A method for establishing a prediction model for organic sorghum yield

Abstract

The present invention particularly relates to a method for establishing an organic sorghum yield prediction model. The method for establishing an organic sorghum yield prediction model first divides the growth cycle of organic sorghum into three periods: the initial planting period, the growth period and the maturity period, and constructs training data sets respectively; preprocesses and screens variables for the three training data sets, screens out the most important characteristic variables, and trains the model; adjusts the model variables and evaluates the feasibility of the model until the modeler's requirements and the judgment reference coefficient are met; finally, the three trained models at different growth stages are saved in a background big data cluster respectively, so as to facilitate the later user to call. The method for establishing an organic sorghum yield prediction model can only predict the yield of organic sorghum at different growth stages, improves the prediction speed and prediction accuracy, and can also provide early warning prompts for users, timely reminding users to take rescue measures, and provide a research and judgment benchmark and reference for organic sorghum.

Description

Technical Field

[0001] The invention relates to the technical field of machine learning algorithms, and in particular to a method for establishing an organic sorghum yield prediction model. Background Art

[0002] Organic sorghum is one of the raw materials frequently used by wineries for brewing, and its output is crucial to the operation of wineries. However, the current prediction of organic sorghum mainly relies on experts to estimate the output of different plots based on their experience. The results of manual prediction are not very accurate and have large errors.

[0003] Based on this, the present invention proposes a method for establishing an organic sorghum yield prediction model. Summary of the invention

[0004] In order to make up for the defects of the prior art, the present invention provides a simple and efficient method for establishing an organic sorghum yield prediction model.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for establishing an organic sorghum yield prediction model, characterized in that it comprises the following steps:

[0007] The first step is to build a training dataset

[0008] The growth cycle of organic sorghum is divided into three periods: the initial planting period, the growth period, and the maturity period. The basic attributes of planting, the natural environment data during the planting period, and the historical data of field management data are used as the initial explanatory variables of the model. The historical yield records of the planting plots are used as the response variables of the model. Three model training data sets are summarized according to the growth stages of organic sorghum.

[0009] The second step is data preprocessing

[0010] Preprocess the three training data sets, including observing the outlier data points in the data sets and removing the information that is suspected to be incorrectly entered by humans;

[0011] Step 3: Variable screening

[0012] In order to improve the stability and prediction accuracy of the model, the preprocessed data was screened for variables, 15 most important characteristic variables were selected, and the characteristic variables were tested for collinearity problems;

[0013] Step 4: Training the model

[0014] The organic sorghum yield was set as the response variable, and other characteristic variables affecting the organic sorghum yield were used as the explanatory variables of the model. The training data of three organic sorghums at different growth stages were used as the training data set to train the model.

[0015] Step 5: Model Diagnosis

[0016] Observe and adjust model variables and evaluate the feasibility of the model, and make adjustments repeatedly until the modeler's requirements and reference coefficients are met;

[0017] Step 6: Model application

[0018] The three trained models of organic sorghum yield in the early planting stage, growth stage and maturity stage are saved in the background big data cluster respectively for the convenience of later users.

[0019] In the first step, the training data set consists of the following three parts:

[0020] Basic properties of planting, including land area, soil temperature, soil moisture, soil pH, and special chemical content information, are obtained through manual records and soil testing;

[0021] Natural environment data during the planting period, including recent temperature changes, light hours, and the impact of special natural disasters, are collected through local meteorological observation stations and base stations where the plots are located;

[0022] Field management data, records of the amount of fertilizer used on the land, the amount of seeds used, the number of pests and diseases, and the number of on-site support provided by agricultural technicians, were obtained through field collection.

[0023] In the second step, when a column of feature data in the training data set is vacant, if the vacant columns account for less than 5% of the total columns, the corresponding column average value is selected to fill in the vacant columns, otherwise the column feature data is deleted or the vacant columns are manually filled.

[0024] In the third step, the concept of variance inflation coefficient in mathematical statistics is used to measure the severity of collinearity of variables in regression modeling, and variables with variance inflation coefficients exceeding 10 are eliminated to obtain three sets of training data of organic sorghum at different growth stages as training data sets for organic sorghum yields at different periods.

