CN110458361A - Grain quality index prediction technique based on BP neural network - Google Patents
Grain quality index prediction technique based on BP neural network Download PDFInfo
- Publication number
- CN110458361A CN110458361A CN201910750723.6A CN201910750723A CN110458361A CN 110458361 A CN110458361 A CN 110458361A CN 201910750723 A CN201910750723 A CN 201910750723A CN 110458361 A CN110458361 A CN 110458361A
- Authority
- CN
- China
- Prior art keywords
- grain
- neural network
- prediction
- quality
- phase
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000012549 training Methods 0.000 claims description 25
- 238000003062 neural network model Methods 0.000 claims description 19
- 239000011159 matrix material Substances 0.000 claims description 15
- 210000002569 neuron Anatomy 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 3
- 210000005036 nerve Anatomy 0.000 claims 1
- 239000000463 material Substances 0.000 abstract description 5
- 235000013339 cereals Nutrition 0.000 description 114
- 239000010410 layer Substances 0.000 description 80
- 240000008042 Zea mays Species 0.000 description 20
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 20
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 20
- 235000005822 corn Nutrition 0.000 description 20
- 241000209094 Oryza Species 0.000 description 9
- 235000007164 Oryza sativa Nutrition 0.000 description 9
- 235000014113 dietary fatty acids Nutrition 0.000 description 9
- 229930195729 fatty acid Natural products 0.000 description 9
- 239000000194 fatty acid Substances 0.000 description 9
- 235000009566 rice Nutrition 0.000 description 9
- 150000004665 fatty acids Chemical class 0.000 description 8
- 108010068370 Glutens Proteins 0.000 description 7
- 238000010521 absorption reaction Methods 0.000 description 7
- 235000021312 gluten Nutrition 0.000 description 7
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 7
- 240000008467 Oryza sativa Japonica Group Species 0.000 description 5
- 235000013305 food Nutrition 0.000 description 5
- 241000209140 Triticum Species 0.000 description 4
- 235000021307 Triticum Nutrition 0.000 description 4
- 239000012535 impurity Substances 0.000 description 4
- 239000002344 surface layer Substances 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 1
- 125000005313 fatty acid group Chemical group 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of grain quality index prediction technique based on BP neural network, comprising: obtains grain prediction phase information t, wherein the grain prediction phase information t indicates to need to predict the index of quality of t phase grain；According to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；BP neural network prediction model is constructed according to the grain quality achievement data got；The index of quality of t phase grain is predicted according to the BP neural network prediction model.The present invention by store the existing related data of grain establish prediction model, can the grain quality index to following a period of time predict, it is high-efficient, save human and material resources.
Description
Technical field
The present invention relates to food storage fields, and in particular to a kind of grain quality index prediction side based on BP neural network
Method.
Background technique
It is well known that ensureing that China's grain reserves are the material items for being related to national economy safely, and how to predict to store up
The grain quality variation deposited is significant to the quality of assessment food storage state, in the prior art, refers to about grain quality
Target research is largely the method based on experiment extraction, but due to the grain initial level of each ecotope, each warehouse
It is different, if the grain quality index of each ecotope, each warehouse is obtained with the method for experiment extraction, it will be greatly reduced effect
Rate, wasting manpower and material resources.
Summary of the invention
In order to overcome the deficiencies of the prior art, the present invention provides a kind of grain quality index prediction side based on BP neural network
Method establishes prediction model by storing the existing related data of grain, can grain quality index to following a period of time into
Row prediction, it is high-efficient, it saves human and material resources.In order to solve the above technical problems, the present invention is realized by following technological means:
Grain quality index prediction technique based on BP neural network, comprising:
Obtain grain prediction phase information t, wherein the grain prediction phase information t indicates to need to predict the product of t phase grain
Matter index；
According to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
BP neural network prediction model is constructed according to the grain quality achievement data got；
The index of quality of t phase grain is predicted according to the BP neural network prediction model.
Further, described according to grain prediction phase information t, obtain the grain quality achievement data packet for meeting time requirement
It includes: obtaining the grain quality achievement data X of t-1 phase from historical data base respectivelyt—1, the t-2 phase grain quality index number
According to Xt-—2With the grain quality achievement data X of t-3 phaset—3。
Further, include: according to the grain quality achievement data building BP neural network prediction model got
Respectively from grain quality achievement data Xt—1, grain quality achievement data Xt—2With grain quality achievement data Xt—3In
Extract a part of data composing training collection XP, XP=[XP t—1, XP t—2, XP t—3]；
To the training set XPX is obtained as normalizedR；
According to XRConstruct BP neural network model.
