Method and terminal for predicting dominant species of red tide organisms

Abstract

The invention discloses a red tide organism dominant species prediction method and terminal, and the method comprises the steps: carrying out the feature screening of collected red tide organism dominant species impact factors through employing a random forest algorithm, and carrying out the modeling based on the screened red tide organism dominant species impact factors through employing a logic regression algorithm, a naive Bayesian algorithm and a neural network algorithm, the method comprises the following steps: respectively obtaining red tide biological dominant species prediction models, respectively determining prediction errors of a plurality of red tide biological dominant species prediction models, respectively endowing the plurality of red tide biological dominant species prediction models with weights according to the prediction errors, and carrying out prediction by using the plurality of red tide biological dominant species prediction models based on the weights. Errors caused by contingency of a certain algorithm are effectively avoided, the prediction precision is improved, the red tide is accurately and rapidly predicted, the red tide risk is reduced, the red tide disaster reduction capacity of marine fishery is improved, and healthy development of fishery economy is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of marine red tide prediction, and in particular to a method and a terminal for predicting dominant species of red tide organisms. Background Art

[0002] There are high-incidence areas of red tides in the existing sea areas. Red tides of varying scales occur during the high-incidence period every year, and some of the red tides that occur are toxic (harmful) red tides. At present, due to the lack of in-depth research on the basic biology and ecology of harmful red tide organisms in the coastal waters of the region, effective methods for the prevention and control of harmful red tides have not yet been established, and it is still impossible to accurately and quickly predict, warn and prevent harmful red tides. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and a terminal for predicting dominant species of red tide organisms, which can accurately and quickly predict red tides.

[0004] In order to solve the above technical problems, a technical solution adopted by the present invention is: A method for predicting dominant species of red tide organisms comprises the following steps: Collect the influencing factors of the dominant species of red tide organisms in the history of the preset sea area, and use the random forest algorithm to perform feature screening on the influencing factors of the dominant species of red tide organisms to obtain the screened influencing factors of the dominant species of red tide organisms; Using a logistic regression algorithm, a naive Bayes algorithm and a neural network algorithm to respectively build models based on the influencing factors of the dominant species of red tide organisms after the screening, respectively obtain prediction models of the dominant species of red tide organisms; Determining prediction errors of a plurality of red tide organism dominant species prediction models respectively, and assigning weights to the plurality of red tide organism dominant species prediction models respectively according to the prediction errors; Predictions are made using the plurality of red tide organism dominant species prediction models based on the weights.

[0005] In order to solve the above technical problems, another technical solution adopted by the present invention is: A terminal for predicting dominant species of red tide organisms includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: Collect the influencing factors of the dominant species of red tide organisms in the history of the preset sea area, and use the random forest algorithm to perform feature screening on the influencing factors of the dominant species of red tide organisms to obtain the screened influencing factors of the dominant species of red tide organisms; Using a logistic regression algorithm, a naive Bayes algorithm and a neural network algorithm to respectively build models based on the influencing factors of the dominant species of red tide organisms after the screening, respectively obtain prediction models of the dominant species of red tide organisms; Determining prediction errors of a plurality of red tide organism dominant species prediction models respectively, and assigning weights to the plurality of red tide organism dominant species prediction models respectively according to the prediction errors; Predictions are made using the plurality of red tide organism dominant species prediction models based on the weights.

[0006] The beneficial effects of the present invention are: using a random forest algorithm to perform feature screening on the collected red tide biological dominant species influencing factors, and then using a logistic regression algorithm, a naive Bayes algorithm and a neural network algorithm to respectively perform modeling based on the screened red tide biological dominant species influencing factors, respectively obtaining red tide biological dominant species prediction models, respectively determining the prediction errors of multiple red tide biological dominant species prediction models, and assigning weights to the multiple red tide biological dominant species prediction models based on the errors, and using the multiple red tide biological dominant species prediction models for prediction based on the weights, effectively avoiding errors caused by the randomness of a certain algorithm, and improving the prediction accuracy, thereby accurately and quickly predicting red tides, reducing red tide risks, improving the red tide disaster reduction capabilities of marine fisheries, and ensuring the healthy development of the fishery economy. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A flowchart of a method for predicting dominant species of red tide organisms according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a terminal for predicting dominant species of red tide organisms according to an embodiment of the present invention; Figure 3 This is a prediction flow chart of the method for predicting dominant species of red tide organisms according to an embodiment of the present invention. DETAILED DESCRIPTION

[0008] In order to explain the technical content, achieved objectives and effects of the present invention in detail, the following is an explanation in combination with the implementation modes and the accompanying drawings.

