Method and system for analyzing marine leisure tourism activity index using mme-based weather forecast observation system

Abstract

A marine leisure tourism activity index analysis system using an MME-based weather forecast observation system may include a memory and a processor for storing instructions. The instructions can control the system to collect weather data from a plurality of weather forecasting models, convert the collected weather data into a standardized format, process the standardized weather data using a Multi-Model Ensemble (MME) model to generate corrected weather forecast data, calculate an activity index for each marine leisure sport based on the corrected weather forecast data, and determine the optimal marine leisure sport at a specific location or determine the optimal location for a specific marine leisure sport based on the calculated activity index and provide it to the user.

Description

Technology Field

[0001] The present invention relates to the fields of weather forecasting, marine observation, and leisure tourism. More specifically, it relates to a system and a method of operation thereof that utilizes a Multi-Model Ensemble (MME) technique to process weather data collected from multiple weather prediction models in an integrated manner, calculates a safety activity index for each marine leisure sport in real time based on this, and provides a customized leisure activity recommendation service reflecting the characteristics of each sea area.

[0002] In addition, the present invention can be applied to the field of intelligent decision-making system technology that supports safe marine leisure activities for users by comprehensively analyzing weather and ocean elements for various marine leisure sports such as fishing, surfing, windsurfing, snorkeling, scuba diving, yachting, and boating, and differentially reflecting the unique environmental characteristics and risk factors of domestic waters such as the West Sea, South Sea, East Sea, and Jeju. Background Technology

[0004] With the recent surge in interest in marine leisure tourism, the number of people enjoying various marine sports such as fishing, surfing, scuba diving, and yachting has increased significantly. However, unlike land, the marine environment is highly sensitive to rapidly changing weather conditions, and engaging in marine activities under inappropriate weather conditions can lead to serious safety accidents that are directly life-threatening.

[0005] Existing weather forecasting systems focus on providing general weather information (e.g., temperature, precipitation, wind speed, etc.) and offer the ability to predict and visually display future weather conditions based on their respective unique algorithms. These systems convey information to users by visualizing wind direction and speed, wave height, and cloud distribution on maps. However, existing weather forecasting systems merely provide weather data and fail to offer customized information or safety assessment indicators that consider the relevance to marine leisure sports activities. In particular, despite the fact that safety and suitability can vary significantly depending on the type of leisure sport even under identical weather conditions, differentiated services that reflect this are rarely found.

[0006] Furthermore, domestic waters possess highly distinct marine environmental characteristics depending on the region, such as the shallow depths and large tidal ranges of the West Sea, the complex straits and currents around islands in the South Sea, the deep waters and swells of the East Sea, and the warm currents and gusts of wind in the waters off Jeju. Existing systems fail to adequately reflect these regional specificities, making it highly likely that individuals will be exposed to unexpected dangerous situations during actual marine leisure activities. The problem to be solved

[0008] The present invention aims to provide a system and method for generating a more accurate and reliable weather forecast model through the MME technique by collecting raw data from different weather prediction models provided by national meteorological agencies and integrating actual weather information including observational data.

[0009] In addition, the present invention aims to provide a system and method for quantifying the influence of meteorological and marine elements, such as wind speed, wind direction, wave height, wave period, current speed and direction, water temperature, turbidity, precipitation, and visibility, on each marine leisure sport, such as fishing, boating, surfing, and scuba diving, and for calculating a comprehensive activity index by differentially applying weights to each element based on this.

[0010] The present invention aims to provide a customized activity index calculation system and method that accurately reflects regional characteristics by identifying region-specific risk factors such as current speed, tidal difference, swells, gusts, water depth, and turbidity for each domestic sea area, including the West Sea, South Sea, East Sea, and Jeju, and by applying differentiated weights and safety standards for each sea area, even for the same leisure sport.

[0011] The present invention aims to provide a service platform that recommends marine leisure sports that can be optimally enjoyed at a specific location when a user selects a specific spot on a map, or conversely, suggests the optimal activity location when a specific marine leisure sport is selected. means of solving the problem

[0013] A marine leisure tourism activity index analysis system using an MME-based weather forecast observation system may include a memory and a processor for storing instructions.

[0014] The instructions can control the system to collect weather data from a plurality of weather forecasting models, convert the collected weather data into a standardized format, process the standardized weather data using a Multi-Model Ensemble (MME) model to generate corrected weather forecast data, calculate an activity index for each marine leisure sport based on the corrected weather forecast data, and determine the optimal marine leisure sport at a specific location or determine the optimal location for a specific marine leisure sport based on the calculated activity index and provide it to the user. Effects of the invention

[0016] The method and system for analyzing marine leisure tourism activity indices using the MME-based weather forecast observation system according to the present invention utilizes an MME technique that integrates multiple weather forecast models to provide significantly improved prediction accuracy compared to a single model, and based on this, can accurately determine the safety and suitability of marine leisure activities in advance.

[0017] In addition, the present invention provides a customized safety index according to the type of leisure activity through the analysis of weather and ocean elements specialized for each marine leisure sport and the application of weights, thereby minimizing instances where users abandon leisure activities or are exposed to danger due to weather conditions, and can provide a safe and satisfying marine leisure experience.

[0018] The present invention applies a region-specific algorithm that quantitatively reflects the unique environmental characteristics and risk factors of domestic sea areas, such as the West Sea, South Sea, East Sea, and Jeju, thereby providing accurate safety assessments differentiated by sea area even under the same weather conditions, and supporting the establishment of leisure activity plans optimized for the marine environment of each region. Brief explanation of the drawing

[0020] FIG. 1 is a diagram illustrating the overall structure of an artificial intelligence-based system according to one embodiment. FIG. 2 is a diagram illustrating the learning of a neural network according to one embodiment. FIG. 3 is a diagram illustrating the configuration of an artificial intelligence model according to one embodiment. FIG. 4 is a block diagram showing the configuration of an MME-based weather forecast observation system according to the present invention in one embodiment. Figure 5 is a flowchart illustrating a method for analyzing a marine leisure tourism activity index using an MME-based weather forecast observation system according to one embodiment. Figure 6 is a flowchart illustrating a method for analyzing a marine leisure tourism activity index using an MME-based weather forecast observation system according to one embodiment. Figure 7 is a flowchart illustrating a method for analyzing a marine leisure tourism activity index using an MME-based weather forecast observation system according to one embodiment. Specific details for implementing the invention

[0021] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0022] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to the specific disclosed forms, and the scope of this specification includes modifications, equivalents, or substitutions that fall within the technical concept.

[0023] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0024] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or coupled with that other component, or that there may be other components in between.

[0025] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0026] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0027] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiments, such detailed description is omitted.

[0028] The embodiments can be implemented in various forms of products such as personal computers, laptop computers, tablet computers, smartphones, televisions, smart home appliances, intelligent automobiles, kiosks, and wearable devices.

[0029] Artificial Intelligence (AI) systems are computer systems that implement human-level intelligence; unlike existing rule-based smart systems, they are systems in which machines learn and make decisions autonomously. As AI systems improve in recognition accuracy and gain a more accurate understanding of user preferences with continued use, existing rule-based smart systems are gradually being replaced by deep learning-based AI systems.

[0030] Artificial intelligence technology consists of machine learning and component technologies utilizing machine learning. Machine learning is an algorithmic technology that autonomously classifies and learns the characteristics of input data, while component technologies are technologies that mimic the cognitive and judgmental functions of the human brain by utilizing machine learning algorithms such as deep learning, and are comprised of technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control.

[0031] The various fields where artificial intelligence technology is applied are as follows. Linguistic understanding refers to technologies that recognize, apply, and process human language and text, including natural language processing, machine translation, dialogue systems, question answering, and speech recognition / synthesis. Visual understanding refers to technologies that perceive and process objects like human vision, including object recognition, object tracking, image search, people recognition, scene understanding, spatial understanding, and image enhancement. Inference and prediction refers to technologies that logically reason and predict by judging information, including knowledge / probability-based inference, optimization prediction, preference-based planning, and recommendation. Knowledge representation refers to technologies that automatically process human experiential information into knowledge data, including knowledge construction (data generation / classification) and knowledge management (data utilization). Motion control refers to technologies that control the autonomous driving of vehicles and the movement of robots, including motion control (navigation, collision, driving) and manipulation control (behavior control).

[0032] Generally, to apply machine learning algorithms to real-world situations, training is performed using a trial-and-error method due to the inherent characteristics of the fundamental methodologies. In particular, deep learning requires hundreds of thousands of iterations. Since it is impossible to execute this in a real physical external environment, training is instead performed through simulations that virtually recreate the actual physical environment on a computer.

[0033] In the present invention, Artificial Intelligence (AI) refers to a technology that imitates human learning ability, reasoning ability, and perceptual ability, and implements them on a computer, and may include concepts such as machine learning and symbolic logic. Machine Learning (ML) is an algorithmic technology that classifies or learns the characteristics of input data on its own. AI technology can analyze input data as a machine learning algorithm, learn from the results of the analysis, and make judgments or predictions based on the results of the learning. Furthermore, technologies that mimic the functions of the human brain, such as cognition and judgment, by utilizing machine learning algorithms can also be understood as falling within the category of AI. For example, technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control may be included.

[0034] Machine learning can refer to the process of training neural network models using experience in processing data. Through machine learning, computer software can improve its own data processing capabilities. A neural network model is constructed by modeling the correlations between data, and these correlations can be expressed by multiple parameters. A neural network model extracts and analyzes features from given data to derive correlations between them; machine learning can be defined as the process of optimizing the model's parameters by repeating this process. For example, a neural network model can learn the mapping (correlation) between inputs and outputs for data given as input-output pairs. Alternatively, even when only input data is provided, a neural network model can derive regularities between the given data and learn those relationships.

[0035] An artificial intelligence learning model or neural network model can be designed to implement the structure of the human brain on a computer and may include multiple network nodes that have weights and simulate neurons of a human neural network. The multiple network nodes may have interconnected relationships by simulating the synaptic activity of neurons, where neurons exchange signals through synapses. In an artificial intelligence learning model, multiple network nodes may be located in layers of different depths and exchange data according to convolutional connections. The artificial intelligence learning model may be, for example, an Artificial Neural Network (ANN) or a Convolutional Neural Network (CNN). As an embodiment, the artificial intelligence learning model may be machine learned according to methods such as supervised learning, unsupervised learning, and reinforcement learning. Machine learning algorithms for performing machine learning may include Decision Tree, Bayesian Network, Support Vector Machine, Artificial Neural Network, Ada-boost, Perceptron, Genetic Programming, and Clustering.

[0036] Among these, CNNs are a type of multilayer perceptron designed to use minimal preprocessing. CNNs consist of one or more convolutional layers and standard artificial neural network layers stacked on top, additionally utilizing weights and pooling layers. Thanks to this structure, CNNs can fully utilize two-dimensional input data. Compared to other deep learning architectures, CNNs demonstrate good performance in both image and audio fields. CNNs can also be trained using standard backpropagation. CNNs have the advantage of being easier to train than other feedforward artificial neural network techniques and using a small number of parameters.

[0037] Convolutional networks are neural networks comprising sets of nodes with bounded parameters. Many computer vision tasks have been significantly improved, driven by the increased size of available training data and the availability of computational power, combined with algorithmic advancements such as discriminative linear units and dropout training. In the case of massive datasets, such as those available for many tasks today, outfitting is not critical, and increasing the network size improves test accuracy. Optimal utilization of computing resources becomes a limiting factor. To address this, distributed, scalable implementations of deep neural networks can be employed.

