Method and system for predicting fatty acid value of canned rice suitable for milling based on BP neural network

Abstract

The invention relates to the technical field of rice storage, in particular to a method and system for predicting the fatty acid value of canned rice suitable for milling based on a BP neural network, and the method comprises the steps: obtaining a canned rice suitable for milling sample to be detected; determining quality parameters of the canned rice sample suitable for milling to be detected; the quality parameters are input into a preset prediction model, fatty acid values of the to-be-detected canned rice samples suitable for milling are output, the prediction model is obtained based on training of a training set, the training set comprises the quality parameters and the corresponding fatty acid values of a plurality of groups of canned rice samples suitable for milling, and the prediction model is constructed by adopting a BP neural network. The method can predict the fatty acid value of the rice suitable for milling in different storage periods, and is rapid in calculation and high in prediction precision.

Description

Technical Field

[0001] The invention relates to the technical field of rice storage, and in particular to a method and system for predicting fatty acid values ​​of canned rice suitable for milling based on a BP neural network. Background Art

[0002] Suitable rice is a kind of rice whose processing precision meets the national standard and whose husk retention is between 2% and 7%. Suitable rice has both the nutrition of brown rice and the taste of polished rice, and is a nutritionally balanced rice product. However, the bran layer and germ retained by suitable rice are rich in fat and enzymes, which are prone to oxidation and rancidity, seriously affecting its storage stability. Therefore, the selection of packaging materials and packaging methods for suitable rice is particularly important. Tinplate packaging has good sealing and opaque properties, and the addition of nitrogen and deoxidizers can improve its storage stability. At present, tinplate canned rice products are increasingly favored by consumers. Exploring the quality changes of canned suitable rice during storage can provide a reference for monitoring its deterioration degree.

[0003] The change in fatty acid value can reflect the degree of quality deterioration of rice products. It is the value of the free fatty acid content. The test result is expressed as the amount of potassium hydroxide required to neutralize the free fatty acids in 100g of sample. The existing fatty acid value determination method has the problems of complex operation, time-consuming and labor-intensive, and benzene is often used as a solvent in the extraction process, which poses potential risks such as explosion and poisoning, which increases the difficulty of fatty acid value determination. In order to solve the problem that fatty acid value determination is cumbersome and has potential safety hazards, the researchers considered relying on models to predict and analyze the changes in fatty acid values ​​of products at different storage stages.

[0004] Traditional reaction kinetic models are often used to predict the changing laws of product quality, but the scope of application of kinetic models is limited. They are not suitable for non-exponential reactions, reactions with complex reaction systems and reactions that do not involve temperature effects. The emergence of machine learning and big data technology in recent years has provided new ideas for grain storage safety. Models such as support vector regression have been widely used in food quality research, but when processing large-scale data sets, the training process is slow and the output results are complicated. BP neural network is a multi-layer feedforward network based on the error back propagation algorithm. It is widely used in the fields of food type and quality classification, element content detection, etc. It has the ability of self-learning and adaptation. It can analyze the input and output data sets, identify the connections and laws between data, construct complex nonlinear system functions, and realize accurate prediction of quality indicators. Therefore, it is feasible to use BP neural network to establish a prediction model to predict the fatty acid value of canned milled rice. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for predicting the fatty acid value of canned rice suitable for milling based on BP neural network, which can predict the fatty acid value of rice suitable for milling in different storage periods, and has rapid calculation and high prediction accuracy.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for predicting fatty acid value of canned rice suitable for milling based on BP neural network, comprising:

[0008] Obtain canned rice samples suitable for milling to be tested;

[0009] Determining the quality parameters of the canned rice sample suitable for milling;

[0010] The quality parameters are input into a preset prediction model, and the fatty acid value of the canned rice sample suitable for milling is output, wherein the prediction model is obtained based on a training set, the training set includes several groups of quality parameters of canned rice samples suitable for milling and corresponding fatty acid values, and the prediction model is constructed using a BP neural network.

[0011] Optionally, the quality parameters include: storage time, storage temperature, initial moisture content, initial fatty acid value. The prediction model includes an input layer, a hidden layer and an output layer, the input layer is used to input the quality parameters and transmit them to the hidden layer, the hidden layer calculates the quality parameters through a number of neurons, and transmits the calculation results to the output layer, and the final fatty acid value is output through the output layer.

