A text structuring processing method and system

Abstract

The application discloses a text structuring processing method and system. After obtaining underwriting information, a structure level table of an insurance to be invested and text key information in an underwriting health expression text are automatically determined, so as to adjust the text key information in the underwriting health expression text into a structured expression text. After the underwriting health expression text is output as the structured expression text, the structured expression text can be automatically analyzed by the system, so as to determine whether the structured expression text meets underwriting rules of the insurance to be invested, without manual participation in underwriting, thereby improving underwriting accuracy and saving labor cost and time cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of text processing, and particularly relates to a text structured processing method and system. BACKGROUND

[0002] With the development of society, more and more people purchase health insurance.

[0003] In the process of purchasing health insurance products, the purchaser needs to provide health information, and the health information provided by the purchaser contains a large amount of unstructured description text, which needs to be manually audited by staff, and the potential risk of the purchaser is evaluated to determine whether to pass the insurance, which leads to a time-consuming and labor-consuming auditing process, and due to the different experiences of the staff, the accuracy of the audit is low, and the user experience is reduced. SUMMARY

[0004] Therefore, the present application provides a text structured processing method and system, and the specific solutions are as follows:

[0005] A text structured processing method, comprising:

[0006] obtaining underwriting information, wherein the underwriting information at least includes underwriting rules and underwriting health description text of an insurance to be purchased;

[0007] determining a structure level table of the insurance to be purchased based on the underwriting rules of the insurance to be purchased, wherein the structure level table at least includes key indicators and attribute information of the insurance to be purchased;

[0008] determining key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table;

[0009] constructing a structured expression text that matches the structure level table of the insurance to be purchased based on the key text information.

[0010] Further, it further comprises:

[0011] determining whether the structured expression text includes at least partially overlapping expression texts;

[0012] If yes, determining a first expression text from the at least partially overlapping expression texts, and replacing the at least partially overlapping expression texts in the structured expression text with the first expression text.

[0013] Further, the determining a first expression text from the at least partially overlapping expression texts comprises:

[0014] determine a representation text containing more parameter information than other representation texts from the plurality of representation texts that at least partially overlap, and determine the representation text containing more parameter information than other representation texts as a first representation text.

[0015] Further, the determining of the key text information in the underwriting health description text that matches the key indicators and attribute information in the structural hierarchical table comprises:

[0016] determining an extended term that matches the key indicators and attribute information;

[0017] determining the key text information in the underwriting health scan text based on the key indicators and attribute information in the structural hierarchical table and the extended term that matches the key indicators and attribute information.

[0018] Further, the determining of the key text information in the underwriting health description text that matches the key indicators and attribute information in the structural hierarchical table comprises:

[0019] performing matching segmentation on each sentence in the underwriting health description text;

[0020] annotating the key text information based on the key indicators and attribute information in the structural hierarchical table.

[0021] Further, the constructing of the structured representation text that matches the structural hierarchical table of the insurance to be applied for based on the key text information comprises:

[0022] determining a representation text template based on the key text information;

[0023] constructing a structured representation text based on the key indicators and attribute information in the structural hierarchical table and the representation text template.

[0024] Further, the method further comprises:

[0025] determining whether the underwriting health description text meets the underwriting rules of the insurance to be applied for based on the structured representation text.

[0026] A text structuring processing system, comprising:

[0027] an obtaining unit configured to obtain underwriting information, the underwriting information at least comprising underwriting rules of an insurance to be applied for and an underwriting health description text;

[0028] a first determining unit configured to determine a structural hierarchical table of the insurance to be applied for based on the underwriting rules of the insurance to be applied for, the structural hierarchical table at least comprising key indicators and attribute information of the insurance to be applied for;

[0029] The second determining unit is configured to determine key text information in the underwriting health description text that matches the key indicators and attribute information in the structure hierarchical table;

[0030] The constructing unit is configured to construct a structured expression text that matches the structure hierarchical table of the insurance to be underwritten based on the key text information.

[0031] An electronic device comprises:

[0032] The processor is configured to obtain underwriting information, the underwriting information at least including underwriting rules of an insurance to be underwritten and an underwriting health description text; determine a structure hierarchical table of the insurance to be underwritten based on the underwriting rules of the insurance to be underwritten, the structure hierarchical table at least including key indicators and attribute information of the insurance to be underwritten; determine key text information in the underwriting health description text that matches the key indicators and attribute information in the structure hierarchical table; and construct a structured expression text that matches the structure hierarchical table of the insurance to be underwritten based on the key text information.