[0039] FIG. 1 is a diagram illustrating the overall structure of an artificial intelligence-based system according to one embodiment.

[0040] As illustrated in FIG. 1, an artificial intelligence-based system (100) may include a plurality of user terminals (110-1, ..., 110-n), a server (120), and a database (130). This system adopts a distributed architecture and operates based on a client-server model, with each component playing a unique role and contributing to increasing the efficiency and scalability of the entire system. According to one embodiment, the database (130) is depicted as being configured separately from the server (120), but is not limited thereto; depending on system design and operational efficiency, the database (130) may be integrated within the server (120). This integrated configuration has the advantage of improving data access speed and reducing system complexity. For example, the server (120) may include a plurality of artificial intelligence models and processing units for performing machine learning algorithms, and these may implement various types of deep learning and machine learning technologies (e.g., CNN, RNN, Transformer, reinforcement learning, etc.) to respond to user requests and provide intelligent services. According to one embodiment, a plurality of user terminals (110-1,..., 110-n), a server (120), and a database (130) can be connected to communicate with each other through a network (N), which enables real-time data exchange and smooth interaction between system components.

[0041] According to one embodiment, the network (N) may perform wireless or wired communication between a plurality of user terminals (110-1, ... 110-n), a server (120), a database (130), etc., and is located in the center of FIG. 1 and serves as a hub connecting all components. For example, the network may perform wireless communication according to methods such as 5G, LTE (Long-Term Evolution), LTE-A (LTE Advanced), CDMA (Code Division Multiple Access), WCDMA (Wideband CDMA), WiBro (Wireless BroadBand), WiFi (Wireless Fidelity), Bluetooth, NFC (Near Field Communication), GPS (Global Positioning System), or GNSS (Global Navigation Satellite System). The 5G network supports high-speed data transmission (up to 20Gbps), ultra-low latency (1ms or less), and large-scale connectivity, making it suitable for real-time processing of large-capacity AI models. For example, the network (N) may be configured to perform wired communication using methods such as USB (Universal Serial Bus), HDMI (High Definition Multimedia Interface), RS-232 (Recommended Standard 232), Ethernet, fiber optic cable, or POTS (Plain Old Telephone Service). In particular, in data center environments, high-performance network technologies such as 400Gbps Ethernet or InfiniBand can be utilized for communication between large-scale AI computing clusters.

[0042] According to one embodiment, user terminals (110-1,..., 110-n) are various client devices that access the system and can be implemented in various forms such as smartphones, tablets, desktop computers, wearable devices, and IoT devices. These terminals are responsible for transmitting service requests through a user interface and receiving and displaying results processed by a server (120). Each terminal may also run a lightweight AI model locally, which can reduce network latency and enhance privacy protection. Direct connection or peer-to-peer (P2P) communication between terminals may be possible, as indicated by the dotted line, which is particularly useful in distributed learning or edge computing scenarios.

[0043] According to one embodiment, the server (120) is the central processing unit of the system, responsible for processing user requests, executing artificial intelligence models, and managing data. The server is equipped with a high-performance processor (CPU, GPU, TPU, NPU, etc.) for large-scale computation, large-capacity memory, and a stable operating system, and may be operated in a virtualized manner on a cloud infrastructure. The server performs training and inference functions for complex deep learning models and transmits the results to user terminals or stores them in a database. In addition, the stability and reliability of the system can be ensured through functions such as load balancing, error recovery, and security management.

[0044] According to one embodiment, a database (130) is a storage facility for storing and managing various data, and is depicted in a cylindrical shape in FIG. 1. Data stored in the database (130) is data acquired, processed, or used by at least one component of a plurality of user terminals (110-1, ..., 110-n) or a server (120), and may include software (e.g., programs), user profiles, training datasets, model weights, log records, etc. Structured relational data may be stored in a SQL-based database (MySQL, PostgreSQL, etc.), and unstructured large-scale data may be stored in a NoSQL database (MongoDB, Cassandra, etc.) or a distributed file system (Hadoop HDFS, etc.). The database (130) may include volatile and / or non-volatile memory, where volatile memory (RAM) is used for caching or temporary data processing requiring fast data access, and non-volatile memory (SSD, HDD, tape drive, etc.) is used for permanent data storage. Databases also ensure the security and integrity of data through data encryption, access control, and backup and recovery mechanisms.

[0045] This artificial intelligence-based system (100) can be utilized in various application fields and can provide services such as user behavior analysis, recommendation systems, natural language processing, image and video recognition, and predictive analysis. In particular, the distributed architecture and scalable design of the system enable stable performance to be maintained even as the number of users and data volume increase.

[0047] FIG. 2 is a diagram illustrating the learning of a neural network according to one embodiment.

[0048] As illustrated in FIG. 2, the learning device can train a neural network (123) to classify review responses received from a user terminal (110-1) by item. Additionally, the learning device can train a neural network (123) to extract user stay history from user movement path information. The neural network (123) is also called an artificial neural network and is a computational model designed inspired by the neural structure of the human brain, specialized in recognizing and learning complex patterns within data. According to one embodiment, the learning device may be a separate entity from the server (120), but it may also be implemented integrated into the same system, so it is not limited thereto.

[0049] According to one embodiment, the neural network (123) includes an input layer (121) into which training samples are input and an output layer (125) that outputs training outputs, and can be learned based on the difference between the training outputs and labels (i.e., actual correct data). Here, the labels are defined based on items corresponding to review responses (e.g., service quality, price satisfaction, cleanliness, etc.) and can be defined based on user dwell history (place visited, time spent, frequency of visit, etc.) corresponding to movement path information. The neural network (123) is connected as a group of multiple nodes and is defined by weights between the connected nodes and an activation function that activates the nodes. Various forms of the activation function may be used, such as sigmoid, hyperbolic tangent (tanh), and ReLU, which enables the network to learn non-linear patterns.

[0050] According to one embodiment, the learning device can train a neural network (123) using a Gradient Descent (GD) technique or a Stochastic Gradient Descent (SGD) technique. The GD technique is a method of updating weights all at once by calculating gradients based on the entire dataset, while the SGD technique is a method of increasing computational efficiency and the possibility of escaping a local optimum by updating weights more frequently using only a randomly selected portion of data (mini-batch). The learning device can use a loss function designed by the outputs and labels of the neural network. The loss function is an important element that quantifies the difference between the model's predicted value and the actual correct answer to suggest the direction of learning.

[0051] The learning device can calculate the training error using a predefined loss function. The loss function can be predefined with labels, outputs, and parameters as input variables, where the parameters can be set by weights within the neural network (123). For example, the loss function can be designed in the form of Mean Square Error (MSE), entropy, etc. MSE is calculated as the squared mean of the difference between the predicted value and the actual value and is mainly used for regression problems, while cross-entropy is a function that measures the difference between the predicted probability distribution and the actual distribution and is suitable for classification problems. In addition, loss functions that are less sensitive to outliers, such as Huber Loss, can be utilized, and various techniques or methods can be employed in the embodiments where the loss function is designed.

[0052] According to one embodiment, the learning device can identify weights that influence the training error using a backpropagation technique. Backpropagation is a process of calculating the degree to which each weight contributes to the final error while propagating the error calculated in the output layer toward the input layer, and is performed through differential calculations using the chain rule. Here, the weights are relationships between nodes within the neural network (123). The learning device may use an SGD technique using labels and outputs to optimize the weights found through the backpropagation technique. For example, the learning device may update the weights of a loss function defined based on labels, outputs, and weights using an SGD technique. This process operates by calculating the gradient (the derivative of the loss function with respect to the weights) and adjusting the weights in the direction of this gradient to gradually decrease the value of the loss function.

[0053] According to one embodiment, the learning device extracts first objects from a review response, obtains first labels which are items corresponding to the first objects, applies the first objects to a first neural network to generate first training outputs corresponding to the first objects, and can train the first neural network based on the first training outputs and the first labels. At this time, the first objects may include key keywords, sentence structures, sentiment expressions, etc. extracted from the review text, and thereby learn the ability to identify the core content of the review and classify it into the corresponding items.

[0054] According to one embodiment, the learning device extracts second objects from movement path information, obtains second labels which are user dwell history corresponding to the second objects, applies the second objects to a second neural network to generate second training outputs corresponding to the second objects, and can train the second neural network based on the second training outputs and the second labels. The second objects may consist of data such as the user's GPS coordinate sequence, movement speed, stopping point, and movement pattern, thereby learning the ability to determine where and for how long the user stayed.

[0055] According to one embodiment, a learning device can generate first training feature vectors based on constituent features (e.g., sentence structure, word frequency, part-of-speech distribution), positional features (e.g., location of important words, location of key sentences), and pattern features (e.g., repetitive expressions, usage patterns of specific phrases) of a review response. These feature vectors are generated through a process of converting raw text data into a numerical form that can be processed by a neural network, and various methods such as TF-IDF, Word2Vec, and BERT may be employed to extract features.

[0056] According to one embodiment, the learning device can generate second training feature vectors based on constituent features of movement path information (e.g., shape of the path, estimation of the means of movement), length features (e.g., total travel distance, travel time, distance between each point), and pattern features (e.g., repeated visit patterns, movement patterns by day of the week / time of the day). Since location data has continuous values ​​over time, time-series data processing techniques or spatial data analysis methodologies may be applied, and various methods may be employed to extract features.

[0057] According to one embodiment, the learning device can obtain training outputs by applying first training feature vectors to the neural network (123). This process is also called feedforward and is a process in which input data passes through each layer of the network to generate a final prediction value. The learning device can train the review item extraction algorithm of the neural network (123) based on the training outputs and first labels. The learning device can train the review item extraction algorithm of the neural network (123) by calculating training errors corresponding to the training outputs and optimizing the connection relationships of nodes within the neural network (123) to minimize the training errors. This process is repeated over several epochs, and in each epoch, the entire training dataset is processed and weights are updated. The server (120) can automatically extract and classify items from new review responses using the first neural network that has been trained.

[0058] According to one embodiment, the learning device can obtain training outputs by applying second training feature vectors to the neural network (123). The learning device can train the user stay history acquisition algorithm of the neural network (123) based on the training outputs and second labels. The learning device can train the user stay history acquisition algorithm of the neural network (123) by calculating training errors corresponding to the training outputs and optimizing the connection relationships of nodes within the neural network (123) to minimize the training errors. Since movement path data has temporal continuity, a recurrent neural network (RNN) structure such as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit) can be applied, which enables effective learning of the patterns of the sequence data. The server (120) can obtain user stay history from movement path information using the second neural network after training is complete, and the information thus obtained can be utilized in various application fields such as improving location-based services, building customized recommendation systems, and analyzing traffic patterns.

[0060] FIG. 3 is a diagram illustrating the configuration of an artificial intelligence model according to one embodiment.

[0061] An artificial intelligence model according to one embodiment may include an input layer, a hidden layer, and an output layer. This multi-layer structure is effective for learning complex patterns and solving non-linear problems, and is utilized as a core structure, particularly in deep learning models. The input layer is a layer related to input values ​​input into the artificial intelligence model and serves to receive data coming from the outside. As illustrated in FIG. 3, the input layer consists of three nodes, and each node may represent a specific feature or attribute of the data. The number of nodes in the input layer may vary depending on the dimension or characteristics of the data to be processed, and may be determined by the number of pixels in the case of image data, the dimension of the word vector in the case of text data, etc.

