A method and system for investigating responsibility
Through image recognition and text recognition technology matching the risk model library, an inquiry speech template was generated, which solved the problem that new employees could find it difficult to identify the type of case, and achieved standardization and efficiency improvement in insurance claims investigations.
Patent Information
- Application Number
- CN202411109756.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2044-08-14
AI Technical Summary
In insurance claims investigation, it is difficult for new employees to identify the inquiry focus of different types of cases, resulting in inquiry irregularities, prolonging the inspection time, and the existing technology lacks a standardized inspection process.
The accident site data is obtained through image recognition and text recognition technology, match the case scenarios in the risk model library, select risk points and generate inquiry speech templates, and realize the automated inquiry process.
The standardization of survey work, intelligence of survey focus and automation of inquiry speech have been achieved, which has reduced survey preparation time, improved survey efficiency, reduced skill requirements, and helped to accurately investigate.
Smart Images

Figure CN118644847B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of insurance investigation, and in particular to a liability investigation method and system. Background Art
[0002] During construction, construction workers and equipment are usually insured. When an accident occurs, the insurance company will be notified to make a claim. Before making an insurance claim, the insurance investigation team usually conducts a series of investigations as the basis and evidence for the insurance claim. Generally speaking, insurance companies have their own claims investigation teams, but investigations often require a lot of manpower and material resources, so some insurance companies also entrust professional investigation companies to do it to save costs. However, in any case, surveyors are required to go to the scene to investigate and collect evidence. In this process, relevant personnel will be questioned. However, different case risk scenarios have different inquiry focuses. Each person may have different inquiry focuses for different types of cases. At the same time, new employees may not be able to identify different cases of the same type. In this way, the inquiry focus of new employees for the same type of cases occurring at different times may also be different, resulting in irregular inquiries. If a second survey is conducted, the survey time may be extended. Therefore, how to standardize the survey process and reduce the survey time is a process that needs to be achieved. Summary of the invention
[0003] The present invention provides a method and system for investigating responsibility to solve the above problems.
[0004] The present invention is achieved through the following technical solutions:
[0005] A method of investigating responsibility, comprising:
[0006] S1. Acquire image data of the accident scene and text data of the time of the accident, the cause of the accident, and the characteristics of the accident subject, and recognize a number of image features in the image data based on image recognition technology, and recognize a number of text features in the text data based on text recognition technology;
[0007] S2. Match and investigate the original case risk scenarios in the risk model library according to all the image features, the text of the cause of the accident, and the features of the accident subject, and find the original case risk scenario with the highest matching degree among all the image features and text features, and make the original case risk scenario the best case risk scenario;
[0008] S3, selecting the risk points corresponding to the optimal case risk scenario, and then matching the corresponding investigation and inquiry script template according to the options of the risk points;
[0009] S4. Conduct inquiries and record them according to the survey inquiry script template.
[0010] As an optimization, the image features include accident location type features and accident occurrence weather type features;
[0011] The accident location type characteristics include the location type characteristics of the accident location and the attribute characteristics of the accident location. The attribute characteristics of the accident location include workplace attributes and non-workplace attributes. The location type characteristics of the accident location include town streets, mountains and forests, rivers, and flat land outside the city.
[0012] The weather type characteristics for the accident include sunny days, rainy days without thunder, snowy days, and rainy days with thunder.
[0013] As an optimization, the text features corresponding to the text data of the accident subject are made the accident subject features, and the accident subject features include personnel identity features, equipment and construction projects, wherein the personnel identity features include employees and third parties.
[0014] As an optimization, the identification constraints of the accident location attributes are as follows:
[0015] Obtain the characteristics of the subject of the accident. If the subject of the accident is an employee, make the following judgment:
[0016] A1. Preliminarily determine whether the accident site is a workplace based on the employee's attire at the time of the accident and whether there are construction equipment around the accident. If the employee is wearing construction clothing and there are construction equipment around the accident, then preliminarily determine that the accident site is a workplace and jump to A2. Otherwise, preliminarily determine that the accident site is not a workplace and jump to A3.
[0017] A2. If the time of the accident is not within the employee's working hours, the employee is determined to have arrived at the workplace during non-working hours, and the attribute of the accident location is a non-workplace attribute; otherwise, the employee is determined to have arrived at the workplace during working hours, and the attribute of the accident location is a workplace attribute;
[0018] A3. If the accident location is within the employee's commute route and the accident occurs within N hours before or after the working hours, the attribute of the accident location is determined to be a workplace attribute;
[0019] Acquiring the characteristics of the subject of the accident, and if the characteristics of the subject of the accident are a third person, determining that the attribute of the accident location is a non-workplace attribute;
[0020] Acquire the main feature of the accident, and if the main feature of the accident is that there is only equipment or only a construction project or only equipment and a construction project, determine that the attribute of the accident location is a workplace attribute;
[0021] The characteristics of the subject of the accident are obtained. If the characteristics of the subject of the accident are a third person and a construction project or a third person and equipment, it is determined that the attribute of the accident location is a non-workplace attribute.