[0033] The memory is configured to store programs for the processor to execute the above-mentioned processing procedures.

[0034] A readable storage medium is configured to store at least a set of instruction sets.

[0035] The instruction sets are configured to be invoked and at least execute the method of text structuring processing according to any one of the above.

[0036] As can be seen from the above technical solutions, the text structuring processing method and system disclosed in the present application obtain underwriting information, the underwriting information at least including underwriting rules of an insurance to be underwritten and an underwriting health description text, determine a structure hierarchical table of the insurance to be underwritten based on the underwriting rules of the insurance to be underwritten, the structure hierarchical table at least including key indicators and attribute information of the insurance to be underwritten, determine key text information in the underwriting health description text that matches the key indicators and attribute information in the structure hierarchical table, and construct a structured expression text that matches the structure hierarchical table of the insurance to be underwritten based on the key text information. After obtaining the underwriting information, the present application automatically determines the structure hierarchical table of the insurance to be underwritten and the text key information in the underwriting health expression text, so as to adjust the text key information in the underwriting health expression text into a structured expression text. After outputting the underwriting health expression text as the structured expression text, the system can automatically analyze the structured expression text, so as to determine whether the structured expression text meets the underwriting rules of the insurance to be underwritten, without human intervention in underwriting, thereby improving the underwriting accuracy and saving the labor cost and time cost. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the accompanying drawings in the following description only only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0038] Figure 1 A flow chart of a text structured processing method disclosed by an embodiment of the present application;

[0039] Figure 2 A flow chart of a text structured processing method disclosed by an embodiment of the present application;

[0040] Figure 3 A flow chart of a text structured processing method disclosed by an embodiment of the present application;

[0041] Figure 4 A structural schematic diagram of a text structured processing system disclosed by an embodiment of the present application;

[0042] Figure 5 A structural schematic diagram of an electronic device disclosed by an embodiment of the present application. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0044] The present application discloses a text structured processing method, a flow chart of which is shown in Figure 1 , including:

[0045] Step S11, obtaining underwriting information, the underwriting information at least including underwriting rules of an insurance to be invested and an underwriting health description text;

[0046] Step S12, determining a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested;

[0047] Step S13, determining key text information in the underwriting health description text matching the key indicators and attribute information in the structure level table;

[0048] Step S14, constructing a structured expression text matching the structure level table of the insurance to be invested based on the key text information.

[0049] In the process of purchasing health insurance services and underwriting, for some insured persons with less health information or no disease, the audit process is relatively simple, and only the online questionnaire needs to be audited. For some insured persons with complex diseases, the insured person needs to provide health information and corresponding medical reports, medical records, medical history and other data. This may result in the problem of complex data format and various types provided by the insured person, which needs to be manually analyzed by the staff to determine whether the insured person meets the underwriting rules of the health insurance they purchase and whether they can continue to insure the health insurance business.

[0050] To avoid the problem of inaccurate analysis and time and labor consumption caused by manual analysis, in the present scheme, the key text information in the underwriting information is adjusted to a structured statement text by analyzing the underwriting information, so that the system can directly determine whether the underwriting health description text meets the underwriting rules of the insurance to be invested based on the structured statement text, thereby determining whether the insured person can continue to insure the health insurance business, realizing automatic underwriting without manual participation, improving the underwriting accuracy through automatic underwriting, and reducing the labor and time cost.

[0051] In the present scheme, first, the underwriting information is obtained, which at least includes the underwriting rules of the insurance to be invested and the underwriting health description text.

[0052] The underwriting health description text is the relevant data of the historical health status submitted by the insured person, such as historical medical reports, medical records and other text data. Through the above historical medical reports, medical records and other text data, the historical health status of the insured person can be determined, such as the problems in the body indicated in the medical report in a certain time period.

[0053] The underwriting rules of the insurance to be invested are the health status that the insured person needs to achieve to insure the insurance to be invested, such as the requirement that the insured person's historical health information cannot exist liver cysts, if the liver cysts exist, the insured person cannot insure the insurance to be invested; or the requirement that the insured person's historical health information can exist liver cysts, but the size of the liver cysts needs to be less than a certain preset value, if the liver cysts are larger than the preset value, the insured person cannot insure the insurance to be invested, if the liver cysts are smaller than the preset value, the insured person can insure the insurance to be invested, etc.

[0054] After obtaining the underwriting rules of the insurance to be invested, a structure hierarchy table of the insurance to be invested needs to be generated based on the underwriting rules, which at least includes the key indicators and attribute information of the insurance to be invested.

