HMM-based port aquatic product transaction comprehensive prediction method and system

Abstract

The invention discloses an HMM-based port aquatic product transaction comprehensive prediction method and system, and the method comprises the steps: continuously collecting fishing boat basic information, port entering and leaving plans, transaction records and market price data, and carrying out the preprocessing; calculating the weight ratio of each aquatic product monthly, and storing the weight ratio and the preprocessed data into a database. A fishing boat transaction state is defined as a multi-dimensional vector, a state transition probability matrix is constructed, and an HMM model is trained. And predicting the fishing boat transaction state in the next month through the trained HMM by using the fishing boat state in the last month and the transfer matrix. And calculating the transaction volume of various aquatic products in the next month by combining the transaction volume in the last month based on the prediction information, and summarizing to the port level. And calculating a monthly average price according to the historical price data, predicting the average transaction price of the aquatic products in the next month by using the time sequence analysis model, calculating the total transaction amount and generating a visual report. Deep fusion of multi-source data is realized, model prediction accuracy is improved, and adaptability and timeliness of the model to external changes are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of aquatic product trading prediction, and in particular to a comprehensive prediction method and system for port aquatic product trading based on HMM. Background Art

[0002] As a distribution center for aquatic products, the trading volume of fishing ports directly affects the income of fishermen and the economic benefits of fishing ports. By predicting the trading volume, fishing ports can better plan resource allocation, such as berth use, cold storage and logistics arrangements, to improve efficiency and reduce costs. The trading price of fish catches is greatly affected by market supply and demand. Predicting the trading price can help fishermen and fishing port managers understand market trends, so as to make more reasonable sales decisions and avoid economic losses caused by price fluctuations. The total transaction volume of the port is an important indicator to measure the economic activities of the fishing port. By predicting the total transaction volume, the fishing port can evaluate its economic contribution and market competitiveness, and then formulate corresponding market strategies and development plans.

[0003] Traditional methods for predicting the trading volume of fish catch mainly rely on expert experience and simple statistical models. These methods often lack sufficient data support and scientificity, resulting in low accuracy and reliability of the prediction results. In the field of port fish catch trading, there have been some attempts to improve the accuracy of trading volume prediction through technical means. For example, some studies use time series analysis methods, such as the ARIMA model, to analyze historical trading data to predict future trading volumes. However, these methods usually only consider the statistical characteristics of time series, while ignoring other influencing factors such as fishing vessel dynamics and weather changes. This scheme aims to use the hidden Markov model (HMM) to predict the trading volume of aquatic products of fishing vessels in the next month, establish a mapping relationship between fishing vessel dynamics and aquatic product trading volume, and realize prediction from individual fishing vessels to the entire port level. Summary of the invention

[0004] The present invention overcomes the defects of the prior art and provides a method and system for comprehensive prediction of port aquatic product trading based on HMM, the important purpose of which is to improve resource allocation efficiency and planning rationality.

[0005] To achieve the above-mentioned purpose, the first aspect of the present invention provides a comprehensive prediction method for port aquatic product trading based on HMM, comprising:

[0006] Continuously collect basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products, and perform data preprocessing on the collected data;

[0007] Calculate the weight percentage of each aquatic product on a monthly basis based on the preprocessed data, and store it in the database together with the preprocessed data;

[0008] Define the monthly trading status of fishing vessels as a multidimensional vector, analyze the monthly status changes of the same fishing vessel, construct the state transition probability matrix, and use historical trading data to train the HMM model;

[0009] Obtain the fishing boat status and corresponding state transfer matrix of last month, use the trained HMM model to predict the fishing boat trading status of next month, and obtain the prediction information of fishing boat trading status of next month;

[0010] Based on the forecast information of the fishing vessel trading status for the next month, the trading volume of various types of aquatic products for the next month is calculated in combination with the total trading volume of the previous month, and the forecast trading volume of each fishing vessel is aggregated to the port level to obtain the forecast total weight of different types of aquatic products;

[0011] The average monthly price is calculated based on historical transaction price data, the average transaction price of aquatic products for next month is predicted through a time series analysis model, and the total transaction amount of each aquatic product for next month is calculated to generate a visual report.

