KR20200123511A - Industry decision system using an artificial neural network - Google Patents

Abstract

The present invention relates to a business type determination system using an artificial neural network, which is capable of automatically determining a business type for calculating an insurance premium rate for a property damage insurance application target. The business type determination system includes: an information transceiving part for receiving business type search information, which is selected and inputted to determine a business type for calculating an insurance premium rate for a property damage insurance application target, from an external computer system; and a business type determination part including at least one artificial neural network trained to a business type, to which a property damage insurance application target belongs, through the deep-learning of a combination of target identification information for identifying each property damage insurance application target written on an insurance premium rate table. The business type determination part transmits and outputs, to the external computer system, the business type determined based on the at least one artificial neural network and the business type search information.

Description

Industry decision system using artificial neural network {INDUSTRY DECISION SYSTEM USING AN ARTIFICIAL NEURAL NETWORK}

The present invention relates to a water insurance contract management system, and in particular, to a business type determination system using an artificial neural network for automatically determining the type of business for calculating the insurance premium rate of a water insurance subscription object.

In the case of long-term insurance, there are human insurance and water insurance, and insurance products that combine water and phosphorus, depending on the object that can be selected. Water insurance is based on the type of building (reinforced steel, concrete, thatched house, etc.) and the building industry (information on business types handled by buildings such as offices and karaoke). And the premium is calculated.

Since there are so many types and numbers of water insurance targets, specialized knowledge and a lot of experience are required in determining the type of business for applying insurance premiums. However, in the field where water insurance is sold and contracted, there are frequent cases where a lot of labor and time are wasted in determining the type of business for water insurance due to lack or absence of regular training on water insurance.

For example, when consulting a water insurance contract, on-site insurance planners enter the name of the business to be searched, such as marts, supermarkets, and restaurants, in their information terminals to search for business types for applying insurance premiums, including electrical and electronic assembly, winding, and insulation tests. It is not easy to enter the name of the business to be searched for for targets that require specialized knowledge, such as industry or chemical or semiconductor related industries.

In this way, if the business type of the factory product is not accurately determined and the type of business is determined at random from the knowledge level of the insurance planner, unnecessary time and loss of work occur because the business type for the target product must be re-determined during the review process. Of course, there is a problem in which accurate industry guidance and prompt response to customers cannot be provided.

In addition, despite the rapid spread of new products and industries developed to meet the needs of consumers, insurance premiums are not updated in a timely manner for these newly developed products or industries. The reviewers in charge of the review work also have difficulties in deciding the type of business for calculating insurance premiums.

Source, http://blog.naver.com/kkts2000?Redirect=Log&logNo=220591001582 (Insurance first step: New contract subscription design work process)

Accordingly, the present invention is an invention invented to solve the above-described problems, and an industry determination system using an artificial neural network that can quickly and accurately determine and provide the business type to which the object of water insurance subscription belongs by only inputting image information. In providing,

Furthermore, another object of the present invention is to provide an artificial neural network that can quickly and accurately search and provide the business type to which the water insurance subscription object belongs by only inputting one or more of video information, audio information, or text information on the object of water insurance subscription. It is to provide the used industry determination system.

In addition, another object of the present invention is to provide a business type determination system using an artificial neural network capable of retraining an artificial neural network model so that a business type for applying insurance premium rates can be automatically determined for newly developed or newly distributed goods or business types. Is in.

The industry determination system using an artificial neural network according to an embodiment of the present invention for solving the above-described technical problem is an operating system that can be built on a computer system (server) operated by an insurance company or executed on a computer system,

An information transmission/reception unit for receiving, from an external computer system, business type search information selected and input to determine a business type for calculating an insurance premium rate for a water insurance subscription object;

Industry determination unit including at least one artificial neural network model trained to determine the industry to which the water insurance subscription object belongs by deep learning a combination of object identification information that can identify each of the water insurance subscription targets listed in the insurance rate table; Including,

The industry type determination unit may transmit and output the business type determined based on the business type search information and the at least one artificial neural network model to the external computer system.

Furthermore, in the industry type determination system using an artificial neural network including the above-described configuration, the at least one artificial neural network model,

It includes a video learning artificial neural network model by deep learning one or more video images that can identify each of the water insurance subscription targets listed in the insurance rate table,

The industry search information is characterized in that it includes one or more images of the object of water insurance subscription,

As another deformable embodiment, the at least one artificial neural network model is at least,

It may include a keyword learning artificial neural network model by deep learning a series of keyword combinations capable of identifying each of the water insurance subscription targets described in the insurance rate table. In this case, the industry search information is characterized in that it is a combination of keywords representing each of raw materials, processes, and products of water insurance subscription targets.