[0022] As an optimization, the causes of the accident include accidental causes and man-made causes, among which the accidental causes include traffic accidents, falls from heights, falls, object strikes, mechanical extrusion, poisoning, explosions, fires, drowning, electric shocks, suffocation, collapse and burial, typhoons, earthquakes, mudslides, landslides, collapses, floods, heavy rains, lightning strikes, snowstorms, tsunamis, the falling of objects in the air, and the collapse of various objects, and the man-made causes include human negligence, intentional acts, and malicious acts.
[0023] As an optimization, the risk points corresponding to the original case risk scenario are obtained based on the risk point configuration model. The first P risk points with the highest scores are selected as the risk points of the original case risk scenario. P is a positive integer, and the risk point configuration model is:
[0024] ;
[0025] in, It represents the correlation of the i-th risk point to the j-th non-accident subject feature and the attribute feature of the non-accident location. , when the i-th risk point is not related to the j-th non-accident subject feature, , when the i-th risk point is related to the j-th non-accident subject feature, , represents the relevance of the ith risk point to the eth accident subject feature, E represents the number of accident subject features contained in the original case risk scenario, , when the i-th risk point is not related to the e-th accident subject feature, then , when the i-th risk point is related to the e-th accident subject feature, then , It represents the correlation between the i-th risk point and the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics. , when the attribute characteristics of the accident location corresponding to the i-th risk point and the e-th accident subject characteristics are not related, , when the i-th risk point is related to the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics, , R represents the number of features that are not accident subject features, Represents the score of the ith risk point for the original case risk scenario.
[0026] As an optimization, in S4, the recording methods of conducting inquiries and recording according to the survey inquiry speech template include text recording, voice recording and picture recording.
[0027] The present invention also discloses a system for investigating responsibility, which uses the aforementioned method for investigating responsibility to obtain an investigation and questioning script template, including:
[0028] Login module, used to collect the identity information of the investigator, the location of inquiry and the date of the accident;
[0029] The evidence collection module obtains image data of the accident scene and text data of the time of the accident, the cause of the accident, and the characteristics of the accident subject, and identifies several image features in the image data based on image recognition technology, and at the same time, identifies several text features in the text data based on text recognition technology;
[0030] Survey risk model library, used to store multiple original case risk scenarios;
[0031] A mapping list module is used to store multiple pointers, each of which is composed of a feature group and an address corresponding to the feature group, the address is the address where the original case risk scenario is stored in the investigation risk model library, and the feature group is composed of various image features and text features of the cause of the accident and the subject of the accident;
[0032] A risk point library is used to store various risk points, each of which is matched with at least one inquiry speech template, and the same risk point is provided with different options for selecting the risk point, and each option of the same risk point corresponds to an inquiry speech template;
[0033] Inquiry script library, used to store inquiry script templates;
[0034] A matching module, used for matching each risk point with the original case risk scenario to find P risk points that best fit the original case risk scenario, and storing the P risk points that match the original case risk scenario, where P is a positive integer;
[0035] A selection module, used to select each option of the P risk points matched to the original case risk scenario, and find a corresponding inquiry speech template in the inquiry speech library according to the selection of each option;
[0036] The query key prompt module is used to display all the query script templates found;
[0037] A reselection module, used to return to the selection module for reselection when the user is not satisfied with all the inquiry speech templates displayed in the inquiry key point prompt module;
[0038] Inquiry module, used to prepare to enter the inquiry interface;
[0039] A recording module, used to record and save inquiry data obtained by inquiring according to the words in the inquiry words template;
[0040] The inquiry record generating module is used to generate a visual interface from the inquiry data.
[0041] As an optimization, it also includes a GPS positioning module and a GPS clock module, which are respectively connected to the login module and are used to provide the login module with query location information and accident date information when the login module logs in;
[0042] It also includes a recording module, which is connected to the inquiry module and is used to record the inquiry content before starting the inquiry;
[0043] The query data recorded by the recording module includes text data and image data, and an electronic signature is performed by the recording module after the recording is completed;
[0044] The evidence collection module includes an image acquisition module and a text acquisition module. The image acquisition module is used to acquire image data of the accident scene, and the text acquisition module is used to acquire text data of the cause of the accident and the characteristics of the accident subject.