[0055] The hierarchical table includes various data, data types, superior-inferior relationships, etc. that need to be focused on, including entities, attributes, and relationships between entities and attributes. Only when the above entities, attributes, and relationships between entities and attributes meet the underwriting rules of the insurance to be invested can the insurance be underwritten and the insurance to be invested be invested.

[0056] The key indicators of the insurance to be invested are the indicators of the health status that the insured person needs to focus on in order to invest in the insurance to be invested. The attribute information is the attribute of the key indicators. Only when the attribute of the key indicators of the insured person meets the underwriting rules can the insurance be invested. As shown in Table 1, it is a schematic table of a structural hierarchical table:

[0057] Table 1 Structural hierarchical table

[0058]

[0059]

[0060] As shown in the above table, the structural hierarchical table can include: type, examination category, examination item, part, item (entity), whether the part is needed, state value type, state value, measured value, etc. Among them, the type can be: based on which type to obtain the health status, such as: examination, or physical examination, etc. The examination category and the examination item can be the data of the health status obtained by the examination item in the form, such as: ultrasonic or blood lipid examination, the part is the part of the examination, such as: kidney, liver, etc. The item is the entity in the underwriting health description text, which can be: cyst, nodule, triglyceride, etc. After determining the entity checked by the examination part, the state value of the entity and the type of the state value need to be determined, such as: the state value can be quantity, single side or double side, increase or decrease, etc. The structural hierarchical table can also have specific measured values to determine that only when the entity of the examination part is within the measured value can the insurance be invested.

[0061] As can be seen from the above table, ultrasonic and blood lipid conditions are the key focus of the insurance to be invested, and kidney cyst and triglyceride are the key item entities, and the state and part conditions of the entity need to be known, such as: size, position, quantity, etc., i.e. attribute relationship and attribute value.

[0062] After determining the structural hierarchical table, it is also necessary to determine whether the item entity specified in the structural hierarchical table exists in the underwriting health description text, i.e. the key indicators, and if so, the size and position data of the item entity, i.e. the attribute information of the key indicators, need to be determined, and the information in the underwriting health description text that matches the above key indicators and attribute information is determined as key text information, and the key text information is constructed into a structured expression text that matches the structural hierarchical table of the insurance to be invested.

[0063] For example, the health description text for underwriting includes: liver small cyst, intrahepatic bile duct stones, mild fatty liver, and prostate small cyst. The B-ultrasound report provides: prostate small cyst 14*12mm, mild fatty liver, liver small cyst 3*3mm. The B-ultrasound report of 202007 provides: liver S4 small cystic lesion, about 3mm in diameter, considered liver cyst, liver S7 small calcification or intrahepatic bile duct stones.

[0064] The above text is a specific description of the health information of the insured person. The structure is relatively complex. Therefore, the key text information in the health description text for underwriting is first determined, and then the key text information is displayed in the form of a structure level table, so that the system can directly match the structured expression text and the structure level table displayed in the form of a structure level table, thereby determining whether the health status of the insured person corresponding to the structured expression text meets the insurance conditions and whether the underwriting can be passed.

[0065] Specifically, based on the structure level table of the insurance to be applied for, the key text information at least includes: cyst, site: liver, type: informed; stones, site: intrahepatic bile duct, type: informed; cyst, site: prostate, type: examination, size: 14*12mm, etc.

[0066] Based on the structure level table, the key indicators of interest in the insurance to be applied for can be determined, so as to determine whether the key indicators of interest in the insurance to be applied for exist in the health description text for underwriting.

[0067] The text structuring processing method disclosed in this embodiment obtains underwriting information, which at least includes the underwriting rules of the insurance to be applied for and the health description text for underwriting. Based on the underwriting rules of the insurance to be applied for, a structure level table of the insurance to be applied for is determined. The structure level table at least includes the key indicators and attribute information of the insurance to be applied for. The key text information in the health description text for underwriting that matches the key indicators and attribute information in the structure level table is determined. Based on the key text information, a structured expression text that matches the structure level table of the insurance to be applied for is constructed. After obtaining the underwriting information, the structure level table of the insurance to be applied for and the key text information in the health description text for underwriting are automatically determined, so as to adjust the key text information in the health description text for underwriting to the structured expression text. After the health description text for underwriting is output as the structured expression text, the system can automatically analyze the structured expression text, so as to determine whether the structured expression text meets the underwriting rules of the insurance to be applied for. This method does not require manual participation in underwriting, improves the accuracy of underwriting, and saves labor cost and time cost.