Further, as another embodiment that can be modified, the at least one artificial neural network model,

An image learning artificial neural network model obtained by deep learning one or more image images capable of identifying each object of water insurance subscription listed in the insurance rate table;

Including a keyword learning artificial neural network model by deep learning a series of keyword combinations that can identify each of the water insurance subscription targets listed in the insurance rate table,

The business type search information is characterized in that it includes one or more images of the object subject to water insurance, or a combination of keywords representing each of raw materials, processes, and products of the object subject to water insurance.

In addition, the industry determination system using the artificial neural network described above,

A neural network retraining unit that retrains the at least one artificial neural network model by setting a combination of object identification information that can identify a new water insurance subscription object that is not listed in the insurance rate table and the industry to which the new water insurance subscription object belongs. Another feature is that it further includes;

Furthermore, in the industry type determination system of the above-described configuration, the industry type determination unit,

Industry search for any one of the main material, material, process, molding method, and production machine of the object subject to water insurance according to the image reading result of the artificial neural network model trained to input the image as the industry search information to determine the industry type Another feature is that additional input of information is transmitted to an external computer system and the response is re-inputted into the artificial new network model to more accurately determine the type of business.

According to the above-described problem solving means, the system for determining a business type using an artificial neural network of the present invention provides a water insurance subscription object by simply receiving one or more images related to a water insurance subscription object or industry search information including a series of keyword combinations. The advantage of improving the overall efficiency of water insurance subscription work as it minimizes the time and human costs invested or required to determine the type of industry to which the insured object belongs, as it quickly and accurately determines and provides the business type to which the insurance belongs. There is this.

In addition, the present invention has the effect of improving the accurate business type guidance of insurance planners, prompt response, and convenience of consultation, and since the artificial neural network model can be retrained from time to time, even if the insurance subscription object not described in the insurance rate table is developed and distributed. Even if it is, there is an advantage of being able to conveniently receive the industry in which the subject of insurance subscription belongs online.

1 is an exemplary diagram of a network configuration in which an industry type determination system using an artificial neural network according to an embodiment of the present invention can be installed or constructed.
2 is an exemplary configuration diagram of a system for determining an industry type using an artificial neural network according to an embodiment of the present invention.
3 is a diagram illustrating a process of forming an artificial neural network model based on deep learning.
4 to 6 are exemplary screens for inputting industry search information according to an embodiment of the present invention.
7 is a view for explaining the operation of the industry type determination system using an artificial neural network according to an embodiment of the present invention.
8 is an exemplary view of a display screen of industry search information and a determined industry type according to an embodiment of the present invention.

Specific structural or functional descriptions of the embodiments according to the concept of the present invention disclosed in the present specification are merely illustrative for the purpose of describing the embodiments according to the concept of the present invention, and the embodiments according to the concept of the present invention are It may be implemented in various forms and is not limited to the embodiments described herein.

In addition, since the embodiments according to the concept of the present invention can apply various changes and have various forms, the embodiments will be illustrated in the drawings and described in detail herein. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes all changes, equivalents, or substitutes included in the spirit and scope of the present invention. In addition, in describing an embodiment of the present invention, when it is determined that a detailed description such as a related known function or configuration may unnecessarily obscure the subject matter of the present invention, a detailed description thereof will be omitted.

In addition, the industry determination system using an artificial neural network according to an embodiment of the present invention may be a set of codes constituting a single application program or a control program of a computer system, and a predetermined code and a hardware resource for executing the predetermined code It may be implemented in the form of a system composed of a number of modules that mean a logical unit.

1 is an illustration of a network configuration diagram in which an industry type determination system using an artificial neural network according to an embodiment of the present invention can be installed or constructed. The industry type determination system using an artificial neural network according to an embodiment of the present invention is an insurance company server 200 Can be installed or built in. Accordingly, in the following, the same reference numeral 200 will be used for each of the insurance company server and the industry determination system using an artificial neural network.

As shown in Fig. 1, the insurance company server 200 on which the business type determination system using an artificial neural network is installed or constructed is based on the business type search information transmitted from the information terminals 100 of the insurance planners accessible through a communication network. The business type information to which the subscription target belongs is determined using an artificial neural network, and is provided to the insurance planner's information terminal 100.

