Data processing method, user identification method and device

Abstract

The present disclosure provides a data processing method, a user identification method and device, the method comprising: obtaining first user data, and extracting first feature data and second feature data; constructing second user data according to the first feature data and the second feature data; determining whether the second user data points to an organization contained in a preset organization name list according to the second user data and the preset organization name list, the preset organization name list comprising a mapping relationship between abbreviations and full names of a plurality of organizations; if not, extracting third feature data from the first user data, the third feature data being used to indicate industry information; updating the third feature data to the second user data, so as to identify a target organization corresponding to a target user from the preset organization name list according to an abbreviation of the target user's organization and the industry information. The present application is beneficial to accurately identifying an organization where a user is located according to data input by the user.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and in particular to a data processing method, a user identification method, and an apparatus. Background Technology

[0002] With the development of internet and big data technologies, the types and quantities of user data are increasing daily. For example, for a user, it is necessary to identify the organization the user belongs to based on the data they input in order to confirm their identity.

[0003] During data processing, user-input data may suffer from various problems, such as non-standard formatting, redundant and complex information, and inaccurate content. Therefore, how to process user-input data to accurately identify the user's organization and other identity information has become a pressing technical challenge. Summary of the Invention

[0004] This disclosure provides a data processing method, a user identification method, and an apparatus that can process user-input data to accurately identify important information such as the organization to which the user belongs.

[0005] Firstly, this disclosure provides a data processing method, including:

[0006] First user data input by the target user is obtained, and first feature data and second feature data are extracted from the first user data. The first feature data is used to indicate the address information of the organization where the target user is located, and the second feature data is used to indicate the keyword information of the organization where the target user is located.

[0007] Second user data is constructed based on the first feature data and the second feature data, wherein the second user data is used to indicate the abbreviation of the organization to which the target user belongs;

[0008] Based on the second user data and a preset list of organization names, determine whether the second user data points to an organization included in the preset list of organization names. The preset list of organization names includes a mapping relationship between the abbreviations and full names of multiple organizations, and each organization abbreviation corresponds to at least one organization full name.

[0009] If not, then extract third feature data from the first user data, the third feature data being used to indicate the industry information of the organization to which the target user belongs;

[0010] The third feature data is updated to the second user data to identify the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the organization to which the target user belongs.

[0011] Secondly, this disclosure provides a user identification method, characterized in that it includes:

[0012] Extract the target organization's metrics data corresponding to the target user from the pre-defined list of organization names;

[0013] The security level of the target user is identified based on the comparison results between the indicator data and the preset standard indicator threshold.

[0014] Based on the target user's security level, determine whether to respond to the target user's service request;

[0015] The indicator data for each organization included in the preset list of organization names are obtained in the following way:

[0016] For the same organization's abbreviation or vector, aggregate the initial indicator data corresponding to multiple user historical data corresponding to the organization's abbreviation or vector; based on the aggregation results, obtain the organization indicator data corresponding to the organization's abbreviation or vector.

[0017] Thirdly, this disclosure provides a data processing apparatus, including:

[0018] The acquisition module is adapted to acquire first user data input by a target user, and extract first feature data and second feature data from the first user data. The first feature data is used to indicate the address information of the organization where the target user is located, and the second feature data is used to indicate the keyword information of the organization where the target user is located.

[0019] The construction module is adapted to construct second user data based on the first feature data and the second feature data, wherein the second user data is used to indicate the abbreviation of the organization to which the target user belongs;

[0020] The determining module is adapted to determine, based on the second user data and a preset list of organization names, whether the second user data points to an organization included in the preset list of organization names, wherein the preset list of organization names includes a mapping relationship between the abbreviations and full names of multiple organizations, and each organization abbreviation corresponds to at least one organization full name;

[0021] The extraction module is adapted to extract third feature data from the first user data if not otherwise specified, the third feature data being used to indicate industry information of the organization to which the target user belongs;

[0022] The identification module is adapted to update the third feature data to the second user data, so as to identify the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the organization to which the target user belongs.

[0023] Fourthly, this disclosure also provides a user identification device based on a data processing device, comprising:

[0024] The data extraction module is suitable for extracting the indicator data of the target organization corresponding to the target user from a preset list of organization names;

[0025] The security level identification module is adapted to identify the security level of the target user based on the comparison result between the indicator data and the preset standard indicator threshold.

[0026] The service response module is adapted to determine whether to respond to the service request of the target user based on the security level of the target user;

[0027] The indicator data for each organization included in the preset list of organization names are obtained in the following way:

[0028] For the same organization's abbreviation or vector, aggregate the initial indicator data corresponding to multiple user historical data corresponding to the organization's abbreviation or vector; based on the aggregation results, obtain the organization indicator data corresponding to the organization's abbreviation or vector.

[0029] Fifthly, this disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores one or more computer programs executable by the at least one processor, the one or more computer programs being executed by the at least one processor to enable the at least one processor to perform the methods described above.

[0030] In a sixth aspect, this disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described method.

[0031] The embodiments provided in this disclosure can extract first feature data and second feature data from first user data input by the user. Based on the first and second feature data, second user data is constructed to indicate a short name of the user's organization. Since the second user data only contains the key feature data extracted from the first user data, redundant and complex interference information is removed, thus enabling faster and more accurate identification of the user's organization. Furthermore, if accurate identification is not possible using the second user data, third feature data indicating the user's industry can be further extracted. This allows for precise determination of the user's organization when organizations with similar names exist in multiple industries, avoiding identification errors caused by organizations with the same name in different industries.

[0032] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0033] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the embodiments of the present disclosure to explain the disclosure and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed description of exemplary embodiments with reference to the accompanying drawings, in which:

[0034] Figure 1 A flowchart of a data processing method provided in an embodiment of this disclosure;

[0035] Figure 2 A flowchart of a user identification method provided in yet another embodiment of this disclosure;

[0036] Figure 3 A flowchart illustrating the preprocessing of historical organization names for historical users is shown.

[0037] Figure 4 This diagram illustrates the process of online querying and risk control based on organization name;

[0038] Figure 5 A block diagram of a data processing apparatus provided in an embodiment of this disclosure;

[0039] Figure 6 A block diagram of a user identification device provided in an embodiment of this disclosure;

[0040] Figure 7 This is a block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0041] To enable those skilled in the art to better understand the technical solutions of this disclosure, exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments of this disclosure to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0042] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.

