JP7325098B2 - Insurance content search system, insurance content search method, and insurance content search program - Google Patents

Description

The present invention relates to an insurance content search system, an insurance content search method, and an insurance content search program.

Recently, many life insurance services with various compensation contents are provided by many insurance companies. There is a need for an insurance content search system that allows the user to easily select the most suitable life insurance service from among such multiple life insurance services.
As such an insurance content search system, Patent Literature 1 below discloses a server that displays life insurance services that a user is interested in in a form that is easy to compare.

JP 2007-299257 A

Generally, such various insurance plans may be restricted in enrollment, for example, by a history of certain diseases.
However, in the above-mentioned insurance content search system, it is necessary for the user to check in detail whether or not the insurance services of the multiple insurance companies displayed for comparison can be subscribed. There was room for

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide an insurance content search system that enables users to search for insurance services to which they can subscribe by a simple operation.

In order to solve the above-mentioned problems, the insurance content search system of the present invention includes a reception unit that receives an input of a speaker's uttered voice, and a voice recognition unit that recognizes the uttered voice received by the reception unit and converts it into text information. , an extraction unit that extracts a disease name from text information, a database that associates the disease name with compensation availability information that indicates whether the disease name is covered by insurance, and the database is referenced based on the disease name. a specifying unit for specifying the corresponding compensation information, an output unit for outputting the compensation information specified by the specifying unit, and a data reading unit for receiving image information including at least the disease name and the compensation information. , a title phrase extraction unit that extracts the disease name indicated as the title from the image information, and a position where the compensation availability information is indicated with respect to the position where the disease name extracted by the title phrase extraction unit is indicated in the image information. is specified, and the compensation availability information extraction unit that extracts compensation availability information, the disease name extracted by the title phrase extraction unit, and the compensation availability information extracted by the compensation availability information extraction unit are registered in a database in association with each other. and

In addition, the compensation availability information extraction unit uses a learning model that has learned the positional relationship between the title phrase and the compensation availability information in a plurality of image information, and places the disease name extracted by the title phrase extraction unit in the position indicated in the image information. On the other hand, the position where the compensation propriety information is indicated may be specified.

Further, the compensation availability information extracting unit may identify an area included in a preset interval with respect to the position where the title phrase in the image information is shown as the position where the compensation availability information is shown.

Further, in the insurance content search method of the present invention, the computer comprises a receiving step of receiving an input of a speaker's uttered voice, a voice recognition step of recognizing the received uttered voice by the receiving step and converting it into text information, a text An extraction step of extracting a disease name from information, and referring to a database that associates a disease name with insurance coverage information indicating whether or not insurance coverage is available for that disease name based on the disease name, to obtain the corresponding compensation. a specifying step of specifying information on whether or not compensation is specified; an output step of outputting information on whether or not compensation is specified by the specifying step; a data reading step of receiving image information including at least a disease name and information on whether or not compensation is available; a title phrase extracting step of extracting a disease name shown as a title from the image information, and a position where compensation availability information is shown with respect to the position where the disease name extracted by the title phrase extraction step is shown in the image information, and compensation availability is determined. a compensation availability information extraction step of extracting information; and a registration step of correlating the disease name extracted by the title phrase extraction step and the compensation availability information extracted by the compensation availability information extraction step and registering them in a database. .

Further, the insurance content search program of the present invention is provided in a computer with a receiving function for receiving an input of a speaker's uttered voice, a voice recognition function for recognizing the uttered voice received by the receiving function and converting it into text information, and a text information. An extraction function that extracts the name of a disease from information, and based on the name of the disease, refers to a database that associates the name of the disease with insurance coverage information that indicates whether or not the disease is covered by insurance. a specific function for specifying information on whether compensation is possible, an output function for outputting information on whether compensation is specified by the specific function, a data reading function for receiving image information including at least a disease name and information on whether or not compensation is possible, A title phrase extraction function that extracts the disease name shown as a title from the image information, and a position where compensation availability information is shown with respect to the position where the disease name extracted by the title phrase extraction function is shown in the image information. A compensation availability information extraction function for extracting information and a registration function for registering in a database the disease name extracted by the title phrase extraction function and the compensation availability information extracted by the compensation availability information extraction function in association with each other are realized. .