Further, to the training set XPX is obtained as normalizedRIt include: to enable XR ij=(XP ij- min (XP ij))/
(max(XP ij)—min(XP ij)), wherein XP ijIndicate column variable XP jIn a data, min (XP ij) indicate column variable XP jIn
Minimum value, max (XP ij) indicate column variable XP jIn maximum value, max (XP ij)—min(XP ij) indicate column variable XP jPole
Difference, j=t-1, t-2, t-3.
Further, according to XRConstructing BP neural network model includes:
BP neural network framework is built, is hidden in the BP neural network framework including input layer, the first hidden layer, second
Layer, Dropout layers, Dense layers and output layer；
By data XRIt is input to input layer；
The output f of the first hidden layer is obtained by formula (1)1, f1Input as the first hidden layer；
The output f of the second hidden layer is obtained by formula (2)2, f2As Dense layers of input；
The output f of output layer is obtained by formula (3)3, f3The BP neural network model completed for building；
f1=ReLU ((Wf1*XR)+b1) (1)
f2=ReLU ((Wf2*f1)+b2) (2)
f3=(Wf3*f2)+b3 (3)
Wherein, Wf1Indicate the weight matrix of the first hidden layer, Wf2Indicate the weight matrix of the second hidden layer, Wf3It indicates
Dense layers of weight matrix, b1Indicate the bias term of the first hidden layer, b2Indicate the bias term of the second hidden layer, b3It indicates
Dense layers of bias term, ReLU indicate activation primitive, and Dropout layers for temporary in undated parameter during model training
The random neuron for disconnecting specified quantity.
Further, the method also includes:
Model verifying is carried out by formula (4)；
Wherein, RE indicates relative error, OtFor t phase grain quality index measured data, PtFor the BP completed by building
The t phase grain quality achievement data of Neural Network model predictive, n are that model verifies sample number.
As another preferred embodiment, it is described according to grain prediction phase information t, obtain the grain quality index for meeting time requirement
Data include: the grain quality achievement data X that the t-4 phase is obtained from historical data baset—4。
Further, include: according to the grain quality achievement data building BP neural network prediction model got
From grain quality achievement data Xt—4It is middle to extract a part of data composing training collection XP t—4；
To the training set XP t—4X is obtained as normalizedR t—4；
According to XR t—4Construct BP neural network model.
Grain quality index prediction technique based on BP neural network, comprising:
Obtain grain prediction phase information t, wherein the grain prediction phase information t indicates to need to predict the matter of t phase grain
Figureofmerit；
According to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
BP neural network prediction model is constructed according to the grain quality achievement data got；
According to the index of quality for the BP neural network prediction model prediction t phase grain that building is completed.
The present invention establishes prediction model by storing the existing related data of grain, can be to the grain of following a period of time
The index of quality is predicted, high-efficient, is saved human and material resources.
Detailed description of the invention
Fig. 1 is a kind of grain quality index prediction technique based on BP neural network shown according to an exemplary embodiment
Flow chart.
Specific embodiment
It is with reference to the accompanying drawing and specific real in order to make those skilled in the art more fully understand technical solution of the present invention
Applying example, the present invention is described in further detail.
Embodiment 1
As shown in Figure 1, the present embodiment provides a kind of grain quality index prediction technique based on BP neural network, comprising:
S1: grain prediction phase information t is obtained, wherein the grain prediction phase information t expression needs to predict t phase grain
The index of quality；
S2: according to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
S3: BP neural network prediction model is constructed according to the grain quality achievement data got；
S4: the index of quality of t phase grain is predicted according to the BP neural network prediction model.
In the present embodiment, grain can be corn, wheat, japonica rice, long-grained nonglutinous rice etc., and grain index includes the index of quality and quality
Index (the present embodiment is that the index of quality is predicted according to the index of quality, and embodiment 2 is to predict the index of quality according to quality index),
In, the index of quality of corn is fatty acid value, and the index of quality of wheat is gluten water absorption, and the index of quality of japonica rice is fatty acid
Value, the index of quality of long-grained nonglutinous rice are fatty acid value, and the quality index of grain includes other in addition to the index of quality such as moisture, impurity
Index when modeling according to quality index, can according to need certain indexs in selection quality index.