[0043] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0044] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Words such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0045] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined herein.

[0046] This disclosure provides a data processing method and a user identification method. The method according to this disclosure can be executed by an electronic device such as a terminal device or a server. The terminal device can be an in-vehicle device, user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, wearable device, etc. The method can be implemented by a processor calling computer-readable program instructions stored in memory. Alternatively, the method can be executed by a server.

[0047] Figure 1 A flowchart illustrating a data processing method provided in yet another embodiment of this disclosure. (Refer to...) Figure 1 The method includes:

[0048] Step S110: Obtain first user data input by the target user, and extract first feature data and second feature data from the first user data. The first feature data is used to indicate the address information of the organization where the target user is located, and the second feature data is used to indicate the keyword information of the organization where the target user is located.

[0049] The first user data input by the target user indicates the organization to which the target user belongs, specifically the organization's full name. However, due to potential redundancy and non-standard input during user input, the full name of the organization entered by the user may not be the correct and standard name. Therefore, to facilitate standardized identification, this step extracts first feature data and second feature data from the first user data. The first feature data indicates the address information of the organization to which the target user belongs, and the second feature data indicates the keyword information of the organization to which the target user belongs. The keyword information can be, for example, unique identifiers or distinctive words that can be extracted from nouns or verbs in the organization's name.

[0050] Therefore, the extracted first feature data and second feature data are the core content contained in the first user data. The specific extraction method can be flexibly set by those skilled in the art. For example, it can be extracted through a pre-set feature data table, or it can be extracted through a pre-trained feature extraction model.

[0051] In one optional implementation, the first user data is input into the trained feature extraction model, and the first feature data and the second feature data are extracted based on the output of the trained feature extraction model. The trained feature extraction model is obtained as follows:

[0052] First, retrieve the historical user data entered by each historical user. This historical user data is used to indicate the full name of the organization to which the user belongs.

[0053] Next, the address-based and keyword-based feature data annotated in the historical user data are obtained. Specifically, for each historical user data, the address-based and keyword-based feature data contained in that historical user data are annotated separately. The address-based feature data indicates the geographical location of the user's organization, such as the province name or county name; the keyword-based feature data indicates the core name of the user's organization, such as highly recognizable core data contained in the organization name.

[0054] Finally, historical user data, along with address-related and keyword-related feature data labeled within it, are used as training samples. An initial feature extraction model is trained based on these samples, resulting in the trained feature extraction model. Thus, the feature extraction model is trained using these training samples. Since the training samples contain address-related feature data (corresponding to the first feature data) and keyword-related feature data (corresponding to the second feature data) labeled with the full names of various organizations, the trained feature extraction model can efficiently extract the first and second feature data from the first user data.

[0055] Step S120: Construct second user data based on the first feature data and the second feature data. The second user data is used to indicate the abbreviation of the organization to which the target user belongs.

[0056] The second user data contains both first and second feature data, used to indicate the abbreviation of the target user's organization. Compared to the first user data, the second user data is less detailed and more refined, removing unnecessary redundancy and interference. Furthermore, with the help of a feature extraction model, it is possible to extract standardized second feature data even when the full organization name in the input first user data is not standardized, thereby improving the accuracy of subsequent identification.

[0057] Step S130: Based on the second user data and the preset list of organization names, determine whether the second user data points to an organization included in the preset list of organization names. The preset list of organization names includes a mapping relationship between the abbreviations and full names of multiple organizations, and each organization abbreviation corresponds to at least one organization full name.

[0058] The list of organization names is compiled in advance based on historical user input data. This list stores the mapping relationships between the abbreviations and full names of multiple organizations. Considering the possibility of inconsistent input of full names, multiple full names may represent the same organization's abbreviation; therefore, each organization abbreviation corresponds to at least one organization's full name.

[0059] In this step, the second user data is matched with the abbreviations of various organizations in the preset list of organization names, and the matching result is used to determine whether the second user data points to an organization in the preset list of organization names.

[0060] Specifically, "the second user data points to an organization within the preset list of organization names" means that the second user data matches the abbreviation of an organization within the list of organization names, and only matches the abbreviation of that one organization. In other words, the second user data has exactly one matching record in the list of organizations.

[0061] If the second user data points to an organization included in the preset list of organization names, the organization pointed to by the second user data is directly identified as the target organization corresponding to the target user, without the need to perform subsequent steps, thus enabling convenient and accurate identification of the organization to which the target user belongs.

[0062] Step S140: If not, extract the third feature data from the first user data. The third feature data is used to indicate the industry information of the organization to which the target user belongs.

[0063] If it is determined that the second user data does not point to an organization included in the preset list of organization names, a third feature data is further extracted from the first user data. This third feature data is used to indicate the industry information of the target user's organization. The industry information includes various types such as finance, real estate, telecommunications, technology, and healthcare, which can effectively distinguish the industry of the user's organization.

[0064] In one alternative implementation, the third feature data is extracted from the first user data in the following way:

[0065] First, extract the auxiliary descriptive information from the first user data, excluding the first and second feature data. This auxiliary descriptive information refers to the remaining content in the first user data after removing the first and second feature data; this content is used to provide an auxiliary description of the user's full organizational name.

[0066] Then, the auxiliary description information is segmented to obtain individual original words. For example, if the auxiliary description information contains multiple words, the auxiliary description information is segmented to obtain multiple original words contained in the auxiliary description information.

[0067] Next, each original word segment is matched with a pre-defined list of industry feature words. The original word segments that successfully match the list of industry feature words are extracted as the third feature data. Since the multiple original word segments contained in the auxiliary description information may be used to provide auxiliary descriptions of the organization name from different dimensions, it is necessary to use the list of industry feature words to filter out the original word segments used for auxiliary descriptions from the industry dimension as the third feature data.

[0068] Since the full name of an organization typically includes terms indicating its industry (e.g., finance, real estate, telecommunications, technology, healthcare), the industry feature word list is determined based on the word frequency of each original word segmentation in the historical user data input by each historical user. Specifically, for multiple historical user data sets, the first and second feature data are extracted from each set to obtain auxiliary descriptive information beyond the first and second feature data. This leads to the original word segments after segmenting the auxiliary descriptive information in the historical user data. Correspondingly, the word frequency of each original word segmentation in the historical user data is statistically analyzed, and the most frequently occurring original word segments are identified as industry feature words and stored in the industry feature word list.