According to the insurance content retrieval system of the present invention, the speech recognition section converts the speech of the speaker received by the reception section into text information. Then, the extraction unit extracts the disease name from the text information. Then, the specifying unit refers to the database and outputs insurance coverage availability information for the disease name. Therefore, for example, even if the user has a history of affliction with the disease, it is possible to immediately search for the insurance content to which the user can subscribe, simply by uttering the name of the disease that the user has a history of affliction with. Therefore, with a simple operation, it is possible to search for insurance services that the user can subscribe to.

1 is a schematic diagram of an insurance content search system of the present invention; FIG. It is a block diagram which shows the structure of an insurance content search system. FIG. 2 is a diagram showing a data structure in a database; FIG. FIG. 10 is a diagram showing contents related to an underwriting standard among image information of materials of an insurance company; It is a figure which shows the teacher data which a compensation propriety information extraction part uses. It is a figure which shows the area|region specification of the compensation propriety information extraction part which concerns on a modification. It is a figure which shows the whole flow among the processing flows of an insurance content search system. It is a figure which shows the database generation step among the processing flows of an insurance content search system. It is a figure which shows the search step among the processing flows of an insurance content search system. It is a figure which shows the search screen in an insurance content search system. It is a figure which shows the selection screen of the insurance company in an insurance content search system. It is a figure which shows the search result screen in an insurance content search system. It is a figure which shows the content of the search result by an insurance content search system.

(First embodiment)
A first embodiment of the present invention will be described below.
FIG. 1 is a schematic diagram of an insurance content search system 1 according to the first embodiment.
The insurance details search system 1 is a system that searches for the best insurance details for the user 20 by the user 20 operating the user terminal 10 . In insurance content search system 1 , information processing server 100 is connected to user terminal 10 via network 40 .

The user terminal 10 includes a display device, an input device, and a processor. The display device is a monitor, and the input devices are a microphone, a keyboard, a mouse, etc. for receiving the user's (speaker's) 20 uttered voice. That is, input may be performed by voice by speaking into a microphone, or input of characters by typing on a keyboard.

The search conditions input to the input device are processed by the information processing server 100 via the processor, and the searched insurance details are displayed on the display device.
The user terminal 10 is, for example, a personal computer as shown. Note that the user terminal 10 may be a smart phone, a mobile terminal, a tablet terminal, or other electronic devices. Note that the user terminal 10 is not limited to the illustrated example, and a plurality of user terminals 10 used by a plurality of users can be connected to the information processing server 100 via the network 40 .

Furthermore, the insurance content search system 1 may use a cloud service (both public cloud and private cloud), or a shared or dedicated information processing server 100 may be physically provided in the target facility to search for insurance content. You may provide a service that searches for
The insurance content search system 1 may be, for example, a device capable of processing a relatively large amount of information suitable for processing so-called big data.
Also, the insurance content search system 1 may be connected to databases of a plurality of insurance companies 30 via a network. The database of the insurance company 30 stores the latest insurance details of each company.

The network 40 is a network for interconnecting the insurance content search system 1 and various devices, such as a wireless network or a wired network.
Specifically, the network 40 includes a wireless LAN (WLAN), a wide area network (WAN), ISDNs (integrated service digital networks), wireless LANs, LTE (long term evolution), LTE-Advanced, Fourth generation (4G), fifth generation (5G), CDMA (code division multiple access), WCDMA (registered trademark), Ethernet (registered trademark), and the like.

In addition, the network 40 is not limited to these examples, for example, public switched telephone network (PSTN), Bluetooth (registered trademark), Bluetooth Low Energy, optical line, ADSL (Asymmetric Digital Subscriber Line) line, LPWAN (Low-Power Wide-Area Network) line, satellite communication network, or any other network.

Also, the network 40 may be, for example, NB-IoT (Narrow Band IoT) or eMTC (enhanced Machine Type Communication). NB-IoT and eMTC are wireless communication schemes for IoT, and are networks capable of long-distance communication at low cost and low power consumption.