[0045] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0046] Through the method and system of the present invention, it is possible to achieve standardization of investigation work content, intelligent reminders of investigation key points, automated generation of interrogation scripts, and electronic delivery of investigation records, thereby achieving the effects of weakening investigation skill requirements, reducing investigation preparation time, improving investigation work efficiency, enabling front-line investigations to reduce burdens, and accurate investigations to help combat counterfeiting. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:
[0048] Figure 1 A flowchart of a method for investigating responsibility according to the present invention;
[0049] Figure 2 This is a schematic diagram of module connections of a responsibility investigation system according to the present invention.
[0050] Marks and corresponding parts names in the attached drawings:
[0051] 1- Login module, 2- Evidence collection module, 2a- Image acquisition module, 2b- Text acquisition module, 3- Mapping list module, 4- Risk model library, 5- Risk key point library, 6- Matching module, 7- Inquiry script library, 8- Click selection module, 9- Inquiry key point prompt module, 10- Reselection module, 11- Inquiry module, 12- Recording module, 13- Recording module, 14- Inquiry record generation module, 15- GPS positioning module, 16- GPS clock module. DETAILED DESCRIPTION
[0052] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0053] Before introducing the specific content, it should be explained that the intention and liability investigation method of the present invention is mainly aimed at the risk investigation of employer insurance, construction employment insurance, construction intention insurance and construction safety insurance cases. Therefore, the subject of the accident may be a person (including employees and third parties, the third party is a person who is not related to the construction project), the entire project, or a certain equipment in the construction project. And because some insurances can protect both projects and personnel at the same time, such as construction safety insurance, it is divided into two categories: material property itself and third-party liability. Among them, the material property itself includes installation projects, civil engineering projects, site cleaning fees, and other properties of the owner or contractor on the construction site; third-party liability refers to the compensation liability that the insured should bear according to law for personal injury or property loss to third parties on the construction site and neighboring areas due to accidents on the construction site during the validity period of the insurance, and the litigation fees paid for this and other expenses agreed in writing by the insurer. Therefore, when conducting an investigation, this aspect needs to be fully considered.
[0054] This embodiment 1 provides a method for investigating responsibility. Figure 1 As shown, including:
[0055] S1. Acquire image data of the accident scene and text data of the time of the accident, the cause of the accident, and the characteristics of the accident subject, and recognize a number of image features in the image data based on image recognition technology, and recognize a number of text features in the text data based on text recognition technology;
[0056] In this embodiment, the image features include accident scene environmental features, wherein the accident scene environmental features include accident location type features and accident occurrence weather type features.
[0057] The text features corresponding to the text data of the accident subject are set as accident subject features, and the accident subject features include personnel identity features, equipment and construction projects, wherein the personnel identity features include employees and third parties.
[0058] More specifically, the accident site type characteristics include the location type characteristics of the accident site and the attribute characteristics of the accident site.
[0059] The location type characteristics of the accident site include town streets, mountains, rivers, and flat land outside the city; construction may be carried out in each type of place, that is, accidents may occur in each type of place, so the place for investigation after the accident may also be anywhere.
[0060] Town streets are a general term here, mainly referring to infrastructure projects within the city. Any building on the street, the road on the street, the park in the city, and the place where accidents occur in rural town streets are all town streets.
[0061] Mountain forest refers to the area where the mountain top height is higher than the set height threshold and the distance to the road edge exceeds the length threshold. It is mainly for construction projects carried out on the mountain, such as the construction of base stations and high-voltage poles. If some towns are located on the mountain, then the location type characteristics of the town located on the mountain belong to the town street, and the area where the distance to the road edge of the town exceeds the length threshold is the mountain forest.
[0062] Rivers are mainly aimed at the construction of bridges and riverbed tunnels.
[0063] The flat land outside the city is mainly for construction projects carried out on flat land that is not surrounded by mountains, forests, rivers and is uninhabited, such as building power stations on grasslands.
[0064] The attribute characteristics of the accident location include workplace attributes and non-workplace attributes. Accidents occurring in the workplace and non-workplace will have different compensation methods. Therefore, when conducting an investigation, it is also necessary to clarify whether the location of the accident is a workplace or a non-workplace. The attribute characteristics of the accident location can be quickly judged through image data and text data of the time of the accident and the characteristics of the accident subject. The judgment constraints (identification constraints) are as follows:
[0065] Obtain the characteristics of the subject of the accident. If the subject of the accident is an employee, make the following judgment:
[0066] A1. Preliminarily determine whether the accident site is a workplace based on the employee's attire at the time of the accident and whether there are construction equipment around the accident. If the employee is wearing construction clothing and there are construction equipment around the accident, then preliminarily determine that the accident site is a workplace and jump to A2. Otherwise, preliminarily determine that the accident site is not a workplace and jump to A3.