[0069] Specifically, the third feature data is added to the second user data, so that the second user data not only contains the first and second feature data, but also further includes the third feature data. Accordingly, combined with the third feature data, the target organization corresponding to the target user can be identified from a pre-defined list of organization names based on the abbreviation and industry information of the target user's organization.

[0070] Step S150: Update the third feature data to the second user data to identify the target organization corresponding to the target user from the preset list of organization names based on the abbreviation of the organization to which the target user belongs and industry information.

[0071] Specifically, the third feature data is added to the second user data, so that the second user data not only contains the first feature data and the second feature data, but also further includes the third feature data. Accordingly, by combining the third feature data, the target organization corresponding to the target user can be identified from a preset list of organization names based on the abbreviation of the target user's organization and industry information. For example, in step S130, if the second user data and the preset list of organization names contain abbreviations that match multiple organization abbreviations, the target organization corresponding to the target user can be identified from the full names of the multiple organizations corresponding to the matched abbreviations by combining industry information.

[0072] Since the second user data includes industry information in addition to the first and second feature data, it can identify the target organization corresponding to the target user from a pre-set list of organization names by combining the industry information. Based on the abbreviation of the target user's organization, the mapping relationship between the abbreviations and full names of multiple organizations included in the pre-set list of organization names is searched to determine the full name of the organization corresponding to the abbreviation of the target user's organization. When the full name of an organization corresponds to multiple organizations, the target organization corresponding to the target user is identified from among these multiple organizations based on the industry information. For example, in some cases, organizations in different industries may have the same name; in this case, the target organization corresponding to the target user can be identified from among multiple organizations by combining industry information.

[0073] In one optional implementation, if the second user data does not point to any organization in the preset list of organization names, firstly, the updated second user data is vectorized to obtain third user data; wherein the updated second user data includes: first feature data, second feature data, and third feature data; the third user data is a vector indicating the organization to which the target user belongs. Thus, the third user data is a vector. Then, the similarity between the third user data and each vector stored in the preset list of organization names is calculated, and the target organization corresponding to the target user is identified based on vectors with similarity values ​​greater than a preset similarity threshold; wherein the preset list of organization names also includes: mapping relationships between vectors and abbreviations of multiple organizations, and / or mapping relationships between at least one full name of an organization corresponding to each organization abbreviation and its vector.

[0074] Since different full names of organizations may correspond to the same organization, the abbreviations constructed will naturally differ depending on the organization's full name. For example, organization A might be abbreviated as "a" under the first organization's full name and as "b" under the second. If the preset list of organization names only stores the abbreviation "a", organization A cannot be matched from the preset list of organization names based on the second organization's full name. Therefore, the preset list of organization names can also include a mapping relationship between at least one full name of an organization and a vector corresponding to each organization's abbreviation, i.e., storing a vector of organization A corresponding to the abbreviation "a". Since the same organization will only correspond to one vector, and the vector uniquely points to one organization, remaining unchanged regardless of the organization's full name or abbreviation, even if organization A cannot be matched based on the second organization's full name, the second organization's full name can be converted into a vector, and organization A can be identified through vector matching.

[0075] Furthermore, since it is rare for the same organization to have multiple different full names, the preset list of organization names only needs to store the mapping relationship between the full name and abbreviation of one organization, and the mapping relationship between the full name and the vector. This is because the organization can be identified by vector matching, and there is no need to store the mapping relationship between the full name and abbreviation of another organization in the preset list of organization names. In this way, the preset list of organization names can be made more concise.

[0076] It is evident that by pre-converting at least one full organization name corresponding to each abbreviation in the list of organization names into a vector, the target organization corresponding to the target user can be directly identified from the list of organization names by calculating vector similarity.

[0077] Therefore, in this embodiment, matching can be performed based on either the vector corresponding to the organization's full name or the vector corresponding to its abbreviation. Each method has its advantages: vector matching based on the organization's full name preserves the original information to the greatest extent possible, facilitating matching using the detailed information within the full name; vector matching based on the organization's abbreviation, since the information in the abbreviation is refined key information, ensures that the vector is more targeted and removes interference from redundant information.

[0078] In one optional implementation, when calculating the similarity between third-party user data and each vector stored in a pre-defined list of organization names, the similarity is calculated based on a pre-trained vector similarity calculation model. The pre-trained vector similarity calculation model is obtained as follows: Multiple labeled positive and negative sample groups are constructed based on historical user data input by each historical user; each positive sample group contains vectors labeled as similar from at least two organizations, and each negative sample group contains vectors labeled as dissimilar from at least two organizations; the vector similarity calculation model is obtained by training on multiple positive and negative sample groups. Through the positive and negative sample groups, the model learns the features of similar and dissimilar vectors, thereby calculating the similarity between vectors based on the learning results.

[0079] Therefore, the list of organization names should contain at least the following information:

[0080] (1) The mapping relationship between the abbreviations and full names of multiple organizations, where each organization abbreviation corresponds to at least one organization full name;

[0081] (2) The mapping relationship between the abbreviations of each organization and the vector; and / or the mapping relationship between the abbreviations of each organization and at least one full name of the organization and the vector.

[0082] In the data processing method provided in this embodiment, multiple matching operations can be performed on the first user data: in the first matching process, abbreviation matching is performed based on the second user data used to represent the organization's abbreviation. This method is more efficient than vector matching and more accurate than matching directly based on the original data. In the second matching process, vector matching is performed based on the vector similarity calculation results, so that if the abbreviation matching fails, the organization with higher vector similarity can be selected as the target organization of the target user.

[0083] Therefore, in this embodiment, first feature data and second feature data can be extracted from the first user data input by the user. Based on these first and second feature data, second user data, which is used to construct a short name indicating the user's organization, can be generated. Since the second user data only contains the extracted feature data, redundant and complex interference information is removed, thus enabling faster and more accurate identification of the user's organization. Furthermore, if accurate identification is not possible using the second user data, third feature data indicating the user's industry can be further extracted. This allows for precise determination of the user's organization when organizations with similar names exist across multiple industries, avoiding identification errors caused by organizations with the same name in different industries.

