Freshwater fishpond water temperature prediction method, system, medium and equipment

Abstract

The invention provides a freshwater fish pond water temperature monitoring method, system, medium and equipment, and relates to the technical field of freshwater fish pond cultivation, a fish pond water temperature data set is formed based on water temperature data of different levels of a fish pond water body and data such as meteorological factors and water quality parameters, and the data set is preprocessed; according to the method, a preprocessed data set is utilized, meteorological factors of the day and a set time period are screened out to serve as forecasting factors, statistical methods such as multiple regression, stepwise regression and fuzzy clustering are adopted, and freshwater fish pond water temperature forecasting models under different weather types such as a low-temperature and low-illumination type, a rapid temperature and pressure reduction type and a high-temperature and stuffiness type are constructed.

Description

Technical Field

[0001] The present invention relates to the technical field of aquaculture, and in particular to a method, system, medium and equipment for predicting water temperature in a fish pond. Background Art

[0002] Currently, many freshwater farmers still rely on traditional, experience-based methods to monitor water temperature, such as using buoys or handheld thermometers. These methods not only consume a lot of time and labor, but also fail to provide continuous and accurate water temperature data, let alone the ability to predict future water temperature changes. At the same time, although conventional weather forecasts can provide reference information for farmers to a certain extent, they generally lack the accuracy and professionalism for the specific needs of aquaculture, and therefore are not very accurate in accurately predicting micro-meteorological changes that have a direct impact on aquaculture production. Summary of the invention

[0003] In order to overcome the shortcomings of the above-mentioned conventional weather forecast in the prior art, which can provide some help but lacks pertinence and professionalism, and has limited ability and low accuracy in accurately predicting micro-meteorological changes that have a direct impact on the aquaculture industry, the main purpose of the present invention is to provide a freshwater fish pond water temperature prediction method, system, medium and equipment.

[0004] To achieve the above object, the present invention adopts the following technical scheme, a method for predicting water temperature in a freshwater fish pond, comprising the following steps:

[0005] Acquire temperature data, meteorological data, and water quality factor data at different levels of freshwater fish pond water bodies, form a data set, and then preprocess it to obtain a preprocessed data set;

[0006] The key features of the preprocessed data set are extracted by using correlation analysis and stepwise regression methods to obtain the key features that affect the water temperature of the freshwater fish pond. The meteorological factors of the set time period are used as forecasting factors. Different prediction models are constructed according to different weather types and in combination with the key features.

[0007] Obtain the meteorological data for the forecast day, confirm the weather type for the forecast day, select the corresponding forecast model according to the weather type for the forecast day, input the meteorological data for the forecast day into the selected forecast model, and obtain the water temperature forecast result of the fish pond water body on the forecast day.

[0008] Building different prediction models includes the following steps:

[0009] The weather types include low temperature and little sunshine, rapid temperature and pressure drop, and high temperature and muggy weather.

[0010] The prediction models include a multivariate linear regression prediction model, a fuzzy clustering prediction model and a stepwise regression prediction model;

[0011] Select the prediction model according to the data characteristics and prediction objectives, train the model using the data subset of each weather type, and adjust the parameters of the prediction model through cross-validation and grid search methods;

[0012] Compare the performance of different models under the same weather type and select the forecast model with the best corresponding effect;

[0013] When the weather type is low temperature and low light, select the multivariate linear regression prediction model;

[0014] When the weather type is a sharp drop in temperature and pressure, the fuzzy clustering prediction model is selected;

[0015] When the weather type is hot and muggy, select the stepwise regression prediction model.

[0016] The multiple linear regression model building step comprises:

[0017] Split the dataset into training and testing sets.

[0018] A multivariate linear regression model was constructed based on the training set, in which water temperature was the dependent variable and air temperature, air pressure, wind speed, etc. were independent variables.

[0019] The regression coefficients were estimated using the least squares method;

[0020] The coefficient of determination R 2 , p-value, F-statistic and other statistical indicators to evaluate the goodness of fit of the model;

[0021] Based on the regression coefficient of each independent variable, its effect on water temperature was obtained.