As an example, current embodiment require that the grain of prediction be that (while the corn of one library point of entrance is same for certain batch corn
A collection of corn), it can be defined as the 1st phase when this batch of corn enters warehouse, be the 2nd phase after the fixed period, it is incremented by successively, the 1st
The corn quality achievement data of phase is X1, the corn quality achievement data of the 2nd phase is X2, and so on, wherein when above-mentioned fixed
Section can be 1 month, 3 months, half a year etc., while needing to record the index of quality and quality index of each issue of corn, pre- when needing
When surveying the index of quality data of this batch of corn t phase, need to obtain in this batch of corn from the existing historical data of this batch of corn
Meet the corn quality achievement data of time requirement to construct BP neural network prediction model.Wherein, each issue of every batch of grain can be with
Multiple index of quality and multiple quality index are recorded, it, can be with when recording certain phase corn achievement data for example, for a collection of corn
The achievement data on surface layer, middle layer and bottom is recorded respectively, and the index of quality on surface layer, middle layer and bottom constitutes the phase corn
One group of index of quality data, the quality index on surface layer, middle layer and bottom constitute one group of quality index data of the phase corn.
When it is implemented, we are within 5%, from actual prediction prediction result to gluten water absorption prediction anticipation error
It sees, mean error is 2.29% or so, far below it is desirable that error；To corn fatty acid value anticipation error be no more than
10%, actual error is 9.1% or so；10% is no more than to paddy fatty acid value anticipation error, long-grained nonglutinous rice actual error is
8.7%, japonica rice actual error is 9.3%.
Preferably, it is described according to grain prediction phase information t, obtain the grain quality achievement data packet for meeting time requirement
It includes: obtaining the grain quality achievement data X of t-1 phase from historical data base respectivelyt—1, the t-2 phase grain quality index number
According to Xt-—2With the grain quality achievement data X of t-3 phaset—3。
Here it is possible to first from historical data base by library point be that data are extracted into Excel table by unit, then from Excel
The achievement data (including the index of quality and quality index) of every a collection of grain is filtered out in table and is sequentially arranged.Due to
Need to meet the data of certain time requirement when constructing BP neural network prediction model, thus can to the data filtered out according to
It is few abandon the principles deleted more and pre-processed, for example, certain a collection of food storage is just sold or changes storehouse in 1 year, with six months for 1
Phase, less than 4 phases, at this time needs to give up this batch of grain then just only saving the achievement data of two phase of this batch of grain in database
The data of food 3 years of certain a collection of food storage, at this moment just save 6 issue evidence of this batch of grain for another example in database, due to
So more issue evidences are not needed when modeling, it can be according to the actual situation by extra rejection of data.Why the present embodiment uses 4
Phase is because finding after data compare and construct model test repeatedly using 4 phases as critical point prediction as critical point
Effect is best.
Preferably, including: according to the grain quality achievement data building BP neural network prediction model got
Respectively from grain quality achievement data Xt—1, grain quality achievement data Xt—2With grain quality achievement data Xt—3In
Extract a part of data composing training collection XP, XP=[XP t—1, XP t—2, XP t—3]；
To the training set XPX is obtained as normalizedR；
According to XRConstruct BP neural network model.
In the present embodiment, after normalized, XRIn each achievement data value between 0-1.
Preferably, enabling XR ij=(XP ij- min (XP ij))/(max(XP ij)—min(XP ij)), wherein XP ijIndicate that column become
Measure XP jIn a data, min (XP ij) indicate column variable XP jIn minimum value, max (XP ij) indicate column variable XP jIn maximum
Value, max (XP ij)—min(XP ij) indicate column variable XP jVery poor, j=t-1, t-2, t-3.Here, one group of grain of t-1 phase
Achievement data may be constructed Xt—1, one group of grain achievement data of t-2 phase may be constructed Xt—2, one group of grain of t-3 phase refers to
Mark data may be constructed Xt—3。
Preferably, according to XRConstructing BP neural network model includes:
BP neural network framework is built, is hidden in the BP neural network framework including input layer, the first hidden layer, second
Layer, Dropout layers, Dense layers and output layer；
By data XRIt is input to input layer；
The output f of the first hidden layer is obtained by formula (1)1, f1Input as the first hidden layer；
The output f of the second hidden layer is obtained by formula (2)2, f2As Dense layers of input；
The output f of output layer is obtained by formula (3)3, f3The BP neural network model completed for building；
f1=ReLU ((Wf1*XR)+b1) (1)
f2=ReLU ((Wf2*f1)+b2) (2)
f3=(Wf3*f2)+b3 (3)
Wherein, Wf1Indicate the weight matrix of the first hidden layer, Wf2Indicate the weight matrix of the second hidden layer, Wf3It indicates
Dense layers of weight matrix, b1Indicate the bias term of the first hidden layer, b2Indicate the bias term of the second hidden layer, b3It indicates
Dense layers of bias term, ReLU indicate activation primitive, and Dropout layers for temporary in undated parameter during model training
The random neuron for disconnecting specified quantity, f3For the index of quality of prediction.