[0012] In this scheme, the basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information and real-time market price data of aquatic products are continuously collected, and data preprocessing is performed on the collected data, specifically including:

[0013] The basic information of the fishing vessel includes: fishing vessel specification information, fishing vessel capacity information and fishing vessel source information;

[0014] The aquatic product transaction record information includes transaction time, transaction type, transaction weight, transaction price, and transaction location;

[0015] The data preprocessing of the collected data includes data cleaning, missing value detection and supplementation, and outlier detection and supplementation.

[0016] In this solution, the weight percentage of each aquatic product is calculated on a monthly basis based on the preprocessed data, and stored in the database together with the preprocessed data, specifically including:

[0017] The weight proportion of each aquatic product is calculated by: based on historical aquatic product transaction data, the ratio of the monthly weight of each aquatic product to the square root of the sum of the squares of its monthly weight is calculated on a monthly basis;

[0018] Before storing the data in the database, the corresponding data is converted into a format suitable for HMM and other analysis models.

[0019] In this scheme, the monthly trading status of the fishing boat is defined as a multidimensional vector, the monthly state changes of the same fishing boat are analyzed, the state transition probability matrix is ​​constructed, and the HMM model is trained using historical trading data, specifically including:

[0020] Based on the fishing vessel information, aquatic product types and weight proportion, the monthly trading status of the fishing vessel is defined as a multidimensional vector, where each dimension represents the proportion of a type of aquatic product in the trading volume of the fishing vessel;

[0021] The port status is defined as the total weight of different aquatic product types, the monthly status changes of the same fishing vessel are analyzed, and the transition probability matrix between fishing vessel states is constructed;

[0022] The HMM model is trained using the historical data of aquatic product transactions as a training data set;

[0023] An HMM model training strategy is constructed based on the sliding window technology, and the target HMM model is retrained within a preset time period through the constructed HMM model training strategy.

[0024] In this solution, the predicted transaction volume of each fishing vessel is aggregated to the port level, including:

[0025] A suspension judgment threshold N is preset, and the entry and exit plan characteristics of the target fishing vessel are extracted based on the entry and exit plan information of the fishing vessel, and the characteristics are judged with the suspension judgment threshold N;

[0026] If it is greater than the out-of-service judgment threshold N, the corresponding fishing vessel will be marked as an out-of-service fishing vessel. After filtering the out-of-service fishing vessels, the predicted transaction volume of the filtered fishing vessels will be aggregated to the port level.

[0027] In this solution, the monthly average price is calculated based on the historical transaction price data, the average transaction price of aquatic products for the next month is predicted through the time series analysis model, and the total transaction amount of each aquatic product for the next month is calculated to generate a visualization report, which specifically includes:

[0028] Obtain historical transaction price data, with the month as the calculation unit, and calculate the average monthly price through the weighted average method;

[0029] The calculated monthly average price is input into the time series analysis model for prediction to obtain the average transaction price of aquatic products for the next month;

[0030] Obtain the predicted total weight of different types of aquatic products, and multiply it by the average transaction price of the aquatic products for the next month to obtain the predicted total transaction amount of various aquatic products for the next month;

[0031] Visualize the predicted total transaction amounts of various aquatic products for next month and generate a visual report for push.

[0032] The second aspect of the present invention provides a comprehensive prediction system for port aquatic product transactions based on HMM, the system comprising: a memory, a processor, a data collection module, a preprocessing module, an HMM training and prediction module, a price prediction module, and a transaction amount calculation module, wherein the memory and the processor store and execute a comprehensive prediction method for port aquatic product transactions based on HMM as described in any one of the above items;

[0033] The data collection module is used to continuously collect basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products;

[0034] The preprocessing module is used to perform data preprocessing and data format conversion on the continuously collected data;

[0035] The HMM training and prediction module is used to train the HMM model and predict the total weight of different types of aquatic products according to the trained HMM model;

[0036] The price prediction module is used to predict the average transaction price of aquatic products in the next month;

[0037] The transaction amount calculation module is used to calculate the total transaction amount of each aquatic product in the next month.

[0038] The third aspect of the present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium includes a comprehensive prediction method program for port aquatic product transactions based on HMM, and when the comprehensive prediction method program for port aquatic product transactions based on HMM is executed by a processor, the steps of the comprehensive prediction method for port aquatic product transactions based on HMM as described in any one of the above items are implemented.

