Data acquisition method and device, computing equipment and computer readable storage medium

Abstract

The embodiment of the invention provides a data acquisition method and device, computing equipment and a computer readable storage medium, and the method comprises the steps: responding to a claim settlement request, and determining a candidate mechanism identifier in a plurality of reference mechanism identifiers based on the similarity with a to-be-analyzed mechanism identifier carried by the claim settlement request; through a text reasoning model, based on the to-be-analyzed mechanism identifier and the candidate mechanism identifier, determining a target mechanism identifier representing the same mechanism with the to-be-analyzed mechanism identifier; and mechanism data corresponding to the target mechanism identifier is acquired, and the mechanism data is used for executing a task corresponding to the claim settlement request. The method can improve the processing effect of the task corresponding to the claim settlement request.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of computer technology, and in particular to a data acquisition method, as well as a data acquisition device, a computing device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the development of computer technology, various data processing tasks can be completed by computers, and the requirements for the processing efficiency and accuracy of various information are becoming increasingly higher.

[0003] For example, in the insurance claim scenario, various claims requests need to be verified accordingly. Usually, the claim information is analyzed and processed based on third-party data to determine whether the claim request is compliant, and then determine whether to provide compensation to the insured.

[0004] However, since the claim information submitted by the insured is usually messy, it is difficult to analyze the claim information, and the efficiency and effect of the task processing corresponding to the claim request are poor. Summary of the invention

[0005] The embodiments of this specification provide a data acquisition method, which can improve the execution effect of tasks corresponding to claim settlement requests. One or more embodiments of this specification also relate to a data acquisition device, a computing device, a computer-readable storage medium, and a computer program product.

[0006] According to one aspect of an embodiment of this specification, a data acquisition method is provided, including: In response to a claim settlement request, determining a candidate institution identifier among a plurality of reference institution identifiers based on a similarity to the institution identifier to be analyzed carried in the claim settlement request; Determine, through a textual reasoning model, a target institution identifier representing the same institution as the institution identifier to be analyzed based on the institution identifier to be analyzed and the candidate institution identifiers; Obtaining the organization data corresponding to the target organization identifier, wherein the organization data is used to execute the task corresponding to the claim settlement request.

[0007] According to another aspect of an embodiment of this specification, there is provided a data acquisition device, including: A first determination module is used to determine, in response to a claim settlement request, a candidate institution identifier from among a plurality of reference institution identifiers based on a similarity with an institution identifier to be analyzed carried in the claim settlement request; A second determination module is used to determine, through a text reasoning model, a target institution identifier representing the same institution as the institution identifier to be analyzed based on the institution identifier to be analyzed and the candidate institution identifiers; The first acquisition module is used to acquire the institutional data corresponding to the target institutional identifier, wherein the institutional data is used to execute the task corresponding to the claim settlement request.

[0008] According to another aspect of the embodiments of this specification, there is provided a computing device, including: a memory and a processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor, the steps of the above method are implemented.

[0009] According to another aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores a computer program / instruction, and the steps of the above method are implemented when the computer program / instruction is executed by a processor.

[0010] According to another aspect of the embodiments of this specification, a computer program product is provided, including a computer program / instruction, and when the computer program / instruction in the computer program product is executed in a processor, the steps of the above method are implemented.

[0011] In one embodiment of this specification, for the identification of the institution to be analyzed carried by the claim request, a similar candidate identification of the institution is first searched in multiple reference identifications, and then the text reasoning model is called to further determine the target identification of the institution that is substantially the same as the identification of the institution to be analyzed based on the candidate identification. In this way, it can be ensured that the target identification of the institution can be determined more accurately among the multiple reference identifications, and the institution data matching the claim request can be accurately obtained accordingly, and the task execution effect for the claim request can be improved based on the institution data. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a flow chart of a data acquisition method provided in an embodiment of this specification; Figure 2 It is a structural schematic diagram of a data acquisition device provided in an embodiment of this specification; Figure 3 It is a structural block diagram of a computing device provided in one embodiment of this specification. DETAILED DESCRIPTION

[0013] Many specific details are described in the following description to facilitate a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of this specification, so this specification is not limited to the specific implementation disclosed below.

[0014] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms of "one", "said" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items. The term "at least one" in one or more embodiments of this specification refers to "one or more", and "multiple" refers to "two or more". The term "including" is an open description and should be understood as "including but not limited to", and may include other content on the basis of the content already described.

