CN111667112A - Fishery resource abundance gray prediction model optimization method and application thereof - Google Patents

Abstract

The invention discloses an optimization method for a grey prediction model of fishery resource abundance and an application thereof. The steps are: selecting a CPUE sequence as a standard sequence for establishing a grey prediction model of resource abundance; using a grey relational analysis method for the standard sequence to calculate resource abundance The gray correlation degree of the influencing factors of high degree of gray correlation is selected as the factor of the gray prediction model of resource abundance; using the discrete GM model, the selected factors are used to establish the gray prediction model of resource abundance, and the established resource abundance gray prediction model The prediction models include GM(0,N) model and GM(1,N) model; the validity of each prediction model is analyzed, and the model with the smallest relative error is selected as the optimal prediction model. In the prediction method of the present invention, the CPUE sequence is optimized to overcome the defect of poor prediction stability of the current gray system model; GM models of different orders are established, and the optimal prediction model is obtained from them, thereby improving the prediction accuracy.

Description

A grey prediction model optimization method of fishery resource abundance and its application

technical field

The invention belongs to the technical field of pelagic fishing flood forecasting, and relates to an optimization method for a grey prediction model of fishery resource abundance and its application, in particular to an optimization method for a grey prediction model of fishery resource abundance based on environmental factors and its application.

Background technique

Pelagic fish is an important economic species, and its resource abundance changes are closely related to factors such as climate and marine environment. Resource abundance prediction is one of the main forecast contents in fishery forecasting. Scientific prediction of resource quantity and assessment of resource abundance are beneficial to the sustainable utilization and scientific management of fishery resources. The establishment of a fishery forecast model, especially the prediction of resource abundance between years, usually encounters the problems of too few samples and too many environmental factors involved in the forecast. These two problems are the difficulties in establishing a resource abundance prediction model.

Compared with other research methods (multiple linear regression, multiple nonlinear regression and habitat index), grey system theory is a subject of uncertain system theory, which has the advantage of allowing a small number of samples and obeying arbitrary distributions. At present, although this method has been applied in the prediction of resource abundance of some economic species, there are still problems such as low prediction accuracy and poor stability.

Therefore, it is of great practical significance to develop a resource abundance prediction method based on grey system theory with high prediction accuracy and high stability.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the defects of low prediction accuracy and poor stability in the prior art, and to provide a resource abundance prediction method based on grey system theory with high prediction accuracy and high stability.

To achieve the above object, the present invention provides the following technical solutions:

An optimization method for a grey prediction model of fishery resource abundance, which is applied to electronic equipment, and the steps are as follows:

(1) For the CPUE sequence of fish C in sea area B in time period A, intercept the CPUE sequence of any time period, and establish a GM(1,1) model corresponding to each intercepted CPUE sequence, and establish m GMs in total (1,1) model, calculate the relative error of each GM(1,1) model separately, and select the CPUE sequence corresponding to the GM(1,1) model with the smallest relative error as the standard sequence for establishing the grey prediction model of resource abundance;

(2) According to the standard sequence selected in step (1), the gray correlation degree of the influencing factor of resource abundance is calculated by using the method of grey correlation analysis;

(3) Select the influencing factor with large grey correlation degree as the factor of resource abundance grey prediction model;

(4) Use the discrete GM model and use the factors selected in step (3) to establish a resource abundance gray prediction model. The established resource abundance gray prediction model includes the GM(0,N) model and the GM(1,N) model. GM The (0,N) model and the GM(1,N) model are both linear dynamic models. The GM(1,N) model is a first-order derivative calculation model, which is more complicated than the GM(0,N) model. Simultaneously constructing GM prediction models of different orders can provide more models for later optimization, and then select the optimal model from them to improve the prediction accuracy of the model;

(5) Carry out validity analysis on the prediction model obtained in step (4). The validity analysis includes relative error analysis. The relative error is obtained by comparing the CPUE value calculated by using the prediction model and the real CPUE value, and the smallest relative error is selected. The model is the optimal prediction model.