Network 40 may also be a combination of these. Network 40 may also include multiple different networks that combine these examples. For example, the network 40 may include a wireless network based on LTE and a wired network such as an intranet, which is a closed network.

Next, the information processing server 100 in the insurance content search system 1 will be described in detail using FIG. FIG. 2 is a block diagram for explaining the configuration of the information processing server 100. As shown in FIG.
As shown in FIG. 2 , the information processing server 100 includes a reception section 110 , a communication section 120 , a database generation section 130 , a search processing section 140 and a storage section 150 .

The accepting unit 110 accepts an input of the uttered voice of the speaker. Specifically, the reception unit 110 receives speech input to the microphone of the user terminal 10 .
The communication unit 120 communicates with the database of the user terminal 10 and the insurance company 30 via the network 40, for example. Under the control of the information processing server 100 , the communication unit 120 acquires, for example, information on the insurance details provided by each insurance company 30 from the database of the insurance company 30 . Also, the communication unit 120 transmits the search conditions from the user 20 input to the user terminal 10 to the search processing unit 140 of the information processing server 100 via the network 40 .

The storage unit 150 is a device that stores various control programs, a database described later, and image information relating to various insurance details provided by the insurance company 30 . An example of the information stored in the database is information in which a disease name is associated with insurance coverage information for the disease name.
The storage unit 150 is realized by various storage media such as HDD, SSD, and flash memory, for example.

Here, the data structure in the database will be described with reference to FIG. FIG. 3 is a diagram showing the data structure in the database. As shown in FIG. 3, data stored in the database is stored in a form in which information IDs, title words, and search target information are associated with each other. Information ID refers to the name of the insurance company in this embodiment. As the name of the insurance company, for example, an identifiable ID number or the like may be stored.

A title term refers to a term used as a keyword for searching. In this embodiment, the title phrase is the disease name. Search target information refers to information specified by search processing. In this embodiment, the search target information is compensation availability information. Compensation availability information is information indicating whether or not the user 20 is subject to compensation in each insurance plan. That is, for example, it is information indicating whether or not a user who has a history of suffering from a certain disease is eligible for each insurance plan, and whether or not the user can subscribe to the insurance plan.

As shown in FIG. 2, the database generator 130 generates a database. The database generation unit 130 includes a data reading unit 131 , a title phrase extraction unit 132 , a compensation availability information extraction unit 133 and a registration unit 134 .
The data reading unit 131 receives image information including at least a disease name and compensation availability information indicating whether insurance compensation is available for the disease name. The image information is information on materials related to insurance services provided by insurance companies.

The title phrase extraction unit 132 extracts the disease name indicated as the title from the image information. Here, the title is text information pointing to a search keyword, and in this embodiment, text information pointing to a disease name.
The title phrase extraction unit 132 can recognize the image information as text information by performing OCR processing on the image information of the insurance document. This point will be described in detail with reference to FIG. FIG. 4 is a diagram showing the contents of the underwriting criteria in the image information of the materials of the insurance company. As shown in FIG. 4, the underwriting standard is information that defines the conditions of the user 20 who can join the insurance plan.

In the image information shown in FIG. 4, the title phrase extraction unit 132 first identifies the positions of the ruled lines of the table. Then, the text information of the "disease name" described in the upper left is specified first. Then, the disease name written vertically or horizontally based on that position is extracted as a title phrase.

Here, it is necessary not to extract the disease name described in the "Notes" column shown in FIG. 4 as a title word. Therefore, when extracting words such as "notes" and "remarks", disease names written vertically or horizontally are not extracted as title words based on their positions. Here, the title phrase extracting unit 132 may perform the above processing based only on the positional relationship of the text information without specifying the ruled lines of the table.

The compensation availability information extraction unit 133 specifies a position where the compensation availability information is indicated with respect to the position where the disease name extracted by the title phrase extraction unit 132 is indicated in the image information, and extracts the compensation availability information.
The registration unit 134 associates the name of the insurance company, the disease name extracted by the title phrase extraction unit 132, and the compensation availability information extracted by the compensation availability information extraction unit 133, and registers them in the database. As a result, the database described above is generated.