[0015] It should be understood that although the terms "first", "second", etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, "first" may also be referred to as "second", and similarly, "second" may also be referred to as "first". Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0016] In addition, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant standards and requirements, and corresponding operation entrances shall be provided for users to choose to authorize or refuse.

[0017] In one or more embodiments of this specification, the model involved may refer to a large model. A large model refers to a deep learning model with large-scale model parameters, which usually contains hundreds of millions, tens of billions, hundreds of billions, trillions or even more than ten trillion model parameters. A large model can also be called a foundation model / foundation model. The large model is pre-trained with a large-scale unlabeled corpus to produce a pre-trained model with more than 100 million parameters. This model can adapt to a wide range of downstream tasks, and the model has good generalization capabilities, such as a large-scale language model (LLM, Large Language Model), a multi-modal pre-training model, etc. When the large model is actually applied, only a small number of samples are needed to fine-tune the pre-trained model and it can be applied to different tasks. The large model can be widely used in natural language processing (NLP), computer vision and other fields, and can be specifically applied to computer vision tasks such as visual question answering (VQA), image description (IC, Image Caption), image generation, and other tasks in the field of computer vision, as well as text-based sentiment classification, text summary generation, machine translation and other natural language processing tasks. The main application scenarios of the large model include digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, etc. The models involved in the embodiments of this specification include but are not limited to the above-mentioned large models, and can be any type of machine learning model. For example, it can be a convolutional neural network (CNN) model, a Transformer model, etc.

[0018] At present, people pay more and more attention to insurance in all aspects, and there are more and more insurance products. Correspondingly, the requirements for insurance claims processing are also getting higher and higher, such as requiring insurance claims to have high execution efficiency and execution effect. In the process of insurance claims, the insured can submit a claim request to the claim platform for the insurance claim case. The claim platform performs the corresponding claim verification task based on the claim request, reviews the insurance claim case, and confirms whether the claim should be compensated, how it should be compensated, or how it should be rejected, so as to realize the processing of the claim request. The claim platform can also provide users or institutions with the risks of insurance based on the analysis of claim information, so as to plan and adjust the insurance products and realize risk control. The claim platform can refer to the service platform of the issuer of the insurance product. In the process of claim verification, the claim platform needs to collect relevant information from various channels, such as combining data from third-party data sources to review the insurance claim case.

[0019] For example, in the insurance claims and risk control scenarios in the medical field, the insured can initiate a claim request for an insurance product for their medical expenses. The claim request carries the information of the collection agency of the medical expenses, such as the name of the hospital. The claims platform needs to match the hospital name carried in the claim request in the third-party data source to obtain the data of the hospital from the third-party data source for claims verification. Accurate matching of the hospital name is crucial for insurance claims. However, in actual scenarios, since the hospital name in the claim request is entered by the user, different users may enter different results for the same hospital. There are many reasons why the hospital name in the claim request cannot be accurately matched to the hospital in the third-party data source.

[0020] For example, the name of the hospital carried in the claim request may be obtained through optical character recognition (OCR), and there may be certain errors in the recognition process. The hospital name may also be manually entered by the user, and the manual entry may also have spelling errors or inconsistent formats. There is also a difference between the full name and the abbreviation of the hospital name. The hospital name carried in the claim request and the hospital name in the third-party data source may use the abbreviation and full name respectively. It is also possible that due to regional differences, there may be slight differences in the names of hospitals in different regions. For example, "City A Hospital B" and "B Hospital" may be two different hospitals. These situations may cause the hospital name carried in the claim request to be unable to be accurately matched in the third-party data source.

[0021] In a related technology, string matching is used to match hospital names. For example, the Levenshtein distance (also known as the edit distance) between two hospital names can be determined, and the similarity between the strings corresponding to the two hospital names can be measured by the edit distance, and a certain threshold can be set to determine whether the two hospital names match. This method is simple to implement, easy to understand and maintain, but the matching effect is poor for complex situations (such as situations where the abbreviation and the full name are inconsistent). In addition, the threshold needs to be set manually, which is prone to mismatching or missing matches.

[0022] In another related technology, a rule matching method is used to match hospital names. A series of rules are formulated according to the characteristics of hospital names, such as replacing common abbreviations, and regular expressions or predefined rule libraries can be used for matching. This method can have a better matching effect for hospital names that meet specific rules, and can handle some common abbreviations that are inconsistent with the full name. However, the maintenance cost of the rule library is high, and it needs to be constantly updated, which consumes a lot of resources, and the ability to handle unforeseen situations is relatively limited.

