Method, device, and computer-readable recording medium for providing a chatbot service that outputs customized trading information

Abstract

To provide a method for providing a chatbot service that outputs customized fishing information. [Solution] When a question about fishing is received from a user account registered as a member of a fishing information providing platform, a service provision process for providing customized fishing information for the question is initiated based on confirmed user information, the question about fishing is analyzed using an artificial intelligence algorithm, and a response message for responding to the question about fishing is generated by reflecting detailed information included in the user information in a vector weight value. When an image corresponding to the question about fishing and the response message is identified, customized fishing information for the question about fishing is generated based on the identified image and the response message, and the customized fishing information is output by a chatbot system.

Description

[Technical Field]

[0001] The present invention relates to a method for providing a chatbot service that outputs customized fishing information. Specifically, the present invention relates to a technology for, when a question about fishing is input from a user account, checking the user information of the account, and based on this, initiating a service provision process for providing customized fishing information, analyzing the question using an artificial intelligence algorithm, generating a response message that reflects the user information, identifying an image corresponding to the question about fishing and the response message, generating customized fishing information based on the image and response message, and providing it to the user account via a chatbot system. [Background technology]

[0002] The global chatbot market is estimated to be worth approximately $7.01 billion in 2024 and is expected to grow at an average annual rate of 24.32% to reach $20.81 billion by 2029. This growth in the chatbot market is driven by increased demand for messaging apps and changes in how companies analyze their customers. In particular, the emergence of ChatGPT (registered trademark), a generative AI technology, is driving the growth of the chatbot market by demonstrating superior performance over existing rule-based chatbots. In response to this trend, companies have been reluctant to enter the generative AI field due to concerns about AI issues such as information security leaks and information distortion. However, as market and consumer demand has surged and service development competition has intensified, ensuring the reliability and ethics of chatbots has become increasingly important.

[0003] As a result, companies are developing technologies to diversify chatbot learning data and utilize it in various industrial sectors.

[0004] As an example, Korean Patent No. 10-2653266 (AI-based chatbot dialogue consultation system and method) discloses a technology for collecting knowledge of a target domain and fine-tuning an AI algorithm.

[0005] However, the prior art only discloses a technology in which knowledge data is simply collected, divided into those for embedding and those for fine tuning, and stored in a database; a training unit generates a custom artificial intelligence model; and a chatbot equipped with the artificial intelligence model outputs a response to a question based on the data stored in the database. However, it does not disclose a technology in which, when a question about fishing is input from a user account, the user information of the account is confirmed, and a service provision process for providing customized fishing information is initiated based on this information; the question is analyzed using an artificial intelligence algorithm, a response message is generated reflecting the user information; an image corresponding to the question about fishing and the response message is identified; customized fishing information is generated based on the image and response message; and the chatbot system provides the customized fishing information to the user account. Therefore, there is a need for a technology that can solve this problem. Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention has been made to solve the problems of the prior art described above, and its purpose is to provide accurate information that meets the needs of the user through the chatbot, and to provide the user with visual information that is easy to understand by providing accurate information that meets the user's needs through the chatbot, when a question about fishing is entered from a user account, the user information of the account is confirmed, and a service provision process for providing customized fishing information based on the user information is initiated based on the user information, the question is analyzed using an artificial intelligence algorithm, a response sentence is generated reflecting the user information, an image corresponding to the question about fishing and the response sentence is identified, customized fishing information is generated based on the image and response sentence, and the chatbot system provides the information to the user account. [Means for solving the problem]

[0007] In accordance with an embodiment of the present invention, there is provided a method for providing a chatbot service for outputting customized change information, which is implemented by a computing device including one or more processors and one or more memories for storing commands executable by the processors, and includes a process initiation step of receiving a question about change from a user account registered as a member of a change information providing platform, checking user information registered in the user account, and initiating a service provision process for providing customized change information for the question about change based on the confirmed user information; and, when the service provision process is initiated, providing the customized change information for the question about change using a stored artificial intelligence algorithm. The method includes a response sentence generating step of analyzing the question and reflecting detailed information included in the user information in a vector weight value to generate a response sentence for responding to the question about fishing; and a customized fishing information providing step of, when an image corresponding to the question about fishing and the response sentence is identified by the stored artificial intelligence algorithm as the generation of the response sentence is completed, generating customized fishing information for the question about fishing based on the identified image and the response sentence, and outputting the customized fishing information through a chatbot system of the fishing information providing platform and providing it to the user account.

[0008] The process initiation step includes a detailed information confirmation step of confirming detailed information included in the user information registered in the user account, such as environmental weather information, aquatic product growth information, marine environment information, aquatic product distribution information, and aquatic product journal information, when a question about fishing is received from the user account; When the detailed information confirmation step is completed, the method further includes an analysis start step of analyzing the question regarding the fishing based on the confirmed detailed information and starting a service provision process for generating and providing customized fishing information.

[0009] The response sentence generation step includes a first tokenization execution step of, when the service provision process is started, tokenizing the question about change using a first model of the stored artificial intelligence algorithm to tokenize a first sentence corresponding to the question about change; and a vector value-based category identification step of, when tokenization of the first sentence is complete, performing a vectorization process using the first model to vectorize a first token of the first sentence, calculating a vector value for the first token quantified based on the position of the first token in the first sentence, and determining which of a plurality of predetermined categories the first sentence belongs to based on the calculated vector value.

[0010] the response sentence generation step includes a second tokenization execution step of, when the service provision process is started, performing a tokenization process on the detailed information included in the user information using a first model of the stored artificial intelligence algorithm, thereby tokenizing each of the second sentences corresponding to the detailed information included in the user information; and a weighted value calculation / classification step of calculating, after completion of tokenization of each of the second sentences, vectorizing the second tokens of each of the second sentences using the first model to calculate vector values ​​for the second tokens quantified based on the positions of the second tokens in each of the second sentences, thereby calculating vector values ​​for each of the second sentences, classifying the vector values ​​for each of the second sentences into weighted values ​​applied to the vector values ​​of the first sentences, and classifying the vector values ​​into a predetermined number of categories.

[0011] The predetermined plurality of categories are categories in which representative vector values ​​are matched for each of the plurality of categories and which include a plurality of reference sentences that are candidate groups for responding to the question about fishing, and the plurality of reference sentences included for each of the plurality of categories are updated by the administrator of the fishing information providing platform along with the representative vector values ​​matched for each of the plurality of categories, forming reference category information.

[0012] The response sentence generation step further includes a response sentence derivation step, which includes a vector value analysis step of starting an analysis of the vector value of the first sentence, the vector value of the second sentence, and the vector value of a reference sentence included in a category to which the first sentence belongs, by a first model of the stored artificial intelligence algorithm when calculation of the vector value of the first sentence and the vector value of the second sentence is completed, and a vector value analysis step of starting an analysis of the vector value of the first sentence, the vector value of the second sentence, and the vector value of a reference sentence included in a category to which the first sentence belongs, by a first model of the stored artificial intelligence algorithm when execution of the function of the vector value analysis step is completed, and the vector value of the first sentence, reflecting the vector value of the second sentence classified by the weighted value on the vector value of the first sentence, thereby identifying a reference sentence having a reference vector value that is highly similar to the vector value of the first sentence reflected by the weighted value; and a sentence derivation completion step, after completing the identification of the reference sentence, checking a portion of the identified reference sentence that is identified as the basis for the question about change, and then proceeding with a summarization process to summarize the reference sentence using the first model, documenting the summarized reference sentence, and completing the derivation of a response sentence to answer the question about change.

[0013] The customized fishing information providing step includes an image information identifying step of, when the generation of the response sentence is completed, analyzing the question about fishing and the response sentence using a second model of the stored artificial intelligence algorithm and identifying at least one image from a plurality of image information stored in an image database that corresponds to the result of analyzing the question about fishing and the response sentence; and a chatbot-based information providing step of, when the identification of the image is completed, generating customized fishing information including the identified image and the response sentence and providing the customized fishing information to a user account using a chatbot system linked to the fishing information providing platform.

[0014] The stored artificial intelligence algorithm includes a first model, which is a large-scale language model that learns a first pattern value derived by natural language processing of reference sentences included in a predetermined plurality of categories, questions about different changings included in a predetermined plurality of categories, other user information registered in other user accounts that have provided questions about other changings, and other response sentences that respond to questions about other changings that reflect other user information, and by analyzing correlations; and a second model, which is an image search model that learns a second pattern value derived by analyzing correlations between reference images included in a predetermined plurality of categories, questions about different changings included in a predetermined plurality of categories, other response sentences that respond to questions about other changings that reflect other user information, other questions about changing, and other images corresponding to the other response sentences.