The search processing unit 140 has a function of performing search processing on the database in response to search conditions input by the user 20 .
The search processing unit 140 includes a speech recognition unit 141 , an extraction unit 142 , a specification unit 143 and an output unit 144 .
The speech recognition unit 141 recognizes the speech received by the reception unit 110 and converts it into text information.

The extraction unit 142 extracts disease names from the text information converted by the speech recognition unit 141 . The extraction unit 142 extracts disease names by performing morphological analysis in natural language processing. Note that the extraction unit 142 may extract the disease name from text information directly input by the user 20 without using the speech recognition unit 141 .
The identifying unit 143 refers to the database based on the name of the disease and identifies corresponding compensation availability information. The output unit 144 outputs the compensation availability information specified by the specifying unit 143 .

In the insurance content search system 1 according to the present embodiment, the disease name extracted by the title phrase extraction unit 132 is displayed in the image information by the compensation availability information extraction unit 133 in the database generation unit 130 using a learning model. The position where the compensation availability information is indicated is specified with respect to the position.
The learning model learns the positional relationship between title words and phrases in image information and compensation availability information using a plurality of pieces of image information.

In FIG. 4, part A indicates the name of the disease, and part B indicates compensation availability information. In other words, in this figure, the places marked with a circle mean that you can subscribe to the insurance plan listed at the top of the column, and the places marked with a cross are listed at the top of the column. This means that you cannot enroll in the insurance plan that was issued.
Specifically, in this figure, for example, if you have had osteosarcoma, you cannot join a cancer-related special contract if it is within 2 years after treatment, but 2 years have passed since treatment. In that case, you will be able to subscribe to the cancer-related special contract according to the conditions described in asterisk 1.

As shown in FIG. 4, for the insurance content proposed by a plurality of insurance companies, each company organizes information on whether or not compensation is possible, and prepares materials as guidelines for underwriting.
In the image information of each insurance document, there is a certain tendency in the relationship between the position (A section) where the disease name is written and the position (B section) where the compensation availability information is written.

Here, FIG. 5 is a diagram showing an example of teacher data used by the compensation availability information extraction unit 133. As shown in FIG.
As shown in FIG. 5, the compensation availability information extraction unit 133 extracts the positional relationship between the disease name as the title word written in the C part and the compensation availability information corresponding to the disease name written in the D part. is training data.
That is, the teacher data is metadata indicating the position range of the disease name (coordinate range in the image) and the position range of the compensation availability information (image information to which metadata indicating the coordinate information in the .

A large amount of such image information is prepared, and the positional relationship between the title words indicating the disease name and the compensation availability information for the large amount of image information is machine-learned using a learning model constructed by, for example, a neural network. .
By using the learning model constructed in this way, if the position where the title word is written for the image information of the insurance document that is first seen, it is possible to grasp the position where the compensation availability information is stated. .

(Modification)
Here, a modified example of the compensation availability information extraction unit 133 will be described.
The compensation propriety information extraction unit 133 according to the modification identifies, as the position where the compensation propriety information is indicated, an area included in a preset interval with respect to the position where the title phrase in the image information is indicated, without using a learning model. . This point will be described in detail with reference to FIG. FIG. 6 is a diagram showing area designation by the compensation availability information extraction unit 133 according to the modification.

As shown in FIG. 6, in the image information, the relationship between the position where the title phrase is described and the position where the compensation availability information is described is generally standardized in the insurance data. be.
Specifically, for example, in FIG. 6, with respect to the position where the title phrase "osteosarcoma" is described, the range is included in the vertical direction with a predetermined interval E, and the predetermined interval is to the right of the title phrase. A range included in a predetermined region G from a position F away can be set as the position where the compensation availability information is described.
Therefore, if a predetermined interval is set in advance for each document of each insurance company, the position where the compensation availability information is described can be specified by extracting the title phrase.

Therefore, first, OCR processing is performed on the image information of the insurance document, the positional relationship in the image information is grasped, and words contained within a predetermined interval from the disease name are read as text information. , the compensation availability information can be extracted.