[0023] In another related technology, matching is performed based on word vectors, such as using a word vector model (such as Word2Vec, GloVe, etc.) to segment the hospital name into multiple words, convert each word into a corresponding word vector, and calculate the similarity between each word vector and the words in the hospital name to be compared in the vector space, so as to find the closest hospital name. This method can match hospital names with certain semantic similarities, and has better robustness than simple string matching methods. However, this method requires training and maintenance of word vector models, which has high maintenance costs, and the processing effect on long texts (such as complete hospital names) is also relatively limited.

[0024] Therefore, there is an urgent need to solve the problem of accurate matching caused by the differences in hospital names between different data sources.

[0025] The embodiments of this specification provide a data acquisition method that can improve the processing effect of tasks corresponding to claim requests. This specification also relates to a data acquisition device, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail one by one in the following embodiments. The data acquisition method can be executed by the data acquisition device or computing device, such as the data acquisition device and computing device can be a server-side device, or can also be a terminal device such as a desktop computer, a laptop computer, or a smart phone.

[0026] Figure 1 This is a flow chart of a data acquisition method provided in an embodiment of this specification, which can be applied to a data acquisition device. For example, the data acquisition device can be a server corresponding to a claims platform, and the user can send a claim request to the claims platform through an application, webpage or applet corresponding to the claims platform; or, the user can submit a claim request to the claims platform through a staff member at an offline service point of the claims platform. After receiving the claim request, the data acquisition device can execute the data acquisition method provided in an embodiment of this specification, and further execute subsequent processes to achieve the task corresponding to the claim request. Figure 1 As shown, the data acquisition method includes the following steps 102 to 106.

[0027] Step 102: In response to a claim settlement request, a candidate institution identifier is determined from among a plurality of reference institution identifiers based on a similarity with the institution identifier to be analyzed carried in the claim settlement request.

[0028] The data acquisition device may receive a claim request triggered by the insured, and the claim request may carry a structure identifier to be analyzed. The identifier of the institution to be analyzed may be the identifier of the collection institution to which the insured has paid or is to pay. For example, in the medical field, the collection institution may be a hospital, and the identifier of the institution to be analyzed may be the name of the hospital. For another example, in the field of auto insurance, the collection institution may be a vehicle repair shop, and the identifier of the institution to be analyzed may be the name of the vehicle repair shop. The embodiments of this specification are explained using the name of a hospital as an example.

[0029] The data acquisition device can obtain the institutional data of various institutions from multiple channels, and then find the institution targeted by the claim request (that is, the institution indicated by the identification of the institution to be analyzed) in each reference institution whose institutional data is obtained, so as to check whether the claim should be paid and how it should be paid based on the institutional data of the reference institution for the claim request. The type of the reference institution is the same type of institution as the institution indicated by the indicator of the institution to be analyzed, such as the reference institutions are all hospitals. The institutional data of the reference institution may include the identification, introduction information, personnel information, transaction records and related claims records of the reference institution. The identification of the reference institution is referred to as the reference institution identification below. The data acquisition device can obtain the identification of each reference institution to obtain multiple reference institution identifications, and by matching the identification of the institution to be analyzed in these reference institution identifications, determine the institutional data required for claim settlement, so as to accurately settle the claim. Optionally, the data acquisition device can continuously obtain the latest institutional data of various institutions related to the insurance products it publishes, so as to efficiently perform related tasks when receiving a claim settlement request.

[0030] The data acquisition device may first determine the similarity between the structure identifier to be analyzed and each reference institution identifier, and then based on the similarity, determine the corresponding reference institution identifier with a similarity higher than a set threshold as a candidate institution identifier. Optionally, a maximum number of candidate institution identifiers may also be set. When the number of reference institution identifiers with a similarity higher than the set threshold exceeds the number, the reference institution identifier with a higher similarity is selected as the candidate institution identifier based on the number. For example, the data acquisition device may use the matching method in the three related technologies described above to determine the candidate institution identifier.