[0068] The text structuring processing method disclosed in this embodiment has a flowchart as shown in Figure 2 , which includes:

[0069] Step S21, obtaining underwriting information, the underwriting information at least including underwriting rules and underwriting health description text of the insurance to be invested;

[0070] Step S22, determining a structure level table of the insurance to be invested based on the underwriting rules, the structure level table at least including key indicators and attribute information of the insurance to be invested;

[0071] Step S23, determining key text information in the underwriting health description text matching the key indicators and attribute information in the structure level table;

[0072] Step S24, constructing a structured expression text matching the structure level table of the insurance to be invested based on the key text information;

[0073] Step S25, determining whether the structured expression text includes at least partially overlapping expression texts;

[0074] Step S26, if yes, determining a first expression text from the at least partially overlapping expression texts, and replacing the at least partially overlapping expression texts in the structured expression text with the first expression text.

[0075] After constructing the structured expression text corresponding to the structure level table based on the key text information, the structured expression text includes project entities related to each item in the structure level table involved in the underwriting health description text, which makes it possible that there are multiple project entities at least partially overlapping in the structured expression text.

[0076] For example, the structured expression text includes a first subset of “liver + cyst + numerical value + state”, a second subset of “liver + cyst + numerical value”, and a third subset of “liver + cyst + state”, wherein the project entities involved in the first subset, the second subset, and the third subset have at least partially overlapping contents, respectively, which requires selecting one subset as the first expression text from the first subset, the second subset, and the third subset, and replacing the first subset, the second subset, and the third subset in the structured expression text with the first expression text, so that the content related to “liver + cyst” in the structured expression text is expressed through only one unique subset, rather than multiple subsets with at least partially overlapping texts, to optimize the structured expression text.

[0077] Further, from the multiple expression texts at least partially overlapping, an expression text containing more parameter information than other expression texts is determined as the first expression text.

[0078] The text with the most contained parameters is selected from the multiple partially overlapping expression texts as the first expression text. The above first subset, second subset, and third subset are taken as an example for description, wherein the same parameters contained in the first subset, second subset, and third subset are "liver + cyst", that is, the overlapping content, and among the other parameters in the three subsets except the same parameters, the first subset includes "numerical value + state" two parameters, the second subset includes "numerical value" one parameter, and the third subset includes "state" one parameter. It can be determined that only the first subset contains more parameters than the other two subsets in the three subsets, and therefore, the first subset is directly determined as the first expression text, one expression text is used to replace the previous three expression texts, the expression texts are simplified, and the data processing efficiency is improved.

[0079] In addition, in the determination of the first expression text, a new expression text can also be generated based on the multiple partially overlapping expression texts and used as the first expression text. That is, any one of the multiple partially overlapping expression texts is not directly used, but a new expression text is generated based on the multiple expression texts.

[0080] For example, the structured expression text includes a first subset of "liver + cyst + numerical value" and a second subset of "liver + cyst + state". Since the overlapping part of the first subset and the second subset is "liver + cyst", in order to express the non-overlapping part in the above two subsets by one expression text, a new subset needs to be generated, which includes both "numerical value" in the first subset and "state" in the second subset. The two partially overlapping expression texts are expressed by one expression text, and the expression mode is simplified.

[0081] Whether the first expression text is generated based on the multiple partially overlapping expression texts or selected from the multiple partially overlapping expression texts needs to be determined based on the judgment result of whether there is an expression text in the multiple partially overlapping expression texts that can include the non-overlapping part of other expression texts.

[0082] The text structuring processing method disclosed in the embodiment obtains underwriting information, the underwriting information at least includes underwriting rules of an insurance to be invested and an underwriting health description text, determines a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least includes key indicators and attribute information of the insurance to be invested, determines key text information in the underwriting health description text matched with the key indicators and attribute information in the structure level table, and constructs a structured expression text matched with the structure level table of the insurance to be invested based on the key text information. After obtaining the underwriting information, the structure level table of the insurance to be invested and the key text information in the underwriting health expression text are automatically determined, so as to adjust the key text information in the underwriting health expression text into the structured expression text. After the underwriting health expression text is output as the structured expression text, the structured expression text can be automatically analyzed by the system, so as to determine whether the structured expression text meets the underwriting rules of the insurance to be invested, without manual participation in underwriting, thereby improving the underwriting accuracy and saving the labor cost and time cost.