[0084] Figure 2 This invention illustrates a user identification method based on yet another embodiment of the present invention. Figure 1 The data processing method shown is implemented as follows:

[0085] Step S210: Extract the indicator data of the target organization corresponding to the target user from the preset list of organization names.

[0086] The list of organization names is the same as in the previous embodiment, and is specifically used to store the following information:

[0087] (1) The mapping relationship between the abbreviations and full names of multiple organizations, where each organization abbreviation corresponds to at least one organization full name;

[0088] (2) The mapping relationship between the abbreviations of each organization and the vector; and / or the mapping relationship between the abbreviations of each organization and at least one full name of the organization and the vector.

[0089] (2) Indicator data of each organization.

[0090] The indicator data for each organization included in the pre-defined list of organization names is obtained in the following way: for the abbreviation and / or vector of the same organization, the initial indicator data corresponding to multiple user historical data corresponding to the abbreviation and / or vector of that organization is summarized; based on the summary results, the organization indicator data corresponding to the abbreviation or vector of that organization is obtained.

[0091] For the same organization, its abbreviation and vector are mutually corresponding. Compared with the full name, which is more affected by the standardization of user input, the abbreviation is more standardized. Therefore, based on the organization's abbreviation and / or vector, the initial indicator data corresponding to multiple user historical data for the same organization can be summarized, thereby making the obtained organizational indicator data of the organization more accurate and avoiding the problem of incorrect indicator calculation due to the existence of multiple non-standard full names for the same organization.

[0092] Step S220: Identify the security level of the target user based on the comparison results between the indicator data and the preset standard indicator threshold.

[0093] The standard indicator thresholds can be set according to actual conditions. Accordingly, the security level of the target user is identified based on the comparison between the indicator data and the preset standard indicator thresholds. For example, the security level can be divided into two levels: secure and insecure, or it can be further divided into multiple levels. This invention does not limit the specific details.

[0094] Step S230: Determine whether to respond to the target user's service request based on the target user's security level.

[0095] If the target user's security level is secure, then the service request from the target user will be responded to; if the target user's security level is insecure, then the service request from the target user will not be responded to. The target user's service request can be any type of business request, such as a service activation request or a service application request.

[0096] To facilitate understanding, the following specific example will be used to describe in detail the implementation details of this embodiment:

[0097] This example primarily focuses on data processing during the user credit investigation process. During the credit investigation, users need to input an organization name, which represents their workplace, school, or residential area. In short, organization names include those for workplaces, companies, hospitals, and schools. Typically, the organization name is the company name. However, due to potential inconsistencies in the organization names input by different users, the same organization name may correspond to different names depending on the user. To address this issue, this example preprocesses a large number of historical organization names input by existing users (also called historical users) offline, generating corresponding indicator data. Then, an online query process is used to determine the indicator data corresponding to the original organization name input by the current user.

[0098] First, we will introduce the process of preprocessing historical organization names. Figure 3 This diagram illustrates the process of preprocessing historical organization names for historical users. Figure 3 As shown, the preprocessing process specifically includes the following steps:

[0099] Step S301: Extract the historical organization names entered by historical users.

[0100] Among them, the historical organization name is the name of the credit reporting organization entered by each historical user during the credit reporting process.

[0101] Step S302: Process the characters contained in the historical organization name to obtain the processed historical organization name.

[0102] For example, removing parentheses, commas, and other symbols from historical organization names yields the processed historical organization names. Alternatively, special characters in organization names can be processed to ensure consistent data format, and the organization name after removing special characters can be used as the full organization name.

[0103] Step S303: Identify the place names and organization names contained in the processed historical organization names using entity recognition. Place names are geographical entities (i.e., the first feature data), and organization names are name entities (i.e., the second feature data). Place names and organization names together constitute the historical target entity.

[0104] Specifically, a portion of the organization name data can be pre-sampled and labeled, including both the place name and the organization name. For example, in "Chongqing XXXX Financial Co., Ltd. Chengdu Branch," 'Chongqing' and 'Chengdu' are labeled as place names, and 'XXXX' as the organization name. This labeled data is then trained using a BERT model for entity recognition. The trained BERT model (also called an entity recognition model) is then used to perform entity recognition on the place name and organization name in the full organization name, extracting the place name and organization name.

[0105] Step S304: Remove place names and organization names from the processed historical organization names, and perform word segmentation on the remaining text to obtain each historical word segment.

[0106] The remaining text consists of historical auxiliary information contained in the historical organization name, excluding the historical target entity. This historical auxiliary information is then segmented into individual historical words. For example, after removing the identified place names and organization names, the remaining text is segmented into words, as in the previous example, resulting in "finance," "shares," "limited," "company," and "branch."

[0107] Step S305: Calculate the word frequency of each historical word segment in the word segmentation results of the entire historical organization names, and extract industry feature words based on the statistical results.

[0108] Specifically, the frequency of occurrence of historical word segments corresponding to each historical organization name is statistically analyzed. Based on the statistical results, industry-specific words are extracted from the historical word segments and stored in an industry-specific word list. For example, word frequency statistics are performed on the word segmentation results of all user organization names, and words with high frequency are selected. Industry-specific words are then manually compiled from these words to form an industry thesaurus, such as "technology," "trade," and "construction," in order to extract industry-specific words from the organization names.

[0109] Step S306: Remove prefixes or suffixes such as province, city, and county from the identified place names. For example, the identified place names may contain words such as province, city, and county. Remove these words to ensure consistency in place names. For example, 'Chongqing City' and 'Chongqing' are both changed to 'Chongqing' to facilitate subsequent matching.

[0110] Step S307: Based on the place name, organization name and industry characteristic words after removing prefix or suffix information such as province, city and county, generate the abbreviation of the historical organization corresponding to the historical organization name, and store each abbreviation of the historical organization to the historical organization abbreviation set.