[0039] The present invention discloses a comprehensive prediction method and system for port aquatic product transactions based on HMM, including: continuously collecting basic information of fishing vessels, port entry and exit plans, transaction records and market price data, and performing preprocessing. Calculate the weight proportion of each aquatic product on a monthly basis and store it in a database together with the preprocessed data. Define the trading status of fishing vessels as a multidimensional vector, construct a state transition probability matrix and train the HMM model. Utilize the fishing vessel status and transition matrix of the previous month to predict the trading status of fishing vessels next month through the trained HMM model. Based on the predicted information, the transaction volume of various types of aquatic products next month is calculated in combination with the transaction volume of the previous month, and summarized to the port level. Calculate the average monthly price based on historical price data, use a time series analysis model to predict the average transaction price of aquatic products next month, calculate the total transaction amount and generate a visual report. Realize the deep integration of multi-source data, improve the prediction accuracy of the model and enhance the adaptability and timeliness of the model to external changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0041] Figure 1 A flow chart of a comprehensive prediction method for port aquatic product trading based on HMM provided in one embodiment of the present invention;

[0042] Figure 2 A flow chart of an HMM model training method provided by one embodiment of the present invention;

[0043] Figure 3 A block diagram of a comprehensive prediction system for port aquatic product trading based on HMM provided by one embodiment of the present invention;

[0044] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0045] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0046] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0047] Figure 1 A flow chart of a comprehensive prediction method for port aquatic product trading based on HMM provided in one embodiment of the present invention;

[0048] like Figure 1 As shown, the present invention provides a flow chart of a comprehensive prediction method for port aquatic product trading based on HMM, including:

[0049] S102, continuously collecting basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products, and performing data preprocessing on the collected data;

[0050] S104, calculating the weight percentage of each aquatic product on a monthly basis based on the preprocessed data, and storing the weight percentage together with the preprocessed data in a database;

[0051] S106, defining the monthly trading status of the fishing boat as a multidimensional vector, analyzing the monthly status changes of the same fishing boat, constructing a state transition probability matrix, and training an HMM model using historical trading data;

[0052] S108, obtaining the fishing boat status and the corresponding state transfer matrix of the previous month, using the trained HMM model to predict the fishing boat trading status of the next month, and obtaining the fishing boat trading status prediction information of the next month;

[0053] S110, based on the forecast information of the fishing vessel trading status for the next month, combined with the total trading volume of the previous month, the trading volume of various types of aquatic products for the next month is calculated, and the forecast trading volume of each fishing vessel is summarized at the port level to obtain the forecast total weight of different types of aquatic products;

[0054] S112, calculate the average monthly price based on historical transaction price data, predict the average transaction price of aquatic products in the next month through a time series analysis model, and calculate the total transaction amount of each aquatic product in the next month to generate a visualization report.

[0055] It should be noted that the present invention provides a comprehensive prediction method and system for port aquatic product transactions based on HMM, which continuously collects basic information of fishing vessels, port entry and exit plans, transaction records and market price data, and pre-processes them. The weight proportion of each aquatic product is calculated monthly and stored in the database together with the pre-processed data. The trading status of the fishing vessel is defined as a multidimensional vector, and a state transition probability matrix is ​​constructed and the HMM model is trained. The state of the fishing vessel and the transition matrix of the previous month are used to predict the trading status of the fishing vessel next month through the trained HMM model. Based on the predicted information, the trading volume of various types of aquatic products next month is calculated in combination with the trading volume of the previous month, and summarized to the port level. The average monthly price is calculated based on the historical price data, and the average transaction price of aquatic products next month is predicted using the time series analysis model. The total transaction amount is calculated and a visual report is generated. The deep fusion of multi-source data is realized to improve the accuracy of model prediction and enhance the adaptability and timeliness of the model to external changes. It not only relies on historical transaction data, but also integrates external factors such as ship plans through multi-source data fusion technology, making the prediction results more comprehensive and accurate. Taking the actual situation of the fishing port into consideration, a mapping relationship between the dynamics of fishing vessels and the trading volume of aquatic products was established, achieving refined predictions from the level of individual fishing vessels to the entire port.