The insurance agent's information terminal 100 inputs business type search information according to the insurance agent's operation and transmits it to the insurance company server 200. For example, an application program for supporting various services related to water insurance subscription by accessing the insurance company server 200 is installed and executed in the information terminal 100 of the insurance agent. This will be referred to as a water insurance consulting system below.

Meanwhile, the'Industry Search Information', which is selected and input by the insurance planner through the above water insurance consulting system, to determine the type of business for calculating the insurance premium rate for the object of water insurance subscription is the object of water insurance subscription according to the type of information learned by the artificial neural network. It may be one or more images for identifying the product, and may be a combination of keywords representing each of raw materials, processes, and products for identifying the object of water insurance subscription.

In addition, depending on the type and number of artificial neural networks provided in the industry type determination system using the artificial neural network, not only the image but also a series of keyword combinations and voice information may be used as industry search information.

The manager terminal 300 not described in FIG. 1 may be used to manage an insurance company server, in particular, an industry type determination system using an artificial neural network. For example, when new water insurance subscription targets are developed and distributed, the industry determination service for the target may be made only after re-learning about them, so that the manager can use his terminal 300 It manages to relearn the subject of insurance.

2 is a diagram illustrating a configuration diagram of an industry type determination system 200 using an artificial neural network according to an embodiment of the present invention, and FIG. 3 is a diagram illustrating an example of an artificial neural network model based on deep learning. will be.

As shown in Figure 2, the industry type determination system 200 using an artificial neural network according to an embodiment of the present invention,

Information for receiving selected and input business search information from an external computer system, for example, information terminal 100 of an insurance planner, to determine the type of business for calculating the insurance premium rate for water insurance subscription objects (goods, factories, buildings, etc.) Transmitting and receiving unit 210,

By deep learning the combination of object identification information (images for the object, keywords representing raw materials/processes/products, voice information, etc.) that can identify each of the water insurance subscription targets listed in the insurance rate table, It includes an industry type determination unit 220 including one or more artificial neural network models (222, 224) trained to determine the type of industry to which it belongs.

The industry type determination unit 220 transmits the industry search information transmitted from the insurance agent information terminal 100 to one or more artificial neural network models 222 or 224, and transfers the business type determined in the artificial neural network model 222 or 224 to the insurance agent. It transmits and outputs to the information terminal 100.

The artificial neural network model included in the industry type determination unit 220 may be an image learning artificial neural network model 222 in which one or more image images capable of identifying each of the water insurance subscription targets described in the insurance rate table are deep-learned. In this case, the industry search information transmitted (input) to the image learning artificial neural network model 222 may be one or more images of the object for water insurance subscription or a keyword that is re-entered after reading the image.

In addition, in order to determine the type of business for calculating the insurance premium rate for the subject of water insurance, if it is a system that can input a combination of keywords input for each search category provided by the information terminal 100 of the insurance agent as business type search information, the industry type determination unit The artificial neural network model included in 220 may be a keyword learning artificial neural network model 224 obtained by deep learning a series of keyword combinations capable of identifying each of the water insurance subscription targets described in the insurance rate table. The search category is a raw material search category, a process search category, and a product search category for distinguishing and identifying targets for water insurance subscription. The insurance planner selects keywords listed in each search category presented, so that a combination of keywords can be generated as industry search information.

Of course, the industry type determination unit 220 may include both the image learning artificial neural network model 222 and the keyword learning artificial neural network model 224, and the water insurance subscription target is based on the industry search information input in the voice form. It may further include a voice learning artificial neural network model that determines the type of business to which it belongs. In this case, the business type search information may be one or more images of the object subject to water insurance, a combination of keywords representing each of the raw materials, processes, and products of the object subject to water insurance, or may be audio information, and includes at least two of the above three types of information. It may be business search information.

As another deformable implementation method, the industry type determination unit 220 inputs an image as the business type search information and the main material of the water insurance subscription object according to the image reading result of the artificial neural network model 222 trained to determine the business type. , Material, process, molding method, sending the additional input of the business search information for any one of the production machine of the object to the insurance agent's information terminal 100, and re-enter the response into the artificial new reputation network model 222 You can also make a final decision. This is because it may not be possible to determine the type of business with only one or more images due to the nature of the object of water insurance subscription, so the re-input of the keyword is requested to further specify the object of water insurance subscription.