[0111] For example, the system combines place names and organization names to form an abbreviation of the organization name. For instance, 'Chongqing XXXX Financial Co., Ltd. Chengdu Branch' is converted to 'Chongqing XXXX Chengdu'. If a branch name is present, its location is included; otherwise, it's omitted. Similarly, if the company name includes a location name, it's included; otherwise, it's omitted. Industry terms are also recorded. This method unifies different organization names entered by the user that actually belong to the same organization. For example, 'China XX Insurance Co., Ltd.', 'China XX', 'China XX Insurance', and 'China XX Company' are all unified as 'China XX'. If the same organization name matches multiple organizations, they are distinguished by industry terms. For example, 'XXXX Group Co., Ltd.', 'XXXX Engineering Design Co., Ltd.', and 'XXXX Intelligent Technology Co., Ltd.' all share the organization name 'XXXX', but are differentiated by the industry terms 'Group', 'Engineering Design', and 'Intelligent Technology'.

[0112] Step S308: Fine-tune the vector generation model used to generate sentence vectors using labeled samples.

[0113] The vector generation model used to generate sentence vectors is also called the vector similarity calculation model. This model has two functions: first, it can generate vectors; second, it can calculate the similarity between the generated vectors and historical vectors. Fine-tuning the vector generation model using labeled samples can improve its accuracy. For example, using a pre-trained model of Sentence-BERT and manually labeling some samples, the model can be fine-tuned to learn simplified names of organizations that are not exactly the same but refer to the same organization. For example, 'China XX' and 'China XX Network Communication' refer to the same company (label 1), and 'ABCD' and 'AC' refer to the same company (label 1). The two names and their labels are then fed into Sentence-BERT for training, and the loss is the difference between the cosine similarity and the labeled similarity. Fine-tuning can adjust some samples that were not originally the most similar to become the most similar, improving accuracy.

[0114] Step S309: Using the fine-tuned vector generation model, vectorize the abbreviations of each historical organization to obtain the vectorized abbreviations of the historical organizations.

[0115] For example, a trained model can be used to vectorize abbreviated organization names. Vectorization means that when an organization name is input into the Sentence-BERT model, the model will obtain a vector representation of the organization name. For example, some models use 383-dimensional vectors, so an organization name is converted into a 383-dimensional vector. The vectors for the same organization name are the same, and the vectors for different organization names are different. However, similar organization names will have very similar cosine similarity scores. The specific similarity criteria can be fine-tuned using defined labeled samples, allowing the model to learn some similar cases. Regarding model fine-tuning, if the organization names are the same, the generated vectors will be the same regardless of the fine-tuning, so fine-tuning will not affect completely identical organization names.

[0116] Step S310: Build an index for each vectorized historical organization abbreviation using Facebook AI Similarity Search.

[0117] Specifically, the vectorized abbreviations of historical organizations (also called historical organization vectors) are stored in a historical organization vector set, and a vector index is built for each vectorized historical organization abbreviation using faiss. For example, the vectorized full set of organization name abbreviations (i.e., historical organization abbreviations) are deduplicated and an index is built using faiss.

[0118] Step S311: Calculate the similarity between each historical organization vector based on the vector index, and merge at least two historical organization vectors with a similarity greater than a preset threshold into one historical organization vector.

[0119] Specifically, for each historical organization vector, a vector search is performed using the vector index to obtain the historical organization vector with the highest similarity to that historical organization vector, and then the two highly similar historical organization vectors are merged.

[0120] In addition to merging vectors, the system can also merge historical organization names and abbreviations based on vector indices. For example, for each historical organization name or abbreviation, the system searches the vector index for the second most similar historical organization name or abbreviation. If the similarity between the second most similar historical organization name or abbreviation and the original historical organization name or abbreviation is greater than a preset threshold, the two historical organization names or abbreviations are merged.

[0121] In one implementation, a vector search is performed on the index for each abbreviated organization name. The second most similar organization name is selected, since the first most similar organization name is always itself. If the similarity exceeds a certain threshold, one of the similar names is removed. Both organization names are represented by retained vectors, indicating that they belong to the same organization. Determining the threshold requires randomly sampling a subset of samples to calculate similarity. Accuracy is verified manually, and a threshold with high precision and recall is selected. For example, 1000 samples are sampled, and the second most similar organization names are selected. Manual verification is performed to determine if two different organization names belong to the same organization and whether the labeling is correct. Then, precision and recall are calculated using a selection threshold. For example, a similarity score above 0.99 is considered similar. The results are compared with the manually labeled verification results to obtain precision and recall, and the threshold with the higher precision and recall is selected. The purpose of vectorizing organization names for similarity comparison is to address situations where organization name abbreviations cannot be matched. For example, consider "China XX Co., Ltd." and "China XX Network Communications Co., Ltd." The former is abbreviated as "China XX" after entity identification, while the latter is abbreviated as "China XX Network Communications". They are actually the same company, but have different abbreviations, so similarity comparison is needed to merge them.

[0122] Step S312: For the merged historical organization vector, construct the indicator data corresponding to the historical organization vector.

[0123] Specifically, the initial indicator data for the historical organization names corresponding to each merged historical organization vector are aggregated to obtain organizational indicator data corresponding to the historical organization vectors. For example, aggregation is performed by vector and industry term to ensure that the vector and industry term uniquely identify an organization, and indicator data such as the delinquency rate of each organization are calculated. The finally determined deduplicated vectors are used to construct a vector index using Faiss for online matching.

[0124] Step S313: Establish a mapping relationship between historical organization vectors and historical organization abbreviations, so that the risk control indicator data corresponding to the historical organization abbreviations are consistent with the risk control indicator data corresponding to the historical organization vectors.

[0125] Specifically, historical organization vectors, historical organization abbreviations, and risk control indicator data can all be stored in a data table for easy subsequent querying. For example, the correspondence between full organization names, abbreviations, organization name vectors, industry terms, and indicators can be constructed into an organization name list table (i.e., the organization name list mentioned above), as shown in the table below. Full organization names are unique, while abbreviations and vectors can be repeated. For repeated abbreviations and vectors, their corresponding indicator values ​​with industry terms are the same. Table 1 shows a schematic diagram of the data table content.

[0126] Table 1

[0127]

[0128] As shown in Table 1, the full organization name is the historical organization name, which is the complete organization name entered by historical users (i.e., the full name of the organization). The historical organization abbreviation is obtained by extracting the location name, organization name, and industry feature words from the historical organization name. The organization name vector index is a vector index constructed based on the historical organization vector. Indicator 1...Indicator n are all indicator data. For example, in one implementation, Indicator 1 is the overdue rate, Indicator 2 is the credit default data, Indicator 3 is the enterprise asset data, etc.