[0067] A2. If the time of the accident is not within the employee's working hours, the employee is determined to have arrived at the workplace during non-working hours, and the attribute of the accident location is a non-workplace attribute; otherwise, the employee is determined to have arrived at the workplace during working hours, and the attribute of the accident location is a workplace attribute;
[0068] During construction, workers (employees) will wear corresponding construction clothing, such as safety helmets. Therefore, if it is detected through image data that there are construction equipment around the accident site and the workers are wearing corresponding construction clothing, then it can be preliminarily determined that the accident site is a construction area and the subject of the accident is a worker in the construction area. Therefore, the accident site is very likely to be the workplace of the subject of the accident. However, there may be situations where the subject of the accident enters the workplace during non-working hours. Therefore, when the subject of the accident enters the workplace during non-working hours and an accident occurs, the work area cannot be determined as a workplace.
[0069] A3. If the accident location is within the employee's commute route and the accident occurs within N hours before or after the working hours, the attribute of the accident location is determined to be a workplace attribute;
[0070] At the same time, according to legal provisions, situations where certain conditions are met during commuting time are considered work-related injuries. Therefore, the location of the accident that meets the requirements of a work-related injury is also considered to be a workplace attribute.
[0071] The characteristics of the accident subject are obtained. If the characteristics of the accident subject are a third person, it is determined that the attribute of the accident location is a non-workplace attribute.
[0072] The main feature of the accident is obtained. If the main feature of the accident is that there are only equipment or only construction projects or only equipment and construction projects, it is determined that the attribute of the accident location is a workplace attribute.
[0073] The characteristics of the accident subject are obtained. If the characteristics of the accident subject are a third person and a construction project or a third person and equipment, the attributes of the accident location are determined to be non-workplace attributes.
[0074] The weather types in which the accident occurred include sunny days, rainy days without thunder, snowy days, and rainy days with thunder.
[0075] If the weather changes while the investigator is conducting an investigation, the type of weather when the accident occurred can be obtained through the Internet. If the weather does not change while the investigator is conducting an investigation, then the type of weather when the accident occurred can be obtained by taking real-time photos.
[0076] The causes of the accident include accidental causes and human causes, among which the accidental causes include traffic accidents, falls from height, falls, object strikes, mechanical extrusion, poisoning, explosions, fires, drowning, electric shocks, suffocation, collapse and burial, typhoons, earthquakes, mudslides, landslides, collapses, floods, rainstorms, lightning strikes, snowstorms, tsunamis, the fall of flying objects, and the collapse of various objects. The human causes include human negligence, intentional acts, and malicious acts. The accident subject characteristics of the same original case risk scenario include at least one at the same time.
[0077] Through the above-mentioned feature extraction, the specific circumstances when the accident occurred can be determined.
[0078] For example, the occurrence of a certain accident can be obtained through image features and text features:
[0079] Accident subject: Employee;
[0080] The type of location where the accident took place: mountain forest;
[0081] Attributes of the accident location: workplace;
[0082] Weather type when the accident occurred: rainy day with thunder;
[0083] Cause of the accident: lightning strike.
[0084] Of course, the location type of the accident site and the weather type where the accident occurred can be divided into more detailed categories, which will not be elaborated here.
[0085] S2. Match and investigate the original case risk scenarios in the risk model library according to all the image features and the text features of the cause of the accident and the subject of the accident, find the original case risk scenario with the highest matching degree among all the image features and text features, and make the original case risk scenario the best case risk scenario;
[0086] Each original case risk scenario has a corresponding matching image feature and text feature. For example, the image feature and text feature are:
[0087] Characteristics of the accident subject 1: third party;
[0088] Accident subject characteristic 2: Construction project;
[0089] Characteristics of the location type of the accident site: mountain forest;
[0090] Attributes of the accident location: 1. Non-workplace;
[0091] Attribute characteristics of the accident location 2: workplace;
[0092] Characteristics of the weather type in which the accident occurred: rainy day with thunder;
[0093] Characteristics of the cause of the accident: lightning strike.
[0094] Corresponding to the original case risk scenario A.
[0095] At this time, the attribute feature of the accident location corresponding to the accident subject feature one is the attribute feature one of the accident location, and the attribute feature of the accident location corresponding to the accident subject feature two is the attribute feature two of the accident location.
[0096] Characteristics of the accident subject: Employee;
[0097] Characteristics of the location type of the accident site: mountain forest;
[0098] Attributes of the accident location: workplace;
[0099] Characteristics of the weather type in which the accident occurred: rainy day with thunder;
[0100] Characteristics of the cause of the accident: lightning strike.
[0101] Corresponding to the original case risk scenario B.
[0102] At this time, there is only one main feature of the accident, so there is also only one attribute feature of the corresponding accident location.