[0031] In one embodiment, the data acquisition device can determine the candidate institution identifier by vectorizing both the institution identifier to be analyzed and the reference institution identifier, and performing similarity analysis on the obtained vectors. Converting the institution identifiers into vectors that are easier to compare can help improve the comparison effect and the accuracy of determining the candidate institution identifiers. Accordingly, step 102 may include: vectorizing the institution identifier to be analyzed carried by the claim request to obtain a vector to be analyzed; determining a candidate vector whose similarity with the vector to be analyzed meets the target condition among multiple vectors corresponding to multiple reference institution identifiers; wherein the vector corresponding to each reference institution identifier is obtained by vectorizing each reference institution identifier; and determining the reference institution identifier corresponding to the candidate vector as the candidate institution identifier.

[0032] The data acquisition device can obtain multiple reference organization identifiers in advance, and vectorize each reference organization identifier to obtain a vector corresponding to each reference organization identifier. These vectors can be stored in a vector database, such as a Milvus vector database. The vectors in the vector database can be continuously updated based on the latest data.

[0033] The reference organization identification in the embodiments of this specification may be derived from a third-party data source. Accordingly, the data acquisition method provided in the embodiments of this specification, before respectively performing vectorization processing on multiple reference organization identifications to obtain multiple vectors, further includes: obtaining multiple reference organization identifications from a third-party data source, wherein the third-party data source includes organization data of multiple organizations, and the organization data of each organization includes the reference organization identification of each organization. By way of example, the data acquisition device may have a third-party data source interface, and the data acquisition device may achieve connection with the third-party data source through the third-party data source interface, thereby acquiring data from the third-party data source.

[0034] In one embodiment, the data acquisition device can use a vector model to implement vectorization processing of the organization logo. The vector model can map any text into a low-dimensional dense vector for tasks such as retrieval, classification, clustering or semantic matching, and can support the call of external knowledge for a large model. For example, the vector model can be a text embedding model. Accordingly, among the multiple vectors corresponding to the above-mentioned multiple reference organization logos, before determining the candidate vector whose similarity with the vector to be analyzed meets the target condition, the method provided in the embodiment of this specification also includes: using the text embedding model to respectively perform vectorization processing on the multiple reference organization logos to obtain multiple vectors.

[0035] The text embedding model can be a pre-trained M3E (Multi-Modal Multi-Task Embedding) model, which is a text embedding method based on contrastive learning. By performing joint training on multiple tasks, richer semantic information can be learned. The M3E model is trained using an in-batch negative sampling contrastive learning method, which can effectively improve the quality of the embedding vector. Optionally, the data acquisition device can also use other text embedding models, such as a BGE (Bilingual General Embedding) model or a text2vec model.

[0036] After receiving the claim request, the data acquisition device can perform vectorization processing on the identification of the institution to be analyzed carried by the claim request. The vectorization processing method is the same as the vectorization processing method for the reference institution identification, so as to ensure that the obtained vector format is consistent and facilitate vector comparison. Accordingly, the above-mentioned vectorization processing of the identification of the institution to be analyzed carried by the claim request to obtain the vector to be analyzed includes: using a text embedding model to vectorize the identification of the institution to be analyzed carried by the claim request to obtain the vector to be analyzed. The text embedding model here is the same as the text embedding model used for the reference institution identification, such as both are M3E models.

[0037] After obtaining the vector to be analyzed, the data acquisition device can search for adjacent vectors for the vector to be analyzed in the vector database. The data acquisition device can compare the vector to be analyzed with each vector in the vector database (that is, the reference vector corresponding to the reference institution identifier), determine the similarity between each vector, and determine whether the similarity meets the target condition. For example, the target condition is that the similarity is greater than a set threshold, or the target condition can include the first n vectors whose similarity is greater than the set threshold, where n is a preset constant. Through this judgment, the vector corresponding to the similarity that meets the target condition can be determined as a candidate vector, and then the reference institution identifiers corresponding to these candidate vectors can be determined as candidate institution identifiers. For example, the candidate institution identifier is a group of hospital names that are closest to the hospital name carried by the claim request (that is, the institution identifier to be analyzed). It is highly likely that the institution identifier to be analyzed is the same as these candidate institution identifiers, and the candidate institution identifiers can be further analyzed later to more accurately determine the reference institution identifier that matches the institution identifier to be analyzed.

[0038] Step 104: Determine a target institution identifier that represents the same institution as the institution identifier to be analyzed based on the institution identifier to be analyzed and the candidate institution identifiers through a text reasoning model.