[0129] Furthermore, as shown in Table 1, even if the full name and abbreviation of an organization are different, the historical organization vectors obtained after vectorization are all X1 because they contain many identical words. Moreover, the indicator data corresponding to the same historical organization vector are obtained by summarizing the initial indicator data of multiple historical organization names corresponding to that historical organization vector. For example, taking indicator 1 as an example, assuming indicator 1 is the organization's delinquency rate indicator, and assuming the initial indicator data for "Chongqing XXXX Financial Co., Ltd. Chengdu Branch" records a delinquency rate of 0.3, the initial indicator data for "Chongqing XXXX Co., Ltd. Chengdu Branch" records a delinquency rate of 0.5, and the initial indicator data for "Chongqing XYXY Financial Co., Ltd. Chengdu Branch" records a delinquency rate of 0.6, then the summarized indicator data corresponding to the historical organization vector X1 is the summarized value obtained by summarizing 0.3, 0.5, and 0.6. The summarization operation can be various operations such as calculating the mean or median, and this invention does not limit this.

[0130] After preprocessing the historical organization names of historical users, the online query process can be realized. Figure 4 This diagram illustrates the process of online querying and risk control based on organization name. Figure 4 As shown, the specific steps include:

[0131] Step S401: Obtain the original organization name entered by the current user.

[0132] The current user is the user currently performing the credit reporting operation, and the original organization name is the credit reporting organization name entered by the current user. Optionally, in this step, symbols can be removed from the original organization name to obtain the original organization name after removing symbols. For example, characters such as parentheses and commas in the original organization name can be removed.

[0133] Step S402: Match the original organization name with all historical organization names. If there is a complete match, directly obtain the indicator data based on the matched historical organization name; if there is no complete match, continue to step S403. Special characters in the organization name can be removed. The processed organization name is then matched with each historical organization name in the data table. Matches that are found are used as the final matching result; otherwise, further processing is performed.

[0134] Step S403: Identify the place names and organization names contained in the original organization name, and extract industry feature words.

[0135] Step S404: Remove the prefix or suffix information such as province, city, and county from the identified place names.

[0136] Step S405: Combine the place name, organization name, and industry characteristic words to form a simplified organization name.

[0137] Step S406: Match the simplified organization name with the set of historical organization abbreviations.

[0138] Specifically, the simplified organization name will be matched with each historical organization abbreviation contained in the historical organization abbreviation set. If there is a complete match, the risk control indicator data will be obtained directly based on the matched historical organization name; if there is no complete match, step S407 will continue.

[0139] For example, the processed simplified organization name is matched against the historical organization abbreviations in the data table. A unique match is used as the final result. If no unique match is found, an industry-specific match is performed. A matching result is used as the final result, while unmatched results are processed further. A unique match means that the organization name abbreviation has only one record. For example, in the previous example, "XXXX" has only one record and can be matched without needing to match the industry term "finance." "YYYY" has multiple records and requires a unique match using industry terms.

[0140] Step S407: Vectorize the simplified organization name to obtain the organization name vector.

[0141] Step S408: Perform a vector search in the vector index using faiss to obtain the historical organization vector with the highest similarity to the organization name vector.

[0142] Specifically, by using vector indexing, the historical organization vector that is most similar to the organization name vector to be matched is obtained, and the organization name corresponding to the historical organization vector is used as the matching organization name.

[0143] Step S409: Obtain the risk control indicator data of the historical organization vector that is most similar to the organization name vector to be matched.

[0144] Step S410: Develop a processing strategy based on the obtained metric data.

[0145] In summary, in this solution, by unifying organization names with different writings into the same organization and calculating various metrics for the corresponding organization, the accuracy of metric calculation is ensured. For example, the organization names filled in by users among the existing users may include 'China XX', 'China XX Network Communications', 'China XX Co., Ltd.', etc. These organization names actually refer to the same organization. If they are not unified, when calculating metrics, they will be calculated separately, and when matching, they will also be matched separately, resulting in possibly different metrics for users who fill in different names. Therefore, it is necessary to unify such organization names. After unification, the number of users in each organization also increases, improving the statistical credibility of the metrics.

[0146] In addition, this solution uses three different methods for matching. The accuracies of the three methods decrease from high to low. First, use the method with the highest accuracy (i.e., directly based on the original organization name) for matching. For those that cannot be matched, then use the method with slightly lower accuracy (i.e., based on the simplified organization name) for matching. If still unable to match, then match according to the organization name vector, thus ensuring the final accuracy of the matching and also ensuring the efficiency during real-time matching. If the organization name can be directly matched, there is no need to use the latter two methods, and the efficiency is the highest. The second method requires entity recognition and consumes a certain amount of time. The third method requires vectorization and vector search, consuming even more time. Therefore, through the three methods, both accuracy and efficiency can be taken into account.

[0147] In addition, this solution performs matching based on the results after entity recognition, solving the problem of unmatched caused by some irregularly filled company names, such as the unmatched problem caused by filling the suffix as 'Co., Ltd.' and 'Limited Liability Company'.

[0148] Moreover, this solution uses the method of comparing sentence vector similarity for matching, which can solve the problem of unmatched caused by some typos or abbreviations. By adjusting the similarity threshold, the accuracy can be controlled within a certain range to meet the final matching requirements. For example, the matching problem between 'China XX' and 'China XX Network Communications'. And use the name after entity recognition for similarity comparison, avoiding the influence of common words such as 'company' and'stock' on similarity comparison. For example, for 'China XX Co., Ltd.' and 'China YY Co., Ltd.', if directly using the full names for similarity comparison, only two characters are different between them, and the similarity is relatively high. However, after entity recognition, when comparing the similarity between 'China XX' and 'China YY', the similarity will be greatly reduced.

[0149] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further. Those skilled in the art will understand that in the above methods of specific implementation, the specific execution order of each step should be determined by its function and possible internal logic.

[0150] It should be noted that all user data involved in this embodiment is data entered by the user themselves. Data acquisition requires the user's authorization and confirmation. Therefore, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution of this disclosure all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0151] In addition, this disclosure also provides a data processing apparatus, an electronic device, and a computer-readable storage medium, all of which can be used to implement any of the data processing methods provided in this disclosure. The corresponding technical solutions and descriptions are described in the corresponding section of the method and will not be repeated here.