In the method for optimizing the grey prediction model of fishery resource abundance of the present invention, before processing (before establishing the grey prediction model of resource abundance), the CPUE sequence is firstly optimized, the CPUE sequence of any time period is intercepted and a GM(1,1) model is established , select the CPUE sequence corresponding to the GM(1,1) model with the smallest relative error as the standard sequence for the subsequent establishment of the gray prediction model of resource abundance, which overcomes the defect of poor prediction stability of the current gray system model to a certain extent. In the grey prediction model of resource abundance, the 0-order and 1-order GM prediction models containing different influencing factors are established, and the model with the smallest relative error is selected as the optimal prediction model from multiple prediction models of different orders. It greatly improves the accuracy of the prediction model and provides a basis for more scientific and efficient fishery production. The method of the invention can be used to predict the middle and pelagic fish in the ocean, has good applicability, can play a good guiding role in marine fishery production, can significantly improve fishing efficiency, reduce fishing cost, and has great application prospects. The obtained optimal prediction model is not static, and the optimal prediction model can be re-obtained according to the latest data obtained in real time. The method of the invention has good adaptability and good application prospect.

As the preferred technical solution:

In the above-mentioned optimization method for grey prediction model of fishery resource abundance, the time period A is from 1998 to 2016; the coordinate range of the sea area B is 35°～45°N, 140°～179°E; Said fish C is the North Pacific squid. The scope of protection of the present invention is not limited to this, only taking the fish in this time period and in this sea area as an example, the prediction method of the present invention can be applied to the resource abundance of pelagic fish in any time period and in any sea area. Degree prediction.

A method for optimizing the grey prediction model of fishery resource abundance as described above, the standard sequence for establishing the grey prediction model of resource abundance is 1998-2005, and the coordinate ranges are 35°～45°N, 140°～179°E CPUE sequences of North Pacific squid. The standard sequence selected here for establishing the grey prediction model of resource abundance is based on the aforementioned data, and those skilled in the art can select the appropriate CPUE sequence as the standard sequence according to the actual data.

In the above-mentioned optimization method for the grey prediction model of fishery resource abundance, the relative errors of the multiple GM(1,1) models obtained in step (1) are the same and are all the minimum values, then the model with the smallest variance of these models is selected The corresponding CPUE sequence is used as the standard sequence for establishing the grey prediction model of resource abundance.

A method for optimizing a fishery resource abundance grey prediction model as described above, the influencing factors of the resource abundance include the Pacific Interdecadal Oscillation Index (PDO), the sea surface temperature of spawning grounds (SGSST), and the sea surface of fattening farms. Temperature (FGSST), chlorophyll concentration in spawning grounds (SGC), and chlorophyll concentration in feedlots (FGC). The scope of protection of the present invention is not limited to this, the impact factor of resource abundance given here only corresponds to the North Pacific squid, and for other fish, those skilled in the art can reasonably select the impact of resource abundance according to the actual situation factor.

A method for optimizing a grey prediction model of fishery resource abundance as described above, the step (2) is specifically: taking the standard sequence selected in step (1) as the parent sequence, and taking the corresponding influence factors as the subsequences, respectively calculating the The gray absolute correlation between the subsequence and the parent sequence.

A method for optimizing a grey prediction model of fishery resource abundance as described above, the factors of the grey prediction model for resource abundance are respectively the sea surface temperature (FGSST) of the fattening farm in October and the Pacific Interdecadal Oscillation Index (PDO) in October. , sea surface temperature (SGSST) of spawning grounds in February, chlorophyll concentration (SGC) of spawning grounds in March and chlorophyll concentration (FGC) of feedlots in August. The prediction model mentioned here and below is only to demonstrate the operation logic of the prediction method of the present invention with partial data, and the protection scope of the present invention is not limited to this. The selected factors of the gray prediction model of resource abundance are not limited to these five factors. Of course, the number of prediction models will also change with the number of factors of the selected gray prediction model of resource abundance.

A kind of fishery resource abundance gray prediction model optimization method as above, the described resource abundance gray prediction model established includes the following models:

Model I, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) and the GM(0,6) model of five factors of feedlot chlorophyll concentration (FGC) in August;

Model II, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration (FGC) in August four GM(0,5) model of the factors;

Model III, including the Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea surface temperature (SGSST) in February, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration (FGC) in August Four-factor GM(0,5) model;

Model IV, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and feedlot chlorophyll concentration (FGC) in August Four-factor GM(0,5) model;

Model V, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) GM(0,5) model with four factors;

Model VI, including feedlot sea surface temperature in October (FGSST), spawning ground sea temperature in February (SGSST), spawning ground chlorophyll concentration in March (SGC) and feedlot chlorophyll concentration in August (FGC) ) GM(0,5) model with four factors;