[0015] According to another aspect of the present invention, there is provided a method for providing a chatbot service for outputting customized fishing information, which is implemented by a computing device including one or more processors and one or more memories for storing commands executable by the processors, and includes a process initiation step of, when a question about fishing is received from a user account registered as a member of a fishing information providing platform, checking user information registered in the user account and initiating a service provision process for providing customized fishing information for the question about fishing based on the checked user information; and, when the service provision process is initiated, analyzing the question about fishing using a stored artificial intelligence algorithm, and reflecting detailed information included in the user information in a vector weight value to provide the customized fishing information. a response sentence generating step of generating a response sentence to respond to a question about fishing; an image normalization proceeding step of identifying an image corresponding to the question about fishing and the response sentence by the stored artificial intelligence algorithm as the generation of the response sentence is completed, and proceeding with a normalization process for the identified image if the identified image satisfies a predetermined tuning condition; and a customized fishing information providing step of generating customized fishing information for the question about fishing based on the normalized image and the response sentence when the normalization process for the image is completed, and outputting the customized fishing information by a chatbot system of the fishing information providing platform and providing it to a user account.

[0016] The process initiation step includes a detailed information confirmation step of confirming detailed information included in the user information registered in the user account, such as environmental weather information, aquatic product growth information, marine environment information, aquatic product distribution information, and aquatic product journal information, when a question about fishing is received from the user account; When the detailed information confirmation step is completed, the method further includes an analysis start step of analyzing the question regarding the fishing based on the confirmed detailed information and starting a service provision process for generating and providing customized fishing information.

[0017] The response sentence generation step includes a first tokenization execution step of, when the service provision process is started, tokenizing the question about change using a first model of the stored artificial intelligence algorithm to tokenize a first sentence corresponding to the question about change; and a vector value-based category identification step of, when tokenization of the first sentence is complete, vectorizing the first token of the first sentence using the first model to calculate a vector value for the first token quantified based on the position of the first token in the first sentence, calculating the vector value for the first sentence, and determining which of a predetermined number of categories the first sentence belongs to based on the calculated vector value.

[0018] The response sentence generation step further includes a second tokenization execution step in which, when the service provision process is started, a first model of the stored artificial intelligence algorithm performs a tokenization process on the detailed information included in the user information, thereby tokenizing each of the second sentences corresponding to the detailed information included in the user information; and a weighted value calculation and classification step in which, when the tokenization of each of the second sentences is completed, the first model performs a vectorization process on the second tokens of each of the second sentences, calculates vector values ​​for the second tokens quantified based on the positions of the second tokens in each of the second sentences, calculates vector values ​​for each of the second sentences, and classifies the vector values ​​for each of the second sentences into weighted values ​​applied to the vector values ​​of the first sentences, and classifies them into a predetermined number of categories.

[0019] The predetermined plurality of categories are categories including a plurality of reference sentences that are candidate groups for responding to the question about fishing, with representative vector values ​​being matched for each of the plurality of categories, and the plurality of reference sentences included for each of the plurality of categories are updated by the administrator of the fishing information providing platform along with the representative vector values ​​being matched for each of the plurality of categories, forming reference category information.

[0020] The response sentence generation step further includes a response sentence derivation step, in which, when calculation of the vector value of the first sentence and the vector value of the second sentence is completed, a vector value analysis step starts analyzing the vector value of the first sentence, the vector value of the second sentence, and the vector value of a reference sentence included in a category to which the first sentence is included by a first model of the stored artificial intelligence algorithm, and, when execution of the function of the vector value analysis step is completed, calculates the reference vector value of each of the reference sentences included in the category to which the first sentence is included from among the predetermined plurality of categories. a reference sentence identification step of comparing the weighted vector value of the first sentence with the vector value of the first sentence, reflecting the vector value of the second sentence classified into the weighted value in the vector value of the first sentence, and identifying a reference sentence having a reference vector value that is highly similar to the vector value of the first sentence reflected with the weighted value; and a sentence derivation completion step of, once the identification of the reference sentence is completed, checking a portion of the identified reference sentence that is identified as the basis for the question about change, and then proceeding with a summarization process to summarize the reference sentence using the first model, documenting the summarized reference sentence, and completing the derivation of a response sentence for answering the question about change.

[0021] The image normalization proceeding step includes a normalization process initiation step of, when the image identification is completed, determining that the predetermined tuning condition is satisfied if an abnormal region exists in the identified image, and starting the normalization process; and a normalization correction completion step of, when the normalization process is initiated, identifying a pixel value distribution for the abnormal region in the image, adjusting brightness and contrast of the identified pixel value distribution, and correcting the abnormal region based on the remaining normal region.

[0022] The stored artificial intelligence algorithm includes a first model, which is a large-scale language model that learns a first pattern value derived by natural language processing of reference sentences included in a predetermined plurality of categories, questions about different changings included in a predetermined plurality of categories, other user information registered in other user accounts that have provided questions about other changings, and other response sentences that respond to questions about other changings that reflect other user information, and by analyzing correlations; and a second model, which is an image search model that learns a second pattern value derived by analyzing correlations between reference images included in a predetermined plurality of categories, questions about different changings included in a predetermined plurality of categories, other response sentences that respond to questions about other changings that reflect other user information, other questions about changing, and other images corresponding to the other response sentences.

[0023] According to another aspect of the present invention, there is provided a chatbot service providing a customized type of fishing information, which is implemented by a computing device including one or more processors and one or more memories for storing commands executable by the processors. The chatbot service includes a process initiation unit that, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checks user information registered in the user account and starts a service providing process for providing customized type of fishing information for the question about fishing based on the checked user information; and, when the service providing process is started, provides the customized type of fishing information for the question about fishing using a stored artificial intelligence algorithm. and a response sentence generation unit that analyzes the question about fishing and generates a response sentence for responding to the question about fishing by reflecting detailed information included in the user information in a vector weight value; and a customized fishing information provision unit that, when an image corresponding to the question about fishing and the response sentence is identified by the stored artificial intelligence algorithm as the generation of the response sentence is completed, generates customized fishing information for the question about fishing based on the identified image and the response sentence, and outputs the customized fishing information through a chatbot system of the fishing information provision platform and provides it to the user account.

[0024] According to another aspect of the present invention, there is provided a chatbot service providing apparatus for outputting customized fishing information, the chatbot service being implemented by a computing device including one or more processors and one or more memories for storing commands executable by the processors. The chatbot service includes a process initiation unit that, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checks user information registered in the user account and starts a service provision process for providing customized fishing information for the question about fishing based on the checked user information. When the service provision process is started, the chatbot service analyzes the question about fishing using a stored artificial intelligence algorithm and reflects detailed information included in the user information in a vector weight value. a response sentence generation unit that generates a response sentence to respond to the question about fishing; an image normalization processing unit that, as the generation of the response sentence is completed, identifies an image corresponding to the question about fishing and the response sentence using the stored artificial intelligence algorithm, and if the identified image satisfies a predetermined tuning condition, performs a normalization process on the identified image; and a customized fishing information provision unit that, when the normalization process on the image is completed, generates customized fishing information for the question about fishing based on the normalized image and the response sentence, and outputs the customized fishing information through a chatbot system of the fishing information provision platform and provides it to a user account.

[0025] A computer-readable recording medium according to another aspect of the present invention stores commands for causing a computing device to perform the following steps, the steps including: upon receiving a question about fishing from a user account registered as a member of a fishing information providing platform, confirming user information registered in the user account, and initiating a service provision process for providing customized fishing information for the question about fishing based on the confirmed user information; and, upon initiation of the service provision process, analyzing the question about fishing using a stored artificial intelligence algorithm and providing customized fishing information for the question about fishing. The method includes a response sentence generating step of generating a response sentence for responding to the question about fishing by reflecting detailed information included in the user information in a vector weight value; and a customized fishing information providing step of, when an image corresponding to the question about fishing and the response sentence is identified by the stored artificial intelligence algorithm as the generation of the response sentence is completed, generating customized fishing information for the question about fishing based on the identified image and the response sentence, and outputting the customized fishing information to the user account through a chatbot system of the fishing information providing platform.

[0026] A computer-readable recording medium according to another aspect of the present invention stores commands for causing a computing device to perform the following steps, the steps including: upon receiving a question about fishing from a user account registered as a member of a fishing information providing platform, confirming user information registered in the user account, and initiating a service providing process for providing customized fishing information for the question based on the confirmed user information; and, upon initiation of the service providing process, analyzing the question about fishing using a stored artificial intelligence algorithm, and responding to the question about fishing by reflecting detailed information included in the user information in a vector weight value. an image normalization proceeding step of identifying an image corresponding to the question about fishing and the response sentence using the stored artificial intelligence algorithm as the generation of the response sentence is completed, and proceeding with a normalization process for the identified image if the identified image satisfies a predetermined tuning condition; and a customized fishing information providing step of generating customized fishing information for the question about fishing based on the normalized image and the response sentence when the normalization process for the image is completed, and outputting the customized fishing information through a chatbot system of the fishing information providing platform and providing it to a user account. [Effects of the Invention]

[0027] The method of providing a chatbot service that outputs customized trading information of the present invention provides accurate information that meets the user's needs, and the user can also be provided with visual information, making it easier to understand.