[0039] In the embodiments of this specification, the data acquisition device can use a large model to further analyze the candidate institution identification, and the above-mentioned text reasoning model is a large model. The large model can be any large model currently available on the market (such as GPT, etc.). Optionally, the large model can also be a large model developed by the claims platform itself. For example, the model can be trained with a large number of institution identifications to obtain a text reasoning model that can be used to determine the target institution identification.

[0040] The data acquisition device can input both the identifier to be analyzed and the candidate organization identifier previously determined into the text reasoning model, and input certain processing instructions into the text reasoning model so that the text reasoning model can perform text reasoning to determine whether each candidate organization identifier is consistent with the organization identifier to be analyzed, that is, whether the candidate organization identifier represents the same organization as the organization identifier to be analyzed. If there is a candidate organization identifier that represents the same organization as the organization identifier to be analyzed, the text reasoning model can output the corresponding target organization identifier; if the candidate organization identifiers are all different in structure from the organization identifier to be analyzed, the text reasoning model can output a prompt message that the target organization information has not been determined. Optionally, the text reasoning model can determine the candidate organization identifier that is closest to the organization identifier to be analyzed as the target organization identifier.

[0041] For example, the text reasoning model can perform semantic recognition on the organization logo to be analyzed and the candidate organization logo, and search for relevant information about the organization logo to be analyzed and the candidate organization logo on the Internet, so as to combine the text reasoning to obtain the target organization logo. In this way, more factors can be considered when determining the match of the organization logo to be analyzed, ensuring that the matching accuracy of the organization logo to be analyzed is high.

[0042] The text reasoning model can receive instructions and output results in the form of questions and answers. The data processing device can input question information related to the organization identifier to be analyzed and the candidate organization identifiers into the text reasoning model, so that the text reasoning model can reply to the question information, and the target organization identifier can be determined based on the reply information. As in the above step 104, the text reasoning model determines the target organization identifier that represents the same organization as the organization identifier to be analyzed based on the organization identifier to be analyzed and the candidate organization identifier, which can include: constructing target question information based on the organization identifier to be analyzed and the candidate organization identifier, wherein the target question information is used to indicate that the organization identifier that is the same as the organization identifier to be analyzed is determined in the candidate organization identifiers; inputting the target question information into the text reasoning module to obtain the reply information output by the text reasoning model; wherein the reply information includes the target organization identifier that represents the same organization as the organization identifier to be analyzed.

[0043] For example, the institution to be analyzed is identified as XX City B Hospital, and the candidate institution identifiers include B Hospital, XXB Hospital, and XX District B Hospital. The question information constructed by the data acquisition device may be "Please determine which of "B Hospital", "XXB Hospital", and "XX District B Hospital" is the same as "XX City B Hospital"? ", or it may be "Please determine whether there is a hospital name that is the same as "XX City B Hospital" among "B Hospital", "XXB Hospital", and "XX District B Hospital"? ". The data acquisition device may call a text reasoning model, and input the question information into the text reasoning model, so that the text reasoning model performs analysis and outputs reply information. For example, the reply information output by the text reasoning model may be ""XXB Hospital" is the same as "XX City B Hospital", and based on the reply information, it can be determined that the target institution is identified as "XXB Hospital". Optionally, the reply information may also include the reasoning process of the text reasoning model to obtain the result.

[0044] In the embodiments of this specification, the text reasoning model can be deployed in the data acquisition device, or can also be deployed in other servers. The data acquisition device can establish a connection with the other server to call the text reasoning model to determine the target organization identifier.

[0045] In some embodiments, when at least two candidate institution identifiers are determined in step 102, step 104 may be performed to avoid wasting computing resources. Accordingly, in step 104, a target institution identifier representing the same institution as the institution identifier to be analyzed is determined based on the institution identifier to be analyzed and the candidate institution identifiers through a text reasoning model, including: when at least two candidate institution identifiers are determined, a target institution identifier representing the same institution as the institution identifier to be analyzed is determined from among at least two candidate institution identifiers through a text reasoning model. If only one candidate institution identifier is determined in step 102, the data acquisition device may directly determine the candidate institution identifier as the target institution identifier, and then perform step 106. Optionally, if only one candidate institution identifier is determined, the text reasoning model may be further used to determine whether the candidate institution identifier is accurate.