[0009] Please refer to Figure 1 , a method for predicting dominant species of red tide organisms, comprising the steps of: Collect the influencing factors of the dominant species of red tide organisms in the history of the preset sea area, and use the random forest algorithm to perform feature screening on the influencing factors of the dominant species of red tide organisms to obtain the screened influencing factors of the dominant species of red tide organisms; Using a logistic regression algorithm, a naive Bayes algorithm and a neural network algorithm to respectively build models based on the influencing factors of the dominant species of red tide organisms after the screening, respectively obtain prediction models of the dominant species of red tide organisms; Determining prediction errors of a plurality of red tide organism dominant species prediction models respectively, and assigning weights to the plurality of red tide organism dominant species prediction models respectively according to the prediction errors; Predictions are made using the plurality of red tide organism dominant species prediction models based on the weights.

[0010] From the above description, it can be seen that the beneficial effects of the present invention are: using the random forest algorithm to perform feature screening on the collected red tide dominant species influencing factors, and then using the logistic regression algorithm, the naive Bayes algorithm and the neural network algorithm to respectively model the red tide dominant species influencing factors, respectively obtain the red tide dominant species prediction models, respectively determine the prediction errors of multiple red tide dominant species prediction models, and assign weights to the multiple red tide dominant species prediction models based on them, and use multiple red tide dominant species prediction models for prediction based on the weights, which effectively avoids the errors caused by the randomness of a certain algorithm and improves the prediction accuracy, thereby accurately and quickly predicting red tides, reducing the risk of red tides, improving the red tide disaster reduction capabilities of marine fisheries, and ensuring the healthy development of the fishery economy.

[0011] Furthermore, the random forest algorithm is used to perform feature screening on the influencing factors of the dominant species of red tide organisms, and the screened influencing factors of the dominant species of red tide organisms include: Taking the influencing factors of the dominant species of red tide organisms as a feature set, and using a random forest algorithm to calculate the importance of each feature in the feature set; Sorting the features in the feature set in descending order according to the importance to obtain a sorted feature set; Eliminate the features in the sorted feature set in order according to a preset elimination ratio to obtain a new feature set; Returning to execute the calculation of the importance of each feature in the feature set until the number of features in the latest feature set reaches a preset number, thereby obtaining each feature set; Obtaining the out-of-bag error corresponding to each feature set; The feature set with the lowest out-of-bag error is determined as the influencing factor of the dominant species of red tide organisms after screening.

[0012] From the above description, it can be seen that determining the feature set with the lowest out-of-bag error as the influencing factor of the dominant species of red tide organisms after screening can ensure the accuracy of the prediction model constructed subsequently.

[0013] Further, the using the random forest algorithm to calculate the importance of each feature in the feature set includes: For each decision tree, select corresponding out-of-bag data and calculate the first out-of-bag data error of the decision tree; After adding random noise interference to the features of all samples of the out-of-bag data, a second out-of-bag data error of the decision tree is calculated; The number of decision trees in the forest is determined, and the importance of the feature is calculated based on the first out-of-bag data error, the second out-of-bag data error, and the number of decision trees.

[0014] From the above description, we can see that if the accuracy of the out-of-bag data drops significantly after adding random noise, it means that this feature has a great impact on the prediction result of the sample, which means that the importance is relatively high. Therefore, calculating the importance of the feature based on the first out-of-bag data error, the second out-of-bag data error and the number of decision trees can effectively determine the importance of the feature.

[0015] Furthermore, the plurality of red tide organism dominant species prediction models include a first red tide organism dominant species prediction model constructed using the logistic regression algorithm, specifically: ; Where P(Y=1|X) represents the probability that a given feature X comes from the positive class, X represents the input feature vector, Y represents the positive class, W represents the feature weight vector, and b represents the bias term.

[0016] From the above description, it can be seen that the first red tide biological dominant species prediction model is constructed using the logistic regression algorithm, which has a faster training speed and better interpretability.

[0017] Furthermore, the plurality of red tide organism dominant species prediction models also include a second red tide organism dominant species prediction model constructed using the naive Bayes algorithm, specifically: ; Where P(C|X) represents the probability that a sample belongs to category C given feature X, P(X|C) represents the probability of observing feature X under category C, P(C) represents the prior probability of category C, and P(X) represents the prior probability of feature X.