Here, why use, Dropout layers refer to for temporarily disconnecting at random during model training in undated parameter
The neuron of fixed number amount is to train the model over-fitting come in order to prevent.
Preferably, the present embodiment further include:
Model verifying is carried out by formula (4)；
Wherein, RE indicates relative error, OtFor t phase grain quality index measured data, PtFor the BP completed by building
The t phase grain quality achievement data of Neural Network model predictive, n are that model verifies sample number.
Embodiment 2
Slightly different with above embodiment, and above-described embodiment needs the grain achievement data of continuous 3 phase, as another kind
It is preferred that the present embodiment only needs the grain achievement data of 1 phase, the grain after predicting for 4 phases by the grain achievement data of this 1 phase refers to
Mark, for example the grain achievement data of the 1st phase is known to predict the grain achievement data of the 5th phase, specifically, the present embodiment includes:
S1: grain prediction phase information t is obtained, wherein the grain prediction phase information t expression needs to predict t phase grain
The index of quality；
S2: according to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
S3: BP neural network prediction model is constructed according to the grain quality achievement data got；
S4: the index of quality of t phase grain is predicted according to the BP neural network prediction model.
It is described according to grain prediction phase information t, it includes: from history that acquisition, which meets the grain quality achievement data of time requirement,
The grain quality achievement data X of t-4 phase is obtained in databaset—4。
As preferred: including: according to the grain quality achievement data building BP neural network prediction model got
From grain quality achievement data Xt—4It is middle to extract a part of data composing training collection XP t—4；
To the training set XP t—4X is obtained as normalizedR t—4；
According to XR t—4Construct BP neural network model.
Here, to the training set XP t—4X is obtained as normalizedR t—4Are as follows: XR i=(XP i- min (XP i))/(max
(XP i)—min(XP i)), wherein XP iIndicate column variable XP t—4In a data, min (XP i) indicate column variable XP t-4In most
Small value, max (XP i) indicate column variable XP t-4In maximum value, max (XP i)—min(XP i) indicate column variable XP t-4It is very poor.
Preferably, according to XR t—4Constructing BP neural network model includes:
BP neural network framework is built, is hidden in the BP neural network framework including input layer, the first hidden layer, second
Layer, Dropout layers, Dense layers and output layer；
By data XR t—4It is input to input layer；
The output y of the first hidden layer is obtained by formula (5)1, y1Input as the first hidden layer；
The output y of the second hidden layer is obtained by formula (6)2, y2As Dense layers of input；
The output y of output layer is obtained by formula (7)3, y3The BP neural network model completed for building；
y1=ReLU ((Wy1*XR t—4)+c1) (5)
y2=ReLU ((Wy2*y1)+c2) (6)
y3=(Wy3*y2)+c3 (7)
Wherein, Wy1Indicate the weight matrix of the first hidden layer, Wy2Indicate the weight matrix of the second hidden layer, Wy3It indicates
Dense layers of weight matrix, c1Indicate the bias term of the first hidden layer, c2Indicate the bias term of the second hidden layer, c3It indicates
Dense layers of bias term, ReLU indicate activation primitive, and Dropout layers for temporary in undated parameter during model training
The random neuron for disconnecting specified quantity.
Preferably, the present embodiment further include:
Model verifying is carried out by formula (4).
When it is implemented, we are within 5%, from actual prediction prediction result to gluten water absorption prediction anticipation error
It sees, mean error is 1.02% or so, far below it is desirable that error；To corn fatty acid value anticipation error be no more than
10%, actual error is 3.98% or so；10% is no more than to paddy fatty acid value anticipation error, long-grained nonglutinous rice actual error is
6.27%, japonica rice actual error is 5.29%.