[0103] S3, selecting the risk points corresponding to the optimal case risk scenario, and then matching the corresponding investigation and inquiry script template according to the options of the risk points;
[0104] The risk points are a summary of the interrogation dialogues corresponding to two different situations, and are also a summary of the situations that must be understood in each case risk scenario. Usually, the risk points are yes-or-no multiple-choice questions, that is, there are generally two options for the risk points, "yes" or "no", "yes" corresponds to one interrogation dialogue, and "no" corresponds to another interrogation dialogue. Of course, the risk points can also be other types of multiple-choice questions, that is, there can be three or more options for the risk points. In this embodiment, there are two options for the risk points, namely "yes" or "no". Different case risk scenarios correspond to different risk points, and different options for the risk points of the same case risk scenario will correspond to different interrogation dialogues, that is, one case risk scenario corresponds to N risk points, and one risk point corresponds to 2 sets of interrogation dialogues, "yes" corresponds to one set of interrogation dialogues, and "no" corresponds to a set of interrogation dialogues.
[0105] Since different situations may lead to different accidents, it is very important to correctly match the corresponding risk points with the corresponding case risk scenarios.
[0106] Since there may be other original case risk scenarios that are not currently thought of, whenever a new original case risk scenario appears, it is necessary to match the corresponding risk points. Specifically,
[0107] The risk points corresponding to the original case risk scenario are obtained based on the risk point configuration model. The first P risk points with the highest scores are selected as the risk points of the original case risk scenario. P is a positive integer. The risk point configuration model is:
[0108] ;
[0109] in, It represents the correlation of the i-th risk point to the j-th non-accident subject feature and the attribute feature of the non-accident location. , when the i-th risk point is not related to the j-th non-accident subject feature, , when the i-th risk point is related to the j-th non-accident subject feature, , represents the relevance of the ith risk point to the eth accident subject feature, E represents the number of accident subject features contained in the original case risk scenario, , when the i-th risk point is not related to the e-th accident subject feature, then , when the i-th risk point is related to the e-th accident subject feature, then , It represents the correlation between the i-th risk point and the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics. , when the attribute characteristics of the accident location corresponding to the i-th risk point and the e-th accident subject characteristics are not related, , when the i-th risk point is related to the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics, , R represents the number of features that are not accident subject features, Represents the score of the ith risk point for the original case risk scenario.
[0110] The characteristics of non-accident subject characteristics and non-accident location attribute characteristics mainly include the location type characteristics of the accident location and the weather type characteristics of the accident.
[0111] If the problem targeted by the risk point is not related to the characteristics of the non-accident subject characteristics and the attribute characteristics of the non-accident location, for example, the risk point is "insurance according to the project cost or building area", and the characteristics of the non-accident subject characteristics and the attribute characteristics of the non-accident location are the location type characteristics of the accident location, specifically "forest", then the correlation between the risk point and the characteristics of the non-accident subject characteristics and the attribute characteristics of the non-accident location is 0; on the contrary, for example, the risk point is "going out on a rainy day", and the characteristics of the non-accident subject characteristics and the attribute characteristics of the non-accident location are the weather type characteristics of the accident, specifically "rainy day with thunder", then the correlation between the risk point and the characteristics of the non-accident subject characteristics and the attribute characteristics of the non-accident location is 1.
[0112] Through the risk point configuration model, the risk points of various original case risk scenarios can be configured at any time. For newly added original case risk scenarios, the corresponding risk points can be quickly configured.
[0113] S4. Conduct inquiries and record them according to the survey inquiry script template.
[0114] In S4, the recording method of the inquiry and recording according to the investigation inquiry speech template includes text recording, voice recording and picture recording.
[0115] Embodiment 2 also discloses a system for investigating responsibility, using the aforementioned method for investigating responsibility to obtain an investigation and questioning script template, such as Figure 2 As shown, including:
[0116] Login module 1, used to collect the identity information of the investigator, the inquiry location information and the accident date information;
[0117] In this embodiment, a GPS positioning module and a GPS clock module are also included. The GPS positioning module and the GPS clock module are respectively connected to the login module 1, and are used to provide the login module 1 with the inquiry location information and the accident date information when the login module 1 logs in; in this way, when the investigator arrives at the accident scene for investigation and enters the login module 1 to input the identity information, the GPS positioning module and the GPS clock module automatically obtain the inquiry location information and the accident date information, thereby avoiding errors in manually entering the inquiry location information and the accident date information.