[0018] From the above description, it can be seen that the second red tide biological dominant species prediction model is constructed using the naive Bayes algorithm, which is simple to implement and can complete the prediction more quickly.

[0019] Further, the respectively determining the prediction errors of the plurality of red tide organism dominant species prediction models comprises: Using the plurality of red tide organism dominant species prediction models respectively to predict the red tide organism dominant species in the historical time period of the preset sea area, and obtaining prediction results corresponding to the red tide organism dominant species prediction models one by one; Obtaining the actual dominant species of red tide organisms in the historical time period of the preset sea area; The prediction errors of the plurality of red tide organism dominant species prediction models are obtained according to the plurality of prediction results and the actual red tide organism dominant species.

[0020] From the above description, it can be known that the prediction errors of the red tide organism dominant species prediction models are obtained according to each prediction result and the actual red tide organism dominant species, so as to adjust the prediction results of each red tide organism dominant species prediction model using weights according to the prediction error.

[0021] Furthermore, assigning weights to the plurality of red tide biological dominant species prediction models respectively according to the prediction errors includes: ; ; ; Wherein, Q1 represents the first weight of the first red tide dominant species prediction model, Q2 represents the second weight of the second red tide dominant species prediction model, Q3 represents the third weight of the third red tide dominant species prediction model, e1 represents the prediction error of the first red tide dominant species prediction model, e2 represents the prediction error of the second red tide dominant species prediction model, and e3 represents the prediction error of the third red tide dominant species prediction model.

[0022] From the above description, it can be seen that by assigning weights to multiple red tide biological dominant species prediction models according to the prediction errors, the prediction errors of each prediction model can be balanced, thereby improving the accuracy of subsequent predictions.

[0023] Further, the prediction based on the weights using the plurality of red tide organism dominant species prediction models comprises: Use the multiple red tide organism dominant species prediction models to predict the red tide organism dominant species in the preset sea area to obtain multiple initial prediction results; The multiple initial prediction results are multiplied by the corresponding weights one by one and then added together to obtain the final prediction result.

[0024] From the above description, it can be seen that the final prediction result is obtained by multiplying multiple initial prediction results one by one with the corresponding weights and then adding them together, which effectively improves the prediction results so as to reduce the risk of red tides and improve the red tide disaster reduction capabilities of marine fisheries.

[0025] Furthermore, after collecting the influencing factors of the dominant species of red tide organisms in the preset sea area history, the method further includes: Preprocessing the influence factor of the dominant species of red tide organisms to obtain the preprocessed influence factor of the dominant species of red tide organisms; The random forest algorithm is used to perform feature screening on the influencing factors of the dominant species of red tide organisms, and the screened influencing factors of the dominant species of red tide organisms include: The random forest algorithm is used to perform feature screening on the pretreated red tide organism dominant species influencing factors to obtain the screened red tide organism dominant species influencing factors.

[0026] From the above description, it can be seen that through preprocessing, outliers and missing values ​​in the influencing factors of dominant species of red tide organisms are removed to ensure the reliability of the prediction model constructed subsequently.

[0027] Please refer to Figure 2 Another embodiment of the present invention provides a terminal for predicting dominant species of red tide organisms, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, each step in the above-mentioned method for predicting dominant species of red tide organisms is implemented.

[0028] The above-mentioned method and terminal for predicting dominant species of red tide organisms of the present invention can be applied to red tide prediction scenarios, and are described below through specific implementation methods: Please refer to Figure 1 and Figure 3 , Embodiment 1 of the present invention is: A method for predicting dominant species of red tide organisms comprises the following steps: S1. Collect the influencing factors of the dominant species of red tide organisms in the preset sea area history, and use the random forest algorithm to perform feature screening on the influencing factors of the dominant species of red tide organisms to obtain the screened influencing factors of the dominant species of red tide organisms, such as Figure 3 As shown, specifically including S11-S17: S11. Collect the influencing factors of dominant red tide species in the preset sea area history.

[0029] In an optional implementation, the influencing factors of dominant species of red tide organisms from observation equipment such as buoys or fish rafts in a preset sea area in the past ten years are collected.