[0083] The text structuring processing method disclosed in the embodiment has a flowchart as shown in Figure 3 The text structuring processing method disclosed in the embodiment has a flowchart as shown in

[0084] In step S31, underwriting information is obtained, the underwriting information at least includes underwriting rules of an insurance to be invested and an underwriting health description text.

[0085] In step S32, a structure level table of the insurance to be invested is determined based on the underwriting rules of the insurance to be invested, the structure level table at least includes key indicators and attribute information of the insurance to be invested.

[0086] In step S33, an extended term matched with the key indicators and attribute information is determined.

[0087] In step S34, key text information in the underwriting health scanning text is determined based on the key indicators and attribute information in the structure level table and the extended term matched with the key indicators and attribute information.

[0088] In step S35, a structured expression text matched with the structure level table of the insurance to be invested is constructed based on the key text information.

[0089] The extension of the term in the structure level table can not only include the key indicators and attribute information, but also other information, such as state values, quantities, etc.

[0090] Based on the conventional or open medical knowledge graph, medical standard word table, synonym word table, entity classification word table and other professional data related to the structure level table can be introduced, and at the same time, based on the conventional corpus or general knowledge, various natural language general term tables with magnetism such as quantity words, symbols, negative words, stop words, mood words and conjunction words can be introduced.

[0091] As shown in Table 2, an extended term vocabulary is expanded:

[0092] Table 2 Extended term vocabulary

[0093]

[0094]

[0095] The introduction of quantity words, symbols, mood words, etc. is to fully tag the description text. The relationship before and after these words can prevent ambiguity and facilitate the production of templates for more accurate extraction of tagged sequences. Mood words may also appear in the description text. Tagging them is to facilitate a better understanding of the grammatical structure of the sentence. The "state value" in the table is the normalized word. The "state value alias" is a synonym that may appear in the description text. The purpose of expanding this vocabulary is to annotate the description text as completely as possible to ensure that no matter what words are used to express the description text, the extended term vocabulary can determine its annotation.

[0096] The extended term vocabulary can be stored according to different tag types.

[0097] Further, the key text information in the health description text for underwriting is determined to match the key indicators and attribute information in the structure hierarchy table, including: performing matching segmentation on each sentence in the health description text for underwriting, and based on the key indicators and attribute information in the structure hierarchy table, annotating the key text information of the segmented text information.

[0098] To determine the key text information in the health description text for underwriting, the health description text for underwriting can be first segmented into sentences, i.e., each sentence in the health description text for underwriting is determined, and the sentence can be segmented by a period as a segmentation point. Each sentence is matched and segmented, and annotated.

[0099] The matching segmentation can be: performing maximum reverse matching segmentation on each sentence to determine the words in each sentence. Of course, other segmentation methods such as jieba segmentation can also be used;

[0100] Tagging can use tag type tagging. When tagging, the normalized word corresponding to each word in the sentence can be first determined based on the extended term vocabulary, and then the tagging is performed. For example, if "cyst" exists in the text, "cyst" is annotated as an entity; if a specific numerical value exists in the text, it is annotated as a size.

[0101] Further, based on the key text information, a structured expression text matching the structure level table of the to-be-insured insurance is constructed, including: determining an expression text model based on the key text information, and constructing a structured expression text according to the expression text template based on the key indicators and attribute information in the structure level table.

[0102] The key text information of the underwriting health description text is different, and the corresponding expression text model is different. After determining the expression text model based on the key text information, the structured expression text is constructed according to the key indicators and attribute information in the structure level table with the expression text model as the model, so as to replace the underwriting health description text with the structured expression text, facilitate automatic underwriting of the structured expression text, realize the process of automatic underwriting, and save labor cost and time cost.

[0103] Specifically, after the underwriting health description text is matched and segmented, it is labeled to determine the type of each word, whether it is an entity, a quantity word, or a state word, etc. After all the words in the underwriting health description text are labeled, statistics are performed, that is, the key text information is counted.

[0104] The statistics are the number of types of expression templates, such as: there are 8 expression templates of the first type, and 1 expression template of other types, then the expression template of the first type is selected as the expression text model, and the structured expression text is constructed according to the expression text model; or, the template with a frequency reaching a preset proportion can be directly selected as the expression model. The parameters in the key text information are filled into the expression model according to the tag type and the underwriting rules to complete the construction of the structured expression text, so that all the information in the underwriting health description text is structured according to the form of the expression text model for expression, to improve the extraction and analysis accuracy of the system for the structured expression text.

