing value conversion on the target asset data and the target equity data based on the conversion value includes: According to the conversion value, the unit values ​​corresponding to the target asset data and the target equity data are converted to obtain the conversion unit value of the corresponding virtual data; Based on the conversion unit value, the unit value of the virtual data is converted to obtain the target value corresponding to the target asset data and the target equity data.

5. A virtual data vectorization management device, characterized in that: The virtual data vectorization management device comprises: A first acquisition module, used to acquire virtual data of a target user, wherein the virtual data includes asset data and equity data; A first processing module, configured to perform feature vectorization processing on the asset data and equity data to obtain an asset feature vector and an equity feature vector; A first calculation module is used to perform similarity calculation on the asset feature vector and the equity feature vector to obtain a target asset feature vector and a corresponding target equity feature vector; A second calculation module is used to perform conversion calculation on the target asset feature vector and the target equity feature vector based on a preset conversion strategy to obtain a conversion value between the target asset feature vector and the target equity feature vector; A first conversion module, configured to perform value conversion on target asset data and target equity data based on the conversion value; The first acquisition module is further used to extract metadata and description information from the asset data to obtain first metadata and first description information; extract metadata and description information from the equity data to obtain second metadata and second description information; store the first metadata and the second metadata in a relational database, which is used to store basic information of virtual data; store the first description information and the second description information in a vector database, which is used to store feature information of virtual data; The first processing module is further used to normalize at least one dimension data in the first metadata and the second metadata to obtain a metadata feature vector in at least one asset feature vector and a rights and interests feature vector; and to process the first description information and the second description information through a preset large language model to obtain a description information feature vector in at least one asset feature vector and a rights and interests feature vector; The first calculation module is also used to determine, based on user needs, a target feature vector combination for conversion, the target feature vector combination including three combinations of an asset feature vector and an equity feature vector, an asset feature vector and an asset feature vector, and an equity feature vector and an equity feature vector; using a cosine similarity algorithm, similarity calculation is performed on the first metadata and the second metadata corresponding to the asset feature vector of the user and the equity feature vector in the target feature vector combination selected for conversion to obtain corresponding metadata similarity; using a preset similarity algorithm, similarity calculation is performed on the first description information and the second description information corresponding to the asset feature vector of the user and the equity feature vector in the target feature vector combination selected for conversion to obtain corresponding description information similarity; based on the metadata similarity and the description information similarity, a target asset feature vector and a corresponding target equity feature vector are determined.

6. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the virtual data vectorization management method as claimed in any one of claims 1 to 4 when executing the computer program.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the virtual data vectorization management method according to any one of claims 1 to 4 are implemented.

Images