[0012] Optionally, the prediction model is:

[0013] Y=2.0815+0.2013H 1 -0.0073H 2 +3.8545H 3 -0.8507H 4 ;

[0014] Among them, Y is the output fatty acid value, H 1 , H 2 , H 3 , H 4 They are initial moisture content, initial fatty acid value, storage time, and storage temperature.

[0015] To further achieve the above-mentioned purpose, the present invention also provides a canned rice suitable for milling fatty acid value prediction system based on BP neural network, comprising: a test sample acquisition module, a quality parameter determination module, and a fatty acid value prediction module, wherein the test sample acquisition module is used to obtain a canned rice suitable for milling sample to be tested; the quality parameter determination module is used to determine the quality parameters of the canned rice suitable for milling sample to be tested; the fatty acid value prediction module is used to input the quality parameters into a preset prediction model, and output the fatty acid value of the canned rice suitable for milling sample to be tested, wherein the prediction model is obtained based on training of a training set, and the training set contains several groups of quality parameters and corresponding fatty acid values ​​of canned rice suitable for milling samples, and the prediction model is constructed using a BP neural network.

[0016] Optionally, the quality parameter determination module determines the quality parameters of the canned rice milling sample to be tested, including determining the storage time, initial moisture content, initial fatty acid value and storage temperature of the canned rice milling sample to be tested.

[0017] Optionally, the prediction model in the fatty acid value prediction module includes an input layer, a hidden layer and an output layer, the input layer is used to input quality parameters and transmit them to the hidden layer, the hidden layer calculates the quality parameters through a number of neurons, and transmits the calculation results to the output layer, and the final fatty acid value is output through the output layer.

[0018] Optionally, the prediction model is:

[0019] Y=2.0815+0.2013H 1 -0.0073H 2 +3.8545H 3 -0.8507H 4 ;

[0020] Among them, Y is the output fatty acid value, H 1 , H 2 , H 3 , H 4 They are initial moisture content, initial fatty acid value, storage time, and storage temperature.

[0021] To further achieve the above object, the present invention further provides a processor, which is used to run a program, wherein the program executes the method for predicting the fatty acid value of canned rice suitable for milling based on the BP neural network when running.

[0022] To further achieve the above objectives, the present invention also provides an electronic device, comprising one or more memories and a processor, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for predicting the fatty acid value of canned rice suitable for milling based on the BP neural network.

[0023] The beneficial effects of the present invention are:

[0024] The present invention uses BP neural network to link the fatty acid value of suitable rice after storage with the quality parameters of suitable rice, and predicts the fatty acid value of suitable rice at different storage periods by measuring the initial moisture content, initial fatty acid value and storage temperature of suitable rice. The method has rapid calculation and high prediction accuracy. By predicting the fatty acid value of suitable rice at different storage periods, the potential shelf life risk can be evaluated in advance, providing strong theoretical support for the storage of suitable rice. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 A schematic diagram of forward signal propagation according to an embodiment of the present invention;

[0027] Figure 2 Schematic diagram of error reverse propagation in an embodiment of the present invention;

[0028] Figure 3 A schematic diagram of the structure of a neural network model according to an embodiment of the present invention;

[0029] Figure 4 A comparison chart of the training set prediction results and actual results during the training process of the model in an embodiment of the present invention;

[0030] Figure 5 This is a comparison chart of the test set prediction results and actual results during the training process of the model in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 are within the scope of protection of the present invention.

[0032] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] This embodiment provides a method for predicting the fatty acid value of canned rice suitable for milling based on a BP neural network, comprising:

[0034] Obtain canned rice samples suitable for milling to be tested;

[0035] Determining the quality parameters of the canned rice sample suitable for milling;

[0036] The quality parameters are input into a preset prediction model, and the fatty acid value of the canned rice sample suitable for milling is output, wherein the prediction model is obtained based on a training set, the training set includes several groups of quality parameters of canned rice samples suitable for milling and corresponding fatty acid values, and the prediction model is constructed using a BP neural network.