[0105] Further, after constructing the structured expression text, it is determined whether the underwriting health description text meets the underwriting rules of the to-be-insured insurance based on the structured expression text, so as to realize the identification and analysis of the underwriting health description text after being structured, and realize the automatic judgment of whether the underwriting health description text can pass the underwriting.

[0106] The text structuring processing method disclosed in the embodiment obtains underwriting information, the underwriting information at least includes underwriting rules of an insurance to be invested and an underwriting health description text, determines a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least includes key indicators and attribute information of the insurance to be invested, determines key text information in the underwriting health description text matched with the key indicators and attribute information in the structure level table, and constructs a structured expression text matched with the structure level table of the insurance to be invested based on the key text information. After obtaining the underwriting information, the text key information in the underwriting health expression text is automatically determined, so as to adjust the text key information in the underwriting health expression text into the structured expression text. After the underwriting health expression text is output as the structured expression text, the structured expression text can be automatically analyzed by the system, so as to determine whether the structured expression text meets the underwriting rules of the insurance to be invested, without manual participation in underwriting, thereby improving the underwriting accuracy and saving the labor cost and time cost.

[0107] The text structuring processing system disclosed in the embodiment has a structure diagram as shown in Figure 4 The text structuring processing system disclosed in the embodiment has a structure diagram as shown in

[0108] The text structuring processing system disclosed in the embodiment has a structure diagram as shown in

[0109] The obtaining unit 41 is configured to obtain underwriting information, the underwriting information at least including underwriting rules of an insurance to be invested and an underwriting health description text;

[0110] The first determining unit 42 is configured to determine a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested;

[0111] The second determining unit 43 is configured to determine key text information in the underwriting health description text matched with the key indicators and attribute information in the structure level table;

[0112] The constructing unit 44 is configured to construct a structured expression text matched with the structure level table of the insurance to be invested based on the key text information.

[0113] Further, the text structuring processing system disclosed in the embodiment can further include:

[0114] The third determining unit is configured to determine whether the structured expression text includes at least partially overlapped expression texts, and if so, determine a first expression text from the at least partially overlapped expression texts and replace the at least partially overlapped expression texts in the structured expression text with the first expression text.

[0115] Further, the third determining unit determines the first expression text from the at least partially overlapping expression texts, including:

[0116] The expression text containing more parameter information than other expression texts is determined from the plurality of expression texts at least partially overlapping, and the expression text containing more parameter information than other expression texts is determined as the first expression text.

[0117] Further, the second determining unit is configured to:

[0118] determine the extended term matching the key indicators and attribute information; and determine the key text information in the underwriting health scan text based on the key indicators and attribute information in the structure hierarchy table and the extended term matching the key indicators and attribute information.

[0119] Further, the second determining unit is configured to:

[0120] perform matching segmentation on each sentence in the underwriting health description text; and perform key text information labeling on the segmented text information based on the key indicators and attribute information in the structure hierarchy table.

[0121] Further, the constructing unit is configured to:

[0122] determine an expression text template based on the key text information; and construct a structured expression text according to the expression text template based on the key indicators and attribute information in the structure hierarchy table.

[0123] Further, the text structuring processing system disclosed in the embodiment can further include:

[0124] a fourth determining unit configured to determine whether the underwriting health description text meets the underwriting rules of the insurance to be applied for based on the structured expression text.

[0125] The text structuring processing system disclosed in the embodiment is implemented based on the text structuring processing method disclosed in the above embodiment, and thus will not be described here.

[0126] The text structuring processing system disclosed by the embodiment obtains underwriting information, the underwriting information at least including underwriting rules of an insurance to be invested and an underwriting health description text, determines a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested, determines key text information in the underwriting health description text matched with the key indicators and attribute information in the structure level table, and constructs a structured expression text matched with the structure level table of the insurance to be invested based on the key text information. After obtaining the underwriting information, the scheme automatically determines the structure level table of the insurance to be invested and the key text information in the underwriting health expression text, so as to adjust the key text information in the underwriting health expression text into the structured expression text. After outputting the underwriting health expression text as the structured expression text, the scheme can analyze the structured expression text automatically, so as to determine whether the structured expression text meets the underwriting rules of the insurance to be invested, without manual participation in underwriting, thereby improving the underwriting accuracy and saving the labor cost and time cost.