[0062] According to one embodiment, a feature map can be output by performing a MAC (multiply-accumulate) operation and an activation operation on the input values ​​in the hidden layer. In FIG. 3, the hidden layer consists of four nodes, and these nodes are connected to all nodes of the input layer to form a fully connected structure. The MAC operation may be an operation that multiplies the input values ​​by their corresponding weights and sums the multiplied values, and this acts as the basic unit of information processing in a neural network.

[0063] According to one embodiment, the activation operation may be an operation that inputs the result of a MAC operation into an activation function and outputs a result value. The activation function is a key element that enables a model to learn complex patterns by introducing non-linearity into the result of a linear operation. The activation function may be of various types. For example, the activation function may include a sigmoid function (which outputs a value between 0 and 1 and is useful for expressing probability), a tangent function (which outputs a value between -1 and 1 and is suitable for balanced data), a ReLU function (which outputs 0 for negative inputs and outputs it as is for positive inputs, mitigating the vanishing gradient problem), a Leaky ReLU function (which applies a linear function with a small gradient to negative inputs), a MaxOut function (which selects the maximum value among several linear functions), and / or an ELU function (which is an exponential linear unit function that provides a smooth saturation curve for negative inputs), but there are no limitations on the types thereof. Recently, new activation functions such as Swish and Mish are also being studied and are contributing to the improvement of model performance.

[0064] According to one embodiment, the hidden layer may be composed of at least one layer, and the more layers there are, the more complex patterns can be learned, resulting in a deep learning structure. For example, when the hidden layer is composed of a first hidden layer and a second hidden layer, the first hidden layer performs MAC operations and activation operations based on the input value of the input layer to output a feature map, and the feature map, which is the result value from the first hidden layer, can become the input value for the second hidden layer. The feature map is an intermediate representation representing features extracted from input data; in image processing, it can represent visual features such as edges, textures, and shapes, and in text processing, it can represent semantic patterns or syntactic structures. The second hidden layer can perform MAC operations and activation operations based on the feature map, which is the result value of the first hidden layer, and through this hierarchical processing, increasingly abstract and high-dimensional features can be extracted.

[0065] According to one embodiment, the output layer may be a layer associated with the result of an operation performed in the hidden layer, and in FIG. 3, it consists of three nodes. The number of nodes in the output layer varies depending on the type of problem to be solved; for binary classification problems, it may be one node, for multi-class classification problems, it may be equal to the number of classes, for regression problems, it may be determined by the dimension of the value to be predicted, and for sequence generation problems, it may be determined by the length of the output sequence. In the output layer, a special activation function such as a softmax function or a sigmoid function is generally used to generate a final predicted value or a probability distribution.

[0066] In one embodiment, the learning model learns syllable (character) patterns that are frequently combined and used in a given corpus to automatically learn the boundaries of compound words and named entities, integrates object information from a first UI source with object information rendered in a browser to create a learning object information file, and uses the learning object information file to generate training data for training a deep learning network. Natural language processing (NLP) technology is utilized in this process, and architectures such as Word Embedding, Recurrent Neural Network (RNN), and Transformer may be applied. Additionally, the learning model receives data from various domains of a support system, standardizes the data from the various domains into an integrated format based on at least one standardization method corresponding to each of the various domains, learns and infers data from a specific domain, determines information to be transmitted for standardization from the data of the specific domain, and performs post-processing on the data from the various domains.

[0067] According to one embodiment, the first UI source includes an XML file, and the training object information file includes an input JSON file for feature learning and an output JSON file that serves as label data during training. This structured data format allows the model to process it easily and is particularly useful for analyzing UI elements of web-based applications. The output JSON file includes a file containing HTML DOM Tree information implemented in compliance with web standards, and the various domains include at least one of a RAN (radio access network), a transport, or a core. A RAN refers to a wireless access network where a mobile device connects to a cellular network; a transport refers to a network layer responsible for data transmission; and a core refers to the central part of a network responsible for central processing and routing. Post-processing may include a correlation function, which plays an important role in analyzing relationships between data collected from various domains and discovering meaningful patterns. Through correlation analysis, the model can identify the causes of anomalies or performance degradation occurring in complex systems and improve prediction accuracy.

[0069] FIG. 4 is a block diagram showing the configuration of an MME-based weather forecast observation system according to the present invention in one embodiment.

[0070] A system (400) according to one embodiment may include a processor (420) and memory (430), and some of the illustrated components may be omitted or substituted. Although the system diagram briefly depicts only the core components, the actual implementation may include various additional components such as an input / output controller, a system bus, a graphics processing unit (GPU), a network interface card (NIC), and a storage device controller. A system (400) according to one embodiment may be a server or a terminal; if implemented as a server, it may be operated in a virtualized environment as part of a cloud infrastructure or implemented as a physical hardware server, and if implemented as a terminal, it may be implemented in various forms such as a desktop computer, laptop, tablet, smartphone, embedded system, or IoT device. According to one embodiment, the processor (420) is a component capable of performing operations or data processing regarding the control and / or communication of each component of the system (400), and may be composed of one or more processors. In modern processor architectures, multi-core designs are common, which improve parallel processing performance by integrating multiple independent processor cores within a single chip.

[0071] The memory (430) can store information related to the method described above or a program in which the method described above is implemented. The memory subsystem is designed with a hierarchical structure and is composed of several levels according to access speed and capacity. The memory (430) may be volatile memory or non-volatile memory. Volatile memory is memory in which stored data is lost when the power is turned off, and is mainly implemented in the form of DRAM (Dynamic RAM) or SRAM (Static RAM). DRAM stores each bit in the form of charge in a capacitor and requires periodic refreshing, whereas SRAM uses flip-flop circuits to maintain data, so it does not require refreshing but has higher cost and power consumption due to the use of more transistors. The memory (430) can store various file data, and the stored file data can be updated according to the operation of the processor (420). The memory management unit (MMU) is responsible for converting virtual memory addresses to physical memory addresses, managing page tables, and maintaining memory protection and cache consistency.

[0072] According to one embodiment, the processor (420) can execute a program and control the device (400). The processor can operate through basic pipeline stages of instruction fetch, decode, execute, and write-back. The code of the program executed by the processor (420) can be stored in memory (430). In addition to the application, the operating system, device driver, middleware, system services, etc. are loaded into memory and executed, and these provide basic system functions such as hardware resource management, process scheduling, interrupt handling, and file system management. The operations of the processor (420) can be performed by loading instructions stored in memory (430). In this process, the program counter (PC) points to the address of the next instruction to be executed, the instruction register (IR) stores the instruction currently being executed, and various general-purpose registers store operation data and intermediate results. The system (400) can be connected to an external device (e.g., a personal computer or a network) through an input / output device (not shown in the drawing) and exchange data. The input / output system can operate on an interrupt-based or polling-based basis and can transfer data directly between memory and the input / output device without the intervention of a processor through a Direct Memory Access (DMA) controller.

[0073] According to one embodiment, data exchange between memory (430) and processor (420) is performed via a system bus, which consists of an address bus, a data bus, and a control bus. The address bus specifies the memory location that the processor intends to access, the data bus transmits actual data, and the control bus transmits control information such as read / write signals. The system (400) may be operated under an operating system that supports advanced features such as multitasking, virtual memory, memory protection, and separation of privileges, and if a real-time operating system (RTOS) is used, it may satisfy deterministic response times and strict time constraints. For artificial intelligence applications, the memory (430) stores a neural network model, weight parameters, a training dataset, intermediate calculation results, etc., and the processor (420) may perform inference and learning by executing forward propagation and backpropagation algorithms. The system (400) can ensure high availability and fault tolerance by applying technologies such as clustering, load balancing, and failover mechanisms for scalability, and can implement technologies such as data encryption, access control, secure booting, and memory protection for security.

[0075] Figure 5 is a flowchart illustrating a method for analyzing a marine leisure tourism activity index using an MME-based weather forecast observation system according to one embodiment.

[0076] The operations described through FIG. 5 can be implemented based on instructions that can be stored in a computer recording medium or memory (e.g., memory (430) of FIG. 4). The order of each operation of FIG. 5 may be changed, some operations may be omitted, and some operations may be performed simultaneously.

[0077] In operation 510, the system (400) can collect and standardize data from multiple weather forecasting models. The system (400) can collect weather and ocean data, such as wind speed, wind direction, wave height, tides, water temperature, and precipitation, in real time from weather forecasting models based on different algorithms operated by national meteorological agencies. The system (400) can normalize the data structure by converting raw data provided in different data formats into a single unified standard format to enable subsequent processing. Through this, the system (400) can consistently utilize weather information from various sources.

[0078] In operation 520, the system (400) can generate corrected weather forecast data through an MME model. The system (400) can input multiple standardized forecast data and observation data such as ECMWF-ERA5 into a Multi-Model Ensemble (MME) algorithm to assign weights to the prediction values ​​of each model and calculate an ensemble average. Based on an analysis of past prediction accuracy, the system (400) evaluates the reliability of each model and reflects this in the calculation of weights to generate a new weather forecast model with improved prediction accuracy compared to a single model. Through this process, the system (400) can obtain more reliable weather forecast information.

[0079] In operation 530, the system (400) can provide an activity index for each marine leisure sport and an optimal recommendation. Based on corrected weather forecast data, the system (400) can calculate the influence of weather elements defined for each marine leisure sport, such as fishing, surfing, scuba diving, and yachting, and apply weights to calculate a comprehensive activity index. The system (400) classifies the calculated activity index into five levels: optimal, good, average, bad, and dangerous, and can determine the optimal marine leisure sport at a specific location selected by the user or recommend the optimal location for a specific sport. Although the classification is set to five levels, this is merely an example and may vary depending on the settings.

[0080] According to one embodiment, the system (400) collects weather data from a plurality of weather forecast models, converts the collected weather data into a standardized format, processes the standardized weather data using a Multi-Model Ensemble (MME) model to generate corrected weather forecast data, calculates an activity index for each marine leisure sport based on the corrected weather forecast data, and determines the optimal marine leisure sport at a specific location or determines the optimal location for a specific marine leisure sport based on the calculated activity index and provides it to the user.

[0081] According to one embodiment, when collecting weather data from a plurality of weather forecasting models, the system (400) may access different numerical forecasting systems operated by meteorological agencies in each country to obtain weather variables such as wind speed, wind direction, wave height, temperature, and humidity in real time, and may utilize HTTP API calls, FTP downloads, or dedicated data streaming protocols for this purpose. The system (400) may recognize the unique data formats and resolution differences used by each forecasting model and may include a retry mechanism and an alternative data source utilization function to prepare for network connection failures or data transmission errors.

[0082] According to one embodiment, the system (400) can analyze the metadata of each model to convert collected weather data into a standardized format and map variable names, units, coordinate systems, time formats, etc., to a unified standard, and in this process, an international standard format such as GRIB or NetCDF can be utilized. The system (400) can perform a function of interpolating data with different spatial resolutions into a common grid and a time alignment function of synchronizing time resolutions.

[0083] According to one embodiment, the system (400) can calculate dynamic weights based on past performance indicators of each prediction model when processing standardized weather data using an MME model, and can generate ensemble prediction values ​​by applying Bayesian model averaging or an optimal weight combination algorithm. The system (400) can perform adaptive weight adjustment by reflecting the characteristic that the reliability of each model varies depending on the prediction time and region, and can also provide uncertainty indicators by analyzing the variance of prediction values ​​between models.