[0037] Specifically, this implementation uses BP neural network to link the fatty acid value of suitable rice with the quality parameters of suitable rice, and predicts the fatty acid value of suitable rice at different storage periods by measuring the initial moisture content, initial fatty acid value and storage temperature of suitable rice. The method has rapid calculation and high prediction accuracy. By predicting the fatty acid value of suitable rice at different storage periods, the potential shelf life risk can be evaluated in advance, providing strong theoretical support for rice storage.

[0038] Furthermore, the quality parameters include: storage time, storage temperature, initial moisture content, and initial fatty acid value.

[0039] Furthermore, the prediction model includes an input layer, a hidden layer and an output layer. The input layer is used to input quality parameters and transmit them to the hidden layer. The hidden layer calculates the quality parameters through a number of neurons and transmits the calculation results to the output layer, and the final fatty acid value is output through the output layer.

[0040] Furthermore, the prediction model is:

[0041] Y=2.0815+0.2013H 1 -0.0073H 2 +3.8545H 3 -0.8507H 4 ;

[0042] Among them, Y is the output fatty acid value, H 1 , H 2 , H 3 , H 4 They are initial moisture content, initial fatty acid value, storage time, and storage temperature.

[0043] Combine the following Figure 1-Figure 5 The prediction model constructed in this embodiment and the construction process are described in detail, which specifically includes the following steps and contents:

[0044] Step 1: Obtain quality indicators suitable for rice milling at different storage periods.

[0045] 1. Materials and methods;

[0046] 1.1. Materials;

[0047] GB / T1354-2018 stipulates that rice with a peel retention rate between 2.0% and 7.0% is suitable for milling. This embodiment selects canned Suijing suitable rice for milling as the research object. The packaging methods of suitable rice for milling in this embodiment are: tinplate cans filled with nitrogen, tinplate cans with deoxidizer, and tinplate cans filled with nitrogen and deoxidizer.

[0048] 1.2. Quality index determination method;

[0049] In this embodiment, the sample quality indicators include initial moisture content and initial fatty acid value.

[0050] 1.2.1. Determination of moisture content;

[0051] The moisture content is determined by the direct drying method in GB5009.3-2016.

[0052] 1.2.2. Determination of fatty acid value;

[0053] The fatty acid value was determined using the method in GB / T20569-2006.

[0054] Step 2: Determine the parameters of the input layer and the output layer.

[0055] There are four parameters input into the input layer, including storage time, initial moisture content, initial fatty acid value and storage temperature;

[0056] The output layer outputs only one parameter - the fatty acid value after storage.

[0057] Step 3: Import and divide.

[0058] The data were imported into MATLAB 2021b software, and the data were randomly divided into training set and test set at a ratio of 8:2. The imported appropriate rice milling parameters are shown in Table 1.

[0059] Table 1

[0060]

[0061]

[0062] Step 4: Determine the neural network structure.

[0063] The preset number of training times is 1000, the learning rate is 0.01, and the minimum error of the training target is 0.00001. The nodes of the hidden layer are usually selected according to the empirical formula as a reference, the formula is as follows: P is the number of input layer nodes, M is the number of output layer nodes, a∈[0,10]. Given the input layer nodes and output layer nodes of the BP neural network model, the number of neuron nodes in the hidden layer is determined to be between 3 and 15, and the optimal choice of the number of neurons is explored.

[0064] Step 5: The operation process of BP neural network.

[0065] The training of BP neural network consists of the following processes:

[0066] (1) Signal forward propagation: The process in which the input signal is transmitted from the input layer to the hidden layer and then to the output layer. The 12 neurons in the input layer directly output to the neurons in the hidden layer without performing any calculations. The input layer neurons are represented by i, where i = 1, 2, ..., 12, and the hidden layer neurons are represented by j, where j = 1, 2, ..., 12. ij is the synaptic connection weight from neuron i to j.

[0067] The hidden layer input is:

[0068] The threshold of the hidden layer is b j , the activation function f of the hidden layer 1 is tansig, then the hidden layer output is:

[0069] The output layer input is:

[0070] The threshold of the output layer is c, and the activation function of the output layer is f 2 is purelin, then the output layer output is:

[0071]

[0072] At this point, the input mode has completed a forward propagation, and the signal forward propagation diagram is as follows Figure 1 shown.

[0073] (2) Backward propagation of error: Calculate the error between the predicted value and the actual value. The error starts from the output layer and propagates backwards. Then calculate the error between the output layer and the hidden layer. Then calculate the error between the hidden layer and the input layer. Finally, the weights are continuously updated according to the error. This process is repeated, and the connection weights and thresholds are continuously updated until the output value of the network is as close to the expected value as possible.