Next, processing of the insurance content search system 1 will be described with reference to FIGS. 7 to 9. FIG. FIG. 7 is a diagram showing the overall flow of the processing flow of the insurance content search system 1. As shown in FIG.
As shown in FIG. 7, as the processing of the insurance content search system 1, first, a database generation step (S10) is executed. In the database search step, the database generator 130 generates a database.
Next, search processing unit 140 executes a search step (S20). In the search step, insurance details suitable for the user 20 are searched based on the user's 20 speech. Each of these steps will now be described in detail.

FIG. 8 is a diagram showing a database generation step in the processing flow of the insurance content search system 1. As shown in FIG.
First, as shown in FIG. 8, in the database generation step, the data reading unit 131 receives image information including at least a disease name and compensation availability information indicating whether insurance coverage is available for the disease name ( Data reading step S11).
Next, the title word/phrase extraction unit 132 extracts the disease name indicated as the title from the image information (title word/phrase extraction step S12).

Next, the compensation availability information extraction unit 133 identifies the position where the compensation availability information is indicated with respect to the position where the disease name extracted by the title phrase extraction unit 132 is indicated in the image information, and extracts the compensation availability information ( Compensation availability information extraction step S13).
Next, the registration unit 134 associates the disease name extracted in the title phrase extraction step with the compensation availability information extracted in the compensation availability information extraction step and registers them in the database (registration step S14).

FIG. 9 is a diagram showing a search step in the processing flow of the insurance content search system 1. As shown in FIG.
Next, as shown in FIG. 9, in the search step, the receiving unit 110 receives an input of the uttered voice of the user 20 (receiving step S21). Instead of the user 20 speaking, the user 20 may input a disease name as a title word that is a search condition by typing on the keyboard.
Next, the voice recognition unit 141 recognizes the uttered voice received in the receiving step and converts it into text information (voice recognition step S22).
Next, the extraction unit 142 extracts disease names from the text information (extraction step S23).

Next, based on the disease name, the identification unit 143 refers to a database that associates the disease name with compensation availability information indicating whether insurance coverage is available for the disease name, and obtains the corresponding compensation availability information. Identify (identification step S24).
Finally, the output unit 144 outputs the compensation availability information specified in the specifying step (output step S25).

Next, operation screens of the insurance content search system 1 will be described with reference to FIGS. 10 to 13. FIG.
FIG. 10 is a diagram showing a search screen in the insurance content search system 1. As shown in FIG. FIG. 11 is a diagram showing an insurance company selection screen in the insurance content search system 1. As shown in FIG. FIG. 12 is a diagram showing a search result screen in the insurance content search system 1. As shown in FIG. FIG. 13 is a diagram showing the contents of search results by the insurance content search system 1. As shown in FIG.

First, as shown in FIG. 10, on the search screen, the user 20 utters a disease name as a search word after clicking the icon of the microphone button. As a result, the search target selection screen shown in FIG. 11 is displayed.
On the search target selection screen shown in FIG. 11, an insurance company to be searched can be selected from among a plurality of insurance companies. Selection of search targets may be performed before searching by speech.

As a result, as shown in FIG. 12, the corresponding insurance plans of each insurance company are searched. In this example, the disease name "depression" is entered as a search word. By selecting one of the insurance companies, the details of the search results can be confirmed.
Thereby, the search result is displayed on the display unit of the user terminal 10 as shown in FIG. Here, FIG. 13 shows search results by the insurance content search system 1. FIG.

As shown in FIG. 13A, when a user 20 who has suffered from depression searches for "depression" as a search word, compensation availability information for depression is displayed. In this example, three insurance plans of "cancer rider", "P side/three major", and "women's disease" were confirmed as insurance plans for the user 20 who has a history of depression. It will happen.

Next, as another example, in the example shown in FIG. 13B, the user 20 who has suffered from subarachnoid hemorrhage is looking for an insurance plan that is covered even if he has a history of suffering from subarachnoid hemorrhage. Assuming the case. Then, when the user 20 utters "subarachnoid hemorrhage" as a search condition, an acceptance standard for subarachnoid hemorrhage is retrieved and displayed. In this example, no insurance plan would cover users 20 with a history of subarachnoid hemorrhage.