[0046] The above embodiments of this specification take the example of calling a text reasoning model to analyze the organization identifier to be analyzed and the candidate organization identifiers. In some methods, the data acquisition device can also enable the text reasoning model to analyze the vector to be analyzed and the determined candidate vector corresponding to the organization identifier to be analyzed, so that the text reasoning model outputs a target vector in the candidate vector that represents the same organization as the vector to be analyzed, and then determines the organization identifier corresponding to the target vector as the target organization identifier. The implementation method of this method can refer to the introduction of the aforementioned analysis of the organization identifier, such as constructing question information based on the vector.

[0047] Step 106: Obtain the organization data corresponding to the target organization identifier, wherein the organization data is used to execute the task corresponding to the claim request.

[0048] Through the aforementioned steps, the data acquisition device can obtain the target institution identifier that accurately matches the institution identifier to be analyzed in the reference institution identifier, and then obtain the institution data corresponding to the target institution identifier, and use the institution data to perform the task corresponding to the claim request, so as to improve the execution effect of the task. By way of example, the task corresponding to the claim request may be a claim verification task. For example, the institution data can be compared with other claim information carried by the claim request, and the payment risk corresponding to the institution, whether there is a possibility of fraud, etc. can be determined based on the institution data. For example, the institution data of the hospital identifier finally determined may include the type of hospital (such as whether it is a tertiary hospital), the division of departments of the hospital, the hospital's medical records, claim records, etc.

[0049] In an embodiment of the present specification, the data acquisition device can acquire the institutional data corresponding to the target institutional identifier from a third-party data source. For example, data acquisition can be achieved by calling an interface corresponding to the third-party data source. Accordingly, acquiring the institutional data corresponding to the target institutional identifier in step 106 includes: adding the target institutional identifier to the calling parameters of the third-party data source interface, wherein the third-party data source interface is connected to the third-party data source; calling the third-party data source interface based on the calling parameters to acquire the institutional data corresponding to the target institutional identifier from the third-party data source.

[0050] For example, when the data acquisition device uses the interface to acquire data, it needs to send corresponding call parameters to the third-party data source, and the call parameters may include the data to be acquired. The data acquisition device may assemble the target organization identifier as a call parameter into the third-party data source interface to achieve data acquisition.

[0051] Optionally, the data acquisition device can utilize a third-party data source interface to maintain a continuous connection with a third-party data source to ensure efficient and rapid acquisition when data is needed; or it can establish a connection only when data is needed, and disconnect after the data is acquired to avoid wasting communication resources.

[0052] In the embodiments of this specification, a large model and a vector library are used to solve the problem of accurate matching of institution identifiers in the claims scenario. Even if the hospital name carried in the claim request is different from the hospital name obtained from the third-party data source due to OCR recognition errors, manual input errors, inconsistencies between the abbreviation and full name of the hospital name, etc., accurate matching of the hospital name can be achieved, ensuring that the hospital name provided in the claims system can accurately correspond to the hospital name in the third-party data source, thereby enabling accurate and efficient execution of claims verification tasks.

[0053] In summary, in the data acquisition method provided in the embodiment of this specification, for the organization identifier to be analyzed carried by the claim request, a similar candidate organization identifier is first queried among multiple reference organization identifiers, and then the text reasoning model is called to further determine the target organization identifier that is substantially the same as the organization identifier to be analyzed based on the candidate organization identifier. In this way, it can be ensured that the target organization identifier is more accurately determined among the multiple reference organization identifiers, and accordingly, the organization data that matches the claim request can be accurately obtained, and the task execution effect for the claim request can be improved based on the organization data.

[0054] Corresponding to the above method embodiment, this specification also provides a data acquisition device embodiment, Figure 2 Schematic diagram of a data acquisition device provided in one embodiment of this specification. Figure 2 As shown, the data acquisition device includes: A first determination module 202 is used to determine, in response to a claim settlement request, a candidate institution identifier from among multiple reference institution identifiers based on a similarity with the institution identifier to be analyzed carried in the claim settlement request; The second determination module 204 is used to determine the target institution identifier that represents the same institution as the institution identifier to be analyzed based on the institution identifier to be analyzed and the candidate institution identifiers through a text reasoning model; The first acquisition module 206 is used to acquire the organization data corresponding to the target organization identifier, wherein the organization data is used to execute the task corresponding to the claim request.