[0056] Furthermore, in a preferred embodiment of the present invention, the continuous collection of basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products, and the pre-processing of the collected data specifically include:

[0057] The basic information of the fishing vessel includes: fishing vessel specification information, fishing vessel capacity information and fishing vessel source information;

[0058] The aquatic product transaction record information includes transaction time, transaction type, transaction weight, transaction price, and transaction location;

[0059] The data preprocessing of the collected data includes data cleaning, missing value detection and supplementation, and outlier detection and supplementation.

[0060] It should be noted that by integrating the basic information of fishing vessels, port entry and exit plans, historical transaction data and aquatic product price data, a deep fusion of multi-source data is achieved. This fusion not only improves the accuracy of the prediction model, but also enhances the adaptability of the model to external changes. First, the character data in the basic information of fishing vessels, such as the name of the fishing vessel, the port of registry and the type of operation, need to be standardized and encoded. This includes removing redundant characters in the data, unifying the naming rules, and converting them into discrete or categorical variables, such as converting text information into numerical form through label encoding or unique hot encoding. Secondly, for numerical data of technical parameters, such as ship tonnage, main engine power and speed, it is necessary to check their integrity and consistency, fill in missing values ​​and normalize them to eliminate the influence of different dimensions and ensure that the data is comparable in the model. In addition, for time-related data, such as construction date, inspection records or time series information of fishing activities, it is necessary to organize and segment them according to the time axis. At the same time, spatially related data, such as fishing vessel operating areas or navigation tracks, need to be geocoded or gridded to represent them as coordinates or area labels that the model can recognize.

[0061] Among them, the fishing vessel specification information includes the fishing vessel identity, fishing vessel specification parameters, etc., the fishing vessel capacity information includes the bearing capacity and driving capacity of the fishing vessel, etc., and the fishing vessel source information includes the operation type of the fishing vessel, fishing vessel ownership information, etc.

[0062] Further, in a preferred embodiment of the present invention, the weight percentage of each aquatic product is calculated on a monthly basis according to the preprocessed data, and stored in the database together with the preprocessed data, specifically including:

[0063] The weight proportion of each aquatic product is calculated by: based on historical aquatic product transaction data, the ratio of the monthly weight of each aquatic product to the square root of the sum of the squares of its monthly weight is calculated on a monthly basis;

[0064] Before storing the data in the database, the corresponding data is converted into a format suitable for HMM and other analysis models.

[0065] It should be noted that before storing the basic information data of fishing vessels in the database, the data needs to be converted in format to adapt to the processing requirements of hidden Markov models (HMM) and other analysis models. To further adapt to the requirements of the analysis model, more expressive variables can be extracted or constructed through feature engineering. For example, comprehensive indicators such as unit power efficiency or tonnage utilization can be calculated based on existing technical parameters to enhance the data's ability to interpret specific analysis tasks. During the entire conversion process, it is also necessary to ensure that the data format is compatible with the structure of the storage database, such as uniformly using JSON or table storage to ensure that the field name is consistent with the input features required by the model, thereby improving the analysis accuracy and calculation efficiency of the model.

[0066] Furthermore, in a preferred embodiment of the present invention, the monthly trading status of the fishing boat is defined as a multidimensional vector, the monthly state changes of the same fishing boat are analyzed, a state transition probability matrix is ​​constructed, and the HMM model is trained using historical trading data, specifically including:

[0067] Based on the fishing vessel information, aquatic product types and weight proportion, the monthly trading status of the fishing vessel is defined as a multidimensional vector, where each dimension represents the proportion of a type of aquatic product in the trading volume of the fishing vessel;

[0068] The port status is defined as the total weight of different aquatic product types, the monthly status changes of the same fishing vessel are analyzed, and the transition probability matrix between the fishing vessel states is constructed;

[0069] The HMM model is trained using the historical data of aquatic product transactions as a training data set;

[0070] An HMM model training strategy is constructed based on the sliding window technology, and the target HMM model is retrained within a preset time period through the constructed HMM model training strategy.