[0025] In the fourth step, the Statsmodels professional regression modeling extension tool language package in Python language is used to write the organic sorghum yield prediction model, and the training data of three organic sorghums at different growth stages are used as training data sets to train the models separately, thereby obtaining three organic sorghum yield prediction models at different growth stages.

[0026] In the fifth step, the variables of the model are adjusted and evaluated, including the following steps:

[0027] S1. Use the residual diagnostic graph to observe the degree of fit between the constructed model and the data, observe the outliers and sample points with strong influence, and remove the outliers with strong influence during the model training process;

[0028] S2. Draw a GAM (generalized additive model) diagram in statistics for the important variable indicators in the model to evaluate whether the variable fitting power is reasonable and add alternative model solutions;

[0029] S3. Eliminate the characteristic variables with T-test P values ​​greater than 0.2 in the statistical diagnosis results of the model, because the T-test shows that there is no significant linear relationship between higher characteristic variables and yield;

[0030] S4. Calculate the adjusted R-Square, AIC (Akaike information criterion) and mean square error of the candidate model in the test data set, and finally compare and select the model with the highest explanatory power and degree of fit for the original data.

[0031] In the sixth step, the basic properties of planting, natural environment data during the planting period and field management data are continuously monitored during the growth stage of sorghum, and the monitoring data is input into the characteristic variables of the model. According to the growth stage input by the user, the organic sorghum yield prediction model of the corresponding growth stage is called to regularly predict the yield of organic sorghum;

[0032] If the predicted value is less than the user's demand or experience value, an early warning message will be issued to remind the user to take timely remedial measures for the abnormal situation.

[0033] In the sixth step, if the purchase volume is more than 15% higher than the predicted output during the grain harvesting stage, it is considered that the farmer may have borrowed or passed off inferior products as good ones, and the model will issue a system reminder.

[0034] The beneficial effects of the present invention are as follows: the method for establishing an organic sorghum yield prediction model can not only predict the yield of organic sorghum at different growth stages, thereby improving the prediction speed and prediction accuracy, but also providing early warning prompts for users, promptly reminding users to take rescue measures, and providing a research and judgment benchmark and reference for organic sorghum. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Attached Figure 1 The present invention is a schematic diagram of a method for establishing an organic sorghum yield prediction model. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0038] The method for establishing an organic sorghum yield prediction model comprises the following steps:

[0039] The first step is to build a training dataset

[0040] This model invention is based on the growth characteristics of organic sorghum. Taking into account the different external factors such as the amount of fertilizer and rainfall required for organic sorghum at different growth stages, the growth cycle of organic sorghum is divided into three periods: the initial planting period, the growth period and the maturity period. The basic attributes of planting, the natural environment data during the planting period and the historical data of field management data are used as the initial explanatory variables of the model, and the historical yield records of the planting plots are used as the response variables of the model. Three model training data sets are summarized according to the growth stages of organic sorghum.

[0041] The second step is data preprocessing

[0042] Preprocess the three training data sets, including observing the outlier data points in the data sets and removing the information that is suspected to be incorrectly entered by humans;

[0043] Step 3: Variable screening

[0044] In order to improve the stability and prediction accuracy of the model, the preprocessed data was screened for variables, 15 most important characteristic variables were selected, and the characteristic variables were tested for collinearity problems;

[0045] Step 4: Training the model

[0046] The organic sorghum yield was set as the response variable, and other characteristic variables affecting the organic sorghum yield were used as the explanatory variables of the model. The training data of three organic sorghums at different growth stages were used as the training data set to train the model.

[0047] Step 5: Model Diagnosis

[0048] Observe and adjust model variables and evaluate the feasibility of the model, and make adjustments repeatedly until the modeler's requirements and reference coefficients are met;

[0049] Step 6: Model application

[0050] The three trained prediction models for organic sorghum yield in the early planting stage, growth stage and maturity stage are saved in the background big data cluster respectively for the convenience of later users.