[0028] Furthermore, by providing customized fishing information in real time using a chatbot, the user can receive the response they desire without being bound by time constraints. [Brief explanation of the drawings]

[0029] [Figure 1] FIG. 1 is a flowchart illustrating a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 2] FIG. 2 is a flowchart illustrating the starting step of a process of a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 3] FIG. 3 is a block diagram illustrating a response sentence generation unit of a device that provides a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 4] FIG. 4 is another block diagram illustrating a response sentence generation unit of a device that provides a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 5] FIG. 5 is a flowchart illustrating a response sentence deriving step of a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 6] FIG. 6 is a flowchart illustrating a step of providing customized change information in a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 7] FIG. 7 is a block diagram illustrating a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 8] FIG. 8 is a flowchart illustrating an image normalization process step of a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention. [Figure 9] FIG. 9 is a diagram illustrating an example of the internal configuration of a computing device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0030] Various embodiments and / or aspects are described below with reference to the drawings. In the following description, for purposes of explanation, numerous specific details are set forth in order to facilitate a general understanding of one or more aspects. However, those skilled in the art will recognize that these aspects may be practiced without such specific details. The following description and the accompanying drawings set forth certain exemplary aspects of one or more aspects in detail. However, such aspects are illustrative, and only a portion of various methods may be utilized in accordance with the principles of the various aspects, and the description is intended to include all such aspects and their equivalents.

[0031] As used herein, "embodiments," "examples," "aspects," "exemplary," and the like may not be construed as constituting any described aspect or design as being better or advantageous over other aspects or designs.

[0032] Additionally, the terms "comprise" and / or "comprising" should be understood to mean that the feature and / or component is present, but not to exclude the presence or addition of one or more other features, components and / or groups thereof.

[0033] Furthermore, terms including ordinal numbers, such as "first," "second," etc., are used to describe various components, but the components are not limited by these terms. These terms are used only to distinguish one component from another. For example, a first component can be referred to as a "second component," and similarly, a second component can be referred to as a "first component," without departing from the scope of the present invention. The term "and / or" includes a combination of multiple related listed items or any of multiple related listed items.

[0034] Furthermore, in the embodiments of the present invention, unless otherwise defined, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art to which the present invention belongs. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning they have in the context of the relevant art, and should not be interpreted as idealized or overly formal unless explicitly defined in the embodiments of the present invention.

[0035] FIG. 1 is a flowchart illustrating a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0036] As shown in FIG. 1, a method for providing a chatbot service that outputs customized change information, implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, includes a process initiation step (S101), a response statement generation step (S103), and a customized change information provision step (S105).

[0037] In the following description, it is understood that the method of providing a chatbot service that outputs customized change information according to each embodiment of the present invention is performed by an apparatus for providing a chatbot service that outputs customized change information according to each embodiment of the present invention shown in Figures 3 to 4 and 7 (hereinafter referred to as the "apparatus of the present invention") and / or a computing device shown in Figure 9. In other words, it is understood that the apparatus of the present invention is embodied by combining one or more of the computing devices of Figure 9.

[0038] In step S101, when the one or more processors (hereinafter referred to as processors) receive a question about fishing from a user account registered as a member of the fishing information provision platform, the processors check the user information registered in the user account and start a service provision process to provide customized fishing information for the question about fishing based on the checked user information.

[0039] According to an embodiment, the fishing information providing platform is a platform linked to a chatbot system, and is a platform for providing customized fishing information to users who are registered as members through the chatbot of the chatbot system.

[0040] According to an embodiment, when the processor receives a question about change from a user account, the processor can check user information registered in the user account.

[0041] In this regard, user information is information containing detailed information created by the user of the user account, including environmental weather information (temperature, humidity, precipitation, and amount of sunlight at the aquaculture site), seafood growth information (growth rate, size, number of fish, etc.), marine environment information (water temperature, salinity, currents, etc.), seafood distribution information (sales price, sales volume), and seafood diary information (recording the aquaculture work content, time, weather, harvest volume, etc.).

[0042] According to an embodiment, after the user information has been confirmed, the processor analyzes the change-related questions received from the user account based on the detailed information included in the confirmed user information, and performs a service provision process to provide customized change information.

[0043] According to an embodiment, the service provision process is a process for providing a service that outputs the customized change information to a user account by a chatbot.

[0044] According to the embodiment, when the service providing process is started, the processor performs the response statement generating step (S103).

[0045] In step S103, when the service provision process is started, the processor can analyze the question about the change using a stored artificial intelligence algorithm, and reflect the detailed information contained in the user information in a vector weight value to generate a response sentence to respond to the question about the change.

[0046] According to an embodiment, the processor analyzes the change question with the stored artificial intelligence algorithm when the service provision process is initiated.

[0047] According to an embodiment, the processor analyzes the change question with the stored artificial intelligence algorithm and calculates a vector value for the change question.

[0048] Here, the processor may convert the detailed information included in the user information into a vector weight value using the stored artificial intelligence algorithm, and reflect the vector weight value in the vector value of the question regarding change.

[0049] That is, the processor analyzes the question regarding change based on the detailed information included in the user information, and generates a response sentence to the question regarding change based on the detailed information included in the user information.

[0050] In this regard, the response sentence is a sentence for responding to the question about change based on the detailed information contained in the user information, and is a sentence that summarizes the answer or related news derived by a stored artificial intelligence algorithm.

[0051] According to the embodiment, when the processor completes the generation of the response sentence, the processor performs a fishing information providing step (S105).

[0052] In step S105, as the generation of the response sentence is completed, if the stored artificial intelligence algorithm identifies an image corresponding to the question about changing and the response sentence, the processor generates customized changing information for the question about changing based on the identified image and the response sentence, and outputs the customized changing information to the user account via a chatbot system of the changing information provision platform.

[0053] According to an embodiment, once the processor has generated the response sentence, the processor analyzes the change question and the response sentence using the stored artificial intelligence algorithm.

[0054] More specifically, the processor can analyze the question about the change and the response sentence using the stored artificial intelligence algorithm to identify images that include objects that contain features of interest that correspond to keywords included in the question about the change and the response sentence.

[0055] According to an embodiment, when the image identification is completed, the processor generates customized change information including the identified image and the response text, where the generated customized change information is information including content and images generated based on a result of analyzing a question about change based on user information.

[0056] According to an embodiment, when the generation of the customized fishing information is completed, the processor outputs the generated customized fishing information through a chatbot of a chatbot system linked to the fishing information provision platform, thereby providing content and images based on the customized fishing information to a user account.

[0057] FIG. 2 is a flowchart illustrating the starting step of a process of a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0058] As shown in FIG. 2, a method for providing a chatbot service that outputs customized change information, implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, includes a process initiation step (e.g., the process initiation step (S101) of FIG. 1).

[0059] According to an embodiment, the process initiation step is a step of, when a question about fishing is received from a user account registered as a member of a fishing information provision platform, checking the user information registered in the user account and starting a service provision process for providing customized fishing information for the question about fishing based on the confirmed user information.

[0060] According to an embodiment, the process initiation step includes a detailed information confirmation step (S201) and an analysis initiation step (S203) as detailed steps for performing the above-mentioned functions.

[0061] In step S201, when the one or more processors (hereinafter referred to as processors) receive a question about fishing from a user account, they can check the detailed information contained in the user information registered in the user account, such as environmental weather information, seafood growth information, marine environment information, seafood distribution information, and seafood diary information.

[0062] According to an embodiment, the user information is information recorded and generated by a user of a user account, and includes environmental weather information, aquatic product growth information, marine environment information, aquatic product distribution information, and aquatic product diary information.

[0063] In this regard, environmental meteorological information is information including the temperature, humidity, precipitation, and amount of sunlight at the farming (or fishing) site; seafood growth information is information including the growth rate, size, and number of seafood to be farmed or fished; marine environment information is information including the water temperature, salinity, and currents at the farming (or fishing) site; seafood distribution information is information including the selling price, sales volume, and distribution route of seafood; and seafood diary information is information that records the farming (or fishing) work content for seafood to be farmed or fished, the time and weather for each work, and the daily harvest volume.

[0064] According to the embodiment, the processor performs an analysis start step (S203) when the detailed information included in the user information has been confirmed.

[0065] In step S203, when the processor completes the function execution of the detailed information confirmation step (S201), it analyzes the question regarding the change based on the confirmed detailed information and starts a service provision process for generating and providing customized change information.

[0066] According to an embodiment, when the processor completes the confirmation of the detailed information included in the user information by executing the function of the detailed information confirmation step (S201), the processor analyzes the change-related question based on the confirmed detailed information using a stored artificial intelligence algorithm, generates customized change information, and initiates a service provision process to provide it to the user account.

[0067] FIG. 3 is a block diagram illustrating a response sentence generation unit of a device that provides a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0068] As shown in FIG. 3, a device for providing a chatbot service that outputs customized change information, which is implemented by a computing device including one or more processors and one or more memories that store commands executable by the processors, includes a response sentence generation unit 300 (e.g., performing the same function as the response sentence generation step (S103) in FIG. 1).

[0069] According to an embodiment, when the service provision process is started by the process initiation unit (e.g., executing the same function as the process initiation step (S101) in FIG. 1), the response statement generation unit 300 analyzes the question about change 301a using a stored artificial intelligence algorithm 305, reflects the detailed information included in the user information in a vector weight value, and generates a response statement to respond to the question about change 301a.

[0070] According to the embodiment, the response sentence generating unit 300 includes a first tokenization executing unit 301 and a vector value-based category identifying unit 303 as detailed components for performing the above-mentioned functions.