As described above, according to the insurance content search system 1 according to the present embodiment, the voice recognition unit 141 converts the speech of the speaker received by the reception unit 110 into text information. Then, the extraction unit 142 extracts the disease name from the text information.
Then, the identifying unit 143 refers to the database and outputs insurance coverage availability information for the disease name. Therefore, for example, even if the user 20 has a history of suffering from a disease, it is possible to immediately search for insurance content that the user 20 can enroll in simply by uttering the name of the disease that the user has a history of suffering from. Therefore, it is possible to search for insurance services to which the user 20 can subscribe through a simple operation.

In addition, the compensation availability information extraction unit 133 uses the learning model to identify the position where the compensation availability information is shown with respect to the position where the disease name extracted by the title phrase extraction unit 132 is shown in the image information. Therefore, by preparing a large amount of image information, it is possible to automatically specify the position where the compensation availability information is indicated in the image information.

Further, in the insurance content search system 1 according to the modified example, the compensation availability information extraction unit 133 designates an area included in a preset interval with respect to the position where the title phrase in the image information is shown as the position where the compensation availability information is shown. Identify. Therefore, it is possible to specify the compensation availability information described in the insurance document with a simple configuration compared to a configuration that performs machine learning.

(Second embodiment)
Next, the insurance content search system 1 according to the second embodiment will be described. In this description, the description of the same configuration as that of the first embodiment is omitted.
In the insurance content search system 1 according to the present embodiment, the reception unit 110 does not receive the speech from the user 20 from the user terminal 10, but the content of the conversation of the user 20 who talks with the operator over the phone is sent to the reception unit 110. is entered.

Thereby, the receiving unit 110 receives the speech from the user. That is, the operator asks the user 20 a question, and while the receiving unit 110 receives the answer, the keyword that becomes the title phrase is extracted. The keyword in this case may be a disease name as in the first embodiment, or may be other text.
By adopting such a configuration, the search can be automatically performed in the conversation between the user 20 and the operator without the operator performing the search processing by himself/herself.

Note that the configuration described above is merely an embodiment, and various modifications, aggregation or separation of functions can be made without departing from the gist of the present invention.
For example, in the present embodiment, the insurance content search system 1 is a separate device from the user terminal 10, but the present invention is not limited to this aspect. A part of the functions of the insurance content search system 1 may be realized by the control unit of the user terminal 10 .

Further, in the above-described embodiment, the title phrase extraction unit 132 has shown a configuration for extracting a disease name as a title phrase, but it is not limited to such a mode. Items other than the disease name may be set as the title phrase. Specifically, instead of information on whether or not insurance coverage is available, you can arbitrarily change the title words and phrases so that you can search for content such as "question items at the time of application," "required documents at the time of application," and "handling regulations." can be done.

In addition, the insurance content search system 1 is not limited to insurance related to diseases such as life insurance, medical insurance, and cancer insurance, and can also be applied to non-life insurance.
Furthermore, the insurance content retrieval system 1 can be applied to business other than insurance, such as business manuals for general retailers such as department stores. In this case, for example, the name of the sales floor may be used as the title phrase, the store name of the department store, the name of the gift certificate, etc. may be used as the title phrase. , information related to the title phrase may be set.

Further, each part of the insurance content search system 1 described above may be implemented as a function of a computer processing unit or the like. That is, the reception unit 110, the communication unit 120, the database generation unit 130, and the search processing unit 140 in the information processing server 100 of the insurance content search system 1 are realized as a reception function, a communication function, a database generation function, and a search processing function, respectively. may be

In such a case, the database generation function has a data reading function, a title phrase extraction function, a compensation availability information extraction function, and a registration function. The search processing function has a speech recognition function, an extraction function, a specific function, and an output function.

In addition, the insurance content search program can cause a computer to implement each function described above. The insurance content search program may be recorded in a non-temporary computer-readable recording medium such as an external memory or an optical disc.

Further, as described above, each part of the insurance content search system 1 may be realized by a computer processing unit or the like. The arithmetic processing unit or the like is configured by an integrated circuit or the like, for example. Therefore, each part of the insurance content search system 1 may be implemented as a circuit that constitutes an arithmetic processing device or the like.