[0030] The influencing factors of the dominant species of red tide organisms include the original monitoring parameters such as dissolved oxygen concentration, chlorophyll, water temperature, salt, pH value, wind field (u), wind field (v), dissolved oxygen saturation, turbidity, tide, air temperature, air pressure, daily increase in water temperature, daily increase in dissolved oxygen, daily increase in chlorophyll, daily increase in air temperature, daily maximum change in water temperature, daily maximum change in dissolved oxygen, daily maximum change in pH value, daily maximum change in chlorophyll, daily maximum change in air temperature, daily maximum change in tide height, etc.

[0031] S12, taking the influencing factors of the dominant species of red tide organisms as a feature set, and using a random forest algorithm to calculate the importance of each feature in the feature set, specifically including S121-S123: S121, taking the influencing factors of the dominant species of red tide organisms as a feature set, and selecting corresponding out-of-bag (OOB) data for each decision tree to calculate the first out-of-bag data error of the decision tree.

[0032] The out-of-bag data refers to the data that is not involved in the establishment of the decision tree each time the decision tree is established. The out-of-bag data is used to evaluate the performance of the decision tree and calculate the prediction error rate of the model.

[0033] S122, after adding random noise interference to the features of all samples of the out-of-bag data, calculate the second out-of-bag data error of the decision tree.

[0034] In an optional implementation, the random noise interference is to randomly change the value of the sample at feature X.

[0035] S123, determining the number of decision trees in the forest, and calculating the importance of the feature according to the first out-of-bag data error, the second out-of-bag data error, and the number of decision trees, specifically: I=∑(errOOB2-errOOB1) / N; Where I represents the importance of the feature, errOOB2 represents the second out-of-bag data error, errOOB1 represents the first out-of-bag data error, and N represents the number of decision trees.

[0036] If the out-of-bag data accuracy drops significantly after adding random noise interference (i.e., errOOB2 increases), it means that this feature has a great impact on the prediction results of the sample, which means that it is more important.

[0037] S13. Sort the features in the feature set in descending order according to the importance to obtain a sorted feature set.

[0038] S14, eliminating features in the sorted feature set in order according to a preset elimination ratio to obtain a new feature set.

[0039] The preset elimination ratio is set according to actual conditions.

[0040] S15. Return to execute S12 until the number of features in the latest feature set reaches a preset number, and each feature set is obtained.

[0041] The feature sets are feature sets obtained in each cycle.

[0042] S16, obtaining the out-of-bag error corresponding to each feature set.

[0043] S17, determining the feature set with the lowest out-of-bag error as the influencing factor of the dominant species of red tide organisms after screening.

[0044] In an optional embodiment, if Figure 3 As shown, S1 is specifically: Collect the influencing factors of the dominant species of red tide organisms in the history of the preset sea area; preprocess the influencing factors of the dominant species of red tide organisms to obtain the preprocessed influencing factors of the dominant species of red tide organisms; use the random forest algorithm to perform feature screening on the preprocessed influencing factors of the dominant species of red tide organisms to obtain the screened influencing factors of the dominant species of red tide organisms, so as to remove outliers and missing values.

[0045] S2, respectively using the logistic regression algorithm, the naive Bayes algorithm and the neural network algorithm to model the influencing factors of the dominant species of red tide organisms after the screening, and respectively obtaining the prediction models of the dominant species of red tide organisms, such as Figure 3 shown.

[0046] Among them, multiple red tide organism dominant species prediction models include a first red tide organism dominant species prediction model constructed using the logistic regression algorithm, a second red tide organism dominant species prediction model constructed using the naive Bayes algorithm, and a third red tide organism dominant species prediction model constructed using the neural network algorithm.

[0047] The logistic regression algorithm uses a linear function to combine the weights of the input features, and then maps the linear combination to a probability value between 0 and 1 through a logistic function (also called a sigmoid function). The first red tide biological dominant species prediction model is: ; Where P(Y=1|X) represents the probability that a given feature X comes from the positive class, X represents the input feature vector, Y represents the positive class, W represents the feature weight vector, and b represents the bias term.

[0048] The naive Bayes algorithm is based on Bayes' theorem and is used to calculate the probability that a sample belongs to a certain category given a certain feature. It assumes that the features are independent of each other (naive assumption). This is a simplified assumption, so it is called "naive" Bayes. The second red tide biological dominant species prediction model is: ; Where P(C|X) represents the probability that a sample belongs to category C given feature X, P(X|C) represents the probability of observing feature X under category C, P(C) represents the prior probability of category C, and P(X) represents the prior probability of feature X.