[0071] According to an embodiment, when the service provision process is started, the first tokenization execution unit 301 performs a tokenization process on the question about change 301a using a first model of the stored artificial intelligence algorithm 305, and tokenizes a first sentence corresponding to the question about change 301a.

[0072] According to an embodiment, the first tokenization execution unit 301 performs natural language processing on the question about change 301a using a first model of the stored artificial intelligence algorithm 305, thereby completing identification of a first sentence that is a sentence corresponding to the question about change 301a.

[0073] According to the embodiment, when the identification of the first sentence is completed, the first tokenization execution unit 301 performs the tokenization process on the first sentence.

[0074] In this regard, when performing the tokenization process, the first tokenization execution unit 301 performs morpheme tokenization rather than word tokenization because, unlike English, Korean is generally an agglutinative language in which morphemes are not composed of only independent words. The first tokenization execution unit 301 recognizes a plurality of morphemes and morpheme types included in the first sentence corresponding to the question 301a about changing money, classifies the morpheme types, recognizes a combination of an independent morpheme and a dependent morpheme as one token, and designates it as one keyword.

[0075] Here, the first tokenization execution unit 301 recognizes the keyword as one token, or recognizes one morpheme as one token.

[0076] According to an embodiment, when tokenization of the first sentence is completed, the vector value-based category identification unit 303 performs a vectorization process on the first token of the first sentence using the first model, calculates a vector value for the first token in the first sentence that has been quantified based on the position of the first token, calculates a vector value for the first sentence, and can determine which of a plurality of predetermined categories the first sentence belongs to based on the calculated vector value.

[0077] According to an embodiment, when the first tokenization execution unit 301 completes its function execution, the vector value-based category identification unit 303 reflects the tokenized morpheme tokens in the first model of the stored artificial intelligence algorithm 305, performs a vectorizing process on the tokens, and digitizes the tokens based on their occurrence frequency and their position in the sentence, thereby calculating a vector value for each of the tokens.

[0078] According to an embodiment, the vectorization process is performed using at least one model selected from the group consisting of a Bag of Words (BoW) model, a TF-IDF model, a Word2Vec model, a GloVe model, and a BERT model, which may also be part of a model included in a stored large-scale language artificial intelligence algorithm.

[0079] In this regard, the Bag of Words (BoW) model is a model that vectorizes words based on their frequency. For example, in the sentence "Please tell me the growth rate of abalone at the one-time farm," the model extracts the words "one-time," "farm," "abalone," "growth," "rate," and "let me know," calculates the frequency of each word, and vectorizes them.

[0080] The TF-IDF model compensates for the shortcomings of BoW by considering both the frequency of words and the importance of the document to create vectors, and the more important a word is in a document, the higher its weighting is.The Word2Vec model learns word similarity and vectorizes it, and generates vectors taking into account the context of the words, resulting in more accurate results than BoW or TF-IDF.

[0081] Furthermore, the GloVe model, like Word2Vec, is a model that learns word similarities and vectorizes them, but unlike Word2Vec, it is a model that is trained using large-scale text data.Finally, the BERT model is a method of vectorizing sentences using a Transformer model, and generates vectors taking context into account, so it is a model that can obtain more accurate results than Word2Vec and GloVe, and is a commonly used model.

[0082] According to an embodiment, when the vector value-based category identification unit 303 completes the calculation of the vector value for the first sentence, it can perform a similarity calculation process between the calculated vector value and representative vector values ​​matched to each of a plurality of predetermined categories to check whether the first sentence is most similar to any one of the plurality of predetermined categories, and classify the first sentence into one of the plurality of predetermined categories that has the highest similarity to the first sentence.

[0083] More specifically, the vector value-based category identification unit 303 normalizes the vector value of the first sentence according to a first model of the stored artificial intelligence algorithm 305, and then calculates similarity by comparing the vector value with the representative vector values ​​matched to the plurality of predetermined categories using a similarity measurement method such as cosine similarity, Euclidean distance, or Manhattan distance. Then, based on the calculated similarity, it can identify the category having the representative vector value most similar to the vector value of the first sentence from among the plurality of predetermined categories.

[0084] In this regard, the predetermined plurality of categories are categories in which representative vector values ​​are matched for each of the plurality of categories and which contain a plurality of reference sentences that are candidate groups for responding to the fishing-related question 301a, and the administrator of the fishing information provision platform updates the plurality of reference sentences included for each of the plurality of categories along with the representative vector values ​​matched for each of the plurality of categories, making up the reference category information.

[0085] In this regard, the representative vector value matched to each of the predetermined plurality of categories is the average value of the vector values ​​of the plurality of reference sentences included in each of the predetermined plurality of categories.

[0086] For example, of the predetermined multiple categories, the first category is a category related to abalone growth, and is a category that includes articles, papers, etc. on abalone growth as multiple reference sentences, and the second category is a weather-related category related to abalone farms, and is a category that includes articles, papers, etc. on the impact of weather on abalone farming as multiple reference sentences.

[0087] According to an embodiment, the first model of the stored artificial intelligence algorithm 305 is a model that performs natural language processing on reference sentences included in a predetermined number of categories, questions about changing that differ from a predetermined number of categories 301a, other user information registered in other user accounts that have provided questions about changing, and other response sentences that respond to questions about changing that reflect other user information, and analyzes correlations to learn a first pattern value that is derived.

[0088] In this regard, the first model is the Large Language Model (LLM), a deep learning algorithm that can recognize, summarize, translate, predict, and generate text and various content based on knowledge obtained from large datasets. It is an advanced artificial intelligence technology that focuses on understanding and analyzing text, and is capable of understanding the complexity of natural language and is more accurate than existing machine learning algorithms.

[0089] In this regard, LLM is a neural network architecture that has revolutionized natural language processing (NLP) work, and is a comprehensive algorithm that includes tokenization, which divides input text into small units such as words or subwords; an encoder (e.g., vectorization), which processes the input sentence and expresses it in vector form; a decoder, which generates output sentences using the vectors output by the encoder; a loss function, which measures the difference between the output sentence generated by the model and the actual sentence as a function used for model training; and a learning algorithm, which adjusts the model parameters to minimize the loss function.

[0090] FIG. 4 is another block diagram illustrating a response sentence generation unit of a device that provides a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0091] As shown in FIG. 4, a device for providing a chatbot service that outputs customized change information, which is implemented by a computing device including one or more processors and one or more memories that store commands executable by the processors, includes a response sentence generation unit 300 (e.g., performing the same function as the response sentence generation step (S103) in FIG. 1).

[0092] According to an embodiment, when the service provision process is started by the process initiation unit (e.g., executing the same function as the process initiation step (S101) in FIG. 1), the response statement generation unit 300 analyzes the question about change 401a using a stored artificial intelligence algorithm 405, reflects the detailed information included in the user information in a vector weight value, and generates a response statement to respond to the question about change 401a.

[0093] According to the embodiment, the response sentence generation unit 400 includes a second tokenization execution unit 401 and a weighted value calculation and classification unit 403 as detailed components for performing the above-mentioned functions.

[0094] According to an embodiment, when the service provision process is started, the second tokenization execution unit 401 performs a tokenization process on the detailed information included in the user information using a first model of the stored artificial intelligence algorithm 405, and tokenizes each of the second sentences corresponding to the detailed information included in the user information.

[0095] According to an embodiment, the detailed information included in the user information is information created and generated by the user of the user account, and the second tokenization execution unit 401 performs natural language processing on the detailed information included in the user information using the first model.

[0096] According to an embodiment, the second tokenization execution unit 401 performs natural language processing on the detailed information included in the user information using the first model, thereby completing identification of a second sentence, which is a sentence corresponding to the detailed information included in the user information.

[0097] According to an embodiment, when the second tokenization execution unit 401 completes identification of a second sentence, which is a sentence corresponding to detailed information included in the user information, it performs a tokenization process on the identified second sentence.

[0098] Here, the second tokenization execution unit 401 recognizes each keyword constituting the second sentence as one token, or recognizes one morpheme as one token.

[0099] According to an embodiment, when tokenization of each of the second sentences is completed, the weighted value calculation and classification unit 403 performs a vectorization process on the second tokens of each of the second sentences using the first model, thereby calculating vector values ​​for the second tokens in each of the second sentences that have been quantified based on the position of the second token, calculating vector values ​​for each of the second sentences, classifying the vector values ​​for each of the second sentences into weighted values ​​applied to the vector values ​​of the first sentences, and classifying them into a predetermined number of categories.

[0100] According to the embodiment, the weight calculation and classification unit 403 performs vectorization processing on each second token of the second sentence using the first model.

[0101] As a result, the weighted value calculation classification unit 403 completes the calculation of the vector value for each of the second sentences by calculating the vector value for the second token contained in each of the second sentences by converting it into a numerical value based on the position in each of the second sentences.

[0102] Here, the calculated vector values ​​of each second sentence are weighted values ​​applied to the vector values ​​of the first sentence, and are classified into the predetermined plurality of categories. The similarity between each vector value of the second sentence and the representative vector values ​​matched to the predetermined plurality of categories is compared to classify each second sentence into the predetermined plurality of categories.