Also, the communication unit 120 and the storage unit 150 of the insurance content search system 1 may be configured by an integrated circuit or the like, for example. Also, the communication unit 120 and the storage unit 150 of the insurance content search system 1 may be configured as a communication device and a storage device by being configured by a plurality of devices, for example.

1 insurance content search system 10 user terminal 100 information processing terminal 110 reception unit 120 communication unit 130 database generation unit 131 data reading unit 132 title phrase extraction unit 133 compensation availability information extraction unit 134 registration unit 140 search processing unit 141 voice recognition unit 142 extraction Unit 143 Identification unit 144 Output unit 150 Storage unit

Claims (5)

a reception unit that receives an input of the speaker's uttered voice;
a speech recognition unit that recognizes the speech received by the reception unit and converts it into text information;
an extraction unit that extracts a disease name from the text information;
a database that associates a disease name with compensation availability information that indicates whether insurance coverage is available for that disease name;
a specifying unit that refers to the database based on the disease name and specifies corresponding compensation availability information;
an output unit for outputting compensation availability information specified by the specifying unit;
a data reading unit that receives image information including at least a disease name and the compensation availability information;
a title phrase extraction unit for extracting a disease name indicated as a title from the image information;
a compensation availability information extraction unit that specifies a position where the compensation availability information is indicated with respect to a position where the disease name extracted by the title phrase extraction unit is indicated in the image information, and extracts the compensation availability information;
An insurance content search system, comprising: a registration unit for registering in the database the disease name extracted by the title phrase extraction unit and the compensation availability information extracted by the compensation availability information extraction unit in association with each other. The compensation propriety information extraction unit uses a learning model that has learned the positional relationship between the title phrase and the compensation propriety information in a plurality of pieces of image information, so that the disease name extracted by the title phrase extraction unit is indicated in the image information. 2. The insurance details retrieval system according to claim 1, wherein the position where the compensation availability information is indicated is specified with respect to the position where the information is indicated. 2. The compensation propriety information extracting unit specifies, as the position where the compensation propriety information is indicated, an area included in a preset interval with respect to a position of the title phrase in the image information. The insurance content search system described in . the computer
a receiving step of receiving an input of speech voice of a speaker;
a voice recognition step of recognizing the uttered voice received by the receiving step and converting it into text information;
an extracting step of extracting a disease name from the text information;
a identifying step of referring to a database in which the name of the disease and information indicating whether the disease is covered by insurance are associated with each other, and identifying the corresponding information about whether the disease is covered by insurance;
an output step of outputting compensation availability information specified in the specifying step;
a data reading step of receiving image information including at least a disease name and the information on whether compensation is possible;
a title phrase extraction step of extracting a disease name indicated as a title from the image information;
a compensation propriety information extraction step of specifying a position where the compensation propriety information is indicated with respect to a position where the disease name extracted by the title phrase extraction step is indicated in the image information, and extracting the compensation propriety information;
and a registration step of correlating the disease name extracted by the title phrase extraction step and the compensation availability information extracted by the compensation availability information extraction step and registering them in the database. to the computer,
a reception function that receives input of the speaker's uttered voice;
a speech recognition function for recognizing the speech received by the reception function and converting it into text information;
an extraction function for extracting a disease name from the text information;
a specific function of referring to a database in which the name of the disease and information indicating whether or not the disease is covered by insurance are associated with each other, and identifying the corresponding information about whether or not the disease is covered;
an output function for outputting compensation availability information specified by the specifying function;
a data reading function for receiving image information including at least a disease name and the compensation availability information;
a title phrase extraction function for extracting a disease name indicated as a title from the image information;
a compensation availability information extracting function for specifying a position where the compensation availability information is indicated with respect to a position where the disease name extracted by the title phrase extraction function is indicated in the image information, and extracting the compensation availability information;
An insurance content search program for realizing a registration function for registering in the database the disease name extracted by the title phrase extraction function and the compensation availability information extracted by the compensation availability information extraction function in association with each other.

Images