[0074] Network expected output value: d;

[0075] Output layer error: e = dY;

[0076] The error from the output layer to the hidden layer:

[0077] The error from hidden layer to input layer:

[0078] Update the weights between the input layer and the hidden layer:

[0079] Update the weights between the hidden layer and the output layer:

[0080] The error back propagation diagram is as follows Figure 2 shown.

[0081] (3) Operation process: Repeatedly calculate the output error of each node in the hidden layer and the output layer, and continuously update the weights and thresholds of each layer. Repeat (1) and (2) until the network error is less than or equal to the predetermined error, and obtain the final result. Figure 4 This is a comparison chart of the training set prediction results and actual results during the training process, and the goodness of fit R between the training set prediction value and the actual value 2 is 0.97; Figure 5 This is a comparison chart of the test set prediction results and actual results during the training process, and the goodness of fit R between the test set prediction value and the actual value 2 It is 0.98.

[0082] Step 6: Determine the prediction model for fatty acid value of rice suitable for milling.

[0083] like Figure 3 As shown, the obtained fatty acid value prediction model is:

[0084] Y=-0.6629+0.5541W 1 -0.0166W 2 +0.8745W 3 ;

[0085] W 1 =3.2297+0.1760×initial moisture content-0.0603×initial fatty acid value+2.0636×storage time-0.8753×storage temperature;

[0086] W 2 =0.8254+2.3027×initial moisture content+0.2086×initial fatty acid value-0.3824×storage time-2.0095×storage temperature;

[0087] W 3=0.3494+0.1624×initial moisture content+0.0338×initial fatty acid value+3.0930×storage time-0.4564×storage temperature;

[0088] Where: Y is the fatty acid value, unit is mg / 100g, W 1 , W 2 , W 3 They are Figure 3 The neurons in .

[0089] The final fatty acid value prediction model is:

[0090] Y=2.0815+0.2013H 1 -0.0073H 2 +3.8545H 3 -0.8507H 4 ;

[0091] Among them, Y is the output fatty acid value, H 1 , H 2 , H 3 , H 4 They are initial moisture content, initial fatty acid value, storage time, and storage temperature.

[0092] In order to further optimize the above technical solution, the present embodiment further provides a canned rice suitable for milling fatty acid value prediction system based on BP neural network, comprising: a test sample acquisition module, a quality parameter determination module, and a fatty acid value prediction module, wherein the test sample acquisition module is used to obtain a canned rice suitable for milling sample to be tested; the quality parameter determination module is used to determine the quality parameters of the canned rice suitable for milling sample to be tested; the fatty acid value prediction module is used to input the quality parameters into a preset prediction model, and output the fatty acid value of the canned rice suitable for milling sample to be tested, wherein the prediction model is obtained based on training of a training set, and the training set contains several groups of quality parameters and corresponding fatty acid values ​​of canned rice suitable for milling samples, and the prediction model is constructed using a BP neural network.

[0093] Furthermore, the quality parameter determination module determines the quality parameters of the canned rice milling sample to be tested, including determining the storage time, initial moisture content, initial fatty acid value and storage temperature of the canned rice milling sample to be tested.

[0094] Furthermore, the prediction model in the fatty acid value prediction module includes an input layer, a hidden layer and an output layer. The input layer is used to input quality parameters and transmit them to the hidden layer. The hidden layer calculates the quality parameters through neurons and transmits the calculation results to the output layer, and the final fatty acid value is output through the output layer.

[0095] Furthermore, the prediction model is:

[0096] Y=-0.6629+0.5541W 1 -0.0166W 2 +0.8745W 3 ;

[0097] Where: Y is the fatty acid value, unit is mg / 100g, W 1 , W 2 , W 3 They are Figure 3 The neurons in .