[0103] More specifically, the weighted value calculation and classification unit 403 normalizes the vector values ​​of the second sentences, and then calculates similarities by comparing them with representative vector values ​​matched to a plurality of predetermined categories using a similarity measurement method such as cosine similarity, Euclidean distance, or Manhattan distance, and then classifies each of the second sentences into a plurality of predetermined categories based on the calculated similarities.

[0104] FIG. 5 is a flowchart illustrating a response sentence deriving step of a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0105] As shown in FIG. 5, a method for providing a chatbot service that outputs customized change information, implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, includes a response sentence derivation step (e.g., the process start step (S101) of FIG. 1).

[0106] According to an embodiment, the response sentence derivation step is a detailed step included in a response sentence generation step (e.g., the response sentence generation step (S103) in FIG. 1) which includes a first tokenization execution step (e.g., performing the same function as the first tokenization execution unit 301 in FIG. 3), a vector value-based category identification step (e.g., performing the same function as the vector value-based category identification unit 303 in FIG. 3), a second tokenization execution step (e.g., performing the same function as the second tokenization execution unit 401 in FIG. 4), and a weight value calculation and classification step (e.g., performing the same function as the weight value calculation and classification unit 403 in FIG. 4).

[0107] According to an embodiment, the response sentence deriving step is a step that is executed after the vector value-based category identification step and the weight value calculation and classification step have been completed.

[0108] According to an embodiment, the response sentence deriving step includes, as detailed steps for performing a function, a vector value analysis step (S501), a reference sentence identification step (S503), and a sentence derivation completion step (S505).

[0109] In step S501, when the one or more processors (hereinafter referred to as processors) have completed calculating the vector value of the first sentence and the vector value of the second sentence, they start analyzing the vector value of the first sentence, the vector value of the second sentence, and the vector values ​​of the reference sentences included in the category to which the first sentence belongs using a first model of the stored artificial intelligence algorithm.

[0110] According to an embodiment, when the processor completes the function execution of the vector value-based category identification step and the weight value calculation and classification step, it starts analyzing the vector value of the first sentence, the vector values ​​of each of the second sentences, and the vector values ​​of the reference sentences included in the category to which the first sentence belongs, using the first model.

[0111] Here, the method by which the processor analyzes the vector value of the first sentence, the vector value of each of the second sentences, and the vector value of the reference sentences included in the category to which the first sentence belongs is to compare the similarity between the vector values ​​of each of the sentences or to perform analysis using an evaluation index included in the first model.

[0112] According to an embodiment, when the processor starts analyzing the vector values ​​of the first sentence, the vector values ​​of each of the second sentences, and the vector values ​​of the reference sentences included in the category to which the first sentence belongs using the first model, it performs a reference sentence identification step (S503).

[0113] In step S503, when the processor completes the execution of the function of the vector value analysis step (S501), it compares the reference vector values ​​of each of the reference sentences included in the category that includes the first sentence among the predetermined plurality of categories with the vector value of the first sentence, reflects the vector value of the second sentence classified as the weighted value in the vector value of the first sentence, and identifies reference sentences having reference vector values ​​that are highly similar to the vector value of the first sentence reflected with the weighted value.

[0114] According to an embodiment, when the processor begins analyzing the vector values ​​of the first sentence, the vector values ​​of each of the second sentences, and the vector values ​​of the reference sentences included in the category to which the first sentence belongs using the first model, the processor identifies the reference vector values ​​of each of the reference sentences included in the category to which the first sentence belongs, among the predetermined plurality of categories.

[0115] According to an embodiment, when the processor completes identifying the reference vector values ​​of each of the reference sentences included in the category to which the first sentence belongs, the processor reflects the vector value of the second sentence in the vector value of the first sentence, and performs a similarity comparison process between the vector value of the first sentence reflected by the vector value of the second sentence and the reference vector values ​​of each of the reference sentences included in the category to which the first sentence belongs.

[0116] According to an embodiment, the processor completes, based on the execution result of the similarity comparison process, identifying the reference sentence that has the highest similarity to the vector value of the first sentence, which reflects the vector value of the second sentence, from among the reference sentences included in the category to which the first sentence belongs.

[0117] In another embodiment, the processor reflects the vector value of the second sentence classified by the weighted value on the vector value of the first sentence, and when identifying a reference sentence having a reference vector value that is highly similar to the vector value of the first sentence reflected by the weighted value, completes the identification of the reference sentence based on the evaluation index included in the first model rather than a similarity comparison process.

[0118] In this regard, evaluation metrics include BLEU (Bilingual Evaluation Understudy), ROUGE (Recall-Oriented Understudy for Gisting Evaluation), METEOR (Metric for Evaluation of Translation with Explicit Ordering), Perplexity, Accuracy, F1 Score, and Human Judgment.

[0119] According to the embodiment, the processor performs the sentence derivation completion step (S505) when it completes identification of a reference sentence having a vector value that is highly similar to the vector value of the first sentence to which the weighted value is reflected.

[0120] In step S505, when the processor has completed identifying the reference sentence, it checks the parts of the identified reference sentence that are identified as the basis for the question about change, and then proceeds with a summarization process to summarize the reference sentence using the first model, documenting the summarized reference sentence and completing the derivation of a response sentence for responding to the question about change.

[0121] According to an embodiment, once the processor has completed identifying a reference sentence having a vector value that is highly similar to the vector value of the first sentence reflecting the weighted value, the processor checks the portion of the identified reference sentence that is identified as the basis for the question about change.

[0122] Here, the processor uses a first model to identify the relationship between keywords that make up the question about changing gear, and identifies keywords included in the identified relationship in the reference sentence, thereby checking the reference sentence for keywords that are based on the keywords included in the identified relationship.

[0123] According to an embodiment, the processor checks the portion of the reference sentence that is identified as the basis for the question about the change, and then proceeds with a summarization process to summarize the reference sentence using the first model.

[0124] In this regard, the processor performs a summarization process using the first model to tokenize and normalize the articles and news articles based on the reference sentences, maintaining consistency of grammar, vocabulary, and semantics of the articles and news articles, where the result of the summarization process is a result that includes the parts identified in the basis for the question about change.

[0125] According to an embodiment, once the summarization process is complete, the processor documents the summarized reference sentence including the portion identified in the rationale for the change question, and completes the derivation of the response sentence.

[0126] Here, as the reference sentences are documented, the response sentences that have been derived (or generated) are grouped by matching questions about change and their corresponding response sentences.

[0127] FIG. 6 is a flowchart illustrating a step of providing customized change information in a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0128] As shown in FIG. 6, a method for providing a chatbot service that outputs customized fishing information, which is implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, includes a customized fishing information providing step (e.g., the customized fishing information providing step (S105) of FIG. 1).

[0129] According to an embodiment, the customized fishing information providing step is a step of, when the generation of the response sentence is completed and an image corresponding to the fishing question and the response sentence is identified by the stored artificial intelligence algorithm, generating customized fishing information for the fishing question based on the identified image and the response sentence, and outputting the customized fishing information to a user account via a chatbot system of the fishing information providing platform.

[0130] According to an embodiment, the step of providing customized fishing information includes, as detailed steps for performing the above-mentioned functions, an image information identification step (S601) and a chatbot-based information provision step (S603).

[0131] In step S601, when the generation of the response sentence is completed, the one or more processors (hereinafter referred to as processors) analyze the question about the change and the response sentence using a second model of the stored artificial intelligence algorithm, and identify at least one image from among the multiple pieces of image information stored in the image database that corresponds to the result of analyzing the question about the change and the response sentence.

[0132] According to an embodiment, once the processor has completed generating the response sentence, it analyzes the question about the change and the response sentence using a second model of the stored artificial intelligence algorithm, and identifies an image corresponding to the analysis result from among multiple pieces of image information stored in an image database.

[0133] More specifically, the processor can identify an image containing an object corresponding to a keyword included in each of the question about fishing and the response sentence from a plurality of images stored in the image database.

[0134] According to an embodiment, the second model of the stored artificial intelligence algorithm is a model that learns second pattern values ​​derived by analyzing correlations between reference images included in a predetermined plurality of categories, questions about changing that differ from a predetermined plurality of categories, other response sentences in response to other questions about changing that reflect other user information, and other images corresponding to other questions about changing and other response sentences.

[0135] In this regard, the second pattern value is a pattern value that calculates the similarity by comparing the feature vector value of the image with the vector value of the sentence, and is a pattern value that identifies an image that contains an object that corresponds to the keyword contained in the sentence.

[0136] According to an embodiment, when the image identification is completed, the processor performs a chatbot-based information providing step (S603).

[0137] In the chatbot-based information provision step (S603), once the processor has completed identifying the image, it generates customized fishing information including the identified image and the response text, and provides the customized fishing information to the user account via a chatbot system linked to the fishing information provision platform.

[0138] According to the embodiment, when the processor completes image identification by executing the function of the image information identification step (S601), it generates customized hanging information including the identified image and the response text.

[0139] In this regard, when the processor completes the generation of the customized fishing information, the processor outputs the generated customized fishing information through a chatbot of a chatbot system linked to the fishing information providing platform, so that the user of the user account can check the image and response text based on the customized fishing information output by the chatbot.