[0127] The electronic device disclosed by the embodiment includes a processor 51 and a memory 52. Figure 5

[0128] The processor 51 is configured to obtain underwriting information, the underwriting information at least including underwriting rules of an insurance to be invested and an underwriting health description text, determine a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested, determine key text information in the underwriting health description text matched with the key indicators and attribute information in the structure level table, and construct a structured expression text matched with the structure level table of the insurance to be invested based on the key text information.

[0129] The memory 52 is configured to store a program for the processor to execute the above processing process.

[0130] The electronic device disclosed by the embodiment is implemented based on the text structuring processing method disclosed by the above embodiment, and details are not described herein.

[0131] The electronic device disclosed by the embodiment is implemented based on the text structuring processing method disclosed by the above embodiment, and details are not described herein.

[0132] ​The electronic device disclosed in the embodiment obtains underwriting information, the underwriting information at least including underwriting rules and an underwriting health description text of an insurance to be invested, determines a structure level table of the insurance to be invested based on the underwriting rules of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested, determines key text information in the underwriting health description text that matches the key indicators and the attribute information in the structure level table, and constructs a structured expression text that matches the structure level table of the insurance to be invested based on the key text information. After obtaining the underwriting information, the electronic device automatically determines the structure level table of the insurance to be invested and the key text information in the underwriting health expression text, so as to adjust the key text information in the underwriting health expression text into the structured expression text. After outputting the underwriting health expression text as the structured expression text, the electronic device can automatically analyze the structured expression text, so as to determine whether the structured expression text meets the underwriting rules of the insurance to be invested, without manual participation in underwriting, thereby improving the underwriting accuracy and saving the labor cost and time cost.

[0133] The application further provides a readable storage medium, which has a computer program stored thereon, the computer program being loaded and executed by a processor to implement the steps of the above-described text structuring processing method. The specific implementation process can refer to the description of the corresponding part of the above-described embodiments, and the embodiment will not be described here.

[0134] The application further provides a computer program product or a computer program, which includes computer instructions stored in a computer readable storage medium. A processor of an electronic device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to enable the electronic device to perform the method provided in the various optional implementation manners of the above-described text structuring processing method aspect or text structuring processing system aspect. The specific implementation process can refer to the description of the corresponding embodiments, and will not be described here.

[0135] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts of each embodiment can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the related parts can be referred to the description of the method part.

[0136] Those skilled in the art will further appreciate that the units and algorithms described in connection with the examples disclosed herein can be embodied directly in hardware, in software, or in a combination of the two. For the sake of brevity, descriptions of a method or an algorithm described in the preceding description will not be repeated in the following description of the examples. For the same reason, not all components and algorithms described in the examples will be repeated. It is to be understood that the above description is intended to be illustrative and not restrictive. Many other embodiments will be apparent to those of skill in the art upon reading the above description. The scope of the application should, therefore, be determined not with reference to the above description, but should instead be determined with reference to the appended claims, along with their full scope of equivalents.

[0137] The steps of a method or algorithm described in connection with the examples disclosed herein can be embodied directly in hardware, in software, or in a combination of the two. A software module can reside in Random Access Memory (RAM), non-volatile memory (e.g., Flash memory, ROM, EEPROM, EPROM, programmable ROM, etc.), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an application-specific integrated circuit (ASIC).

[0138] The above description is intended to be illustrative and not restrictive. Many other embodiments will be apparent to those of skill in the art upon reading the above description. The scope of the application should, therefore, be determined not with reference to the above description, but should instead be determined with reference to the appended claims, along with their full scope of equivalents. The disclosure of all articles and references are incorporated by reference in their entirety.

Claims

1. A method of text structuring processing, characterized by, The method comprises the following steps: obtaining underwriting information, wherein the underwriting information at least comprises underwriting rules of an insurance to be underwritten and an underwriting health description text; determining a structure level table of the insurance to be underwritten based on the underwriting rules, wherein the structure level table at least comprises key indicators and attribute information of the insurance to be underwritten; determining key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table; constructing a structured expression text that matches the structure level table of the insurance to be underwritten based on the key text information; the determination of the key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table comprises: determining extended terms that match the key indicators and attribute information, which comprises: introducing professional data related to the structure level table based on a medical knowledge graph, wherein the professional data comprises a medical standard word table, a synonym word table and an entity classification word table, and introducing various natural language general terms with word types based on general corpus or general knowledge, including quantity words, symbols, negative words, stop words, mood words and conjunction words; determining the key text information in the underwriting health description text based on the key indicators and attribute information in the structure level table and the extended terms that match the key indicators and attribute information; the method further comprises: determining whether the structured expression text comprises at least partially overlapping expression texts; if yes, determining an expression text that contains more parameter information than other expression texts from the at least partially overlapping expression texts, and determining the expression text that contains more parameter information than other expression texts as a first expression text; or, if yes, regenerating a new expression text based on the at least partially overlapping expression texts and taking the new expression text as the first expression text; wherein whether to directly select an expression text from the at least partially overlapping expression texts as the first expression text or to regenerate an expression text is determined based on whether one expression text in the at least partially overlapping expression texts can contain a non-overlapping part of other expression texts.