[0084] According to one embodiment, the system (400) may refer to a weighting table for each weather element suitable for the characteristics of each sport to calculate an activity index for each marine leisure sport based on corrected weather forecast data, and may apply a mathematical model that calculates a score by matching real-time weather conditions with the corresponding weights and synthesizes them. The system (400) may additionally apply a correction coefficient reflecting regional characteristics of each sea area, and may derive a more accurate activity index by considering fluctuation factors according to season or time of day.

[0085] According to one embodiment, when determining the optimal marine leisure sport at a specific location based on a calculated activity index, the system (400) can identify the sport with the highest value by comparing the activity indices for all sports at that location, and in the event of a tie, apply a decision logic that prioritizes safety. Conversely, when determining the optimal location for a specific marine leisure sport, the system (400) can calculate and rank the activity index of the sport at all candidate locations within the area of ​​interest to recommend the top locations, and can also consider additional information such as accessibility and amenities. The system (400) can provide these analysis results and recommendation information to the user in an intuitive and easy-to-understand form through a web interface, mobile application, or API, and can utilize various presentation methods such as map visualization and graph charts. Here, the types of weather forecast models and marine leisure sports are merely examples and are not limited thereto, and may vary depending on the settings.

[0087] Figure 6 is a flowchart illustrating a method for analyzing a marine leisure tourism activity index using an MME-based weather forecast observation system according to one embodiment.

[0088] The operations described through FIG. 6 can be implemented based on instructions that can be stored in a computer recording medium or memory (e.g., memory (430) of FIG. 4). The order of each operation of FIG. 6 may be changed, some operations may be omitted, and some operations may be performed simultaneously.

[0089] In operation 610, the system (400) can collect national meteorological agencies and ECMWF-ERA5 data. The system (400) can collect raw data from different weather forecasting models provided by meteorological agencies of each country, including the Korea Meteorological Administration, in real time via API or data transmission protocol. The system (400) can additionally collect actual observed weather and ocean data, including global reanalysis data such as ECMWF-ERA5, and use them for verification and correction of the forecasting model. Through the collection of such diverse data, the system (400) can improve the accuracy of the forecast.

[0090] In operation 620, the system (400) can apply GRIB standard format conversion and an MME model. The system (400) can convert collected weather data of different forms into the GRIB (GRIdded Binary) standard format to unify the data structure and ensure interoperability. The system (400) can input standardized forecast data and observation data into the MME model to apply weights differentially based on the results of past performance analysis of each forecast model and calculate an ensemble average. Through this, the system (400) can generate a new corrected weather forecast model that complements the limitations of individual models and reduces forecast uncertainty.

[0091] In operation 630, the system (400) can define weather elements and assign scores for each marine leisure sport. The system (400) can define major weather and ocean elements, including wind speed, wind direction, wave height, wave period, current speed and direction, water temperature, turbidity, and precipitation, for each marine leisure sport, including fishing, boating, surfing, and scuba diving. The system (400) assigns an impact score ranging from -2 to +2 according to the degree of impact on the sport for each element, and stores this in a database to be used for calculating an activity index. Through this systematic score assignment, the system (400) can provide customized evaluation criteria that quantitatively reflect the characteristics of each sport. The range of the impact score and the types of weather and ocean elements are merely examples and may vary depending on the settings.

[0092] In operation 640, the system (400) can identify region-specific risk factors for each sea area and apply weights. The system (400) can identify region-specific risk factors, including current speed, tidal difference, swells, gusts, water depth, and turbidity, for each domestic sea area including the West Sea, South Sea, East Sea, and Jeju, and can database the environmental characteristics of each sea area. The types of sea areas are merely examples and may vary depending on the settings. Even for the same marine leisure sport, the system (400) can accurately reflect regional characteristics by applying differentiated weights for each sea area; for example, it can increase the current-related weight for the West Sea region by 30% compared to the reference value or adjust the wave height-related weight for the East Sea region upward. Through this, the system (400) can provide a more accurate safety assessment that takes into account the specificity of the marine environment for each region.

[0093] According to one embodiment, the system (400) collects raw data from different weather forecast models of national meteorological agencies and additionally collects actual weather data including ECMWF-ERA5 observation data, converts the collected data into a GRIB standard format to unify the data structure, inputs the standardized forecast data and observation data into an MME model to assign weights to the forecast values ​​of each model and calculates an ensemble average to create a new corrected weather forecast model, defines weather elements including wind speed, wind direction, wave height, wave period, current speed and direction, water temperature, turbidity, and precipitation for each marine leisure sport including fishing, boating, surfing, and scuba diving, assigns an impact score ranging from negative to positive for each element, identifies region-specific risk elements including current speed, tidal difference, swell, gust, water depth, and turbidity for each sea area including the West Sea, South Sea, East Sea, and Jeju, and controls to reflect regional characteristics by applying differentiated weights for each sea area even for the same leisure sport.

[0094] According to one embodiment, the system (400) stores weight mapping rule sets for leisure sports, sea areas, and weather elements in JSON format or in a database, matches them with weather and ocean data received in real time, automatically calculates a comprehensive activity index as the sum of values ​​obtained by multiplying the scores of each element by weights, compares the calculated comprehensive activity index with a plurality of preset threshold values ​​to classify it into a plurality of activity suitability stages including optimal, good, average, bad, and dangerous, measures underwater turbidity and reflects it in the snorkeling and scuba diving activity index, calculates the probability of swell wave occurrence by comprehensively analyzing wave direction and wave period as well as wave height, monitors gust conditions for each sea area in real time and reflects them in the activity index, calculates the activity index for each marine leisure sport based on real-time and predicted weather data of the location when a user selects a specific spot on a map and recommends the sport with the highest score as the optimal activity, or when a user selects a specific marine leisure sport, calculates the activity index for a plurality of spots and ranks the suitability to recommend the optimal location, and controls all of the analysis results and recommendation information to be provided to an external application through an API.

[0095] According to one embodiment, the system (400) can store weight mapping rule sets by leisure sport, sea area, and weather element in a structured JSON format or in a relational database, thereby enabling efficient management of complex multidimensional weight systems and rapid searching and updating. The system (400) can ensure data integrity through a JSON schema verification function and can track rule set change history and roll back to a previous version if necessary by applying a version control system. The system (400) can quickly connect weather and ocean data received in real time with corresponding weight rules using a hash table-based matching algorithm and improve processing performance by minimizing repetitive database access through memory caching.

[0096] According to one embodiment, the system (400) can derive a comprehensive activity index by calculating a weighted sum of values ​​obtained by multiplying the scores of each element by a weight, and can ensure stable calculation results by including exception handling logic for missing data or outliers in this process. When calculating the comprehensive activity index, the system (400) can model not only simple linear combinations but also non-linear relationships such as multiplication effects under specific conditions or minimum value constraints, thereby enabling realistic safety assessments. The system (400) can automatically perform a five-stage classification of optimal, good, average, bad, and dangerous by sequentially comparing the calculated comprehensive activity index with a predefined threshold array, and can adjust sensitivity to minute fluctuations by applying a probabilistic classification technique near the boundary value.

[0097] According to one embodiment, the system (400) may utilize turbidity sensor data in units of Nephelometric Turbidity Units (NTU) or a turbidity estimation algorithm through satellite image analysis for measuring underwater turbidity, and may provide underwater visibility information essential for snorkeling and scuba diving activities by modeling the inverse correlation between turbidity and underwater visibility distance. Underwater visibility distance may refer to the distance of visibility underwater. The system (400) may process wave height, wave direction, and wave period data using time-series analysis techniques to detect long-period waves and consistent wave directions, which are characteristic patterns of swell waves, and may quantify the probability of occurrence by calculating swell wave energy density through spectral analysis. The system (400) may perform particularly precise monitoring in the East Sea and Jeju regions, where the risk of swell waves is high, and may provide an early warning function by detecting the influence of distant typhoons or low-pressure systems in advance.

[0098] According to one embodiment, the system (400) can analyze the hourly rate of change of wind speed and the pattern of sudden change in wind direction to monitor conditions for gust occurrence by sea area in real time, and can predict the risk of gusts by applying a local weather model that considers pressure gradients and topographic effects. The system (400) can detect orographic gusts, such as mountain winds and valley winds, occurring particularly around mountainous terrain like Jeju Island, as well as sudden change phenomena at the point of transition between sea and land winds, and can ensure user safety by reflecting this information in the activity index in real time. When a user clicks or touches a specific point on the map interface, the system (400) can extract latitude and longitude information of the corresponding coordinates, collect current and predicted weather data at that location in real time, and calculate the activity index for each marine leisure sport in parallel.

[0099] According to one embodiment, the system (400) can identify the sport that recorded the highest score by comparing calculated activity indices, and in the event of a tie, apply tie-breaking logic that prioritizes the user's past preferences or safety, and can provide detailed scores for each sport and major influencing factors along with the recommendation results. Conversely, when a user selects a specific marine leisure sport, the system (400) can calculate activity indices based on a grid for a predefined area of ​​interest or the entire sea area, rank suitability using a sorting algorithm, and visually display the top recommended locations on a map. The system (400) can provide practical recommendations by also considering additional information such as travel distance, transportation, and local amenities, and can even suggest the optimal activity time by predicting changes over time.

[0100] According to one embodiment, the system (400) can provide all analysis results and recommendation information to external applications through a standard web service interface such as a RESTful API or GraphQL, and can ensure compatibility across various platforms by responding with structured data in JSON or XML format. The system (400) may include security and performance optimization features such as API key-based authentication, request frequency limiting, and data compression, and may provide detailed API documentation and SDKs so that external developers can easily integrate them. Here, the data storage format and API protocol are merely examples and are not limited thereto, and may vary depending on the settings.

[0102] Figure 7 is a flowchart illustrating a method for analyzing a marine leisure tourism activity index using an MME-based weather forecast observation system according to one embodiment.

[0103] The operations described through FIG. 7 can be implemented based on instructions that can be stored in a computer recording medium or memory (e.g., memory (430) of FIG. 4). The order of each operation of FIG. 7 may be changed, some operations may be omitted, and some operations may be performed simultaneously.

[0104] In operation 710, the system (400) can set the range of weather elements for each marine leisure sport and classify them into five stages. The system (400) can set multiple ranges for weather elements such as wind speed, wave height, precipitation, tide, water temperature, and visibility for each marine leisure sport and classify them into five stages: optimal, good, normal, bad, and dangerous. The system (400) can store threshold values ​​for each stage in a database and automatically classify the current state by comparing real-time weather data with the corresponding threshold values. Through this systematic classification, the system (400) can provide safety assessment information that the user can intuitively understand. The types and number of weather elements and stages are merely examples and are not limited thereto.

[0105] In operation 720, the system (400) can define an optimal weather condition range for each sport. The system (400) can optimally set a relatively low wind speed range (e.g., 0–5 m / s) for fishing, and an intermediate wind speed range (e.g., 8–12 m / s) and a specific wave height range (e.g., 1.5–2.5 m) for surfing and windsurfing. The system (400) can define a unique optimal weather condition range for each sport, including optimally setting a low wave height range (e.g., 0–0.5 m) and high underwater visibility (e.g., 10 m or more) for snorkeling and scuba diving. Through this, the system (400) can apply differentiated evaluation criteria tailored to the characteristics of each sport. The wind speed range, wave height range, and visibility are merely examples and are not limited to these, and may vary depending on the settings.