[0118] Evidence collection module 2, used to obtain image data of the accident scene and text data of the time of the accident, the cause of the accident, and the characteristics of the accident subject, and to identify several image features in the image data based on image recognition technology, and at the same time, to identify several text features in the text data based on text recognition technology;
[0119] Specifically, the evidence collection module 2 includes an image acquisition module 2a and a text acquisition module 2b. The image acquisition module 2a is used to acquire image data of the accident scene, and the text acquisition module 2b is used to acquire text data of the cause of the accident and the characteristics of the subject of the accident. In the text acquisition module 2b, there are a column for inputting the cause of the accident and a column for inputting the characteristics of the subject of the accident, so that the cause of the accident and the characteristics of the subject of the accident can be quickly identified and distinguished.
[0120] Image recognition technology and text recognition technology are existing technologies. Recognizing the weather and location type in an image through image recognition technology is an existing technology and will not be elaborated here.
[0121] Investigation risk model library 4, used to store multiple original case risk scenarios;
[0122] If there is a new original case risk scenario, the new original case risk scenario is stored in the investigation risk model library 4.
[0123] A mapping list module 3 is used to store multiple pointers, each of which is composed of a feature group and an address corresponding to the feature group, the address is the address of the original case risk scenario stored in the investigation risk model library 4, and the feature group is composed of various image features and text features of the cause of the accident and the subject of the accident;
[0124] That is, the format of the feature group can be: accident subject feature | accident location location type feature | accident location attribute feature | accident weather type feature | accident cause feature. For example, a feature group for an accident subject feature and an accident location attribute feature can be expressed as:
[0125] Employees|forests|workplace|rainy day with thunder|lightning strikes.
[0126] The feature group for the two accident subject features and the attribute features of the two accident locations can be expressed as:
[0127] Third party & construction project | forest | non-workplace & workplace | rainy day with thunder | lightning strike.
[0128] That is, different features are distinguished by a distinguishing symbol, such as "|", and different contents of the same feature are distinguished by another distinguishing symbol, such as "&".
[0129] Feature group A-scenario A in the mapping list module means that the original case risk scenario A corresponds to feature group A.
[0130] A risk point library 5 is used to store various risk points, each of which is matched with at least one inquiry speech template, and the same risk point is provided with different options for selecting the risk point, and each option of the same risk point corresponds to an inquiry speech template;
[0131] The risk key point library 5 can add new risk key points.
[0132] Inquiry script library 7, used to store inquiry script templates;
[0133] Matching module 6, used for matching each risk point with the original case risk scenario to find P risk points that best fit the original case risk scenario, and storing the P risk points that match the original case risk scenario, where P is a positive integer;
[0134] The matching module 6 stores a risk point configuration model. When a certain original case risk scenario has not yet matched a corresponding risk point, the original case risk scenario is input into the matching module 6. The risk point configuration model searches for the risk point with the highest score as the risk point of the original case risk scenario according to the characteristics of the original case risk scenario. , , Once formed, a binding is formed between the risk points and the characteristics, so the score of the risk points for the original case risk scenario can be quickly obtained.
[0135] The selection module 8 is used to select each option of the P risk points matched to the original case risk scenario, and find the corresponding inquiry speech template in the inquiry speech library 7 according to the selection of each option.
[0136] After selecting the risk key points option in the interface of the selection module 8, click on the inquiry prompt to enter the interface of the query key points prompt module 9 to check the inquiry script template.
[0137] The query key point prompt module 9 is used to display all the query script templates found;
[0138] After checking the inspection key points prompts (interrogation script template), if you are not satisfied with the interrogation script template, you can reselect it and re-enter the selection module 8 through the reselection module 10. Of course, you can also reopen the system and re-collect evidence through the evidence collection module 2; if you think the interrogation script template is acceptable, enter the interrogation module 11.
[0139] A reselection module 10, for returning to the selection module 8 for reselection when the user is not satisfied with all the inquiry speech templates displayed in the inquiry key point prompt module 9;
[0140] The inquiry module 11 is used to prepare to enter the inquiry interface. On the inquiry screen, the information collected by the login module 1 is displayed, and a user can click whether to ask the user himself or others. The inquiry module 11 also includes a recording module 12. The recording module 12 is connected to the inquiry module 11 and is used to record the inquiry content before starting the inquiry. After selecting, click "Start Inquiry" to enter the recording module 13.
[0141] A recording module 13, used to record and save inquiry data obtained by inquiring according to the words in the inquiry words template;
[0142] Specifically, the query data recorded by the recording module 13 includes text data and image data, and an electronic signature is performed by the recording module 13 after the recording is completed;
[0143] The inquiry record generating module 14 is used to generate a visual interface from the inquiry data.