Meanwhile, in addition to the above-described configurations, the industry determination system 200 using an artificial neural network according to another embodiment of the present invention includes a combination of object identification information capable of identifying a new water insurance subscription object not described in the insurance rate table. It may further include a neural network retraining unit 240 for retraining the one or more artificial neural network models 222 and 224 by setting a business type to which the new water insurance subscription target belongs.

The industry type determination system according to the embodiment of the present invention by the neural network relearning unit 240 newly updates the artificial neural network model for the newly developed and distributed target or industry, which is not already described in the insurance rate table used in the industry. For the target object, it can automatically determine the type of business and provide it to insurance planners or insurance contract reviewers.

For reference, the storage 230, which is not described in FIG. 2, includes data (object identification information, industry search table, a series of keyword combinations, etc.) necessary for training the artificial neural network models 222 and 224, as well as uploaded industry search information, Information related to the insurance contract is stored, and these information are accessed by the industry determination system 200 using artificial neural networks as well as other functional modules constituting the insurance company server.

Hereinafter, a process of forming an artificial neural network model used in the industry type determination system 200 using an artificial neural network according to an embodiment of the present invention will be further described with reference to FIG. 3.

First, FIG. 3 is a diagram illustrating a process of forming an artificial neural network model based on deep learning, and FIGS. 4 to 6 illustrate a screen for inputting industry search information according to an embodiment of the present invention.

In general, an artificial neural network consists of three layers: an input layer that receives an input, one or more hidden layers that actually learn, and an output layer that returns the result of an operation, as shown in FIG. 3. Is composed.

In the artificial neural network of this structure, each node represents a neuron, and when data is input, each node calculates a weighted sum and passes the calculation result to the next node through an activation function. At this time, the weight has a value for each node through learning. As an input of each node (neuron), the result values of the preceding node (neuron) are weighted and input, and the sum is calculated by adding a bias to it.

The general deep learning learning process is a forward propagation process that calculates a deep learning parameter and an objective function from the input layer to the output layer through a number of hidden layers, and the error is reflected from the output layer. It is performed by repeating the back propagation process of modifying the weights from the layer to the input layer. The weights modified in this deep learning learning process are repeatedly updated until errors are minimized, and all computers share weights and feature values. As a result, in a deep learning learning algorithm, an artificial neural network model with optimized parameters is generated after a learning process using the training data is performed, so an artificial neural network model capable of accurate prediction can be formed as the number of training data increases.

Accordingly, the industry type determination system 220 according to the embodiment of the present invention includes information that can identify the object of water insurance subscription (this information may be referred to as object identification information when viewed in the learning process, and insurance planners, etc. In terms of input, it can be referred to as industry search information), and an artificial neural network is used to accurately predict the type of industry to which the water insurance policy belongs.

In other words, since one or more images can be used as business type search information to distinguish and recognize the targets of water insurance, first of all, by deep learning a combination of target identification information that can identify each of the water insurance targets listed in the insurance rate table. By training to determine one type of industry to which the object of water insurance subscription belongs, one image learning artificial neural network model 222 can be formed, and the present invention uses the image learning artificial neural network model 222 formed in this way. You can decide the type of industry.

The combination of the object identification information may be defined as a set of one or more images or one or more images and keywords (or keywords) necessary to distinguish and recognize a water insurance subscription object.

Furthermore, since the keyword to identify each of the raw materials, processes, and products of the water insurance subscription object can be used in the industry search information sheet to recognize the object of water insurance subscription, it is possible to identify each water insurance subscription object listed in the insurance rate table. One keyword learning artificial neural network model 224 can be formed by deep learning learning a series of possible keyword combinations (this also corresponds to the object identification information) to determine the type of industry to which the water insurance subscription target belongs. The invention can determine the business type of the water insurance subscription object using the keyword learning artificial neural network model 224 thus formed.

For reference, the insurance planner may select and input keywords for each raw material search category, process search category, and product search category through the industry search information input screen displayed on their information terminal 100 as shown in FIGS. 4 to 6. .

Hereinafter, a process of determining the type of industry to which a water insurance subscription target belongs using one or more artificial neural network models 222 and 224 formed through deep learning learning will be described with reference to FIG. 7.

First, FIG. 7 shows a diagram for explaining the operation of the industry type determination system 200 using an artificial neural network according to an embodiment of the present invention, and FIG. 8 is a diagram of industry type search information and a determined industry type according to an embodiment of the present invention. The display screen is illustrated.