[0140] FIG. 7 is a block diagram illustrating a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0141] As shown in FIG. 7, the method for providing a chatbot service for outputting customized fishing information, which is implemented in a computing device including one or more processors and one or more memories for storing commands executable by the processors, includes a process initiation unit 701 (e.g., performing the same function as the process initiation step (S101) of FIG. 1), a response statement generation unit 703 (e.g., performing the same function as the response statement generation step (S103) of FIG. 1), an image normalization progress unit 705, and a customized fishing information provision unit 707 (e.g., performing the same function as the customized fishing information provision step (S105) of FIG. 1).

[0142] According to an embodiment, when the process initiation unit 701 receives a question about fishing 701a from a user account registered as a member of a fishing information provision platform, the process initiation unit 701 checks the user information registered in the user account and starts a service provision process for providing customized fishing information for the question about fishing 701a based on the checked user information.

[0143] According to an embodiment, the fishing information providing platform is a platform linked to a chatbot system, and is a platform for providing customized fishing information to users who are registered as members through the chatbot of the chatbot system.

[0144] According to the embodiment, when the process initiation unit 701 receives a question 701a about change from a user account, the process initiation unit 701 checks the user information registered in the user account.

[0145] In this regard, user information is information containing detailed information created by the user of the user account, including environmental weather information (temperature, humidity, precipitation, and amount of sunlight at the aquaculture site), seafood growth information (growth rate, size, number of fish, etc.), marine environment information (water temperature, salinity, currents, etc.), seafood distribution information (sales price, sales volume), and seafood diary information (recording the aquaculture work content, time, weather, harvest volume, etc.).

[0146] According to an embodiment, once the process initiation unit 701 has completed the confirmation of the user information, it analyzes the change-related question 701a received from the user account based on the detailed information included in the confirmed user information, and performs a service provision process to provide customized change information 707a.

[0147] According to the embodiment, the service provision process is a process for providing a service that outputs the customized change information 707a to a user account by a chatbot.

[0148] According to an embodiment, when the service provision process is started, the response sentence generation unit 703 analyzes the question about change 701a using a stored artificial intelligence algorithm 709, reflects detailed information included in the user information in a vector weight value, and generates a response sentence to respond to the question about change 701a.

[0149] According to an embodiment, when the service provision process is initiated, the response sentence generator 703 analyzes the change question 701a using the stored artificial intelligence algorithm 709.

[0150] According to an embodiment, the response sentence generator 703 analyzes the question about change 701a using the stored artificial intelligence algorithm 709 and calculates a vector value for the question about change 701a.

[0151] Here, the response sentence generation unit 703 converts the detailed information included in the user information into a vector weight value using the stored artificial intelligence algorithm 709, and reflects the vector weight value in the vector value of the question about change 701a.

[0152] That is, the response sentence generation unit 703 analyzes the question 701a about change based on the detailed information included in the user information, and generates a response sentence to the question 701a about change based on the detailed information included in the user information.

[0153] In this regard, the response sentence is a sentence for responding to the question 701a about change based on the detailed information contained in the user information, and is a sentence that summarizes the answer and related news derived by the stored artificial intelligence algorithm 709.

[0154] According to an embodiment, as the generation of the response sentence is completed, the image normalization proceeding unit 705 identifies an image 705a corresponding to the question 701a about change and the response sentence from the stored artificial intelligence algorithm 709, and if the identified image 705a satisfies a predetermined tuning condition, it can proceed with a normalization process for the identified image. According to an embodiment, after completing identification of the image 705a corresponding to the question 701a about the change and the response sentence, the image normalization proceeding unit 705 determines whether a region included in the image 705a satisfies the predetermined tuning condition.

[0155] Accordingly, when the image normalization proceeding unit 705 determines that one area included in the image 705a satisfies the predetermined tuning condition, it proceeds with the normalization process for the identified image 705a.

[0156] In this regard, the normalization process includes at least one of the following processes: mean normalization process, which divides the pixel values ​​of the image by the mean value and then multiplies them by the original size again to normalize; standard deviation normalization process, which divides the pixel values ​​of the image by the standard deviation and then multiplies them by the original size again to normalize; min-max normalization process, which divides the pixel values ​​of the image into minimum and maximum values ​​and then multiplies them by the original size again to normalize; normalization map process, which converts each pixel value of the image to a value between 0 and 1; normalization map process, which converts the pixel values ​​of the image to the mean and standard deviation; and PCA (Principal Component Analysis) process, which extracts and normalizes the main components of the image.

[0157] According to an embodiment, when the normalization process for the image 705a is completed, the customized fishing information providing unit 707 generates customized fishing information 707a for the question about fishing based on the normalized image and the response text, and outputs the customized fishing information 707a through a chatbot system of the fishing information providing platform and provides it to the user account.

[0158] According to an embodiment, when the normalization process of the image and the generation of the response sentence are completed, the customized trading information providing unit 707 generates customized trading information 707a including the image for which the normalization process has been completed and the response sentence.

[0159] Here, the generated customized changing information 707a is information that includes content generated based on the results of analyzing the question 701a about changing with the user information as a reference, and a normalized image.

[0160] According to an embodiment, when the customized fishing information provision unit 707 completes the generation of the customized fishing information 707a, it outputs the generated customized fishing information 707a through a chatbot of a chatbot system linked to the fishing information provision platform, thereby providing content and images based on the customized fishing information to a user account.

[0161] That is, when the function of the image normalization progress unit 705 is completed, the customized fishing information providing unit 707 generates customized fishing information 707a including the normalized image and the response text, and completes providing the customized fishing information 707a to the user account through a chatbot system linked to the fishing information providing platform.

[0162] FIG. 8 is a flowchart illustrating an image normalization process step of a method for providing a chatbot service that outputs customized change information according to an embodiment of the present invention.

[0163] As shown in FIG. 8, a method for providing a chatbot service that outputs customized trading information, implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, includes an image normalization process step (e.g., performing the same function as the image normalization process unit 705 of FIG. 7).

[0164] According to an embodiment, the image normalization proceeding step is a step of identifying an image corresponding to the question about change and the response sentence using the stored artificial intelligence algorithm as the generation of the response sentence is completed, and proceeding with a normalization process for the identified image if the identified image satisfies a predetermined tuning condition.

[0165] According to an embodiment, the image normalization process step includes a normalization process start step (S801) and a normalization correction completion step (S803) as detailed steps for performing the above-mentioned functions.

[0166] In step S801, when the one or more processors (hereinafter referred to as processors) complete the identification of the image, if an abnormal area exists in the identified image, they determine that the specified tuning condition is met and start the normalization process.

[0167] According to an embodiment, once the processor has completed identifying the image, it may analyze the image with a second model of a stored artificial intelligence algorithm to determine whether an abnormal region exists within the identified image.

[0168] In this regard, the abnormal region includes a resolution abnormal region, a brightness abnormal region, an object blur region, etc. located in the image.

[0169] According to an embodiment, the processor analyzes not only the multiple images stored in the image database but also the images included in the papers and news articles based on the response sentence, and when predetermined tuning conditions are met, initiates a normalization process for abnormal regions included in the images included in the papers and news articles based on the response sentence.

[0170] According to the embodiment, when the normalization process is started, the processor performs the normalization correction completion step (S803).

[0171] In step S803, when the normalization process begins, the processor identifies pixel value distributions for abnormal regions in the image and adjusts the brightness and contrast of the identified pixel value distributions to correct the abnormal regions based on the remaining normal regions.

[0172] In an embodiment, when the normalization process for the image is initiated, the processor identifies a pixel value distribution for an abnormal region in the image and adjusts brightness and contrast of the identified pixel value distribution, where the adjusted setting values ​​can be derived using a second model.

[0173] According to an embodiment, the processor performs a normalization process corresponding to a tuning condition among predetermined tuning conditions that satisfies an abnormal area included in the image, and performs correction on the abnormal area, thereby correcting the abnormal area based on a normal area of ​​the image.

[0174] FIG. 9 is a diagram illustrating an example of the internal configuration of a computing device according to an embodiment of the present invention.

[0175] FIG. 9 illustrates an example of the internal configuration of a computing device according to an embodiment of the present invention, and in the following description, unnecessary descriptions that overlap with the descriptions of FIGS. 1 to 8 will be omitted.

[0176] 9, the computing device 10000 includes at least one processor 11100, a memory 11200, a peripheral device interface 11300, an input / output subsystem 11400, a power circuit 11500, and a communication circuit 11600. Here, the computing device 10000 corresponds to a user terminal (A) connected to a haptic interface device or the computing device (B).

[0177] The memory 11200 may include, for example, high-speed random access memory, a magnetic disk, an SRAM, a DRAM, a ROM, a flash memory, or a non-volatile memory. The memory 11200 may include software modules, command sets, or various other data required for the operation of the computing device 10000.

[0178] Here, access to memory 11200 from other components such as processor 11100 and peripheral device interface 11300 is controlled by processor 11100.

[0179] The peripheral device interface 11300 couples input and / or output peripheral devices of the computing device 10000 to the processor 11100 and memory 11200. The processor 11100 executes software modules or sets of commands stored in the memory 11200 to perform various functions and process data for the computing device 10000.