[0098] W 1 =3.2297+0.1760×initial moisture content-0.0603×initial fatty acid value+2.0636×storage time-0.8753storage temperature;

[0099] W 2 =0.8254+2.3027×initial moisture content+0.2086×initial fatty acid value-0.3824×storage time-2.0095storage temperature;

[0100] W 3 =0.3494+0.1624×initial moisture content+0.0338×initial fatty acid value+3.0930×storage time-0.4564×storage temperature;

[0101] The final fatty acid value prediction model is:

[0102] Y=2.0815+0.2013H 1 -0.0073H 2 +3.8545H 3 -0.8507H 4 ;

[0103] Among them, Y is the output fatty acid value, H 1 , H 2 , H 3 , H 4 They are initial moisture content, initial fatty acid value, storage time, and storage temperature.

[0104] And a processor, the processor is used to run a program, wherein when the program is run, the method for predicting the fatty acid value of canned rice suitable for milling based on the BP neural network is executed.

[0105] And an electronic device, comprising one or more memories and a processor, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for predicting the fatty acid value of canned rice suitable for milling based on the BP neural network.

[0106] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for predicting fatty acid value of canned rice suitable for milling based on BP neural network, characterized in that: include: Obtain a canned rice sample suitable for milling to be tested; Determining the quality parameters of the canned rice sample suitable for milling; The quality parameters are input into a preset prediction model, and the fatty acid value of the canned rice sample suitable for milling is output, wherein the prediction model is obtained based on a training set, the training set includes several groups of quality parameters of canned rice samples suitable for milling and corresponding fatty acid values, and the prediction model is constructed using a BP neural network.

2. The method for predicting fatty acid value of canned rice suitable for milling based on BP neural network according to claim 1, characterized in that: The quality parameters include: storage time, storage temperature, initial moisture content, and initial fatty acid value.

3. The method for predicting fatty acid value of canned rice suitable for milling based on BP neural network according to claim 1, characterized in that: The prediction model includes an input layer, a hidden layer and an output layer. The input layer is used to input quality parameters and transmit them to the hidden layer. The hidden layer calculates the quality parameters through a number of neurons and transmits the calculation results to the output layer, and the final fatty acid value is output through the output layer.

4. The method for predicting fatty acid value of canned rice suitable for milling based on BP neural network according to claim 3, characterized in that: The prediction model is: Y=2.0815+0.2013H1-0.0073H2+3.8545H3-0.8507H4; Among them, Y is the output fatty acid value, H1, H2, H3, and H4 are the initial moisture content, initial fatty acid value, storage time, and storage temperature, respectively.

5. A system for predicting fatty acid value of canned rice suitable for milling based on BP neural network, characterized in that: include: A test sample acquisition module, a quality parameter determination module, and a fatty acid value prediction module, wherein the test sample acquisition module is used to obtain a canned rice sample suitable for milling to be tested; the quality parameter determination module is used to determine the quality parameters of the canned rice sample suitable for milling to be tested; the fatty acid value prediction module is used to input the quality parameters into a preset prediction model and output the fatty acid value of the canned rice sample suitable for milling to be tested, wherein the prediction model is obtained based on a training set, the training set contains several groups of quality parameters of canned rice samples suitable for milling and the corresponding fatty acid values, and the prediction model is constructed using a BP neural network.

6. The fatty acid value prediction system of canned rice suitable for milling based on BP neural network according to claim 5 is characterized in that: The quality parameter determination module determines the quality parameters of the canned rice milling sample to be tested, including determining the storage time, initial moisture content, initial fatty acid value and storage temperature of the canned rice milling sample to be tested.

7. The fatty acid value prediction system of canned rice suitable for milling based on BP neural network according to claim 5, characterized in that: The prediction model in the fatty acid value prediction module includes an input layer, a hidden layer and an output layer. The input layer is used to input quality parameters and transmit them to the hidden layer. The hidden layer calculates the quality parameters through a number of neurons and transmits the calculation results to the output layer, and the final fatty acid value is output through the output layer.

8. The fatty acid value prediction system of canned rice suitable for milling based on BP neural network according to claim 7, characterized in that: The prediction model is: Y=2.0815+0.2013H1-0.0073H2+3.8545H3-0.8507H4; Among them, Y is the output fatty acid value, H1, H2, H3, and H4 are the initial moisture content, initial fatty acid value, storage time, and storage temperature, respectively.

9. A processor, characterized in that: The processor is used to run a program, wherein the program executes the method according to any one of claims 1 to 4 when running.

10. An electronic device, characterized in that: The invention comprises one or more memories and processors, wherein the memories are used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 4.

Images