[0071] It should be noted that in order to use the fishing vessel trading data for analysis and construct a hidden Markov model (HMM), it is necessary to accurately model the monthly trading status and port status of the fishing vessel, and train the model through historical data to capture its dynamic change pattern. First, the monthly trading status of the fishing vessel is defined as a multidimensional vector, in which each dimension represents the proportion of a certain aquatic product in the trading volume of the fishing vessel. This representation normalizes the trading structure of different aquatic products, making it have good comparability and feature expression capabilities. At the same time, the port status is defined as the total trading weight of different types of aquatic products in the port, thereby reflecting the characteristics of the overall trading activities of the port. On this basis, by analyzing the monthly trading status changes of the same fishing vessel, the transition probability matrix between its trading states can be constructed. The transition probability matrix describes the possibility of a fishing vessel transferring between different trading states, reflecting the dynamic patterns and trends of its trading activities. The training data set is constructed using the historical trading data of the fishing vessel, and the HMM model is trained to capture the implicit laws and trends in the trading state sequence, effectively learning the implicit state and transition relationships from the time series data. In order to adapt to the possible changes in trading patterns over time, the sliding window technology is further used to optimize the training strategy of the HMM. In this method, historical data is divided into several sliding windows according to a preset time period, and the data in each window is used to train or update the HMM model. This dynamic training method can update the model parameters according to the latest data, thereby maintaining the HMM's ability to accurately predict fishing vessel trading behavior. In practical applications, by retraining the target HMM model in each time period, the model's sensitivity to short-term changes can be improved while ensuring its stability to long-term patterns.

[0072] Furthermore, in a preferred embodiment of the present invention, the predicted transaction volume of each fishing vessel is aggregated to the port level, specifically including:

[0073] A suspension judgment threshold N is preset, and the entry and exit plan characteristics of the target fishing vessel are extracted based on the entry and exit plan information of the fishing vessel, and the characteristics are judged with the suspension judgment threshold N;

[0074] If it is greater than the out-of-service judgment threshold N, the corresponding fishing vessel will be marked as an out-of-service fishing vessel. After filtering the out-of-service fishing vessels, the predicted transaction volume of the filtered fishing vessels will be aggregated to the port level.

[0075] Furthermore, in a preferred embodiment of the present invention, the monthly average price is calculated based on the historical transaction price data, the average transaction price of aquatic products for the next month is predicted by the time series analysis model, and the total transaction amount of each aquatic product for the next month is calculated to generate a visualization report, which specifically includes:

[0076] Obtain historical transaction price data, with the month as the calculation unit, and calculate the average monthly price through the weighted average method;

[0077] The calculated monthly average price is input into the time series analysis model for prediction to obtain the average transaction price of aquatic products for the next month;

[0078] Obtain the predicted total weight of different types of aquatic products, and multiply it by the average transaction price of the aquatic products for the next month to obtain the predicted total transaction amount of various aquatic products for the next month;

[0079] Visualize the predicted total transaction amounts of various aquatic products for next month and generate a visual report for push.

[0080] It should be noted that, firstly, based on the historical transaction price data and with the month as the calculation unit, the average monthly price of each aquatic product is calculated by the weighted average method. The weight is usually allocated according to the number of transactions, so as to ensure that the calculation result can accurately reflect the changing trend of the market price level. The calculation result of the monthly average price is time series data, which can be used for further trend analysis and prediction. The calculated monthly average price is input into the time series analysis model, such as the SARIMA model or the LSTM model, to predict the future price trend. Next, the predicted total weight of each aquatic product is multiplied by its corresponding predicted average transaction price to calculate the predicted total transaction amount of each aquatic product in the next month, thereby reflecting the changing trend and specific distribution of the total transaction amount in the market. Finally, the calculated predicted total transaction amount of various aquatic products is visualized and displayed in charts, curves or other forms. Visual reports can not only intuitively display data, but also reveal structural changes and potential business opportunities in the aquatic product market. The generated visualization report can be pushed electronically to relevant users, such as management departments, enterprises or industry analysts, to help better decision-making and planning.