[0152] Figure 5 This is a block diagram of a data processing apparatus provided in an embodiment of the present disclosure.

[0153] Reference Figure 5 This disclosure provides a data processing apparatus 50, which includes:

[0154] The acquisition module 51 is adapted to acquire first user data input by the target user, and extract first feature data and second feature data from the first user data. The first feature data is used to indicate the address information of the organization where the target user is located, and the second feature data is used to indicate the keyword information of the organization where the target user is located.

[0155] Construction module 52 is adapted to construct second user data based on the first feature data and the second feature data, wherein the second user data is used to indicate the abbreviation of the organization to which the target user belongs;

[0156] The determining module 53 is adapted to determine, based on the second user data and a preset list of organization names, whether the second user data points to an organization included in the preset list of organization names, wherein the preset list of organization names includes a mapping relationship between the abbreviations and full names of multiple organizations, and each organization abbreviation corresponds to at least one organization full name;

[0157] Extraction module 54 is adapted to extract third feature data from the first user data if no, the third feature data being used to indicate industry information of the organization to which the target user belongs;

[0158] The identification module 55 is adapted to update the third feature data to the second user data, so as to identify the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the organization to which the target user belongs.

[0159] Optionally, the acquisition module is specifically adapted to:

[0160] The first user data is input into the trained feature extraction model, and the first feature data and the second feature data are extracted based on the output of the trained feature extraction model.

[0161] The trained feature extraction model is obtained in the following way:

[0162] Retrieve historical user data entered by each historical user;

[0163] Obtain the address-type feature data and keyword-type feature data marked in the historical user data;

[0164] The historical user data, along with the address-type and keyword-type feature data annotated within the historical user data, are used as training samples. An initial feature extraction model is trained based on these training samples to obtain the trained feature extraction model.

[0165] Optionally, the extraction module is specifically adapted to:

[0166] Extract auxiliary descriptive information from the first user data, excluding the first feature data and the second feature data;

[0167] The auxiliary description information is segmented to obtain the original words;

[0168] Each original word segment is matched with a preset list of industry feature words, and the original word segments that successfully match the list of industry feature words are extracted as the third feature data;

[0169] The industry feature word list is determined based on the word frequency of each original word segment contained in the historical user data input by each historical user.

[0170] Optionally, the identification module is specifically adapted to:

[0171] Based on the abbreviation of the organization to which the target user belongs, the mapping relationship between the abbreviations and full names of multiple organizations included in the preset list of organization names is found, and the full name of the organization corresponding to the abbreviation of the organization to which the target user belongs is determined;

[0172] When the full name of the organization corresponds to multiple organizations, the target organization corresponding to the target user is identified from the multiple organizations based on the industry information.

[0173] Optionally, the identification module is specifically adapted to:

[0174] If the second user data does not point to any organization in the preset list of organization names, the updated second user data is vectorized to obtain third user data; wherein, the updated second user data includes: first feature data, second feature data, and third feature data; the third user data is used to indicate the vector of the organization to which the target user belongs;

[0175] The similarity between the third user data and each vector stored in the preset list of organization names is calculated, and the target organization corresponding to the target user is identified based on the vectors whose similarity values ​​are greater than a preset similarity threshold; wherein, the preset list of organization names also includes the mapping relationship between at least one full name of an organization corresponding to each organization abbreviation and the vector.

[0176] Optionally, the identification module is specifically adapted to:

[0177] Based on the pre-trained vector similarity calculation model, the similarity between the third user data and each vector stored in the preset list of organization names is calculated.

[0178] The pre-trained vector similarity calculation model is obtained in the following way:

[0179] Based on the historical user data input by each historical user, multiple labeled positive sample groups and negative sample groups are constructed; each positive sample group contains vectors of at least two organizations labeled as similar, and each negative sample group contains vectors of at least two organizations labeled as dissimilar.

[0180] The vector similarity calculation model is obtained by training based on the multiple positive sample groups and negative sample groups.

[0181] Figure 6 This diagram illustrates a user identification device implemented based on the above-described data processing apparatus according to another embodiment of the present invention, specifically including:

[0182] The data extraction module 61 is adapted to extract the indicator data of the target organization corresponding to the target user from a preset list of organization names;

[0183] The security level identification module 62 is adapted to identify the security level of the target user based on the comparison result between the indicator data and the preset standard indicator threshold.

[0184] The service response module 63 is adapted to determine whether to respond to the service request of the target user based on the security level of the target user.

[0185] The indicator data for each organization included in the preset list of organization names are obtained in the following way:

[0186] For the same organization's abbreviation or vector, aggregate the initial indicator data corresponding to multiple user historical data corresponding to the organization's abbreviation or vector; based on the aggregation results, obtain the organization indicator data corresponding to the organization's abbreviation or vector.

[0187] Figure 7 This is a block diagram of an electronic device provided in an embodiment of the present disclosure.

[0188] Reference Figure 7 This disclosure provides an electronic device 70, which includes: at least one processor 501; at least one memory 502; and one or more I / O interfaces 503 connected between the processor 501 and the memory 502; wherein the memory 502 stores one or more computer programs that can be executed by the at least one processor 501, and the one or more computer programs are executed by the at least one processor 501 to enable the at least one processor 501 to perform the above-described risk control method based on organization name.

[0189] This disclosure also provides a computer-readable storage medium storing a computer program thereon, wherein the computer program implements the above-described method when executed by a processor / processor core. The computer-readable storage medium may be a volatile or non-volatile computer-readable storage medium.

[0190] This disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the above-described method.

[0191] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0192] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable program instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0193] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0194] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0195] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0196] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0197] These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable organization-based risk control device to produce a machine such that, when executed by the processor of the computer or other programmable organization-based risk control device, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, a programmable organization-based risk control device, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0198] Computer-readable program instructions may also be loaded onto a computer, other programmable organization-name-based risk control device, or other device to cause a series of operational steps to be performed on the computer, other programmable organization-name-based risk control device, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable organization-name-based risk control device, or other device to perform the functions / actions specified in one or more boxes of the flowchart and / or block diagram.