[0022] The stepwise regression star construction includes the following steps:

[0023] Multiple univariate linear regression models were built for each predictor;

[0024] Evaluate the explanatory power of multiple univariate linear regression models, rank the variables according to their explanatory power, and obtain single-variable ranking results;

[0025] Start with the variables with the strongest explanatory power and gradually add them to the model. After each variable is added, use the AIC criterion to evaluate the model effect.

[0026] If the model effect is not significantly improved after adding a new variable, delete the variable;

[0027] Use the test set to evaluate the prediction accuracy of the model, such as R-squared and RMSE;

[0028] If you are not satisfied with the prediction results, return to the previous step and adjust the variables and re-model.

[0029] The fuzzy clustering construction includes the following steps:

[0030] Extract key variables representing water temperature variation characteristics from the preprocessed data and extract features:

[0031] Calculate the fuzzy similarity between samples and construct the fuzzy similarity matrix;

[0032] Based on the fuzzy similarity matrix and the preset number of clusters, the cluster centers are determined;

[0033] According to the fuzzy similarity between the sample and the cluster center, fuzzy classification is performed. ;

[0034] The clustering results are iteratively optimized by updating the cluster centers and sample memberships.

[0035] The clustering results were sorted out, the characteristics of water temperature changes under different clusters were analyzed, and the water temperature prediction results were obtained.

[0036] A fish pond water temperature prediction system, comprising:

[0037] Data acquisition and processing module: obtain temperature data, meteorological data, and water quality factor data at different levels of the fish pond water body, form a data set, pre-process the data set, and obtain the pre-processed data set;

[0038] Prediction model analysis module: obtain the water temperature, air temperature, air pressure and wind speed of the fish pond as data sets, and use the meteorological factors of the forecast day and the previous three days as forecast factors, and build corresponding prediction models according to different weather types; among them, the weather types include low temperature and low light type, rapid temperature and pressure drop type, and high temperature and sultry type; the prediction models include multivariate linear regression model, univariate linear regression model, and fuzzy clustering prediction model;

[0039] The water temperature prediction module obtains the water temperature data to be measured, confirms the weather type, selects the corresponding prediction model, and obtains the water temperature prediction result.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention obtains temperature data, meteorological data, and water quality factor data at different levels of the water body. Comprehensively consider various factors affecting water temperature to improve the accuracy and reliability of prediction. Data cleaning and classification processing improve data quality, reduce data noise, and ensure the efficiency of model training and the accuracy of prediction. Different prediction models are selected according to different weather types, and the most suitable model is used for prediction according to the characteristics of water temperature changes under different conditions to improve the accuracy and adaptability of prediction. The regression coefficient is estimated by the least squares method, the model goodness of fit is evaluated, and the influence of independent variables on water temperature is analyzed. The model structure is gradually optimized by ranking the explanatory power of the univariate linear regression model, and the model effect is evaluated using the AIC criterion. The water temperature change characteristics are analyzed by fuzzy similarity calculation and cluster center iterative optimization. Ensure that each prediction model can perform at its best under its applicable conditions, and improve the accuracy of prediction through model optimization. Input the data to be tested into the selected prediction model to obtain the water temperature prediction result. Providing accurate water temperature prediction for aquaculture farmers helps to reasonably arrange breeding activities, reduce the risks caused by water temperature changes, and improve the economic benefits of aquaculture.

[0041] The accuracy of water temperature prediction is improved by comprehensively analyzing multiple data and adopting appropriate prediction models. Different prediction models can be selected according to different weather types to adapt to complex and changeable natural conditions. Accurate water temperature prediction helps farmers optimize breeding management, reduce risks, and improve production and quality. Scientific predictions can guide breeding activities, reduce costs, and improve the overall economic benefits of aquaculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0043] Figure 1 It is a schematic diagram of the process structure of the present invention. DETAILED DESCRIPTION

[0044] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0045] Embodiment 1:

[0046] In today's aquaculture field, traditional monitoring methods such as buoy or handheld thermometers can no longer meet farmers' needs for accurate, continuous and predictive water temperature data. These methods are not only inefficient, but also unable to provide predictions of future water temperature changes, which is crucial for the refined management of the aquaculture industry. Although conventional weather forecasts provide some help, their ability to predict micro-meteorological changes is limited, and it is difficult to meet the aquaculture industry's needs for accurate predictions of specific environmental factors. In response to these problems, the present invention proposes an innovative freshwater fish pond water temperature monitoring and forecasting method, which integrates modern sensor technology, data analysis, meteorological science and information technology, and aims to provide farmers with a solution for real-time monitoring and accurate prediction of water temperature changes.