Embodiment 3
Grain quality index prediction technique the present embodiment provides another kind based on BP neural network, comprising:
A1: grain prediction phase information t is obtained, wherein the grain prediction phase information t expression needs to predict t phase grain
Quality index；
A2: according to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
A3: BP neural network prediction model is constructed according to the grain quality achievement data got；
A4: according to the quality index for the BP neural network prediction model prediction t phase grain that building is completed；
A5: the index of quality of t phase grain is assessed according to the quality index of the t phase grain predicted.
What needs to be explained here is why the present embodiment can predict the index of quality by quality index, it is because of matter
Figureofmerit or the index of quality are all certain features of grain itself, artificially according to people need to be divided into quality index and
The index of quality, be between various features in fact it is associated, in biological field, this connection can use experience or routine side
Method obtains, for example the height of quality index moisture of wheat will affect the height of index of quality gluten water absorption, by predicting
Quality index moisture can predict the height of index of quality gluten water absorption, so the present embodiment can first pass through quality and refer to
Mark carrys out forecast quality index, then assesses the index of quality according to the quality index of prediction.
In the present embodiment, the quality index of grain includes the moisture of grain, impurity, perfect kernel rate, head rice rate and yellow grain
Rice rate, wherein the grain quality achievement data that acquisition meets time requirement is the grain quality achievement data for obtaining the t-1 phase
Zt-1, Zt-1=(x1t-1, x2t-1, x3t-1, x4t-1, x5t-1), x1t-1, x2t-1, x3t-1, x4t-1, x5t-1Refer respectively to the t-1 phase
Moisture, impurity, perfect kernel rate, head rice rate and the yellow rice kernel rate of grain.What needs to be explained here is that different grain kinds is being predicted
The quality index used when the index of quality is different, for example, can be commented by this quality index of moisture for wheat
Estimate this index of quality of gluten water absorption, then can enable Zt-1=x1t-1, similarly, moisture and impurity are used if necessary, then may be used
To enable Zt-1=(x1t-1, x2t-1)。
As preferred: including: according to the grain quality achievement data building BP neural network prediction model got
From grain quality achievement data Zt—1It is middle to extract a part of data composing training collection ZP t—1；Here, refer to respectively from
x1t-1, x2t-1, x3t-1, x4t-1, x5t-1It is middle to extract a part of data composing training collection ZP t—1；
To the training set ZP t—1Z is obtained as normalizedR t—1；
According to ZR t—1Construct BP neural network model.
Here, to the training set ZP t—1Z is obtained as normalizedR t—1Are as follows: ZR i=(ZP i- min (ZP i))/(max
(ZP i)—min(ZP i)), wherein ZP iIndicate column variable ZP t—4In a data, min (ZP i) indicate column variable ZP t-4In most
Small value, max (ZP i) indicate column variable ZP t-4In maximum value, max (ZP i)—min(ZP i) indicate column variable ZP t-4It is very poor.
Preferably, according to ZR t—1Constructing BP neural network model includes:
BP neural network framework is built, is hidden in the BP neural network framework including input layer, the first hidden layer, second
Layer, Dropout layers, Dense layers and output layer；
By data ZR t—1It is input to input layer；
The output v of the first hidden layer is obtained by formula (8)1, v1Input as the first hidden layer；
The output v of the second hidden layer is obtained by formula (9)2, v2As Dense layers of input；
The output v of output layer is obtained by formula (10)3, v3,The BP neural network model completed for building；
v1=ReLU ((Wv1*ZR t—4)+d1) (1)
v2=ReLU ((Wv2*v1)+d2) (2)
v3=(Wv3*v2)+d3 (3)
Wherein, Wv1Indicate the weight matrix of the first hidden layer, Wv2Indicate the weight matrix of the second hidden layer, Wv3It indicates
Dense layers of weight matrix, d1Indicate the bias term of the first hidden layer, d2Indicate the bias term of the second hidden layer, d3It indicates
Dense layers of bias term, ReLU indicate activation primitive, and Dropout layers for temporary in undated parameter during model training
The random neuron for disconnecting specified quantity.
Preferably, the present embodiment further includes carrying out model verifying by formula (4).