[0106] In operation 730, the system (400) can apply special adjustment elements specific to each sea area. For the West Sea region, the system (400) can increase the weight related to current speed by a specified level (e.g., 30%) relative to the reference value and add mudflat turbidity conditions, considering shallow water depth and large tidal differences; for the South Sea region, it can add current direction stability conditions, considering complex straits and currents around islands. For the East Sea region, the system (400) can increase the weight related to wave height by a specified level (e.g., 30%) relative to the reference value and add swell index conditions, considering deep water depth and swell wave occurrence; and for the Jeju region, it can add current-related weights and wind speed gust conditions, considering warm currents and gusts. Through these special adjustments specific to each sea area, the system (400) can accurately reflect the unique marine environmental characteristics of the region. The values ​​of the weights and reference values ​​are merely examples and may vary depending on the settings.

[0107] In operation 740, the system (400) can calculate a final activity index that reflects regional characteristics. The system (400) can calculate a final activity index for each marine leisure sport by comprehensively applying the previously defined optimal condition range for each sport and special adjustment elements for each sea area. The system (400) can apply a formula that collects real-time weather and ocean data, calculates a score for each element based on this data, and then calculates a total sum by multiplying it by a weight for each sea area. The system (400) can classify the calculated final activity index into five safety grades by comparing it with a preset threshold value and visually display this to the user to support safe marine leisure activity decision-making.

[0108] According to one embodiment, the system (400) sets multiple ranges for each weather element for each marine leisure sport and classifies them into optimal, good, normal, bad, and dangerous stages. It defines the optimal weather condition range for each sport, including setting a relatively low wind speed range optimally for fishing, setting a medium wind speed range and a specific wave height range optimally for surfing and windsurfing, and setting a low wave height range and high underwater visibility optimally for snorkeling and scuba diving. It also controls the calculation of an activity index that reflects regional characteristics by applying special adjustments for each sea area, such as increasing the weight related to current speed relative to a reference value and adding mudflat turbidity conditions in the case of the West Sea region considering shallow water depth and large tidal difference, adding current direction stability conditions in the case of the South Sea region considering complex straits and currents around islands, increasing the weight related to wave height relative to a reference value and adding swell index conditions in the case of the East Sea region considering deep water depth and swell wave occurrence, and adding current-related weights and wind speed gust conditions in the case of the Jeju region considering warm currents and gusts.

[0111] According to one embodiment, the system (400) divides underwater turbidity into multiple turbidity unit ranges, maps the underwater visibility distance corresponding to each range, applies an inverse correlation such that the underwater visibility distance increases as the turbidity decreases, classifies it into multiple stages from optimal to dangerous, and reflects this in the snorkeling and scuba diving activity index; comprehensively analyzes wave height, wave direction, and wave period data to calculate the possibility of swell wave occurrence when a pattern is detected in which the wave period is longer than a preset threshold and the wave height increases rapidly; if such pattern is detected in the East Sea and Jeju regions, the swell wave risk is additionally reflected in the activity index; and calculates the rate of change over time of wind speed and wind direction to determine as a gust condition when the wind direction changes by more than a preset angle within a short period of time or when the wind speed increases by more than a preset ratio relative to the average value.

[0112] According to one embodiment, the system (400) can measure underwater turbidity in units of Nephelometric Turbidity Units (NTU) and classify it into multiple ranges of 0 to 1, 1 to 2, 2 to 4, 4 to 6, and 6 or more, and can map the underwater visibility distance corresponding to each turbidity range based on a physical optical model. The system (400) can apply an inverse correlation equation in which the visibility distance decreases exponentially as the turbidity increases, such as setting the underwater visibility distance to 10m or more when the turbidity is 0 to 1 NTU and the underwater visibility distance to 1m or less when the turbidity is 6 NTU or more. The system (400) can reclassify the calculated underwater visibility distance into five stages—optimal, good, average, bad, and dangerous—and reflect it as a key variable when calculating the snorkeling and scuba diving activity index, and can particularly be utilized for safety evaluation of underwater sports where securing underwater visibility is important.

[0113] According to one embodiment, the system (400) can detect characteristic patterns of swell waves by processing wave height, wave direction, and wave period data using a time series analysis technique, and can set a pattern in which the wave period has a long period of 8 seconds or more and the wave height increases by more than 50% compared to the average value within a short period as a condition for the occurrence of swell waves. The system (400) can analyze the consistency and persistence of the wave direction to evaluate the likelihood of swell waves highly when waves coming from the same direction are maintained for a certain period of time or longer, and can detect an increase in low-frequency energy components in the wave spectrum through Fast Fourier Transform (FFT) analysis. The system (400) can enhance safety by applying an additional risk factor to the activity index of all marine leisure sports when such patterns are detected, by reflecting the regional characteristics in which swell waves frequently occur due to the influence of distant typhoons or low pressure systems, particularly in the East Sea and Jeju regions.

[0114] According to one embodiment, the system (400) can detect rapid weather changes in real time by calculating the rate of change of wind speed and wind direction over time at intervals of 1 minute or 5 minutes, and can set a threshold condition for the occurrence of a gust when the wind direction changes by more than 45 degrees within 10 minutes or when the wind speed increases by more than 30% compared to the average value. The system (400) can distinguish between simple measurement noise and actual weather changes by applying a moving average filter, and can more accurately predict the probability of a gust occurring by considering additional weather elements such as the rate of change of atmospheric pressure and the rate of change of temperature. The system (400) can separately model orographic gusts occurring around mountainous terrain such as Jeju Island and rapid change phenomena at the point of transition between sea and land breezes, and can ensure user safety by immediately raising the warning level for all marine leisure activities in the area when a gust risk is detected. Here, the turbidity unit range and threshold setting are merely examples and are not limited thereto, and may vary depending on the settings.

[0116] According to one embodiment, the system (400) can control the activity index by building a database of appropriate water temperatures for each fish species for fishing activities, matching it with the current water temperature, and reflecting the influence of current speed and direction on the activity of fish species; for surfing activities, evaluating the quality of waves by analyzing the consistency of wave period and wave direction as well as wave height, and increasing the activity index when there is a tailwind or crosswind condition; for yacht and boat activities, classifying it as a danger stage when visibility is below a preset threshold, and calculating the difficulty of navigation by analyzing the combination of current speed and wind direction, and generating specific risk factors and precautions as text along with the activity index calculated for each marine leisure sport and providing them to the user.

[0117] According to one embodiment, the system (400) may build a database containing optimal water temperature ranges for major fish species for fishing activities, and may store appropriate water temperature information based on the biological characteristics of each fish species, such as 15 to 20 degrees Celsius for flounder and 18 to 25 degrees Celsius for sea bass. Here, the types of fish species and water temperature ranges are merely examples and are not limited thereto, and may vary depending on the settings. The system (400) may calculate an activity index of a specific fish species by matching the current water temperature measured in real time with the database, and may apply a relationship in which the probability of successful fishing decreases as the water temperature deviates from the appropriate range. The system (400) may model the influence of current speed and direction on the feeding activity and movement patterns of fish species, and may dynamically adjust the fishing activity index by reflecting the characteristics that fish species activity increases at an appropriate current speed (e.g., 0.5 to 1.5 knots) and fishing conditions improve at the point of change in current direction.

[0118] According to one embodiment, the system (400) can evaluate the quality of waves by comprehensively analyzing the consistency of wave period and wave direction as well as the height of the waves for surfing activities, and can assign a high score when the wave period is maintained consistently within the range of 8 to 12 seconds and the wave direction approaches the coastline at an appropriate angle. The system (400) can apply a wave deformation model that considers water depth data and coastline shape to predict the shape and breaking pattern of the waves, and can identify clean wave conditions suitable for surfing. The system (400) can analyze wind direction conditions and adjust the activity index upward by reflecting that the wave surface becomes smooth and a favorable environment for surfing is created when there is an offshore wind or crosswind condition, and conversely, can consider the effect of waves becoming rough when there is an onshore wind condition.

[0119] According to one embodiment, the system (400) can automatically classify a danger level when the visibility for yacht and boat activities is below a specified level (e.g., 5 km). The system (400) can calculate the difficulty of navigation by analyzing a combination of current speed and wind direction, and can implement physical characteristics in a mathematical model where wave height increases rapidly and vessel maneuverability becomes difficult when the current and wind are in opposite directions. The system (400) can apply differential evaluation criteria based on the size and performance of the vessel and can provide a customized safety index by considering the difference in wave resistance and maneuverability between small boats and large yachts.

[0120] According to one embodiment, the system (400) can automatically generate text using a natural language generation algorithm containing specific risk factors and precautions expected under current weather conditions, along with an activity index calculated for each marine leisure sport, and can provide customized guidance messages such as, for example, "The current wave height is 2.5m, so it may be dangerous for beginner surfers; please accompany an experienced surfer" or "The current speed is 2 knots; please be careful when anchoring while fishing." The system (400) can adjust the intensity of the warning message according to the level of risk and can generate strong recommendation messages such as "Evacuate to land immediately" in urgent dangerous situations. The system (400) can also include situational precautions reflecting regional characteristics and seasonal factors, and can provide personalized safety guides by considering the user's experience level and preferences.

[0122] According to one embodiment, the system (400) constructs a dataset by matching the user's past marine leisure activity history data with actual weather condition data at the corresponding time in chronological order, analyzes the dataset to extract patterns of weather conditions preferred by each user to create a personalized preferred weather condition profile, classifies the user's proficiency level into multiple stages through a machine learning classification algorithm that analyzes the user's activity frequency and the difficulty of weather conditions during the activity, calculates a risk tolerance score according to each level to derive a personalized activity index adjustment parameter, stores real-time marine accident location and time information in a database in conjunction with the weather conditions at the corresponding time, statistically analyzes the correlation between specific weather patterns and accident occurrences to calculate the probability of accident occurrence by weather condition, automatically lowers the activity index of the corresponding sport if the calculated probability exceeds a preset risk threshold, and sends a risk alert message to the user's mobile device, analyzes time-based weather change prediction data to identify the sport optimal for each time period for a user who wishes to enjoy multiple marine leisure sports continuously, calculates the travel time and distance required for activity switching to determine the optimal sports combination and order, and generates and provides an integrated schedule including travel paths between each activity location. It can be controlled.

[0123] According to one embodiment, the system (400) can collect data on a user's past marine leisure activity history as information such as GPS logs, activity start and end times, and types of sports, and can construct an integrated dataset by matching actual weather condition data at that time with weather agency observation data and time stamps. The system (400) can accurately link the activity time and weather conditions using a time-series data alignment algorithm, and can apply interpolation and error correction functions to resolve issues of data omission or inconsistency. The system (400) can process the constructed dataset using clustering analysis and pattern recognition algorithms to identify the range of weather conditions, such as wind speed, wave height, and water temperature, preferred by each user, and save this as an individual preferred weather condition profile.

[0124] According to one embodiment, the system (400) can analyze the frequency of user activity by month / season and quantify the difficulty of weather conditions during activity as an objective indicator to use as input data for a machine learning classification algorithm, and can automatically classify users into proficiency levels such as beginner, intermediate, and expert through supervised learning techniques such as decision trees or support vector machines. The system (400) can calculate risk-taking propensity scores according to each proficiency level through statistical analysis, and based on this, can derive activity index adjustment parameters differentiated for each user even under the same weather conditions to provide personalized safety evaluation criteria.