[0180] The I / O subsystem 11400 couples various I / O peripherals to the peripheral interface 11300. For example, the I / O subsystem 11400 includes controllers for coupling peripherals such as a monitor, keyboard, mouse, printer, and, if necessary, a touch screen or sensor to the peripheral interface 11300. In another aspect, an I / O peripheral can also be coupled to the peripheral interface 11300 without going through the I / O subsystem 11400.

[0181] The power circuit 11500 may provide power to all or some of the components of the terminal, and may include, for example, a power management system, one or more power sources such as a battery or alternating current (AC), a charging system, a power failure detection circuit, a power converter or inverter, a power status indicator, or any other component for power generation, management, or distribution.

[0182] The communications circuitry 11600 allows for communication with other computing devices using at least one external port.

[0183] Alternatively, as previously mentioned, if desired, the communications circuitry 11600 may include RF circuitry to enable communication with other computing devices by sending and receiving RF signals, also known as electromagnetic signals.

[0184] The embodiment of Figure 9 is merely one example of computing device 10000, and computing device 10000 may omit some components in Figure 9, include additional components not shown, or have a configuration or arrangement that combines two or more components. For example, a computing device for a communication terminal in a mobile environment may further include a touch screen and sensors in addition to the components in Figure 9, and communication circuit 11600 may include circuits for RF communication of various communication methods (WiFi, 3G, LTE, Bluetooth, NFC, Zigbee, etc.). The components included in computing device 10000 may be embodied as hardware, software, or a combination of both hardware and software, including one or more integrated circuits specialized for signal processing or applications.

[0185] Methods according to embodiments of the present invention may be implemented in the form of program instructions executed by various computing devices and recorded on a computer-readable medium. In particular, the program according to this embodiment may be implemented as a PC-based program or an application dedicated to mobile terminals. The application to which the present invention is applied is installed in a user terminal through a file provided by a file distribution system. For example, the file distribution system may include a file transfer unit (not shown) that transfers the file in response to a request from the user terminal.

[0186] The devices described above may be implemented using hardware components, software components, and / or a combination of hardware and software components. For example, the devices and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device executes an operating system (OS) and one or more software applications that run on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, although a single processing device may be described as being used, those skilled in the art will recognize that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, the processing device may include multiple processors or one processor and one controller. Other processing configurations, such as parallel processors, are also possible.

[0187] Software includes computer programs, codes, instructions, or a combination of one or more of these, which can configure a processing device to perform a desired operation or independently or jointly instruct the processing device. The software and / or data can be permanently or temporarily embodied in any type of machine, component, physical device, virtual device, computer storage medium, or device to be analyzed by the processing device or to provide instructions or data to the processing device. The software can also be distributed across computing devices connected by a network and stored or executed in a distributed manner. The software and data are stored on one or more computer-readable recording media.

[0188] Methods according to the embodiments may be embodied in the form of program instructions executed by various computer means and recorded on a computer-readable medium. The computer-readable medium may include, alone or in combination, program instructions, data files, data structures, and the like. The program instructions recorded on the medium may be those specially designed and constructed for the embodiments, or those known and available to those skilled in the art of computer software. Computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape; optical media such as CD-ROMs and DVDs; magneto-optical media such as protocol disks; and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, and flash memory. Program instructions include not only machine language code, such as produced by a compiler, but also high-level language code executed by a computer using an interpreter, for example. The hardware devices may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

[0189] Although the embodiments have been described above using limited examples and drawings, those skilled in the art will appreciate that various modifications and variations may be made from the above description. For example, the described techniques may be performed in a different order than described, and / or the described system, structure, device, circuit, or other components may be combined or combined in a different manner than described, or may be substituted or replaced with other components or equivalents, while still achieving suitable results. Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims. [Explanation of symbols]

[0190] 300...Response sentence generation unit, 301...First tokenization execution unit, 301a...Question about change, 303...Vector value based category identification unit, 305...Stored artificial intelligence algorithm, 10000...Computing device, 11100...Processor, 11200...Memory, 11300...Peripheral device interface, 11400...Input / output subsystem, 11500...Power circuit, 11600...Communication circuit

Claims

1. A method for providing a chatbot service that outputs customized change information, the chatbot service being implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, comprising: a process initiation step of, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checking the user information registered in the user account and starting a service provision process for providing customized fishing information for the question about fishing based on the checked user information; a response sentence generation step of analyzing the question about the change using a stored artificial intelligence algorithm when the service provision process is started, and generating a response sentence for responding to the question about the change by reflecting detailed information included in the user information in a vector weight value; and a customized fishing information providing step of generating customized fishing information for the question about fishing based on the identified image and the response sentence when the stored artificial intelligence algorithm identifies an image corresponding to the question about fishing and the response sentence as the generation of the response sentence is completed, and outputting the customized fishing information through a chatbot system of the fishing information providing platform and providing it to a user account.

2. The process initiation step includes: a detailed information confirmation step of confirming, when a question about fishing is received from the user account, detailed information included in the user information registered in the user account, such as environmental weather information, marine product growth information, marine environment information, marine product distribution information, and marine product journal information; and an analysis initiation step of analyzing the question regarding the change based on the confirmed detailed information and initiating a service provision process for generating and providing customized change information, when the function execution of the detailed information confirmation step is completed.

3. The response statement generating step a first tokenization execution step of, when the service provision process is started, tokenizing the question about change by using a first model of the stored artificial intelligence algorithm to tokenize a first sentence corresponding to the question about change; a vector value-based category identification step of: when tokenization of the first sentence is completed, performing a vectorization process on the first token of the first sentence using the first model, thereby calculating a vector value for the first token quantified based on the position of the first token in the first sentence, thereby calculating a vector value for the first sentence; and determining which of a plurality of predetermined categories the first sentence belongs to based on the calculated vector value.

4. The response statement generating step a second tokenization execution step of, when the service provision process is started, performing a tokenization process on the detailed information included in the user information using a first model of the stored artificial intelligence algorithm, thereby tokenizing each second sentence corresponding to the detailed information included in the user information; 4. The method for providing a chatbot service that outputs customized ticket information according to claim 3, further comprising: a weighted value calculation and classification step in which, after tokenization of each of the second sentences is completed, vectorization processing is performed on the second tokens of each of the second sentences using the first model to calculate vector values ​​for the second tokens quantified based on the position of the second token in each of the second sentences, thereby calculating vector values ​​for each of the second sentences; and classifying the vector values ​​for each of the second sentences into weighted values ​​applied to the vector values ​​of the first sentences and classifying them into a predetermined number of categories.

5. 5. The method for providing a chatbot service that outputs customized fishing information according to claim 4, wherein the predetermined plurality of categories are categories in which representative vector values ​​are matched for each of the plurality of categories and which include a plurality of reference sentences that are a group of candidates for responding to the question about fishing, and the plurality of reference sentences included for each of the plurality of categories are updated together with the representative vector values ​​matched for each of the plurality of categories by an administrator of the fishing information providing platform.

6. the response sentence generating step further includes a response sentence deriving step, The response sentence deriving step a vector value analysis step of starting an analysis of the vector value of the first sentence, the vector value of the second sentence, and the vector values ​​of reference sentences included in the category to which the first sentence belongs, using a first model of the stored artificial intelligence algorithm, when the calculation of the vector value of the first sentence and the vector value of the second sentence is completed; a reference sentence identification step, when the execution of the function of the vector value analysis step is completed, of comparing the reference vector values ​​of each reference sentence included in a category including the first sentence among the predetermined plurality of categories with the vector value of the first sentence, and reflecting the vector value of the second sentence classified into the weighted value in the vector value of the first sentence to identify a reference sentence having a reference vector value that is highly similar to the vector value of the first sentence reflected with the weighted value; and a sentence derivation completion step of, when the identification of the reference sentence is completed, checking the portion of the identified reference sentence that is identified as the basis for the question about change, and then proceeding with a summarization process to summarize the reference sentence using the first model, documenting the summarized reference sentence, and completing derivation of a response sentence to answer the question about change. The method for providing a chatbot service that outputs customized change information according to claim 4, further comprising:

7. The customized fishing information providing step includes: an image information identification step of analyzing the question about the change and the response sentence using a second model of the stored artificial intelligence algorithm when the generation of the response sentence is completed, and identifying at least one image corresponding to the result of analyzing the question about the change and the response sentence from among a plurality of pieces of image information stored in an image database; and a chatbot-based information providing step of generating customized fishing information including the identified image and the response text when the image identification is completed, and providing the customized fishing information to a user account through a chatbot system linked to the fishing information providing platform.

8. The stored artificial intelligence algorithm comprises: a first model, which is a large-scale language model that performs natural language processing on reference sentences included in a predetermined plurality of categories, questions about different changing spellings in a predetermined plurality of categories, other user information registered in other user accounts that have provided questions about other changing spellings, and other response sentences that respond to questions about other changing spellings that reflect other user information, and analyzes correlations to learn first pattern values ​​derived from the natural language processing; A method for providing a chatbot service that outputs customized changing information, as described in claim 1, characterized in that it includes: reference images included in a predetermined plurality of categories; questions about changing that differ according to a predetermined plurality of categories; other response sentences that respond to other questions about changing that reflect other user information; and a second model that is an image search model that learns second pattern values ​​derived by analyzing correlations between other images corresponding to other questions about changing and other response sentences.