[0049] In the neural network algorithm, in order to avoid a large number of neuron deaths while having nonlinear processing capabilities, the activation function in the neural network algorithm is modified to Leaky-Relu. At the same time, in order to solve overfitting, L1 regularization is added to the loss function to increase the running speed while reducing overfitting, and a Dropout layer is added to each hidden layer to improve the generalization ability of the model. Therefore, the activation function is: ; In the formula, x represents the input vector and α represents the hyperparameter, which makes the model more robust. When there is noise in the training samples, the LeakyReLU function can effectively reduce the overfitting of the model to the training data.

[0050] S3, respectively determining the prediction errors of a plurality of red tide organism dominant species prediction models, and assigning weights to the plurality of red tide organism dominant species prediction models according to the prediction errors, such as Figure 3 As shown, specifically including S31-S34: S31. Use the plurality of red tide organism dominant species prediction models respectively to predict the red tide organism dominant species in the historical time period of the preset sea area, and obtain prediction results corresponding to the red tide organism dominant species prediction models one by one.

[0051] In an optional implementation, the historical time period is the past five years.

[0052] S32, obtaining the actual dominant species of red tide organisms in the historical time period of the preset sea area.

[0053] S33. Obtain prediction errors of the plurality of red tide organism dominant species prediction models according to the plurality of prediction results and the actual red tide organism dominant species.

[0054] S34, assigning weights to the plurality of red tide organism dominant species prediction models respectively according to the prediction errors, specifically: ; ; ; Wherein, Q1 represents the first weight of the first red tide dominant species prediction model, Q2 represents the second weight of the second red tide dominant species prediction model, Q3 represents the third weight of the third red tide dominant species prediction model, e1 represents the prediction error of the first red tide dominant species prediction model, e2 represents the prediction error of the second red tide dominant species prediction model, and e3 represents the prediction error of the third red tide dominant species prediction model.

[0055] S4, using the plurality of red tide organism dominant species prediction models to make predictions based on the weights, specifically including S41-S42: S41, using the plurality of red tide organism dominant species prediction models to predict the red tide organism dominant species in the preset sea area, and obtaining a plurality of initial prediction results.

[0056] S42: multiply the multiple initial prediction results by the corresponding weights one by one and then add them up to obtain the final prediction result.

[0057] Please refer to Figure 2 , Embodiment 2 of the present invention is: A terminal for predicting dominant species of red tide organisms comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, each step of the method for predicting dominant species of red tide organisms in the first embodiment is implemented.

[0058] In summary, the present invention provides a method and terminal for predicting dominant species of red tide organisms, which use a random forest algorithm to perform feature screening on the collected dominant species influencing factors of red tide organisms, and then use a logistic regression algorithm, a naive Bayes algorithm and a neural network algorithm to respectively perform modeling based on the screened dominant species influencing factors of red tide organisms, respectively obtain dominant species prediction models of red tide organisms, respectively determine the prediction errors of multiple dominant species prediction models of red tide organisms, and assign weights to multiple dominant species prediction models based on them, and use multiple dominant species prediction models of red tide organisms for prediction based on the weights, which effectively avoids the errors caused by the randomness of a certain algorithm, improves the prediction accuracy, and thus accurately and quickly predicts red tides, reduces the risk of red tides, improves the disaster reduction capacity of marine fisheries red tides, and ensures the healthy development of the fishery economy; and, according to each prediction result and the actual dominant species of red tide organisms, respectively obtains the prediction error of the dominant species prediction model of red tide organisms, so as to adjust the prediction results of each dominant species prediction model of red tide organisms using weights according to the prediction error.

[0059] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for predicting dominant species of red tide organisms, characterized in that: Includes steps: Collect the influencing factors of the dominant species of red tide organisms in the history of the preset sea area, and use the random forest algorithm to perform feature screening on the influencing factors of the dominant species of red tide organisms to obtain the screened influencing factors of the dominant species of red tide organisms; Using a logistic regression algorithm, a naive Bayes algorithm and a neural network algorithm to respectively build models based on the influencing factors of the dominant species of red tide organisms after the screening, respectively obtain prediction models of the dominant species of red tide organisms; Determining prediction errors of a plurality of red tide organism dominant species prediction models respectively, and assigning weights to the plurality of red tide organism dominant species prediction models respectively according to the prediction errors; Predictions are made using the plurality of red tide organism dominant species prediction models based on the weights.