Model VII, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) and the GM(1,6) model of five factors of feedlot chlorophyll concentration (FGC) in August;

Model VIII, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration in August (FGC) IV GM(1,5) model of the factors;

Model IX, including October Pacific Decadal Oscillation Index (PDO), spawning ground sea surface temperature (SGSST) in February, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration (FGC) in August Four-factor GM(1,5) model;

Model X, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and feedlot chlorophyll concentration (FGC) in August Four-factor GM(1,5) model;

Model XI, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) GM(1,5) model with four factors;

Model XII, including feedlot sea surface temperature in October (FGSST), spawning ground sea temperature in February (SGSST), spawning ground chlorophyll concentration in March (SGC) and feedlot chlorophyll concentration in August (FGC) ) four-factor GM(1,5) model.

In the above-mentioned optimization method for grey prediction model of fishery resource abundance, model IV is the optimal prediction model. On the basis of conforming to common knowledge in the art, the above preferred conditions can be combined arbitrarily to obtain preferred examples of the present invention.

The present invention also provides an electronic device, comprising one or more processors, one or more memories, one or more programs and a data collection device;

The data collection device is used to obtain the CPUE sequence of the fishes C in the sea area B within the time period A, and the one or more programs are stored in the memory, and when the one or more programs are executed by the processor At the time, the electronic device is made to execute the above-mentioned method for optimizing the grey prediction model of fishery resource abundance.

Beneficial effects:

(1) In the method for optimizing the grey prediction model of fishery resource abundance of the present invention, before processing (before establishing the grey prediction model of resource abundance), the CPUE sequence is first optimized, the CPUE sequence of any time period is intercepted and the GM (1, 1) Model, the CPUE sequence corresponding to the GM(1,1) model with the smallest relative error is selected as the standard sequence for the subsequent establishment of the gray prediction model of resource abundance, which overcomes the defect of poor prediction stability of the current gray system model to a certain extent;

(2) In the method for optimizing the gray prediction model of fishery resource abundance of the present invention, when establishing the gray prediction model of resource abundance, 0-order and 1-order GM prediction models containing different influencing factors are simultaneously established. Among the prediction models, the model with the smallest relative error is selected as the optimal prediction model, which can greatly improve the accuracy of the prediction model and provide a basis for more scientific and effective fishery production;

(3) The electronic device of the present invention has simple structure and low cost, and can quickly realize the prediction of the abundance of marine fishery resources.

Description of drawings

Fig. 1 is the schematic flow sheet of the grey prediction model optimization method of fishery resource abundance of the present invention;

Fig. 2 is the relative error diagram of multiple GM(1,1) models in step (1);

Fig. 3 is the average relative error comparison diagram of each prediction model in step (6);

Figure 4 is a fitting diagram of the predicted value of the optimal prediction model;

FIG. 5 is a schematic structural diagram of an electronic device of the present invention.

Detailed ways

In order to make the purposes, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention.

Example 1

An optimization method for grey prediction model of fishery resource abundance, as shown in Figure 1, is as follows:

(1) CPUE (unit fishing effort per year) sequence of North Pacific squid in the coordinate range of 35°～45°N and 140°～179°E from 1998 to 2016 , latitude, daily production (in tons) and fishing effort (in number of boats) from 1998 to 2016, the coordinates range from 35° to 45°N and 140° to 179°E for the North Pacific squid. fishery production statistics), intercept the CPUE sequence of any time period, and establish a GM(1,1) model corresponding to each intercepted CPUE sequence. A total of m GM(1,1) models are established, respectively. Calculate the relative error of each GM(1,1) model, and select the CPUE sequence corresponding to the GM(1,1) model with the smallest relative error as the standard sequence for establishing the gray prediction model of resource abundance, such as multiple GM(1,1) ) models have the same relative error and are all the minimum values, then select the corresponding CPUE sequence of the model with the smallest variance of these models as the standard sequence for establishing the grey prediction model of resource abundance. Selected multiple GM(1,1) models The relative error map of , is shown in Figure 2. Therefore, the CPUE sequence of the North Pacific squid from 1998 to 2005 with the coordinates ranging from 35° to 45°N and 140° to 179°E was selected (the average relative error was the smallest, 6.28%). ) as the standard sequence for establishing the grey prediction model of resource abundance;