[0125] According to one embodiment, the system (400) can accumulate information on the location and time of marine accidents collected in real time through the Korea Coast Guard, the Ministry of Oceans and Fisheries, media reports, etc., in a database, and can construct an integrated accident analysis dataset by linking weather condition data at the time of the accident. The system (400) can quantify the causal relationship between specific weather patterns (e.g., wave height 3m or more, wind speed 15m / s or more) and the occurrence of accidents by applying statistical methods such as regression analysis and correlation analysis, and can calculate the probability of accident occurrence for each weather condition through a Bayesian probability model. If the calculated accident probability exceeds a preset risk threshold (e.g., 10%), the system (400) can automatically lower the activity index of the corresponding sport and immediately send a risk warning message to the user's mobile device through a push notification service.

[0126] According to one embodiment, the system (400) can identify the optimal sport for each time interval by analyzing time-based weather forecast data for a user who wishes to enjoy multiple marine leisure sports continuously, and can predict changes in conditions such as fishing in the morning and surfing in the afternoon. The system (400) can calculate travel time and distance between activity locations in real time by linking a geographic information system and a traffic information API, and can provide an accurate travel time prediction function that considers traffic congestion, road conditions, and means of transportation. The system (400) can determine the optimal combination and order of sports considering the user's time constraints and preferences by applying optimization techniques such as genetic algorithms or dynamic programming, and can automatically generate detailed travel plans and timetables, including the shortest route or scenic route between each activity location, and provide them in the form of an integrated schedule. Here, the classification of skill levels and accident data sources are merely examples and are not limited thereto, and may vary depending on the settings.

[0128] According to one embodiment, the system (400) calculates the difference between a predicted value at a specific past point in time and an actual observed value at the same point in time to derive an error value and stores the error values ​​as a time series to verify the prediction accuracy of the MME model; analyzes the temporal and spatial patterns of the error to identify prediction biases that repeatedly appear in specific weather conditions or specific regions; calculates a correction coefficient according to the magnitude and direction of the identified bias and updates adaptive weights applied to the predicted value of the corresponding condition or region in real time; receives data including acceleration, gyroscope, and heart rate from a sensor device worn by a user engaged in marine leisure activities in real time; analyzes the user's heart rate variability and skin electrical response collected from a wearable device to quantify the actual activity intensity and stress level; calculates the degree of agreement between the predicted value and the actual value by comparing this with the activity index predicted by the system; adjusts the parameters of the prediction model if the degree of agreement is lower than a specified level; collects input from users via a mobile application regarding marine conditions and activity satisfaction, including wave height, wind speed, and underwater visibility actually experienced by users, on a 5-level scale; utilizes the collected data as a verification dataset for the prediction model; and [in a specific region or specific weather] If the difference between user feedback and the predicted value in a condition consistently exceeds a preset threshold, the weights and algorithm coefficients for that condition can be automatically readjusted to improve prediction accuracy.

[0129] According to one embodiment, the system (400) can systematically compare predicted values ​​generated, for example, 24 hours, 48 ​​hours, or 72 hours prior with actual observed values ​​at that time to verify the prediction accuracy of the MME model, and can calculate quantitative error values ​​using statistical indicators such as Mean Absolute Error (MAE) or Root Mean Square Error (RMSE). The system (400) can store the calculated error values ​​in a database in chronological order to construct time series data, thereby allowing tracking of the temporal change trend and seasonal patterns of prediction performance. The system (400) can automatically identify prediction biases that appear repeatedly under specific weather conditions (e.g., when a typhoon approaches) or specific regions (e.g., complex coastline regions) by applying time series analysis and spatial statistical techniques, and can quantify the magnitude and direction of the bias to dynamically calculate correction coefficients by region and condition.

[0130] According to one embodiment, the system (400) can execute an adaptive weight adjustment algorithm based on identified prediction bias and can update weights in real time whenever new observation data is collected using Bayesian updates or Kalman filtering techniques. The system (400) can establish a multi-layer correction system that reflects regional characteristics and weather pattern characteristics, and can prioritize safety by applying conservative weights under conditions of high prediction uncertainty. The system (400) can receive data from a 3-axis accelerometer, gyroscope, and heart rate sensor in real time via Bluetooth or Wi-Fi from a smartwatch or fitness tracker worn by a user during marine leisure activities, thereby objectively measuring the user's actual physical activity intensity and movement patterns.

[0131] According to one embodiment, the system (400) can utilize advanced sensors of a wearable device to measure autonomic nervous system activity through heart rate variability (HRV) analysis and evaluate stress levels through skin electrical response (GSR) analysis, and can quantify biosignals into activity intensity and stress levels by applying machine learning algorithms. The system (400) can calculate correlation coefficients or agreement indicators by comparing and analyzing biosignal data collected in real time with activity indices predicted in advance by the system, and can quantitatively evaluate the difference between the predicted safety level and the activity intensity experienced by the actual user. The system (400) can perform a self-learning function that automatically readjusts weight parameters of the prediction model or threshold settings when the agreement falls below a preset threshold (e.g., 80%).

[0132] According to one embodiment, the system (400) can receive input from users through a user interface of a mobile application regarding marine conditions, such as wave height, wind speed, and underwater visibility, and overall activity satisfaction experienced directly by users after the activity ends, on a scale from 1 point (very dissatisfied) to 5 points (very satisfied), and can utilize this as verification data in a crowdsourcing manner. The system (400) can preprocess the collected user feedback data to remove outliers and select only reliable data, and can identify systematic differences between system prediction values ​​and user experience values ​​through statistical analysis. If a pattern is detected in which the difference between user feedback and prediction values ​​continuously exceeds a threshold value in a specific region or under specific weather conditions, the system (400) can automatically readjust the weight matrix for the corresponding condition and the coefficients of the activity index calculation algorithm using optimization techniques such as gradient descent or genetic algorithms, thereby achieving continuous improvement in prediction accuracy and user satisfaction. Here, the sensor type and feedback scale are merely examples and are not limited thereto, and may vary depending on the settings.

[0134] According to one embodiment, the system (400) analyzes weather data over a specified period to learn the trend of change in weather patterns by season and time of day, generates a time-dependent weighting function that applies differentiated weights according to season and time of day even under the same weather conditions, particularly calculates the rate of change in visibility during the time of sunrise and sunset, statistically models the probability of local gusts occurring during the afternoon hours in summer, calculates the risk of freezing due to a drop in water temperature during the dawn hours in winter as a temperature function and reflects it in the activity index, performs vector operations to calculate the rate of increase in wave height when the wind speed vector and the current vector have opposite directions, applies a regression model to calculate the rate of decrease in visibility when the combination of temperature and humidity is within a specific range, calculates the gust risk when the rate of change in atmospheric pressure and the rate of change in wind direction simultaneously exceed a threshold, builds a biological risk factor database by collecting data on the location and concentration of red tide occurrence, data on the frequency and types of jellyfish appearance, and plankton density data from a marine ecosystem monitoring system, quantifies the impact on marine leisure sports for each biological risk factor, and in the case of snorkeling and scuba diving, marine organisms The risk level can be controlled to be calculated as a separate sub-index and provided along with the overall activity index.

[0135] According to one embodiment, the system (400) can learn the trends of change in weather patterns by season (e.g., spring, summer, autumn, winter) and by time of day (e.g., dawn, morning, afternoon, evening) by processing weather data over a specified period (e.g., the last 5 years) using a time series analysis technique, and can automatically generate a time-dependent weighting function that reflects the characteristics that safety assessments vary depending on the season and time of day, even under the same wind speed or wave height conditions, through a machine learning algorithm. In particular, the system (400) can monitor rapid changes in visibility caused by changes in atmospheric water vapor and temperature during specified times (e.g., 1 to 2 hours) before and after sunrise and sunset, and can quantify the hourly rate of change in visibility through differential calculations to reflect it in the safety assessment of yacht or boat activities. The system (400) can model the probability of occurrence of convective gusts caused by land heating during the afternoon hours (14-18 hours) in summer through statistical analysis of past data, and can apply a probability distribution function with variables such as temperature, humidity, and solar radiation as inputs.

[0136] According to one embodiment, the system (400) can calculate the risk of freezing due to a rapid drop in water temperature during the early morning hours (04-08 hours) in winter using an exponential or logistic function with the difference between air temperature and water temperature as a variable, and can also consider the effect of raising the freezing point due to a decrease in salinity concentration, especially in coastal areas with a large inflow of freshwater. The system (400) can calculate the rate of increase in wave height based on physical wave theory when the two vectors are in opposite directions (negative dot product value) by calculating the vector dot product of the wind speed vector and the current vector in a two-dimensional coordinate system, which may be a quantification of the wave steepening phenomenon caused by the interaction of wind and current. The system (400) can apply a multiple regression model that predicts a decrease in visibility due to the occurrence of fog or mist when the combination of air temperature and relative humidity is in a specific range (e.g., air temperature 10-20 degrees Celsius, humidity 80% or higher), thereby allowing for the prior evaluation of the risk to marine activities due to deterioration in visibility.

[0137] According to one embodiment, the system (400) may apply a composite indicator model that calculates the rate of change in atmospheric pressure (hPa / hour) and the rate of change in wind direction (degrees / hour) in real time and evaluates the risk of gusts as high when both indicators simultaneously exceed a preset threshold (e.g., rate of change in atmospheric pressure of 3 hPa / hour or higher, rate of change in wind direction of 30 degrees / hour or higher), which may detect rapid weather change patterns that occur when a front passes or a trough approaches. The system (400) may build a database of biological risk factors by sea area by regularly collecting data on the location of red tide occurrence and chlorophyll concentration, information on the frequency of jellyfish appearance and types of toxicity, and density data of phytoplankton and zooplankton from marine ecosystem monitoring systems such as the National Institute of Fisheries Science and the Korea Marine Environment Management Corporation. The system (400) may quantify the impact of each biological risk factor on marine leisure sports through expert opinions and analysis of past accident cases, and may evaluate, for example, the risk of skin contact for toxic red tide, the risk of stinging for large jellyfish, and the water quality pollution for harmful plankton.

[0138] According to one embodiment, the system (400) can calculate a marine life risk level as a separate sub-index, taking into account the high possibility of direct contact with marine life, particularly in the case of underwater activities such as snorkeling and scuba diving, and can provide this along with a weather condition-based main activity index so that the user can check comprehensive safety information. The system (400) can intuitively display the marine life risk level sub-index by visualizing it with color coding (green: safe, yellow: caution, red: danger) or separate icons, and can also provide detailed information and countermeasures for each risk factor. Here, the distinction between seasons and time zones and the types of biological risk factors are merely examples and are not limited thereto, and may vary depending on the settings.

[0140] According to one embodiment, the system (400) executes an equipment recommendation algorithm using a calculated activity index and predicted weather conditions as input values ​​to map the buoyancy grade of a buoyancy vest suitable for a given range when the wave height falls within a preset range, calculates the thickness of a wetsuit capable of maintaining body temperature at a given temperature based on the water temperature falling within a specific range, calculates the sail size capable of stable navigation based on the wind speed falling within a specific range, determines the luminous intensity of the necessary lighting equipment when visibility is below a specific value, and automatically generates an equipment configuration optimized for each marine leisure sport; when real-time weather changes are detected and the risk level rises, it searches for nearby safe points based on the user's current GPS coordinates and the type of sport being engaged in, calculates the risk level during movement by simulating the distance to each safe point and predicted weather conditions along the movement path, selects the path with the lowest risk as the optimal evacuation path, generates and provides the estimated time required to move along the path and precautions for each section, collects location information and activity status of users active in the same area in real time to form a proximity user network, and transmits information on local weather changes or dangerous situations detected through one user's sensor or input to other users within the network It automatically sends push notifications to users and statistically processes weather information collected from multiple users to identify highly reliable subtle weather changes and controls the real-time reflection of these in the activity index calculation.