As shown in FIG. 7, the industry determination system 200 using an artificial neural network according to an embodiment of the present invention is a combination of object identification information capable of identifying each of the water insurance subscription objects described in the insurance rate table as described above. It is assumed that it includes one or more artificial neural network models 222 and 224 formed (step S120) by training to determine the type of industry to which the water insurance subscription target belongs by deep learning learning (step S110) of (combination of image/keyword).

Under this assumption, an insurance agent who consults on a water insurance contract executes a water insurance consulting system installed in his or her information terminal 100 to access information necessary for a water insurance contract and explain it to a customer.

In order to determine the type of business for the water insurance subscription object presented by the customer during consultation, the insurance planner takes pictures of the water insurance subscription object through'photographing' provided by the water insurance consulting system according to the instructions of the water insurance consulting system. Take one or more shots. An example of such a picture taking screen is shown in FIG. 8.

For reference, photography can be performed in various ways, such as the exterior, side, bottom, and...

In this way, when one or more images are inputted as business type search information by taking a picture of the water insurance subscription object (step S100), the insurance designer transmits it to the business type determination system 200 using an artificial neural network (step S130). Accordingly, the information transmission/reception unit 210 of the business type determination system 200 using an artificial neural network transmits the business type search information to the business type determination unit 220.

Accordingly, the image learning artificial neural network model 222 of the business type determination unit 220 recognizes the transmitted business type search information, determines the type of business to which the recognized water insurance subscription object belongs (step S140), and transmits the result to the information transmission/reception unit 210 Through the transmission to the insurance agent's information terminal 100 (step S150). The type of business transmitted to the information terminal 100 of the insurance agent is displayed on the information terminal 100.

Accordingly, the insurance planner can guide the business type information displayed on the information terminal 100 to the consulting customer, and accurately describe the business type for the water insurance subscription purpose in other documents provided on the insurance contract provided to the insurance company server side. will be.

For reference, in the plastics industry, there are industries that can be specified only with finished products, and there are cases where it is impossible. Looking at plastic products, humans can also be judged to be one of the industries called non-foaming molding (other use work) or non-foaming molding (working using A/B-class dangerous goods). In order to classify whether it is a used work, the raw material is photographed again, and the shape of the raw material is recognized, and it can be divided into non-foaming, other used work, and A/B-class dangerous goods used work.

Therefore, the artificial neural network model 222 of the industry type determination system 200 according to an embodiment of the present invention is trained to receive one or more images as business type search information and determine the type of business to which the object of water insurance subscription belongs.

On the other hand, taking the textile industry as an example, it is difficult for even a person to distinguish through photographs based on the state of the finished fabric. Accordingly, when the fabric is recognized through the picture, the image learning artificial neural network model 222 is trained to narrow it down to industries such as woven fabric manufacturing, knitted fabric (knit, meriyasu) manufacturing, fabric cutting and sewing, dyeing, bleaching, and processing, and the above 4 Training is given to request additional photographing (ie, additional input of industry search information) for the production machine of the object to determine one of the branches.

According to the training result, if the insurance agent additionally takes a picture of the fabric manufacturing machine or shoots and transmits it at one time, the image learning artificial neural network model 222 is used to manufacture a woven fabric if the recognized image is a weaving machine. If it is a loom, it is necessary to add one or more images, one or more images, and industry search information by training to determine the type of business by manufacturing, dyeing, bleaching and processing if it is a cutting machine, and fabric cutting and sewing if it is a cutting machine. Through input, etc., it accurately determines and presents the business type to which the object of water insurance subscription belongs.

In addition, the image learning artificial neural network model 222 according to the embodiment of the present invention may be trained to recognize that the group of industries is'metal industry' through the first image of the object, and in this state, the'metal By training to recognize industries such as'processing, assembly and design work, metal plate printing work, painting work', the optimal business type can be determined and presented based on industry search information including one or more images.

Therefore, the industry type determination unit 220 of the present invention is the main material, material, raw material, and process of the water insurance subscription object according to the image reading result of the artificial neural network model 222 trained to input an image as industry search information to determine the industry type. It is possible to finally determine the type of business by transmitting the additional input of the business type search information for any one of the method of manufacturing, the molding method, or the production machine of the object (product) to an external computer system, and re-entering the response into the artificial new reputation network model.

In the above embodiment, the image has been described as business type search information, but a series of keyword combinations may be used as business type search information to determine the type of business for which the water insurance is to be subscribed. For this, the keyword learning artificial neural network model 224 must be included in the industry type determination unit 220.