2. A method for predicting dominant species of red tide organisms according to claim 1, characterized in that: The random forest algorithm is used to perform feature screening on the influencing factors of the dominant species of red tide organisms, and the screened influencing factors of the dominant species of red tide organisms include: Taking the influencing factors of the dominant species of red tide organisms as a feature set, and using a random forest algorithm to calculate the importance of each feature in the feature set; Sorting the features in the feature set in descending order according to the importance to obtain a sorted feature set; Eliminate the features in the sorted feature set in order according to a preset elimination ratio to obtain a new feature set; Returning to execute the calculation of the importance of each feature in the feature set until the number of features in the latest feature set reaches a preset number, thereby obtaining each feature set; Obtaining the out-of-bag error corresponding to each feature set; The feature set with the lowest out-of-bag error is determined as the influencing factor of the dominant species of red tide organisms after screening.

3. A method for predicting dominant species of red tide organisms according to claim 2, characterized in that: The use of the random forest algorithm to calculate the importance of each feature in the feature set includes: For each decision tree, select corresponding out-of-bag data and calculate the first out-of-bag data error of the decision tree; After adding random noise interference to the features of all samples of the out-of-bag data, a second out-of-bag data error of the decision tree is calculated; The number of decision trees in the forest is determined, and the importance of the feature is calculated based on the first out-of-bag data error, the second out-of-bag data error, and the number of decision trees.

4. A method for predicting dominant species of red tide organisms according to claim 1, characterized in that: The plurality of red tide organism dominant species prediction models include a first red tide organism dominant species prediction model constructed using the logistic regression algorithm, specifically: ； Where P(Y=1|X) represents the probability that a given feature X comes from the positive class, X represents the input feature vector, Y represents the positive class, W represents the feature weight vector, and b represents the bias term.

5. A method for predicting dominant species of red tide organisms according to claim 4, characterized in that: The plurality of red tide organism dominant species prediction models also include a second red tide organism dominant species prediction model constructed using the naive Bayes algorithm, specifically: ； Where P(C|X) represents the probability that a sample belongs to category C given feature X, P(X|C) represents the probability of observing feature X under category C, P(C) represents the prior probability of category C, and P(X) represents the prior probability of feature X.

6. A method for predicting dominant species of red tide organisms according to claim 1, characterized in that: The prediction errors of the prediction models for respectively determining the dominant species of red tide organisms include: Using the plurality of red tide organism dominant species prediction models respectively to predict the red tide organism dominant species in the historical time period of the preset sea area, and obtaining prediction results corresponding to the red tide organism dominant species prediction models one by one; Obtaining the actual dominant species of red tide organisms in the historical time period of the preset sea area; The prediction errors of the plurality of red tide organism dominant species prediction models are obtained according to the plurality of prediction results and the actual red tide organism dominant species.

7. A method for predicting dominant species of red tide organisms according to claim 1, characterized in that: The step of assigning weights to the plurality of red tide biological dominant species prediction models respectively according to the prediction errors comprises: ； ； ； Wherein, Q1 represents the first weight of the first red tide dominant species prediction model, Q2 represents the second weight of the second red tide dominant species prediction model, Q3 represents the third weight of the third red tide dominant species prediction model, e1 represents the prediction error of the first red tide dominant species prediction model, e2 represents the prediction error of the second red tide dominant species prediction model, and e3 represents the prediction error of the third red tide dominant species prediction model.

8. A method for predicting dominant species of red tide organisms according to claim 1, characterized in that: The prediction using the plurality of red tide organism dominant species prediction models based on the weights comprises: Use the multiple red tide organism dominant species prediction models to predict the red tide organism dominant species in the preset sea area to obtain multiple initial prediction results; The multiple initial prediction results are multiplied by the corresponding weights one by one and then added together to obtain the final prediction result.

9. A method for predicting dominant species of red tide organisms according to claim 1, characterized in that: After collecting the influencing factors of the dominant species of red tide organisms in the preset sea area history, it also includes: Preprocessing the influence factor of the dominant species of red tide organisms to obtain the preprocessed influence factor of the dominant species of red tide organisms; The random forest algorithm is used to perform feature screening on the influencing factors of the dominant species of red tide organisms, and the screened influencing factors of the dominant species of red tide organisms include: The random forest algorithm is used to perform feature screening on the pretreated red tide organism dominant species influencing factors to obtain the screened red tide organism dominant species influencing factors.

10. A terminal for predicting dominant species of red tide organisms, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, each step of the method for predicting dominant species of red tide organisms described in any one of claims 1 to 9 is implemented.

Images