[0144] The system of the present invention focuses on risk screening, risk warnings, investigation scripts, and one-click generation of electronic transcripts for employer insurance, construction employment insurance, construction malpractice insurance, and construction safety insurance cases. It has a built-in investigation risk model library, covering 1024 types of case risk scenarios, and accurately matching 1024 versions of risk warning points and 2048 sets of investigation and questioning scripts. You only need to check 10 major risk questionnaire questions, and the system will automatically match the most suitable case handling risk warnings. It can automatically generate investigation and questioning script templates based on the checked risk points, support recording, manual voice input, and electronic signatures, and generate electronic transcripts with one click, breaking down the barriers of professional skills and experience, intelligently conducting pre-risk screening, and improving the standardization of investigation operations.
[0145] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for investigating responsibility, characterized in that: include: S1. Acquire image data of the accident scene and text data of the time of the accident, the cause of the accident, and the subject of the accident, and recognize a number of image features in the image data based on image recognition technology, and recognize a number of text features in the text data based on text recognition technology; S2. Match and investigate the original case risk scenarios in the risk model library according to all the image features and the text features of the cause of the accident and the subject of the accident, find the original case risk scenario with the highest matching degree among all the image features and text features, and make the original case risk scenario the best case risk scenario; The risk points corresponding to the original case risk scenario are obtained based on the risk point configuration model. The first P risk points with the highest scores are selected as the risk points of the original case risk scenario. P is a positive integer. The risk point configuration model is: Among them, s i,j represents the correlation of the i-th risk point to the j-th non-accident subject feature and the attribute feature of the non-accident location, s i,j = {0, 1}, when the i-th risk point is not related to the j-th non-accident subject feature, s i,j = 0, when the i-th risk point is related to the j-th non-accident subject feature, s i,j =1,s′ i,e represents the relevance of the ith risk point to the eth accident subject feature, E represents the number of accident subject features contained in the original case risk scenario, and s′ i,e = {0, 1}, when the i-th risk point is not related to the e-th accident subject feature, then s′ i,e = 0, when the i-th risk point is related to the e-th accident subject feature, then s′ i,e =1,s″ i,e represents the correlation between the i-th risk point and the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics, s″ i,e = {0, 1}, when the attribute characteristics of the accident site corresponding to the i-th risk point and the e-th accident subject feature are not related, s″ i,e = 0, when the i-th risk point is related to the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics, s″ i,e = 1, R represents the number of features that are not accident subject features, s i represents the score of the i-th risk point for the original case risk scenario; S3, selecting the risk points corresponding to the optimal case risk scenario, and then matching the corresponding investigation and inquiry script template according to the options of the risk points; S4. Conduct inquiries and record them according to the survey inquiry script template.
2. A method for investigating responsibility according to claim 1, characterized in that: The image features include accident location type features and accident occurrence weather type features; The accident location type features include the location type features of the accident location and the attribute features of the accident location. The attribute features of the accident location include workplace attributes and non-workplace attributes. The location type features of the accident location include town streets, mountains and forests, rivers, and flat land outside the city. The weather type characteristics for the accident include sunny days, rainy days without thunder, snowy days, and rainy days with thunder.
3. A method for investigating responsibility according to claim 2, characterized in that: The text features corresponding to the text data of the accident subject are set as accident subject features, and the accident subject features include personnel identity features, equipment and construction projects, wherein the personnel identity features include employees and third parties.
4. A method for investigating responsibility according to claim 3, characterized in that: The identification constraints of the attribute features of the accident site are as follows: Obtain the characteristics of the subject of the accident. If the subject of the accident is an employee, make the following judgment: A1. Preliminarily determine whether the accident site is a workplace based on the employee's attire at the time of the accident and whether there are construction equipment around the accident. If the employee is wearing construction clothing and there are construction equipment around the accident, then preliminarily determine that the accident site is a workplace and jump to A2. Otherwise, preliminarily determine that the accident site is not a workplace and jump to A3. A2. If the time of the accident is not within the working hours of the employee, the employee is determined to have arrived at the workplace during non-working hours, and the attribute of the accident location is a non-workplace attribute; otherwise, the employee is determined to have arrived at the workplace during working hours, and the attribute of the accident location is a workplace attribute; A3. If the accident location is within the employee's commute route and the accident occurs within N hours before or after the working hours, the attribute of the accident location is determined to be a workplace attribute; Acquiring the characteristics of the subject of the accident, and if the characteristics of the subject of the accident are a third person, determining that the attribute of the accident location is a non-workplace attribute; Acquire the main feature of the accident, and if the main feature of the accident is that there is only equipment or only a construction project or only equipment and a construction project, determine that the attribute of the accident location is a workplace attribute; The characteristics of the subject of the accident are obtained. If the characteristics of the subject of the accident are a third person and a construction project or a third person and equipment, it is determined that the attribute of the accident location is a non-workplace attribute.