2. The method of claim 1, wherein, the determination of the key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table comprises: performing matching segmentation on each sentence in the underwriting health description text; annotating the key text information based on the key indicators and attribute information in the structure level table.

3. The method of claim 1, wherein, the construction of the structured expression text that matches the structure level table of the insurance to be underwritten based on the key text information comprises: determining an expression text template based on the key text information; constructing the structured expression text according to the expression text template based on the key indicators and attribute information in the structure level table.

4. The method of claim 1, wherein, the method further comprises: determining whether the underwriting health description text meets the underwriting rules of the insurance to be underwritten based on the structured expression text.

5. A text structuring processing system characterized by comprising: The method comprises the following steps: an obtaining unit is configured to obtain underwriting information, wherein the underwriting information at least comprises underwriting rules of an insurance to be underwritten and an underwriting health description text; The first determining unit is configured to determine a structure level table of the insurance to be invested based on the underwriting rule of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested; The second determining unit is configured to determine key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table; The constructing unit is configured to construct a structured expression text that matches the structure level table of the insurance to be invested based on the key text information; The determination of the key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table includes: Determination of extended terms that match the key indicators and attribute information includes: based on a medical knowledge graph, introduction of professional data related to the structure level table, the professional data including a medical standard word table, a synonym word table, and an entity classification word table, and based on a general corpus or general knowledge, introduction of various natural language general term tables with word features, including quantity words, symbols, negative words, stop words, mood words, and conjunction words; Determination of the key text information in the underwriting health description text based on the key indicators and attribute information in the structure level table and the extended terms that match the key indicators and attribute information; The third determining unit is configured to determine whether the structured expression text includes at least partially overlapping expression texts, and if so, determine an expression text containing more parameter information than other expression texts from the at least partially overlapping expression texts, and determine the expression text containing more parameter information than other expression texts as a first expression text. Or, if so, a new expression text is regenerated based on the at least partially overlapping expression texts, and the new expression text is taken as the first expression text.

6. An electronic device, comprising: The processor is configured to obtain underwriting information, the underwriting information at least including an underwriting rule and an underwriting health description text of an insurance to be invested; The first determining unit is configured to determine a structure level table of the insurance to be invested based on the underwriting rule of the insurance to be invested, the structure level table at least including key indicators and attribute information of the insurance to be invested; The second determining unit is configured to determine key text information in the underwriting health description text that matches the key indicators and attribute information in the structure level table; The constructing unit is configured to construct a structured expression text that matches the structure level table of the insurance to be invested based on the key text information; The third determining unit is configured to determine whether the structured expression text includes at least partially overlapping expression texts, and if so, determine an expression text containing more parameter information than other expression texts from the at least partially overlapping expression texts, and determine the expression text containing more parameter information than other expression texts as a first expression text. Or, if included, a new expression text is regenerated based on multiple expression texts that at least partially overlap, and the new expression text is taken as the first expression text; wherein whether to directly select an expression text from the multiple expression texts that at least partially overlap as the first expression text or to regenerate an expression text is determined based on a result of judging whether there is an expression text in the multiple expression texts that can include a non-overlapping part in other expression texts A memory for storing programs for the processor to execute the above-mentioned processing procedures; The determining of the key text information in the health description text that matches the key indicators and attribute information in the structure hierarchy table comprises: Determining the extended terms that match the key indicators and attribute information comprises: introducing professional data related to the structure hierarchy table based on a medical knowledge graph, wherein the professional data comprises a medical standard word table, a synonym word table, and an entity classification word table, and introducing various natural language general term tables with word types based on general corpus or general knowledge, wherein the natural language general term tables comprise quantity words, symbols, negative words, stop words, mood words, and conjunction words; Determining the key text information in the health description text based on the key indicators and attribute information in the structure hierarchy table and the extended terms that match the key indicators and attribute information.

7. A readable storage medium for storing at least a set of instruction sets; The instruction sets are used to be invoked and at least execute the method of text structuring processing according to any one of claims 1-4.

Images