For example, if the keyword in the raw material search category is "aluminum ingot", the keyword in the process search category is "forging/stretching/cooling/cutting", and the keyword in the product search category is "bearing/steel plate", If the keyword learning artificial neural network model 224 trained as "metal processing" is included in the industry type determination unit 224,

The insurance agent enters "aluminum ingot" as the keyword of the raw material search category, "forging/drawing/cooling/cutting" as the keyword of the process search category, and "bearing/steel plate" as the keyword of the product search category to provide industry search information. When configured, the keyword learning artificial neural network model 224 transmits the industry type information "metal processing" to the information terminal 100 of the insurance agent. Even when a keyword is used as business type search information, the type of business may be determined according to the input of one keyword, the type of business may be determined according to the second keyword input according to the request to input additional business type search information, or the type of business may be determined according to the third keyword input. It can be said that this can also be implemented by learning an artificial neural network so that the industry can be determined.

According to the embodiment of the present invention described above, the insurance planner can quickly search the business type to which the object of water insurance subscription belongs by simply inputting the business type search information including one or more images or a series of keyword combinations through the water insurance consulting system. Since it can be provided accurately and provided to counseling or insurance company servers, it not only improves the overall efficiency of water insurance work for industry determination, but also improves accurate business type guidance, prompt response, and convenience for consultation by insurance planners. There is.

On the other hand, step S170, which is not described in FIG. 7, shows a re-learning step for the artificial neural network models 222 and 224. Since new types of targets or types of industries can be developed and distributed over time, it is necessary to relearn the artificial neural network models 22 and 224 so that the industry type can be determined quickly.

Therefore, if there is a new water insurance target that is not listed in the insurance premium rate table, the manager sets the combination of target identification information that can identify the target and the business type to which the new water insurance target object belongs, and sets the artificial neural network model (222,224). ), the business type to which the new water insurance subscription target belongs in steps S100, S130, and S140 may be determined by the business type determination unit 220 and transmitted to the information terminal 100 of the insurance planner.

The above has been described with reference to the embodiments shown in the drawings, but these are merely exemplary, and those of ordinary skill in the art will understand that various modifications and other equivalent embodiments are possible therefrom. Therefore, the true technical scope of the present invention should be determined only by the appended claims.

Claims (6)

An information transmission/reception unit for receiving, from an external computer system, business type search information selected and input to determine a business type for calculating an insurance premium rate for a water insurance subscription object;
Industry determination unit including at least one artificial neural network model trained to determine the industry to which the water insurance subscription object belongs by deep learning a combination of object identification information that can identify each of the water insurance subscription targets listed in the insurance rate table; Including,
The industry type determination unit transmits and outputs the industry type determined based on the industry type search information and the at least one artificial neural network model to the external computer system. The method according to claim 1, wherein the one or more artificial neural network models at least,
It includes a video learning artificial neural network model by deep learning one or more video images that can identify each of the water insurance subscription targets listed in the insurance rate table,
The industry search information is a business type determination system using an artificial neural network, characterized in that it includes one or more images of the object of water insurance subscription. The method according to claim 1, wherein the one or more artificial neural network models at least,
Including a keyword learning artificial neural network model by deep learning a series of keyword combinations that can identify each of the water insurance subscription targets listed in the insurance rate table,
The industry search information is a business type determination system using an artificial neural network, characterized in that the combination of keywords representing raw materials, processes, and products of a water insurance subscription object. The method of claim 1, wherein the at least one artificial neural network model,
An image learning artificial neural network model obtained by deep learning one or more image images capable of identifying each object of water insurance subscription listed in the insurance rate table;
Including a keyword learning artificial neural network model by deep learning a series of keyword combinations that can identify each of the water insurance subscription targets listed in the insurance rate table,
The industry search information includes one or more images of the object of water insurance subscription or a combination of keywords representing each of raw materials, processes, and products of the water insurance subscription object. The method according to any one of claims 1 to 4, wherein a combination of object identification information that can identify a new water insurance subscription object not listed in the insurance rate table and a business type to which the new water insurance subscription object belongs is set. A neural network retraining unit for retraining the one or more artificial neural network models. A system for determining a business type using an artificial neural network, further comprising. The method of claim 1 or 2, wherein the industry type determination unit,
Industry search for any one of the main material, material, process, molding method, and production machine of the object subject to water insurance according to the image reading result of the artificial neural network model trained to input the image as the industry search information to determine the industry type An industry type determination system using an artificial neural network, characterized in that an additional input of information is transmitted to the external computer system and the response is re-inputted into an artificial new reputation network model to finally determine the type of business.

Images