(2) For the standard sequence selected in step (1), use the method of gray correlation analysis to calculate the gray correlation degree of the influencing factor of resource abundance. Specifically, the standard sequence selected in step (1) is used as the parent sequence, and the Corresponding to each impact factor as a sub-sequence, the gray absolute correlation degree of each sub-sequence and the parent sequence is calculated respectively (the gray correlation coefficient of each impact factor sub-sequence and the CPUE parent sequence is shown in Table 1). Influencing factors include Pacific Decadal Oscillation Index (PDO), spawning ground sea surface temperature (SGSST), feedlot sea surface temperature (FGSST), spawning ground chlorophyll concentration (SGC) and feedlot chlorophyll concentration (FGC) );

Table 1 The grey correlation coefficient of each impact factor subsequence and CPUE parent sequence

(3) It can be seen from Table 1 that the influencing factors with large gray correlation are the sea surface temperature (FGSST) of the fattening field in October, the Pacific Interdecadal Oscillation Index (PDO) in October, and the sea surface temperature of the spawning ground in February (SGSST). ), the chlorophyll concentration (SGC) of the spawning grounds in March and the chlorophyll concentration (FGC) of the fattening farms in August, so the above influencing factors were selected as the factors of the grey prediction model of resource abundance;

(4) Using the discrete GM model and using the factors selected in step (3) to establish a resource abundance gray prediction model, the established resource abundance gray prediction model includes a GM(0,N) model and a GM(1,N) model. Includes the following models:

Model I, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) and the GM(0,6) model of five factors of feedlot chlorophyll concentration (FGC) in August;

Model II, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration (FGC) in August four GM(0,5) model of the factors;

Model III, including the Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea surface temperature (SGSST) in February, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration (FGC) in August Four-factor GM(0,5) model;

Model IV, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and feedlot chlorophyll concentration (FGC) in August Four-factor GM(0,5) model;

Model V, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) GM(0,5) model with four factors;

Model VI, including feedlot sea surface temperature in October (FGSST), spawning ground sea temperature in February (SGSST), spawning ground chlorophyll concentration in March (SGC) and feedlot chlorophyll concentration in August (FGC) ) GM(0,5) model with four factors;

Model VII, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) and the GM(1,6) model of five factors of feedlot chlorophyll concentration (FGC) in August;

Model VIII, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration in August (FGC) IV GM(1,5) model of the factors;

Model IX, including October Pacific Decadal Oscillation Index (PDO), spawning ground sea surface temperature (SGSST) in February, spawning ground chlorophyll concentration (SGC) in March, and feedlot chlorophyll concentration (FGC) in August Four-factor GM(1,5) model;

Model X, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and feedlot chlorophyll concentration (FGC) in August Four-factor GM(1,5) model;

Model XI, including feedlot sea surface temperature (FGSST) in October, Pacific Decadal Oscillation Index (PDO) in October, spawning ground sea temperature in February (SGSST), and spawning ground chlorophyll concentration in March (SGC) ) GM(1,5) model with four factors;

Model XII, including feedlot sea surface temperature in October (FGSST), spawning ground sea temperature in February (SGSST), spawning ground chlorophyll concentration in March (SGC) and feedlot chlorophyll concentration in August (FGC) ) GM(1,5) model with four factors;

(5) Analyze the validity of each prediction model obtained in step (4), and use the fishery production data in 2006 for verification. The verification results are shown in Tables 2 and 3. The average relative error of each prediction model is compared in Figure 3. As shown, the models 1 to 6 on the left in FIG. 3 are in one-to-one correspondence with the models I to VI of step (4), respectively, and the models 1 to 6 on the right are in one-to-one correspondence with models VII to XII of step (4). ;

Table 2 Relative error of GM(0,N) prediction model for North Pacific softfish resource abundance

Table 3 The relative errors of the GM(1,N) prediction model for the abundance of squid in the North Pacific

"Error" in Table 2 and Table 3 is the average relative error, "Validation" is the error rate between the 2006 fishery production data predicted by the model and the actual fishery production data in 2006, and the unit in the table is %;

It can be seen from Figure 2 that the GM(0,N) and GM(1,N) prediction models with environmental factors are almost all more accurate than (except for model XII) the GM(1,1) model and have a 0th order. The gray prediction models of all are more accurate than the first-order fitting.