[0047] The method of the present invention comprises the following key steps:

[0048] Data collection and processing, the data collection module is the cornerstone of the system, and its functions include data entry, modification, addition, deletion, as well as data cleaning and quality control. Users can import observation data by uploading TXT, EXCEL, and CSV format files. The system will identify and read data according to the preset file naming rules, data attribute rules, arrangement order rules, and separator rules. Cleaning and quality control of raw data are the key to ensuring data accuracy, including deleting duplicate records, checking time sequence, deleting records with missing key attributes, unit conversion, and data format unification. Ultimately, qualified data will be organized into standardized data sets and uploaded to the designated database.

[0049] The data elements collected and stored include date and time, air temperature, relative humidity, water level, atmospheric pressure, water temperature at different depths, wind speed, rainfall, pH value, ammonia nitrogen, turbidity, dissolved oxygen and sunshine time.

[0050] Data preprocessing, in the data preprocessing stage, the system expands its functions to process and analyze data more accurately, automatically detecting and correcting abnormal data caused by collection errors, human errors or transmission errors. This step includes splitting data by season, splitting data by time series, calculating daily difference and daily average, and processing null value data to ensure data consistency and accuracy.

[0051] The modeling and analysis phase is the core of the present invention, which mainly studies the relationship between water temperature and air temperature, air pressure, wind speed and other factors. This phase uses statistical methods such as multiple regression, stepwise regression, and fuzzy clustering, combined with relevant meteorological factors of the day and the previous 1-3 days, to establish a water temperature forecast model under different weather types.

[0052] In the construction of the multiple regression model, the linear equation is established, and the regression coefficient is calculated using the least squares method and other techniques to evaluate the goodness of fit and predictive ability of the model. The stepwise regression algorithm traverses the variables, gradually establishes the multiple linear regression model, evaluates the explanatory power of each variable, and finally determines the prediction model.

[0053] The fuzzy clustering method is used to extract variables representing the characteristics of water temperature changes from the preprocessed data. By calculating the fuzzy similarity between samples, the fuzzy classification of samples is achieved, thereby optimizing the prediction model.

[0054] Each forecast model operation will generate a detailed calculation conclusion, including the data set of selected sample data, the correlation strength of the calculated factor set with time or temperature, and the text description of the significance level. The conclusion is accompanied by factor analysis charts, such as line charts and bar charts, which intuitively display the results of the correlation analysis and the time series changes of the data of each factor.

[0055] Through the detailed implementation of the above steps, the present invention can not only help farmers manage the breeding environment more effectively and improve breeding efficiency, but also reduce the risks and losses caused by environmental changes. The method can process and analyze the relationship between multiple variables, revealing how factors such as air temperature, sunshine duration, rainfall, humidity and wind speed affect water temperature alone or together, providing strong technical support for the sustainable development of the breeding industry.

[0056] When implementing the method of the present invention, special attention is paid to the comprehensiveness and accuracy of data collection. In order to ensure the integrity of the data, a variety of data verification mechanisms are added to the data collection module to prevent data entry errors or omissions. In addition, a flexible data import and export system is designed to facilitate users to easily migrate data between different platforms.

[0057] In the data preprocessing stage, advanced data cleaning technologies were used, including automated outlier detection and correction, and intelligent alignment of time series data. The application of these technologies significantly improved the quality of the data and laid a solid foundation for subsequent modeling and analysis.

[0058] In the modeling and analysis phase, we not only used traditional multiple regression and stepwise regression methods, but also explored the application of machine learning algorithms in the prediction model. By comparing the prediction effects of different models, we can provide users with more accurate water temperature prediction results.