[0081] Figure 2 A flow chart of an HMM model training method provided by one embodiment of the present invention;

[0082] like Figure 2 As shown, the present invention provides a flow chart of an HMM model training method, comprising:

[0083] S202, based on the fishing vessel information, the types of aquatic products and the weight proportion, define the monthly trading status of the fishing vessel as a multi-dimensional vector, where each dimension represents the proportion of a type of aquatic product in the trading volume of the fishing vessel;

[0084] S204, defining the port status as the total weight of different aquatic product types, analyzing the monthly status changes of the same fishing vessel, and constructing the transition probability matrix between fishing vessel states;

[0085] S206, using the historical data of aquatic product transactions to form a training data set to train the HMM model;

[0086] S208, constructing an HMM model training strategy based on the sliding window technology, and retraining the target HMM model within a preset time period through the constructed HMM model training strategy.

[0087] It is worth mentioning that the HMM-based comprehensive prediction method and system for aquatic product trading provided in the present invention realizes the deep fusion of multi-source data by integrating the basic information of fishing vessels, port entry and exit plans, historical transaction data and aquatic product price data. This fusion not only improves the accuracy of the prediction model, but also enhances the adaptability of the model to external changes. Furthermore, considering the actual situation of the fishing port, a mapping relationship between the dynamics of fishing vessels and the trading volume of aquatic products is established, and a refined prediction from individual fishing vessels to the entire port level is realized. The HMM model is used to predict the trading volume of fishing vessels. The model can capture the random process of the trading volume of fishing vessels changing over time, and provide more accurate prediction results by defining the state of the fishing vessel and the transition probability between different states. Traditional statistical models are often difficult to handle nonlinear relationships and multivariate effects, while the HMM model used in the present invention can effectively capture and process these complex relationships and improve the generalization ability of the model. Compared with machine learning models that require a large amount of labeled data, it reduces the dependence on labeled data, reduces the difficulty and cost of data collection, and makes the model more practical.

[0088] Figure 3 A port aquatic product transaction comprehensive prediction system 3 based on HMM is provided in one embodiment of the present invention, the system comprises: a memory 31, a processor 32, a data collection module 33, a preprocessing module 34, an HMM training and prediction module 35, a price prediction module 36, and a transaction amount calculation module 37, wherein the memory 31 and the processor 32 store and execute a port aquatic product transaction comprehensive prediction method based on HMM as described in any one of the above items;

[0089] The data collection module is used to continuously collect basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products;

[0090] The preprocessing module is used to perform data preprocessing and data format conversion on the continuously collected data;

[0091] The HMM training and prediction module is used to train the HMM model and predict the total weight of different types of aquatic products according to the trained HMM model;

[0092] The price prediction module is used to predict the average transaction price of aquatic products in the next month;

[0093] The transaction amount calculation module is used to calculate the total transaction amount of each aquatic product in the next month.

[0094] The third aspect of the present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium includes a comprehensive prediction method program for port aquatic product transactions based on HMM, and when the comprehensive prediction method program for port aquatic product transactions based on HMM is executed by a processor, the steps of the comprehensive prediction method for port aquatic product transactions based on HMM as described in any one of the above items are implemented.

[0095] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0096] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0097] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0098] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0099] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0100] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A comprehensive prediction method for port aquatic product trading based on HMM, characterized in that: include: Continuously collect basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products, and perform data preprocessing on the collected data; Calculate the weight percentage of each aquatic product on a monthly basis based on the preprocessed data, and store it in the database together with the preprocessed data; Define the monthly trading status of fishing vessels as a multidimensional vector, analyze the monthly status changes of the same fishing vessel, construct the state transition probability matrix, and use historical trading data to train the HMM model; Obtain the fishing boat status and corresponding state transfer matrix of last month, use the trained HMM model to predict the fishing boat trading status of next month, and obtain the prediction information of fishing boat trading status of next month; Based on the forecast information of the fishing vessel trading status for the next month, the trading volume of various types of aquatic products for the next month is calculated in combination with the total trading volume of the previous month, and the forecast trading volume of each fishing vessel is aggregated to the port level to obtain the forecast total weight of different types of aquatic products; The average monthly price is calculated based on historical transaction price data, the average transaction price of aquatic products for next month is predicted through a time series analysis model, and the total transaction amount of each aquatic product for next month is calculated to generate a visual report.

2. The method for comprehensive prediction of port aquatic product trading based on HMM according to claim 1 is characterized in that: The continuous collection of basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products, and pre-processing of the collected data specifically include: The basic information of the fishing vessel includes: fishing vessel specification information, fishing vessel capacity information and fishing vessel source information; The aquatic product transaction record information includes transaction time, transaction type, transaction weight, transaction price, and transaction location; The data preprocessing of the collected data includes data cleaning, missing value detection and supplementation, and outlier detection and supplementation.