[0199] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0200] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in connection with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in connection with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this disclosure as set forth by the appended claims.

Claims

1. A data processing method, characterized by, include: First user data input by the target user is obtained, and first feature data and second feature data are extracted from the first user data. The first feature data is used to indicate the address information of the organization where the target user is located, and the second feature data is used to indicate the keyword information of the organization where the target user is located. Second user data is constructed based on the first feature data and the second feature data, wherein the second user data is used to indicate the abbreviation of the organization to which the target user belongs; Based on the second user data and a preset list of organization names, determine whether the second user data uniquely matches an organization contained in the preset list of organization names. The preset list of organization names includes a mapping relationship between the abbreviations and full names of multiple organizations, and each organization abbreviation corresponds to at least one organization full name. If not, then extract third feature data from the first user data, the third feature data being used to indicate the industry information of the organization to which the target user belongs; The third feature data is updated to the second user data to identify the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the organization to which the target user belongs.

2. The method of claim 1, wherein, The step of extracting the first feature data and the second feature data from the first user data includes: The first user data is input into the trained feature extraction model, and the first feature data and the second feature data are extracted based on the output of the trained feature extraction model. The trained feature extraction model is obtained in the following way: Retrieve historical user data entered by each historical user; Obtain the address-type feature data and keyword-type feature data marked in the historical user data; The historical user data, along with the address-type and keyword-type feature data annotated within the historical user data, are used as training samples. An initial feature extraction model is trained based on these training samples to obtain the trained feature extraction model.

3. The method of claim 1, wherein, The step of extracting the third feature data from the first user data includes: extracting auxiliary descriptive information contained in the first user data other than the first feature data and the second feature data; The auxiliary description information is segmented to obtain the original words; Each original word segment is matched with a preset list of industry feature words, and the original word segments that successfully match the list of industry feature words are extracted as the third feature data; The industry feature word list is determined based on the word frequency of each original word segment contained in the historical user data input by each historical user.

4. The method of claim 1, wherein, The step of identifying the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the target user's organization includes: Based on the abbreviation of the organization to which the target user belongs, the mapping relationship between the abbreviations and full names of multiple organizations included in the preset list of organization names is found, and the full name of the organization corresponding to the abbreviation of the organization to which the target user belongs is determined; When the full name of the organization corresponds to multiple organizations, the target organization corresponding to the target user is identified from the multiple organizations based on the industry information.

5. The method of claim 1, wherein, The step of identifying the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the target user's organization includes: If the second user data does not point to any organization in the preset list of organization names, the updated second user data is vectorized to obtain third user data; wherein, the updated second user data includes: first feature data, second feature data, and third feature data; the third user data is used to indicate the vector of the organization to which the target user belongs; Calculate the similarity between the third user data and each vector stored in the preset list of organization names, and identify the target organization corresponding to the target user based on the vector with a similarity value greater than a preset similarity threshold; wherein, the preset list of organization names also includes: the mapping relationship between vectors and abbreviations of multiple organizations, and / or the mapping relationship between at least one full name of an organization corresponding to each organization abbreviation and a vector.

6. The method of claim 5, wherein, The calculation of the similarity between the third user data and each vector stored in the preset list of organization names includes: Based on the pre-trained vector similarity calculation model, the similarity between the third user data and each vector stored in the preset list of organization names is calculated. The pre-trained vector similarity calculation model is obtained in the following way: Based on the historical user data input by each historical user, multiple labeled positive sample groups and negative sample groups are constructed; each positive sample group contains vectors of at least two organizations labeled as similar, and each negative sample group contains vectors of at least two organizations labeled as dissimilar. The vector similarity calculation model is obtained by training based on the multiple positive sample groups and negative sample groups.

7. A user identification method based on the data processing method according to any one of claims 1-6, characterized in that, include: Extract the target organization's metrics data corresponding to the target user from the pre-defined list of organization names; The security level of the target user is identified based on the comparison results between the indicator data and the preset standard indicator threshold. Based on the target user's security level, determine whether to respond to the target user's service request; The indicator data for each organization included in the preset list of organization names are obtained in the following way: For the same organization's abbreviation or vector, aggregate the initial indicator data corresponding to multiple user historical data corresponding to the organization's abbreviation or vector; based on the aggregation results, obtain the organization indicator data corresponding to the organization's abbreviation or vector.

8. A data processing apparatus, characterized in that, include: The acquisition module is adapted to acquire first user data input by a target user, and extract first feature data and second feature data from the first user data. The first feature data is used to indicate the address information of the organization where the target user is located, and the second feature data is used to indicate the keyword information of the organization where the target user is located. The construction module is adapted to construct second user data based on the first feature data and the second feature data, wherein the second user data is used to indicate the abbreviation of the organization to which the target user belongs; The determining module is adapted to determine, based on the second user data and a preset list of organization names, whether the second user data uniquely matches an organization contained in the preset list of organization names, wherein the preset list of organization names includes a mapping relationship between the abbreviations and full names of multiple organizations, and each organization abbreviation corresponds to at least one organization full name; The extraction module is adapted to extract third feature data from the first user data if not otherwise specified, the third feature data being used to indicate industry information of the organization to which the target user belongs; The identification module is adapted to update the third feature data to the second user data, so as to identify the target organization corresponding to the target user from the preset list of organization names based on the abbreviation and industry information of the organization to which the target user belongs.

9. A user identification device implemented based on the data processing apparatus of claim 8, characterized in that, include: The data extraction module is suitable for extracting the indicator data of the target organization corresponding to the target user from a preset list of organization names; The security level identification module is adapted to identify the security level of the target user based on the comparison result between the indicator data and the preset standard indicator threshold. The service response module is adapted to determine whether to respond to the service request of the target user based on the security level of the target user; The indicator data for each organization included in the preset list of organization names are obtained in the following way: For the same organization's abbreviation or vector, aggregate the initial indicator data corresponding to multiple user historical data corresponding to the organization's abbreviation or vector; based on the aggregation results, obtain the organization indicator data corresponding to the organization's abbreviation or vector.

10. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs that can be executed by the at least one processor, the one or more computer programs being executed by the at least one processor to enable the at least one processor to perform the method as described in any one of claims 1-7.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when executed by a processor, implements the method as described in any one of claims 1-7.

Images