[0059] In order to make the prediction results more intuitive and easy to understand, a chart generation system was developed, which can automatically generate line charts, bar charts, scatter charts and other charts based on the model calculation results. These charts not only show the trend of water temperature changes, but also reveal the correlation between various environmental factors, providing farmers with intuitive data analysis and decision-making support.

[0060] In short, the water temperature monitoring and forecasting method of the present invention is an advanced system integrating modern technology and data analysis. It has brought revolutionary changes to the aquaculture industry, improved the level of intelligence in aquaculture management, and brought tangible economic benefits to farmers.

[0061] Example 2

[0062] The following is a specific implementation case, detailing the steps of the freshwater fish pond water temperature prediction method:

[0063] In a freshwater fish farming area, farmers hope to predict changes in pond water temperature in advance in order to better manage their farming activities.

[0064] Step 1: Data collection and preprocessing

[0065] Data collection: Install temperature sensors at different depths of the fish pond to automatically record water temperature data every hour. Obtain weather data such as air temperature, air pressure, and wind speed through the local weather station API. Use water quality testing equipment to collect water quality data such as pH value and dissolved oxygen.

[0066] Data preprocessing, check and fill missing data, remove outliers, and ensure data consistency. The data is divided into spring data sets by season, and the data sets of the day and the previous 1-3 days are split by time series.

[0067] Step 2: Key feature extraction

[0068] The Pearson correlation coefficient was used to analyze the correlation between water temperature and various meteorological factors and water quality factors. It was found that air temperature, air pressure, and wind speed were highly correlated with water temperature, so these were selected as key features. The stepwise regression method was applied to further screen and retain the features that had the greatest impact on water temperature.

[0069] Step 3: Build a prediction model

[0070] Low temperature and low light type: Select a multiple linear regression model, with air temperature, air pressure, and wind speed as independent variables and water temperature as the dependent variable. Divide the data set into a training set (70%) and a test set (30%), and use the training set to build the model. 2 , p-value, and F-statistic were used to evaluate the goodness of fit of the model.

[0071] Sharp cooling and pressure reduction type: Select the fuzzy clustering model, extract the characteristics of water temperature changes, calculate the fuzzy similarity between samples, determine the cluster center, and perform classification. Analyze the clustering results to verify the accuracy of the model in predicting water temperature changes. Select the stepwise regression model, start with single variable linear regression, and gradually add variables. Use the AIC criterion to evaluate the model effect after each variable is added. Use the test set to evaluate the prediction accuracy of the model.

[0072] Step 4: Water Temperature Prediction

[0073] Obtain the current and previous days' meteorological data and water temperature data, confirm the weather type as low temperature and low sunshine according to the weather forecast, select the prediction model as the multiple linear regression model, input the data into the multiple linear regression model. The model outputs the water temperature prediction results for the next 24 hours.

[0074] It is a sunny day in autumn, and the weather forecast shows that low temperatures and little sunshine will continue in the next few days. The temperature, air pressure, and wind speed data for the day and the previous 1-3 days were collected. After data preprocessing, a multivariate linear regression model was used for prediction. After inputting the data, the model predicted that the water temperature would drop by 1.5℃ tomorrow. Based on this prediction, the farmer adjusted the use of the aerator in advance to prevent the low water temperature from affecting the fry.

[0075] Through this implementation case, we can see that the use of freshwater fish pond water temperature prediction methods can help farmers better cope with water temperature changes, thereby improving breeding efficiency and management level.

[0076] It should be noted that, in the present invention, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0077] The above embodiments are merely examples of the present invention and do not limit the protection scope of the present invention. All designs that are the same or similar to the present invention fall within the protection scope of the present invention.

Claims

1. A method for predicting water temperature in a freshwater fish pond, characterized in that: The following steps are involved: Acquire temperature data, meteorological data, and water quality factor data at different levels of freshwater fish pond water bodies, form a data set, and then preprocess it to obtain a preprocessed data set; The key features of the pre-processed data set are extracted by using correlation analysis and stepwise regression method to obtain the key features that affect the water temperature of the freshwater fish pond. Different prediction models are constructed by combining the meteorological factors of the set time period with the key features according to different weather types; Obtain the meteorological data for the forecast day, confirm the weather type for the forecast day, select the corresponding forecast model according to the weather type for the forecast day, input the meteorological data for the forecast day into the selected forecast model, and obtain the water temperature forecast result of the fish pond water body on the forecast day.