5. The method for investigating responsibility according to claim 3, characterized in that: The causes of the accident include accidental causes and man-made causes. The accidental causes include traffic accidents, falls from heights, falls, object strikes, mechanical squeezing, poisoning, explosions, fires, drowning, electric shocks, suffocation, collapse and burial, typhoons, earthquakes, mudslides, landslides, collapses, floods, heavy rains, lightning strikes, snowstorms, tsunamis, the falling of objects in the air, and the collapse of various objects. The man-made causes include human negligence, intentional acts, and malicious acts.
6. The method for investigating responsibility according to claim 1, characterized in that: In S4, the recording method of the inquiry and recording according to the investigation inquiry speech template includes text recording, voice recording and picture recording.
7. A responsibility investigation system, using a responsibility investigation method as described in any one of claims 1 to 6 to obtain an investigation and questioning script template, characterized in that: include: Login module, used to collect the identity information of the investigator, the location of inquiry and the date of the accident; The evidence collection module obtains image data of the accident scene and text data of the time of the accident, the cause of the accident, and the characteristics of the accident subject, and identifies several image features in the image data based on image recognition technology, and at the same time, identifies several text features in the text data based on text recognition technology; Survey risk model library, used to store multiple original case risk scenarios; A mapping list module is used to store multiple pointers, each of which is composed of a feature group and an address corresponding to the feature group, the address is the address where the original case risk scenario is stored in the investigation risk model library, and the feature group is composed of various image features and text features of the cause of the accident and the subject of the accident; A risk point library is used to store various risk points, each of which is matched with at least one inquiry speech template, and the same risk point is provided with different options for selecting the risk point, and each option of the same risk point corresponds to an inquiry speech template; Inquiry script library, used to store inquiry script templates; A matching module, used for matching each risk point with the original case risk scenario to find P risk points that best fit the original case risk scenario, and storing the P risk points that match the original case risk scenario, where P is a positive integer; The risk points corresponding to the original case risk scenario are obtained based on the risk point configuration model. The first P risk points with the highest scores are selected as the risk points of the original case risk scenario. P is a positive integer. The risk point configuration model is: Among them, s i,j represents the correlation of the i-th risk point to the j-th non-accident subject feature and the attribute feature of the non-accident location, s i,j = {0, 1}, when the i-th risk point is not related to the j-th non-accident subject feature, s i,j = 0, when the i-th risk point is related to the j-th non-accident subject feature, s i,j =1,s′ i,e represents the relevance of the ith risk point to the eth accident subject feature, E represents the number of accident subject features contained in the original case risk scenario, and s′ i,e = {0, 1}, when the i-th risk point is not related to the e-th accident subject feature, then s′ i,e = 0, when the i-th risk point is related to the e-th accident subject feature, then s′ i,e =1,s″ i,e represents the correlation between the i-th risk point and the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics, s″ i,e = {0, 1}, when the attribute characteristics of the accident site corresponding to the i-th risk point and the e-th accident subject feature are not related, s″ i,e = 0, when the i-th risk point is related to the attribute characteristics of the accident location corresponding to the e-th accident subject characteristics, s″ i,e = 1, R represents the number of features that are not accident subject features, s i represents the score of the i-th risk point for the original case risk scenario; A selection module, used to select each option of the P risk points matched to the original case risk scenario, and find a corresponding inquiry speech template in the inquiry speech library according to the selection of each option; The query key prompt module is used to display all the query script templates found; A reselection module, used to return to the selection module for reselection when the user is not satisfied with all the inquiry speech templates displayed in the inquiry key point prompt module; Inquiry module, used to prepare to enter the inquiry interface; A recording module, used to record and save inquiry data obtained by inquiring according to the words in the inquiry words template; The inquiry record generating module is used to generate a visual interface from the inquiry data.
8. The responsibility investigation system according to claim 7, characterized in that: It also includes a GPS positioning module and a GPS clock module, which are respectively connected to the login module and are used to provide the login module with query location information and accident date information when the login module logs in; It also includes a recording module, which is connected to the inquiry module and is used to record the inquiry content before starting the inquiry; The query data recorded by the recording module includes text data and image data, and an electronic signature is performed by the recording module after the recording is completed; The evidence collection module includes an image acquisition module and a text acquisition module. The image acquisition module is used to acquire image data of the accident scene, and the text acquisition module is used to acquire text data of the cause of the accident and the characteristics of the accident subject.
Citation Information
Patent Citations
Automobile insurance claim settlement method and system
CN107679996A
An automobile insurance claim settlement anti-fraud system
CN109272413A
Questionnaire data processing method and device based on data analysis, server and medium
CN112100461A
Data processing method and device, computer equipment and storage medium
CN114170600A
Insurance verbal skill generation method and device, electronic equipment and storage medium
CN116702736A