3. The method for comprehensive prediction of port aquatic product trading based on HMM according to claim 1 is characterized in that: The weight percentage of each aquatic product is calculated on a monthly basis according to the pre-processed data, and stored in the database together with the pre-processed data, specifically including: The weight proportion of each aquatic product is calculated by: based on historical aquatic product transaction data, the ratio of the monthly weight of each aquatic product to the square root of the sum of the squares of its monthly weight is calculated on a monthly basis; Before storing the data in the database, the corresponding data is converted into a format suitable for HMM and other analysis models.

4. The method for comprehensive prediction of port aquatic product trading based on HMM according to claim 1 is characterized in that: The monthly trading status of the fishing boat is defined as a multidimensional vector, the monthly status changes of the same fishing boat are analyzed, the state transition probability matrix is ​​constructed, and the HMM model is trained using historical trading data, specifically including: Based on the fishing vessel information, aquatic product types and weight proportion, the monthly trading status of the fishing vessel is defined as a multidimensional vector, where each dimension represents the proportion of a type of aquatic product in the trading volume of the fishing vessel; The port status is defined as the total weight of different aquatic product types, the monthly status changes of the same fishing vessel are analyzed, and the transition probability matrix between fishing vessel states is constructed; The HMM model is trained using the historical data of aquatic product transactions as a training data set; An HMM model training strategy is constructed based on the sliding window technology, and the target HMM model is retrained within a preset time period through the constructed HMM model training strategy.

5. The method for comprehensive prediction of port aquatic product trading based on HMM according to claim 1 is characterized in that: The forecasted trading volume for each fishing vessel is aggregated to the port level, including: A suspension judgment threshold N is preset, and the entry and exit plan characteristics of the target fishing vessel are extracted based on the entry and exit plan information of the fishing vessel, and the characteristics are judged with the suspension judgment threshold N; If it is greater than the out-of-service judgment threshold N, the corresponding fishing vessel will be marked as an out-of-service fishing vessel. After filtering the out-of-service fishing vessels, the predicted transaction volume of the filtered fishing vessels will be aggregated to the port level.

6. The method for comprehensive prediction of port aquatic product trading based on HMM according to claim 1 is characterized in that: The monthly average price is calculated based on the historical transaction price data, the average transaction price of aquatic products for the next month is predicted through the time series analysis model, and the total transaction amount of each aquatic product for the next month is calculated to generate a visualization report, which specifically includes: Obtain historical transaction price data, with the month as the calculation unit, and calculate the average monthly price through the weighted average method; The calculated monthly average price is input into the time series analysis model for prediction to obtain the average transaction price of aquatic products for the next month; Obtain the predicted total weight of different types of aquatic products, and multiply it by the average transaction price of the aquatic products for the next month to obtain the predicted total transaction amount of various aquatic products for the next month; Visualize the predicted total transaction amounts of various aquatic products for next month and generate a visual report for push.

7. A comprehensive prediction system for port aquatic product trading based on HMM, characterized in that: The system comprises: a memory, a processor, a data collection module, a preprocessing module, an HMM training and prediction module, a price prediction module, and a transaction amount calculation module, wherein the memory and the processor store and execute a comprehensive prediction method for port aquatic product transactions based on HMM as described in any one of claims 1 to 6; The data collection module is used to continuously collect basic information of fishing vessels, port entry and exit plan information, aquatic product transaction record information, and real-time market price data of aquatic products; The preprocessing module is used to perform data preprocessing and data format conversion on the continuously collected data; The HMM training and prediction module is used to train the HMM model and predict the total weight of different types of aquatic products according to the trained HMM model; The price prediction module is used to predict the average transaction price of aquatic products in the next month; The transaction amount calculation module is used to calculate the total transaction amount of each aquatic product in the next month.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a comprehensive prediction method program for port aquatic product transactions based on HMM. When the comprehensive prediction method program for port aquatic product transactions based on HMM is executed by a processor, the steps of the comprehensive prediction method for port aquatic product transactions based on HMM as described in any one of claims 1 to 6 are implemented.

Images