2. The method for predicting water temperature in a freshwater fish pond according to claim 1, wherein: The construction of different prediction models includes the following steps: The weather types include low temperature and little sunshine, rapid temperature and pressure drop, and high temperature and muggy weather. The prediction models include a multivariate linear regression prediction model, a fuzzy clustering prediction model and a stepwise regression prediction model; Select the prediction model according to the data characteristics and prediction objectives, train the model using the data subset of each weather type, and adjust the parameters of the prediction model through cross-validation and grid search methods; Compare the performance of different models under the same weather type and select the forecast model with the best corresponding effect; When the weather type is low temperature and low light, select the multivariate linear regression prediction model; When the weather type is a sharp drop in temperature and pressure, the fuzzy clustering prediction model is selected; When the weather type is hot and muggy, select the stepwise regression prediction model.

3. The freshwater fish pond water temperature prediction method as claimed in claim 2, characterized in that: The multiple linear regression model building step comprises: Divide the dataset into training and testing sets; A multivariate linear regression model was constructed based on the training set, where water temperature was the dependent variable and air temperature, air pressure, wind speed, etc. were independent variables; The regression coefficients were estimated using the least squares method; The coefficient of determination R 2 , p-value, and F-statistic were used as statistical indicators to evaluate the goodness of fit of the model; Based on the regression coefficient of each independent variable, its effect on water temperature was obtained.

4. The method for predicting water temperature in a freshwater fish pond as claimed in claim 2, wherein: The stepwise regression model construction includes the following steps: Multiple univariate linear regression models were built for each predictor; Evaluate the explanatory power of multiple univariate linear regression models, rank the variables according to their explanatory power, and obtain single-variable ranking results; Start with the variables with the strongest explanatory power and gradually add them to the model. After each variable is added, use the AIC criterion to evaluate the model effect. If the model effect is not significantly improved after adding a new variable, delete the variable; Use the test set to evaluate the prediction accuracy of the model, such as R-squared and RMSE; If you are not satisfied with the prediction results, return to the previous step and adjust the variables and re-model.

5. The method for predicting water temperature in a freshwater fish pond as claimed in claim 2, wherein: The fuzzy clustering construction includes the following steps: Extract key variables representing water temperature variation characteristics from the preprocessed data and extract features: Calculate the fuzzy similarity between samples and construct the fuzzy similarity matrix; Based on the fuzzy similarity matrix and the preset number of clusters, the cluster centers are determined; Fuzzy classification is performed based on the fuzzy similarity between the sample and the cluster center; Iteratively optimize the clustering results by updating the cluster centers and sample memberships; The clustering results were sorted out, the characteristics of water temperature changes under different clusters were analyzed, and the water temperature prediction results were obtained.

6. The method for predicting water temperature in a freshwater fish pond as claimed in claim 1, wherein: The preprocessing includes: data cleaning and classification processing; The data cleaning includes: missing value processing, outlier detection and consistency processing; The classification processing includes splitting data by season, splitting data by time series, calculating daily difference, daily average value, and processing null value data.

7. A freshwater fish pond water temperature prediction system, characterized in that: include: Data acquisition and processing module; Acquire temperature data, meteorological data, and water quality factor data at different levels of freshwater fish pond water bodies, form a data set, and then preprocess it to obtain a preprocessed data set; Prediction model analysis module: extract key features of the preprocessed data set using correlation analysis and stepwise regression methods to obtain key features that affect the water temperature of freshwater fish ponds, use meteorological factors in a set time period as forecasting factors, and build different prediction models based on different weather types and key features; The water temperature prediction module obtains the meteorological data of the prediction day, confirms the weather type of the prediction day, selects the corresponding prediction model according to the weather type of the prediction day, inputs the meteorological data of the prediction day into the selected prediction model, and obtains the water temperature prediction result of the fish pond water body on the prediction day.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method described in any one of claims 1 to 6 is implemented.

Images