[0141] According to one embodiment, the system (400) can execute an equipment recommendation algorithm using a calculated activity index and current weather conditions as input values, and can utilize a buoyancy class mapping table that meets safety standards for each wave height range, such as recommending a 50N class buoyancy vest when the wave height is in the range of 1-2m and a 100N class buoyancy vest when the wave height is in the range of 2-3m. The system (400) can calculate the required thickness of a wetsuit by applying a body heat loss equation based on water temperature data, and can use a temperature-specific thermal insulation effect model that recommends, for example, a wetsuit with a thickness of 5mm when the water temperature is 15 degrees Celsius or lower and a thickness of 3mm when the water temperature is 20 degrees Celsius or higher.

[0142] According to one embodiment, the system (400) determines an emergency situation when the activity index drops sharply or exceeds a risk threshold through real-time weather monitoring, and can search for safe points (ports, beaches, rescue bases, etc.) within a radius of 10-20 km centered on the user's current GPS coordinates in a geographic information system database. The system (400) can calculate the straight-line distance to each safe point and the actual travel path, and can evaluate additional risk factors (strong winds, high waves, etc.) during travel by simulating predicted weather conditions on the travel path, and can automatically select the optimal evacuation path with the lowest risk by applying shortest path search techniques such as the Dijkstra algorithm. The system (400) can calculate the estimated time required for each means of transportation (e.g., walking, vehicle, small boat) for the selected evacuation path, and can automatically generate and provide a detailed evacuation guide including expected weather risk factors and countermeasures for each section of the path.

[0143] According to one embodiment, the system (400) can form a proximity user network by collecting GPS location information of users active in the same sea area in real time and automatically grouping users within close proximity (e.g., within 5 km), and can apply a location-based clustering algorithm for this purpose. When the system (400) detects a sudden weather change (e.g., sudden increase in wind speed, rise in wave height) from a sensor device worn by a user or when a user manually reports a dangerous situation, the system (400) can immediately transmit the information to other users in the network via a push notification service, and the information may include detailed information such as location, type of danger, and urgency. The system (400) can evaluate reliability by cross-validating weather information collected from multiple users, and when multiple users report similar weather changes in the same area, it can identify highly reliable micro-weather changes through weighted average or median calculation.

[0144] According to one embodiment, the system (400) can remove measurement errors and noise by processing real-time weather information collected through user crowdsourcing using Bayesian filtering or Kalman filtering techniques, and can recalculate a more accurate local activity index in real time by combining verified micro-weather change information with existing weather agency data. To evaluate the quality of information collected through the user network, the system (400) can score each user's past reporting accuracy and use it as a weight, and can implement a system that assigns greater weight to information from users with high reliability. Through such real-time information sharing and collaborative monitoring, the system (400) can complement the spatial limitations of existing weather observation networks and significantly improve the safety of marine leisure activities. Here, the types of equipment and safety points are merely examples and are not limited thereto, and may vary depending on the settings.

[0146] According to one embodiment, the system (400) executes an anomaly detection algorithm that compares patterns learned from past weather data with real-time data currently being collected to identify early warning signals of weather patterns or extreme weather phenomena that deviate from the normal range, increases the safety factor when an anomaly is detected that deviates from the prediction confidence interval of the existing MME model, and conservatively recalculates all marine leisure activity indices, analyzes sea surface temperature distribution images received from a satellite in pixel units to calculate the temperature gradient, identifies the location of the sea-land breeze boundary, detects the intersection point where cold and warm currents meet, recognizes the vortex generation pattern around an island to classify such areas as high-risk zones, applies an additional reduction factor that reflects uncertainty when calculating the activity index of the zone, trains a neural network model that takes long-term accumulated user activity history data and accident data as input to classify the user's experience level into beginner, intermediate, and expert, generates a personalized safety threshold that sets different acceptable risk levels for each level even under the same weather conditions, and controls to generate and provide differentiated activity indices and detailed safety recommendations tailored to the user's level.

[0147] According to one embodiment, the system (400) can build a machine learning model that learns the range and variability of normal weather patterns from weather data accumulated over the past several years, and can execute a statistical anomaly detection algorithm (e.g., Isolation Forest, LSTM-based Anomaly Detection) by comparing currently collected real-time data with the model. The system (400) can identify an anomaly pattern when weather variables such as wind speed, wave height, and atmospheric pressure exceed 2-3 times the standard deviation compared to the same period in the past or show a rapid rate of change, and can detect early warning signals of extreme weather phenomena such as typhoons and sudden storms. The system (400) can monitor the prediction confidence interval provided by the existing MME model in real time, and if an anomaly is detected where the observed value deviates from the interval, it determines that the prediction uncertainty has increased and performs a conservative evaluation by applying a safety factor in the range of 0.8-0.9 to all marine leisure activity indices. The criteria for identifying anomaly patterns are merely examples and may vary depending on the settings.

[0148] According to one embodiment, the system (400) can analyze sea surface temperature (SST) distribution images received from a weather satellite or an ocean observation satellite using digital image processing techniques, and can perform differential operations to calculate spatial temperature gradients by extracting temperature values ​​of each pixel. The system (400) can recognize areas where the temperature gradient changes rapidly as boundaries to automatically identify areas where sea and land breezes occur or fronts where different water masses meet, and in particular, can detect characteristic temperature patterns appearing at the intersection of cold and warm currents using a pattern matching algorithm. The system (400) can recognize vortices or eddy patterns occurring around islands through vorticity calculation or streamline analysis, automatically classify areas where such complex ocean dynamic phenomena occur as high-risk zones, and apply an additional reduction factor in the range of 0.7-0.8 when calculating the activity index of the zone to reflect uncertainty.

[0149] According to one embodiment, the system (400) can build a deep neural network model using long-term accumulated user-specific activity history data (activity frequency, activity time, selected weather conditions, etc.) and marine accident occurrence data (accident type, occurrence conditions, user characteristics, etc.) as input variables and train it using a supervised learning method. The system (400) can analyze the activity patterns of new users through the trained neural network model and automatically classify them into experience levels such as beginner (e.g., less than 1 year of experience), intermediate (e.g., 1-5 years of experience), and expert (e.g., 5 years or more of experience), and can also provide a confidence score for each classification result. The system (400) can differentiate and set acceptable risk levels for each experience level even under the same weather conditions, and can generate a personalized safety threshold that classifies wave heights of 1.5m or more as risk for beginners and wave heights of 3m or more as risk for experts. The safety threshold is merely an example and may vary depending on the settings.

[0150] According to one embodiment, the system (400) can calculate differentiated activity indices even under the same weather conditions according to the user's experience level, and can provide a more conservative evaluation to beginners and a relatively lenient evaluation to experts. The system (400) can automatically generate safety recommendations specialized for each experience level using natural language generation technology. For example, the system (400) can provide detailed precautions to beginners, such as "Accompanied by an experienced person" and "Wear safety equipment," and concise information to experts, such as "Advanced skills are required under current conditions." The types of information are merely examples and are not limited thereto, and may vary depending on the settings.

[0151] According to one embodiment, the system (400) can continuously collect the results of the user's actual activities and feedback to update the individual experience level classification model, thereby dynamically reflecting the user's proficiency as it improves over time. Here, the distinction between experience levels and safety thresholds are merely examples and are not limited thereto, and may vary depending on the settings.

[0153] According to one embodiment, the system (400) can continuously monitor the prediction accuracy within the last 72 hours by applying a sliding window technique to evaluate the real-time performance of each weather prediction model, and can implement an adaptive ensemble system that automatically reduces the weight for models with increasing prediction errors and increases the weight for models with improved accuracy. The system (400) can provide individual model correction functions, such as learning the systematic bias patterns exhibited by each model according to weather conditions using a deep learning algorithm, and adjusting the wind speed prediction value of model A upward by 10% when a typhoon approaches, and adjusting the wave height prediction value of model B downward by 15% when a high pressure is under influence.

[0154] According to one embodiment, the system (400) can quantify the current stress level and physical condition by analyzing biosignals such as heart rate, skin temperature, and galvanic skin response collected in real time from the user's wearable device, and can reflect this as an individual correction factor when calculating an activity index. For example, when the stress index is high, the risk level can be adjusted upward even under the same weather conditions to recommend safer activities. The system (400) can learn the individual risk tolerance and preferred weather conditions by analyzing the user's past activity selection patterns using a reinforcement learning algorithm, and can implement a situation-aware personalization system that applies different recommendation criteria depending on the purpose of the activity, the presence or absence of a companion, and the time of day, even for the same user.

[0155] According to one embodiment, the system (400) can execute an optimal scheduling algorithm that considers time-based weather change predictions and transition times between activities for a user who wishes to enjoy multiple marine leisure sports continuously, and can automatically generate an optimal activity sequence that comprehensively considers the user's physical exhaustion, equipment preparation time, and travel distance by applying a genetic algorithm or particle swarm optimization technique. The system (400) can dynamically calculate travel time between activity locations by linking with a real-time traffic information API, predict the possibility of delays due to traffic congestion or adverse weather conditions, and provide schedule adjustment notifications, and can compare and analyze optimal routes and travel times for each mode of transportation, such as public transportation, private cars, and rental cars.

[0156] According to one embodiment, the system (400) can collect ultra-high-resolution fine weather data at 1-minute intervals from IoT weather sensor networks distributed regionally, thereby compensating for the spatial resolution limitations of existing weather observation networks and precisely detecting local weather changes. The system (400) can directly measure real-time weather conditions in dangerous waters or hard-to-access areas by utilizing unmanned drones or autonomous navigation marine buoys, and can continuously improve prediction accuracy by utilizing such field verification data for real-time correction of the MME model. The system (400) can collect perceived weather information (actually felt wind speed, wave height, water temperature, etc.) from users engaged in marine leisure activities through a crowdsourcing method, and build a correlation model between human subjective perception and objective measurements by comparing and analyzing this with official weather data.

[0157] According to one embodiment, the system (400) can automatically analyze real-time wave conditions, changes in seawater color, and coastline conditions from coastal-installed CCTVs or satellite images using computer vision technology, and can convert visually observed marine conditions into quantitative indicators through an image processing algorithm and reflect them in the calculation of activity indices. The system (400) can detect abnormal weather patterns or marine conditions different from normal conditions at an early stage through a machine learning-based anomaly detection system, and can provide a safety management system that automatically recalculates all activity indices conservatively and immediately sends a warning message to the user in the event of sudden weather changes or a sharp drop in the reliability of the prediction model.

[0158] According to one embodiment, the system (400) can provide an immersive user experience that visualizes weather information such as wave height, wind speed, and tide direction in real time by linking with augmented reality (AR) technology when a user points a smartphone camera at the coast, and intuitively displays the suitability of each marine leisure sport through color coding.

[0159] According to one embodiment, the system (400) can utilize blockchain technology to store and verify actual weather information and activity reviews provided by users in a decentralized manner, and can simultaneously improve the quality and participation of crowdsourced data through an incentive system that pays token rewards to users who provide accurate information. The system (400) can record users' activity history and safety records on the blockchain and link with marine leisure insurance companies, and can implement smart contract functions that provide insurance premium discounts to users who comply with safety recommendations provided by the system or enable rapid compensation processing in the event of an accident.