It can be seen from Table 2 and Table 3 that the average fitting errors of the models in each GM(0,N) model are in descending order: Model I > Model III > Model II > Model V > Model IV > Model VI; 1,N) The average fitting error of the models in the model is from small to large: model X > model VII > model VIII > model IX > model XI > model XII; from the verification results, model IV and model X are much higher than other models , the relative errors were 1.18% (the lowest) and 1.20%, so Model IV was selected as the optimal prediction model for the abundance of North Pacific squid.

Model IV is used for prediction, and its prediction fitting diagram is shown in Figure 4. It can be seen from the figure that the variation trend of CPUE is basically the same, and the variation range of the fitting value predicted by the model is small. From the perspective of the parameters of the prediction model, - The value of a is -1.71 (Table 4), which satisfies the condition of the medium and long-term forecast model (-a<0.3).

Table 4 Parameter values for the factors of Model IV

It has been verified that the prediction method of the present invention first optimizes the CPUE sequence before processing (before establishing the resource abundance gray prediction model), intercepts the CPUE sequence of any time period and establishes a GM(1,1) model, and selects the relative error. The CPUE sequence corresponding to the smallest GM(1,1) model is used as the standard sequence for the subsequent establishment of the gray prediction model of resource abundance, which overcomes the defect of poor prediction stability of the current gray system model. When establishing the gray prediction model of resource abundance, At the same time, the 0th-order and 1st-order GM prediction models containing different influencing factors are established, and the model with the smallest relative error is selected as the optimal prediction model from multiple prediction models of different orders, which can greatly improve the accuracy of the prediction model. , which provides a basis for more scientific and efficient fishery production, and has great application prospects.

Example 2

An electronic device, as shown in Figure 5, includes one or more processors, one or more memories, one or more programs and a data collection device;

The data collection device is used to obtain the CPUE sequence of fish C in sea area B in time period A (that is, from 1998 to 2016 in Example 1, the coordinates range from 35° to 45°N and 140° to 179°E in the North Pacific Ocean. The CPUE sequence of fish), one or more programs are stored in the memory, and when the one or more programs are executed by the processor, make the electronic device execute the gray prediction model optimization method of fishery resource abundance as described in Embodiment 1.

Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that these are only examples, and various changes may be made to these embodiments without departing from the principle and essence of the present invention. Revise.

Claims (10)