[0051] In the first step, the training data set consists of the following three parts:

[0052] Basic properties of planting, including land area, soil temperature, soil moisture, soil pH, and special chemical content information, are obtained through manual records and soil testing;

[0053] Natural environment data during the planting period, including recent temperature changes, light hours, and the impact of special natural disasters, are collected through local meteorological observation stations and base stations where the plots are located;

[0054] Field management data, records of the amount of fertilizer used on the land, the amount of seeds used, the number of pests and diseases, and the number of on-site support provided by agricultural technicians, were obtained through field collection.

[0055] In the second step, when a column of feature data in the training data set is vacant, if the vacant columns account for less than 5% of the total columns, the corresponding column average value is selected to fill in the vacant columns, otherwise the column feature data is deleted or the vacant columns are manually filled.

[0056] In the third step, the concept of variance inflation coefficient in mathematical statistics is used to measure the severity of collinearity of variables in regression modeling, and variables with variance inflation coefficients exceeding 10 are eliminated to obtain three sets of training data of organic sorghum at different growth stages as training data sets for organic sorghum yields at different periods.

[0057] In the fourth step, the Statsmodels professional regression modeling extension tool language package in Python language is used to write the organic sorghum yield prediction model, and the training data of three organic sorghums at different growth stages are used as training data sets to train the models separately, thereby obtaining three organic sorghum yield prediction models at different growth stages.

[0058] In the fifth step, the variables of the model are adjusted and evaluated, including the following steps:

[0059] S1. Use the residual diagnostic graph to observe the degree of fit between the constructed model and the data, observe the outliers and sample points with strong influence, and remove the outliers with strong influence during the model training process;

[0060] S2. Draw a GAM (generalized additive model) diagram in statistics for the important variable indicators in the model to evaluate whether the variable fitting power is reasonable and add alternative model solutions;

[0061] S3. Eliminate the characteristic variables that are higher (P value greater than 0.2) in the T test in the statistical diagnosis results of the model, because the T test shows that there is no significant linear relationship between the higher characteristic variables and the yield;

[0062] S4. Calculate the adjusted R-Square, AIC (Akaike information criterion) and mean square error of the candidate model in the test data set, and finally compare and select the model with the highest explanatory power and degree of fit for the original data.

[0063] The user infers the demand for organic sorghum in the next year based on the winemaking plan for the next year, and thus formulates a planting plan. To ensure that the demand for usage is met, in the sixth step, the basic properties of planting, the natural environment data during the planting period, and the field management data are continuously monitored during the growth stage of sorghum. The monitoring values ​​of the variables are entered in the page according to the characteristic variables in the model, and the growth stage of the organic sorghum under the current calculation plot is added; after reading the necessary parameters, the background task calls the organic sorghum yield prediction model of the corresponding growth stage to regularly predict the yield of organic sorghum;

[0064] If the predicted value is less than the user's demand or experience value, an early warning message will be issued to remind the user to take timely remedial measures for the abnormal situation.

[0065] In the sixth step, if the purchase volume is more than 15% higher than the predicted output during the grain harvesting stage, it is considered that the farmer may have borrowed or passed off inferior products as good ones, and the model will issue a system reminder.

[0066] Compared with the current existing technology, the method for establishing an organic sorghum yield prediction model has the following characteristics:

[0067] First, considering the growth stages of organic sorghum, three training data sets were constructed respectively, and models were built for the yield of organic sorghum in three different periods, so that the model can learn the impact of organic sorghum in different periods on various indicators in a targeted manner.

[0068] Second, a result comparison function has been added. That is, when the user predicts the future yield of organic sorghum through the constructed model, if the result is lower than the user's demand or experience value, an early warning prompt will be provided to the user, which can promptly remind the user to take rescue measures for the insufficient yield of the corresponding plot.

[0069] Third, through machine learning and statistical regression modeling, we provide customers with a set of rules for judging the yield of organic sorghum, that is, we determine the main factors affecting the yield of organic sorghum. After strict screening, these main factors have the characteristics of being observable, adjustable, and simulatable, providing a benchmark and reference for subsequent research on organic sorghum and other crops.