[0160] According to one embodiment, the system (400) can build a deep learning-based prediction model to learn the individual user's past activity patterns, preferences, and physical fitness levels, and predict the optimal activity time and intensity based on this. The system (400) can provide a personalized detailed guide, for example, "Based on your patterns, surfing conditions are expected to be best around 10:00 AM tomorrow, and we recommend you be active for about 2 hours." For users planning group activities, the system (400) can recommend an optimal common denominator activity that all participants can enjoy safely and satisfactorily by comprehensively analyzing each member's experience level and preferences, and can provide a group safety management function that applies a conservative evaluation based on the safety standards of the weakest member in the group.

[0161] According to one embodiment, the system (400) can recommend activity methods and time periods that can minimize the impact on the marine ecosystem by reflecting an environmental protection perspective, and can induce sustainable marine leisure activities, such as suggesting information on fishing ban zones to protect spawning fish or yacht routes that avoid the migration routes of marine mammals. The system (400) can collect real-time status information from the user's wetsuit, life vest, fishing rod, etc. by linking with smart marine equipment equipped with IoT sensors, and can provide a predictive maintenance service that detects signs of equipment performance degradation or failure in advance to prevent safety accidents and informs the user of the timing for replacing equipment for optimal performance.

[0162] According to one embodiment, the system (400) can provide customized safety guidance by taking into account the user's health status and medication information, and can perform a personalized risk assessment from a medical perspective, such as recommending that a person with heart disease avoid activities involving sudden changes in water temperature, or restricting boat activities in high wave conditions for a person taking motion sickness medication. The system (400) can utilize machine learning to analyze the correlation between specific weather conditions and user behavior patterns and can provide advice based on individual behavior prediction.

[0164] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, 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 unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0165] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, 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 unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0166] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave in order to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0167] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

Claims

Claim 1 A marine leisure tourism activity index analysis system utilizing an MME-based weather forecast observation system comprises a memory and a processor for storing instructions, wherein, when the instructions are executed by the processor, the system collects weather data from multiple weather forecast models, converts the collected weather data into a standardized format, processes the standardized weather data using a Multi-Model Ensemble (MME) model to generate corrected weather forecast data, calculates an activity index for each marine leisure sport based on the corrected weather forecast data, determines the optimal marine leisure sport at a specific location or determines the optimal location for a specific marine leisure sport based on the calculated activity index and provides it to the user, in the case of fishing activities, constructs a database of appropriate water temperatures for each fish species to match with the current water temperature and adjusts the activity index by reflecting the influence of current speed and direction on the activity of fish species, in the case of surfing activities, evaluates wave quality by analyzing the consistency of wave period and direction as well as wave height and adjusts the activity index upward under tailwind or crosswind conditions, and in the case of yacht and boat activities, classifies it as a danger stage if visibility is below a preset threshold. A system that calculates navigation difficulty by analyzing a combination of current speed and wind direction, and controls the generation of specific risk factors and precautions as text, along with activity indices calculated for each marine leisure sport, to provide to the user. Claim 2 In claim 1, the instructions, when executed by the processor, cause the system to collect raw data from different weather forecast models of national meteorological agencies and additionally collect actual weather data including ECMWF-ERA5 observation data, convert the collected data into the GRIB standard format to unify the data structure, input the standardized forecast data and observation data into an MME model to assign weights to the forecast values ​​of each model and calculate an ensemble average to generate a new corrected weather forecast model, define meteorological elements including wind speed, wind direction, wave height, wave period, current speed and direction, water temperature, turbidity, and precipitation for each marine leisure sport including fishing, boating, surfing, and scuba diving, and assign an impact score ranging from negative to positive to each element, identify region-specific risk elements including current speed, tidal difference, swell, gust, water depth, and turbidity for each sea area including the West Sea, South Sea, East Sea, and Jeju, and control the system to reflect regional characteristics by applying differentiated weights by sea area even for the same leisure sport. System. Claim 3 In claim 2, the above instructions, when executed by the processor, control the system to store weight mapping rule sets for leisure sports, sea areas, and weather elements in JSON format or in a database, match them with weather and ocean data received in real time to automatically calculate a comprehensive activity index as the sum of values ​​obtained by multiplying the scores of each element by weights, classify the calculated comprehensive activity index into multiple activity suitability stages by comparing it with a plurality of preset thresholds, measure underwater turbidity and reflect it in the snorkeling and scuba diving activity index, calculate the probability of swell wave occurrence by comprehensively analyzing wave direction and wave period as well as wave height, monitor gust conditions for each sea area in real time and reflect them in the activity index, calculate an activity index for each marine leisure sport based on real-time and predicted weather data of the location when a user selects a specific spot on a map and recommend the sport with the highest score as the optimal activity, or calculate an activity index for multiple spots and rank suitability to recommend the optimal location when a user selects a specific marine leisure sport, and provide all the above analysis results and recommendation information to an external application via an API. Claim 4 In claim 1, the above instructions, when executed by the processor, define the optimal weather condition range for each sport by applying special adjustments specific to each sea area, such as: setting multiple ranges for each weather element for each marine leisure sport and classifying them into optimal, good, moderate, bad, and dangerous stages; optimally setting a relatively low wind speed range for fishing; optimally setting a moderate wind speed range and a specific wave height range for surfing and windsurfing; and optimally setting a low wave height range and high underwater visibility for snorkeling and scuba diving; and calculating an activity index reflecting regional characteristics by applying special adjustments for each sea area, such as increasing current speed-related weights relative to a reference value and adding mudflat turbidity conditions for the West Sea region considering shallow water depth and large tidal differences; adding current direction stability conditions for the South Sea region considering complex straits and currents around islands; increasing wave height-related weights relative to a reference value and adding swell index conditions for the East Sea region considering deep water depth and swell wave occurrence; and adding current-related weights and wind speed gust conditions for the Jeju region considering warm currents and gusts. Controlling system. Claim 5 In claim 1, the above instructions, when executed by the processor, control the system to classify underwater turbidity into multiple turbidity unit ranges, map underwater visibility distances corresponding to each range, apply an inverse correlation in which underwater visibility distance increases as turbidity decreases, classify into multiple stages from optimal to dangerous, and reflect in the snorkeling and scuba diving activity index; comprehensively analyze wave height, wave direction, and wave period data to calculate the possibility of swell wave occurrence when a pattern is detected in which the wave period is longer than a preset threshold and the wave height increases rapidly; additionally reflect the swell wave risk in the activity index when such pattern is detected in the East Sea and Jeju regions; and calculate the rate of change of wind speed and wind direction over time to determine as a gust condition when the wind direction changes by more than a preset angle within a short period of time or the wind speed increases by more than a preset ratio relative to the average value. Claim 6 delete Claim 7 In claim 1, when the instructions are executed by the processor, the system constructs a dataset by matching the user's past marine leisure activity history data with actual weather condition data at the corresponding point in time in chronological order, analyzes the dataset to extract patterns of weather conditions preferred by each user to generate an individual preferred weather condition profile, classifies the user's proficiency level into multiple stages through a machine learning classification algorithm that analyzes the user's activity frequency and the difficulty of weather conditions during the activity, calculates a risk tolerance score according to each level to derive personalized activity index adjustment parameters, stores real-time marine accident location and time information in a database in conjunction with the weather conditions at the corresponding point in time, statistically analyzes the correlation between specific weather patterns and accident occurrences to calculate the probability of accident occurrence by weather condition, automatically lowers the activity index of the corresponding sport if the calculated probability exceeds a preset risk threshold, and sends a risk alert message to the user's mobile device, analyzes time-based weather change prediction data for users who wish to enjoy multiple marine leisure sports continuously to identify the sport optimal for each time period, calculates the travel time and distance required for activity switching, and determines the optimal combination and sequence of sports A system that determines and controls the generation and provision of an integrated schedule including movement paths between each activity location. Claim 8 In claim 1, when the instructions are executed by the processor, the system calculates the difference between a predicted value at a specific past point in time and an actual observed value at the same point in time to derive an error value and stores the error values ​​as a time series to verify the prediction accuracy of the MME model; analyzes the temporal and spatial patterns of the error to identify prediction biases that repeatedly appear under specific weather conditions or in specific regions; calculates correction coefficients based on the magnitude and direction of the identified biases; updates adaptive weights in real time that apply to the predicted values ​​of the conditions or regions; receives data including acceleration, gyroscope, and heart rate in real time from a sensor device worn by a user engaged in marine leisure activities; analyzes the user's heart rate variability and skin electrical response collected from wearable devices to quantify the actual activity intensity and stress level; receives input from users via a mobile application regarding marine conditions and activity satisfaction, including wave height, wind speed, and underwater visibility experienced in real time, on a 5-level scale; utilizes the collected data as a validation dataset for the prediction model; and if the difference between user feedback and the predicted value continuously exceeds a preset threshold under specific regions or specific weather conditions A system that controls the weights and algorithm coefficients for the corresponding conditions to automatically readjust them to improve prediction accuracy. Claim 9 In claim 1, the instructions, when executed by the processor, enable the system to analyze weather data for a specified period to learn the trends of change in weather patterns by season and time of day, generate a time-dependent weighting function that applies differentiated weights according to season and time of day even under the same weather conditions, particularly calculate the rate of change in visibility during the time periods before and after sunrise and sunset, statistically model the probability of local gusts occurring during the afternoon hours in summer, calculate the risk of freezing due to water temperature drop during the dawn hours in winter as a temperature function and reflect it in the activity index, perform vector operations to calculate the rate of increase in wave height when the wind speed vector and the current vector have opposite directions, apply a regression model to calculate the rate of decrease in visibility when the combination of temperature and humidity is within a specific range, calculate the gust risk when the rate of change in atmospheric pressure and the rate of change in wind direction simultaneously exceed a threshold, collect data on the location and concentration of red tide occurrences, data on the frequency and types of jellyfish appearances, and plankton density data from a marine ecosystem monitoring system to construct a biological risk factor database, and quantify the impact on marine leisure sports for each biological risk factor. A system that controls the calculation of marine biological risk for snorkeling and scuba diving as a separate sub-index and provides it along with the overall activity index. Claim 10 In claim 1, the instructions, when executed by the processor, execute an equipment recommendation algorithm using the system's calculated activity index and predicted weather conditions as input values ​​to map the buoyancy grade of a buoyancy vest suitable for a given range when the wave height falls within a preset range, calculate the thickness of a wetsuit capable of maintaining body temperature based on the water temperature falling within a specific range, calculate the sail size capable of stable navigation based on the wind speed falling within a specific range, determine the luminous intensity of necessary lighting equipment when visibility is below a specific value, and automatically generate an equipment configuration optimized for each marine leisure sport; when real-time weather changes are detected and the risk level rises, search for nearby safe points based on the user's current GPS coordinates and the type of sport being engaged in, calculate the risk level during movement by simulating the distance to each safe point and predicted weather conditions along the movement path, select the path with the lowest risk as the optimal evacuation route, generate and provide the estimated time required to move along the route and precautions for each section, collect location information and activity status of users active in the same area in real time to form a proximity user network, and a user's sensor or input A system that automatically transmits information on local weather changes or hazardous situations detected through it to other users within the network as push notifications, and statistically processes weather information collected from multiple users to identify highly reliable minute weather changes and controls them to reflect them in real-time in the activity index calculation.

Images