1. a fishery resource abundance grey prediction model optimization method, applied to electronic equipment, is characterized in that, step is as follows: (1) For the CPUE sequence of fish C in sea area B in time period A, intercept the CPUE sequence of any time period, and establish a GM(1,1) model corresponding to each intercepted CPUE sequence, and establish m GMs in total (1,1) model, calculate the relative error of each GM(1,1) model separately, and select the CPUE sequence corresponding to the GM(1,1) model with the smallest relative error as the standard sequence for establishing the grey prediction model of resource abundance; (2) According to the standard sequence selected in step (1), the gray correlation degree of the influencing factor of resource abundance is calculated by using the method of grey correlation analysis; (3) Select the influencing factor with large grey correlation degree as the factor of resource abundance grey prediction model; (4) Using the discrete GM model and using the factors selected in step (3) to establish a resource abundance gray prediction model, the established resource abundance gray prediction model includes a GM(0,N) model and a GM(1,N) model; (5) Carry out validity analysis on the prediction model obtained in step (4). The validity analysis includes relative error analysis. The relative error is obtained by comparing the CPUE value calculated by using the prediction model and the real CPUE value, and the smallest relative error is selected. The model is the optimal prediction model. 2 . The method for optimizing a grey prediction model of fishery resource abundance according to claim 1 , wherein the time period A is from 1998 to 2016; the coordinate range of the sea area B is from 35° to 45°N. 3 . , 140°～179°E; the fish C is North Pacific squid. 3 . The method for optimizing a grey prediction model of fishery resource abundance according to claim 2 , wherein the standard sequence for establishing the grey prediction model for resource abundance is from 1998 to 2005, and the coordinate range is from 35° to 45° 3 . °N, 140°～179°E CPUE sequence of North Pacific squid. 4. a kind of fishery resource abundance grey prediction model optimization method according to claim 3 is characterized in that, the relative errors of the multiple GM (1,1) models obtained in step (1) are the same and are all minimum values, Then, the corresponding CPUE sequence of the model with the smallest variance of these models is selected as the standard sequence for establishing the grey prediction model of resource abundance. 5. a kind of fishery resource abundance grey prediction model optimization method according to claim 4, is characterized in that, the influence factor of described resource abundance comprises Pacific interdecadal shock index, the sea surface temperature of spawning ground, fattening ground sea surface temperature, chlorophyll concentration in spawning grounds, and chlorophyll concentration in feedlots. 6. A kind of fishery resource abundance grey prediction model optimization method according to claim 5, is characterized in that, described step (2) is specifically: take the standard sequence selected in step (1) as the mother sequence, to correspond to each The impact factor is the subsequence, and the gray absolute correlation between each subsequence and the parent sequence is calculated separately. 7. A kind of fishery resource abundance grey prediction model optimization method according to claim 6, is characterized in that, the factor of described resource abundance grey prediction model is respectively the sea surface temperature of the fattening farm in October, the age of the Pacific Ocean in October. Oscillation index, sea surface temperature of spawning grounds in February, chlorophyll concentration of spawning grounds in March and chlorophyll concentration of feedlots in August. 8. a kind of fishery resource abundance grey prediction model optimization method according to claim 7, is characterized in that, described resource abundance grey prediction model of establishment comprises following model: Model I, including the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the sea surface temperature of the spawning ground in February, the chlorophyll concentration of the spawning ground in March, and the chlorophyll concentration of the feedlot in August. GM(0,6) model; Model II, including the GM(0,5) model of four factors including the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the chlorophyll concentration of the spawning ground in March and the chlorophyll concentration of the feedlot in August; Model III, including the GM(0,5) model of four factors including the Pacific Interdecadal Oscillation Index in October, the sea surface temperature of the spawning grounds in February, the chlorophyll concentration of the spawning grounds in March and the chlorophyll concentration of the fattening grounds in August; Model IV, including the feedlot sea surface temperature in October, the Pacific decadal shock index in October, the sea surface temperature of the spawning ground in February (and the chlorophyll concentration of the feedlot in August), a four-factor GM(0,5) model ; Model V, including the GM(0,5) model of four factors including the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the sea surface temperature of the spawning ground in February, and the chlorophyll concentration of the spawning ground in March ; Model VI, including the sea surface temperature of the feedlot in October, the sea surface temperature of the spawning ground in February, the chlorophyll concentration of the spawning ground in March, and the GM(0,5) model of the four factors of the chlorophyll concentration of the feedlot in August ; Model VII, including five factors: the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the sea surface temperature of the spawning ground in February, the chlorophyll concentration of the spawning ground in March and the chlorophyll concentration of the feedlot in August The GM(1,6) model of ; Model VIII, including the GM(1,5) model of four factors including the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the chlorophyll concentration of the spawning ground in March and the chlorophyll concentration of the feedlot in August; Model IX, including the GM(1,5) model of four factors including the Pacific Decadal Oscillation Index in October, the sea surface temperature of the spawning grounds in February, the chlorophyll concentration of the spawning grounds in March and the chlorophyll concentration of the fattening grounds in August; Model X, including the GM(1,5) model of four factors including the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the sea surface temperature of the spawning ground in February and the chlorophyll concentration of the feedlot in August; Model XI, a GM(1,5) model with four factors including the sea surface temperature of the feedlot in October, the Pacific decadal shock index in October, the sea surface temperature of the spawning ground in February, and the chlorophyll concentration of the spawning ground in March ; Model XII, including the sea surface temperature of the feedlot in October, the sea surface temperature of the spawning ground in February, the chlorophyll concentration of the spawning ground in March and the chlorophyll concentration of the feedlot in August. The GM(1,5) model . 9 . The method for optimizing a grey prediction model of fishery resource abundance according to claim 8 , wherein the model IV is an optimal prediction model. 10 . 10. An electronic device, characterized in that it comprises one or more processors, one or more memories, one or more programs and a data collection device; The data collection device is used to obtain the CPUE sequence of the fishes C in the sea area B within the time period A, and the one or more programs are stored in the memory, and when the one or more programs are executed by the processor At the time, the electronic device is made to execute the method for optimizing the gray prediction model of fishery resource abundance according to any one of claims 1 to 9.

Images