[0055] Optionally, the second determining module 204 is used to: Based on the identification of the institution to be analyzed and the identification of the candidate institutions, constructing target question information, wherein the target question information is used to indicate that an institution identification identical to the identification of the institution to be analyzed is determined among the candidate institution identifications; The target question information is input into the text reasoning module to obtain the reply information output by the text reasoning model; wherein the reply information includes the target institution identifier representing the same institution as the institution identifier to be analyzed.

[0056] Optionally, the second determining module 204 is used to: When at least two candidate institution identifiers are determined, a target institution identifier that represents the same institution as the institution identifier to be analyzed is determined from among the at least two candidate institution identifiers through a textual reasoning model.

[0057] Optionally, the first determining module 202 includes: A processing submodule is used to vectorize the identification of the institution to be analyzed carried in the claim request to obtain a vector to be analyzed; A first determination submodule is used to determine, among a plurality of vectors corresponding to a plurality of reference organization identifiers, a candidate vector whose similarity with the vector to be analyzed meets a target condition; wherein the vector corresponding to each reference organization identifier is obtained by vectorizing each reference organization identifier; The second determination submodule is used to determine the reference organization identifier corresponding to the candidate vector as the candidate organization identifier.

[0058] Optionally, the data acquisition device further includes: A processing module, used to use a text embedding model to perform vectorization processing on the multiple reference organization identifiers to obtain multiple vectors before determining the candidate vector whose similarity with the vector to be analyzed meets the target condition among the multiple vectors corresponding to the multiple reference organization identifiers; The processing submodule is used to: The text embedding model is used to vectorize the identification of the institution to be analyzed carried in the claim request to obtain the vector to be analyzed.

[0059] Optionally, the data acquisition device further includes: A second acquisition module is used to obtain multiple reference institution identifiers from a third-party data source before respectively performing vectorization processing on multiple reference institution identifiers to obtain multiple vectors using a text embedding model, wherein the third-party data source includes institution data of multiple institutions, and the institution data of each institution includes a reference institution identifier of each institution; and / or, The first acquisition module 206 is used for: Adding the target organization identifier to the calling parameters of the third-party data source interface, wherein the third-party data source interface is connected to the third-party data source; The third-party data source interface is called based on the calling parameters to obtain the institution data corresponding to the target institution identifier from the third-party data source.

[0060] In summary, in the data acquisition device provided by the embodiment of this specification, for the identification of the institution to be analyzed carried by the claim request, the first determination module first searches for similar candidate institution identifications among multiple reference institution identifications, and the second determination module then calls the text reasoning model to further determine the target institution identification that is substantially the same as the identification of the institution to be analyzed based on the candidate institution identification. In this way, it can be ensured that the target institution identification is more accurately determined among the multiple reference institution identifications, and accordingly, the first acquisition module can accurately obtain the institution data that matches the claim request, and based on the institution data, the task execution effect for the claim request can be improved.

[0061] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. As for the data acquisition device, since it is basically similar to the data acquisition method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the data acquisition method embodiment.

[0062] Figure 3 300 is a block diagram of a computing device provided in an embodiment of the present specification. The components of the computing device 300 include but are not limited to a memory 310 and a processor 320. The processor 320 is connected to the memory 310 via a bus 330, and the database 350 is used to store data.

[0063] The computing device 300 also includes an access device 340 that enables the computing device 300 to communicate via one or more networks 360. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 340 may include one or more of any type of network interface (e.g., a network interface card (NIC)) of wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a world-wide interoperability for microwave access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, and a near field communication (NFC).

[0064] In one embodiment of the present specification, the above components of the computing device 300 and Figure 3 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 3 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0065] The computing device 300 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or a personal computer (PC). The computing device 300 may also be a mobile or stationary server.

[0066] The processor 320 is used to execute computer programs / instructions, which, when executed by the processor, implement the above Figure 1 The method shown.

[0067] As for the computing device embodiment, since it is basically similar to the data acquisition method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the data acquisition method embodiment.

[0068] One embodiment of the present specification also provides a computer-readable storage medium, which stores computer instructions, and when the computer instructions are executed by the processor, the steps of the above-mentioned data acquisition method are implemented. The computer instructions include computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable storage medium may include: any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electric carrier signals and telecommunication signals.

[0069] An embodiment of the present specification also provides a computer program product, including a computer program / instruction, which implements the steps of the above data acquisition method when the computer program / instruction is executed in a processor.

[0070] As for the computer-readable storage medium embodiment and the computer program product embodiment, since they are basically similar to the data acquisition method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the data acquisition method embodiment.