9. A method for providing a chatbot service that outputs customized change information, the chatbot service being implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, comprising: a process initiation step of, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checking the user information registered in the user account and starting a service provision process for providing customized fishing information for the question about fishing based on the checked user information; a response sentence generation step of analyzing the question about the change using a stored artificial intelligence algorithm when the service provision process is started, and generating a response sentence for responding to the question about the change by reflecting detailed information included in the user information in a vector weight value; an image normalization proceeding step of identifying an image corresponding to the question about the change and the response sentence by the stored artificial intelligence algorithm as the generation of the response sentence is completed, and proceeding with a normalization process for the identified image if the identified image satisfies a predetermined tuning condition; a customized fishing information providing step of generating customized fishing information for the question about fishing based on the normalized image and the response text when the normalization process for the image is completed, and outputting the customized fishing information through a chatbot system of the fishing information providing platform and providing it to the user account.

10. The process initiation step includes: a detailed information confirmation step of confirming, when a question about fishing is received from the user account, detailed information included in the user information registered in the user account, such as environmental weather information, marine product growth information, marine environment information, marine product distribution information, and marine product journal information; 10. The method for providing a chatbot service that outputs customized changing information according to claim 9, further comprising: an analysis initiation step of analyzing the question regarding the change based on the confirmed detailed information and initiating a service provision process for generating and providing customized changing information, when the function execution of the detailed information confirmation step is completed.

11. The response statement generating step a first tokenization execution step of, when the service provision process is started, tokenizing the question about change by using a first model of the stored artificial intelligence algorithm to tokenize a first sentence corresponding to the question about change; 10. The method for providing a chatbot service that outputs customized ticket information according to claim 9, further comprising: a vector value-based category identification step of: when tokenization of the first sentence is completed, performing a vectorization process on the first token of the first sentence using the first model, thereby calculating a vector value for the first token quantified based on the position of the first token in the first sentence; calculating a vector value for the first sentence; and determining which of a plurality of predetermined categories the first sentence belongs to based on the calculated vector value.

12. The response statement generating step a second tokenization execution step of, when the service provision process is started, performing a tokenization process on the detailed information included in the user information using a first model of the stored artificial intelligence algorithm, thereby tokenizing each second sentence corresponding to the detailed information included in the user information; 12. The method for providing a chatbot service that outputs customized ticket information according to claim 11, further comprising: a weighted value calculation and classification step of, after tokenization of each of the second sentences is completed, performing a vectorization process on the second tokens of each of the second sentences using the first model; calculating vector values ​​for the second tokens quantified based on the positions of the second tokens in each of the second sentences to calculate vector values ​​for each of the second sentences; classifying the vector values ​​for each of the second sentences into weighted values ​​applied to the vector values ​​of the first sentences; and classifying the vector values ​​into a predetermined plurality of categories.

13. 13. The method for providing a chatbot service that outputs customized fishing information according to claim 12, wherein the predetermined plurality of categories are categories including a plurality of reference sentences that are candidate groups for responding to the fishing-related question by matching representative vector values ​​for each of the plurality of categories, and the plurality of reference sentences included for each of the plurality of categories are updated by an administrator of the fishing information providing platform along with the representative vector values ​​matched for each of the plurality of categories.

14. the response sentence generating step further includes a response sentence deriving step, The response sentence deriving step a vector value analysis step of starting an analysis of the vector value of the first sentence, the vector value of the second sentence, and the vector values ​​of reference sentences included in the category to which the first sentence belongs, using a first model of the stored artificial intelligence algorithm, when the calculation of the vector value of the first sentence and the vector value of the second sentence is completed; a reference sentence identification step, when the execution of the function of the vector value analysis step is completed, of comparing the reference vector values ​​of each of the reference sentences included in the category including the first sentence among the predetermined plurality of categories with the vector value of the first sentence, and reflecting the vector value of the second sentence classified into the weighted value in the vector value of the first sentence to identify a reference sentence having a reference vector value that is highly similar to the vector value of the first sentence reflected with the weighted value; 13. The method for providing a chatbot service that outputs customized change information according to claim 12, further comprising: a sentence derivation completion step of, when the identification of the reference sentence is completed, checking the identified portion of the reference sentence that is identified as the basis for the question about change, and then proceeding with a summarization process to summarize the reference sentence using the first model, documenting the summarized reference sentence, and completing the derivation of a response sentence for answering the question about change.

15. The image normalization proceeding step includes: a normalization process initiation step of determining, upon completion of the image classification, that the predetermined tuning condition is satisfied and initiating the normalization process if an abnormal region exists in the classified image; 10. The method for providing a chatbot service that outputs customized type-finishing information, as described in claim 9, further comprising: a normalization correction completion step of identifying a pixel value distribution for an abnormal region in the image when the normalization process is started, adjusting the brightness and contrast of the identified pixel value distribution, and correcting the abnormal region based on the remaining normal region.

16. The stored artificial intelligence algorithm comprises: a first model, which is a large-scale language model that performs natural language processing on reference sentences included in a predetermined plurality of categories, questions about different changing spellings in a predetermined plurality of categories, other user information registered in other user accounts that have provided questions about other changing spellings, and other response sentences that respond to questions about other changing spellings that reflect other user information, and analyzes correlations to learn first pattern values ​​derived from the natural language processing; and a second model, which is an image search model that learns second pattern values ​​derived by analyzing correlations between reference images included in a predetermined plurality of categories, questions about different changes in a predetermined plurality of categories, other response sentences that respond to other questions about changes that reflect other user information, and other questions about changes and other images corresponding to the other response sentences.

17. An apparatus for providing a chatbot service that outputs customized change information, the chatbot service being implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, a process initiation unit that, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checks user information registered in the user account and starts a service provision process for providing customized fishing information for the question about fishing based on the checked user information; a response sentence generation unit that, when the service provision process is started, analyzes the question about the change using a stored artificial intelligence algorithm, reflects detailed information included in the user information in a vector weight value, and generates a response sentence to respond to the question about the change; and a customized fishing information providing unit that, when an image corresponding to the question about fishing and the response sentence is identified by the stored artificial intelligence algorithm as the generation of the response sentence is completed, generates customized fishing information for the question about fishing based on the identified image and the response sentence, and outputs the customized fishing information through a chatbot system of the fishing information providing platform and provides it to a user account.

18. An apparatus for providing a chatbot service that outputs customized change information, the chatbot service being implemented in a computing device including one or more processors and one or more memories that store commands executable by the processors, a process initiation unit that, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checks user information registered in the user account and starts a service provision process for providing customized fishing information for the question about fishing based on the checked user information; a response sentence generation unit that, when the service provision process is started, analyzes the question about the change using a stored artificial intelligence algorithm, reflects detailed information included in the user information in a vector weight value, and generates a response sentence to respond to the question about the change; an image normalization processing unit that, as the generation of the response sentence is completed, identifies an image corresponding to the question about the change and the response sentence using the stored artificial intelligence algorithm, and if the identified image satisfies a predetermined tuning condition, performs a normalization process on the identified image; and a customized fishing information providing unit that, when the normalization process for the image is completed, generates customized fishing information for the question about fishing based on the normalized image and the response text, and outputs the customized fishing information through a chatbot system of the fishing information providing platform and provides it to a user account.

19. A computer-readable recording medium, The computer-readable medium stores instructions that cause a computing device to perform the following steps, the steps comprising: a process initiation step of, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checking the user information registered in the user account and starting a service provision process for providing customized fishing information for the question about fishing based on the checked user information; a response sentence generation step of analyzing the question about the change using a stored artificial intelligence algorithm when the service provision process is started, and generating a response sentence for responding to the question about the change by reflecting detailed information included in the user information in a vector weight value; and a customized fishing information providing step of, when the stored artificial intelligence algorithm identifies an image corresponding to the question about fishing and the response sentence as the generation of the response sentence is completed, generating customized fishing information for the question about fishing based on the identified image and the response sentence, and outputting the customized fishing information to a user account through a chatbot system of the fishing information providing platform.

20. A computer-readable recording medium, The computer-readable medium stores instructions that cause a computing device to perform the following steps, the steps comprising: a process initiation step of, when receiving a question about fishing from a user account registered as a member of a fishing information providing platform, checking the user information registered in the user account and starting a service provision process for providing customized fishing information for the question about fishing based on the checked user information; a response sentence generation step of analyzing the question about the change using a stored artificial intelligence algorithm when the service provision process is started, and generating a response sentence for responding to the question about the change by reflecting detailed information included in the user information in a vector weight value; an image normalization proceeding step of identifying an image corresponding to the question about the change and the response sentence by the stored artificial intelligence algorithm as the generation of the response sentence is completed, and proceeding with a normalization process for the identified image if the identified image satisfies a predetermined tuning condition; and a customized fishing information providing step of generating customized fishing information for the fishing question based on the normalized image and the response text when the normalization process for the image is completed, and outputting the customized fishing information through a chatbot system of the fishing information providing platform and providing it to a user account.

Images