[0070] The embodiment described above is only one specific implementation of the present invention. Common changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for establishing an organic sorghum yield prediction model, characterized in that: The following steps are involved: The first step is to build a training dataset The growth cycle of organic sorghum is divided into three periods: the initial planting period, the growth period, and the maturity period. The basic attributes of planting, the natural environment data during the planting period, and the historical data of field management data are used as the initial explanatory variables of the model. The historical yield records of the planting plots are used as the response variables of the model. Three model training data sets are summarized according to the growth stages of organic sorghum. The training dataset consists of the following three parts: Basic properties of planting, including land area, soil temperature, soil moisture, soil pH, and special chemical content information, are obtained through manual records and soil testing; Natural environment data during the planting period, including recent temperature changes, light hours, and the impact of special natural disasters, are collected through local meteorological observation stations and base stations where the plots are located; Field management data, records of the amount of fertilizer used on the land, the amount of seeds used, the number of pests and diseases, and the number of on-site support from agricultural technicians, were obtained through field collection; The second step is data preprocessing Preprocess the three training data sets, including observing the outlier data points in the data sets and removing the information that is suspected to be incorrectly entered by humans; When there is a vacant feature data in a column in the training data set, if the vacant column accounts for less than 5% of the total columns, the corresponding column average value is selected to fill in the vacant column, otherwise the feature data of the column is deleted or the vacant column is filled manually; Step 3: Variable screening In order to improve the stability and prediction accuracy of the model, the preprocessed data was screened for variables, 15 most important characteristic variables were selected, and the characteristic variables were tested for collinearity problems; The concept of variance inflation coefficient in mathematical statistics was used to measure the severity of collinearity of variables in regression modeling. Variables with variance inflation coefficients exceeding 10 were eliminated, and three sets of training data of organic sorghum at different growth stages were obtained as training data sets for organic sorghum yields at different periods. Step 4: Training the model The organic sorghum yield was set as the response variable, and other characteristic variables affecting the organic sorghum yield were used as the explanatory variables of the model. The training data of three organic sorghums at different growth stages were used as the training data set to train the model. Step 5: Model Diagnosis Observe and adjust model variables and evaluate the feasibility of the model, and make adjustments repeatedly until the modeler's requirements and reference coefficients are met; The variable adjustment and evaluation of the model includes the following steps: S1. Use the residual diagnostic graph to observe the fit between the constructed model and the data, observe the outliers and sample points with strong influence, and remove the outliers with strong influence during the model training process; S2. Draw a GAM diagram in statistics for the important variable indicators in the model to evaluate whether the variable fitting power is reasonable and add alternative model solutions; S3. Eliminate the characteristic variables with T-test P values ​​greater than 0.2 in the statistical diagnosis results of the model, because the T-test shows that there is no significant linear relationship between higher characteristic variables and yield; S4. Calculate the corrected determination coefficient, AIC Akaike information criterion and the mean square error of the candidate model in the test data set, and finally compare and select the model with the highest explanatory power and fit degree of the original data; Step 6: Model application The three trained models of organic sorghum yield in the early planting period, growth period and maturity period are saved in the backend big data cluster for easy access by users later; During the growth stage of sorghum, the basic properties of planting, the natural environment data during the planting period and the field management data are continuously monitored, and the monitoring data is input into the characteristic variables of the model. According to the growth stage input by the user, the organic sorghum yield prediction model of the corresponding growth stage is called to make regular predictions on the yield of organic sorghum; If the predicted value is less than the user's demand or experience value, an early warning message will be issued to remind the user to take timely remedial measures for the abnormal situation.

2. The method for establishing an organic sorghum yield prediction model according to claim 1, characterized in that: In the fourth step, the Statsmodels professional regression modeling extension tool language package in Python language is used to write the organic sorghum yield prediction model, and the training data of three organic sorghums at different growth stages are used as training data sets to train the models separately, thereby obtaining three organic sorghum yield prediction models at different growth stages.

3. The method for establishing an organic sorghum yield prediction model according to claim 1, characterized in that: In the sixth step, if the purchase volume is more than 15% higher than the predicted output during the grain harvest stage, it is considered that the farmer may be borrowing or passing off inferior products as good ones, and the model will issue a system reminder.

Images