[0071] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0072] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0073] It should be noted that the above is a description of a specific embodiment of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of the present specification.

[0074] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0075] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that technicians in the relevant technical field can understand and use this specification well.

Claims

1. A data acquisition method, characterized in that: include: In response to a claim settlement request, determining a candidate institution identifier among a plurality of reference institution identifiers based on a similarity to the institution identifier to be analyzed carried in the claim settlement request; Determine, through a textual reasoning model, a target institution identifier representing the same institution as the institution identifier to be analyzed based on the institution identifier to be analyzed and the candidate institution identifiers; Obtaining the organization data corresponding to the target organization identifier, wherein the organization data is used to execute the task corresponding to the claim settlement request.

2. The method according to claim 1, characterized in that The determining, through the textual reasoning model, based on the to-be-analyzed institution identifier and the candidate institution identifier, of a target institution identifier that represents the same institution as the to-be-analyzed institution identifier, comprises: Based on the to-be-analyzed institution identifier and the candidate institution identifiers, construct target question information, wherein the target question information is used to indicate that an institution identifier identical to the to-be-analyzed institution identifier is to be determined among the candidate institution identifiers; The target question information is input into a text reasoning module to obtain answer information output by the text reasoning model; wherein the answer information includes a target institution identifier that represents the same institution as the institution identifier to be analyzed.

3. The method according to claim 1, characterized in that The determining, through the textual reasoning model, based on the to-be-analyzed institution identifier and the candidate institution identifier, of a target institution identifier that represents the same institution as the to-be-analyzed institution identifier, comprises: When at least two candidate institution identifiers are determined, a target institution identifier that represents the same institution as the institution identifier to be analyzed is determined from among the at least two candidate institution identifiers through a textual reasoning model.

4. The method according to any one of claims 1 to 3, characterized in that: The step of determining a candidate institution identifier among the multiple reference institution identifiers based on the similarity with the institution identifier to be analyzed carried in the claim settlement request comprises: Vectorizing the identification of the institution to be analyzed carried in the claim settlement request to obtain a vector to be analyzed; Determine, among multiple vectors corresponding to multiple reference organization identifiers, a candidate vector whose similarity with the vector to be analyzed meets a target condition; wherein the vector corresponding to each reference organization identifier is obtained by performing the vectorization process on each reference organization identifier; The reference organization identifier corresponding to the candidate vector is determined as the candidate organization identifier.

5. The method according to claim 4, characterized in that Before determining, among the multiple vectors corresponding to the multiple reference mechanism identifiers, a candidate vector whose similarity with the vector to be analyzed meets the target condition, the method further includes: Using the text embedding model, vectorization is performed on multiple reference institution identifiers to obtain multiple vectors; The vectorization processing of the identification of the institution to be analyzed carried in the claim settlement request to obtain the vector to be analyzed includes: The text embedding model is used to vectorize the identification of the institution to be analyzed carried in the claim request to obtain a vector to be analyzed.

6. The method according to claim 5, characterized in that Before the text embedding model is used to perform vectorization processing on the multiple reference organization identifiers to obtain multiple vectors, the method further includes: Acquire multiple reference institution identifiers from a third-party data source, wherein the third-party data source includes institution data of multiple institutions, and the institution data of each institution includes a reference institution identifier of each institution; and / or, The obtaining of the organization data corresponding to the target organization identifier includes: Adding the target organization identifier to the calling parameters of a third-party data source interface, wherein the third-party data source interface is connected to a third-party data source; The third-party data source interface is called based on the calling parameters to obtain the institution data corresponding to the target institution identifier from the third-party data source.

7. A data acquisition device, characterized in that: include: A first determination module is used to determine, in response to a claim settlement request, a candidate institution identifier from among a plurality of reference institution identifiers based on a similarity with an institution identifier to be analyzed carried in the claim settlement request; A second determination module is used to determine a target institution identifier that represents the same institution as the institution identifier to be analyzed based on the institution identifier to be analyzed and the candidate institution identifiers through a text reasoning model; The first acquisition module is used to acquire the institutional data corresponding to the target institutional identifier, wherein the institutional data is used to execute the task corresponding to the claim settlement request.

8. A computing device, characterized in that include: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: A computer program / instruction is stored, and when the computer program / instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The method comprises a computer program / instruction, which implements the method according to any one of claims 1 to 6 when the computer program / instruction is executed in a processor.

Images