When it is implemented, we are within 5%, from actual prediction prediction result to gluten water absorption prediction anticipation error
It sees, mean error is 3.6% or so, far below it is desirable that error；To corn fatty acid value anticipation error be no more than 10%,
Actual error is 8.5% or so；10% is no more than to paddy fatty acid value anticipation error, long-grained nonglutinous rice actual error is 8.7%, japonica rice
Actual error is 6.4%.
The above is only the preferred embodiment of the present invention, it is noted that above-mentioned preferred embodiment is not construed as pair
Limitation of the invention, protection scope of the present invention should be defined by the scope defined by the claims..For the art
For those of ordinary skill, without departing from the spirit and scope of the present invention, several improvements and modifications can also be made, these change
It also should be regarded as protection scope of the present invention into retouching.
Claims (9)
1. the grain quality index prediction technique based on BP neural network characterized by comprising
Obtain grain prediction phase information t, wherein the grain prediction phase information t expression needs to predict that the quality of t phase grain refers to
Mark；
According to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
BP neural network prediction model is constructed according to the grain quality achievement data got；
The index of quality of t phase grain is predicted according to the BP neural network prediction model.
2. the grain quality index prediction technique according to claim 1 based on BP neural network, which is characterized in that described
According to grain prediction phase information t, it includes: respectively from historical data base that acquisition, which meets the grain quality achievement data of time requirement,
Obtain the grain quality achievement data X of t-1 phaset—1, the t-2 phase grain quality achievement data Xt-—2With the grain of t-3 phase
Index of quality data Xt—3。
3. the grain quality index prediction technique according to claim 2 based on BP neural network, which is characterized in that according to
The grain quality achievement data got constructs BP neural network prediction model
Respectively from grain quality achievement data Xt—1, grain quality achievement data Xt—2With grain quality achievement data Xt—3Middle extraction
A part of data composing training collection XP, XP=[XP t—1, XP t—2, XP t—3]；
To the training set XPX is obtained as normalizedR；
According to XRConstruct BP neural network model.
4. the grain quality index prediction technique according to claim 3 based on BP neural network, which is characterized in that institute
State training set XPX is obtained as normalizedRIt include: to enable XR ij=(XP ij- min (XP ij))/(max(XP ij)—min(XP ij)),
Wherein, XP ijIndicate column variable XP jIn a data, min (XP ij) indicate column variable XP jIn minimum value, max (XP ij) indicate
Column variable XP jIn maximum value, max (XP ij)—min(XP ij) indicate column variable XP jVery poor, j=t-1, t-2, t-3.
5. the grain quality index prediction technique according to claim 3 based on BP neural network, which is characterized in that according to
XRConstructing BP neural network model includes:
Build BP neural network framework, include in the BP neural network framework input layer, the first hidden layer, the second hidden layer,
Dropout layers, Dense layers and output layer；
By data XRIt is input to input layer；
The output f of the first hidden layer is obtained by formula (1)1, f1Input as the first hidden layer；
The output f of the second hidden layer is obtained by formula (2)2, f2As Dense layers of input；
The output f of output layer is obtained by formula (3)3, f3The BP neural network model completed for building；
f1=ReLU ((Wf1*XR)+b1) (1)
f2=ReLU ((Wf2*f1)+b2) (2)
f3=(Wf3*f2)+b3 (3)
Wherein, Wf1Indicate the weight matrix of the first hidden layer, Wf2Indicate the weight matrix of the second hidden layer, Wf3Indicate Dense layers
Weight matrix, b1Indicate the bias term of the first hidden layer, b2Indicate the bias term of the second hidden layer, b3Indicate Dense layers inclined
Item is set, ReLU indicates activation primitive, and Dropout layers specified for temporarily disconnecting at random during model training in undated parameter
The neuron of quantity.
6. the grain quality index prediction technique according to claim 5 based on BP neural network, which is characterized in that also wrap
It includes:
Model verifying is carried out by formula (4)；
Wherein, RE indicates relative error, OtFor t phase grain quality index measured data, PtFor the BP nerve completed by building
The t phase grain quality achievement data of network model prediction, n are that model verifies sample number.
7. the grain quality index prediction technique according to claim 1 based on BP neural network, it is characterised in that: described
According to grain prediction phase information t, it includes: to obtain from historical data base that acquisition, which meets the grain quality achievement data of time requirement,
The grain quality achievement data X of t-4 phaset—4。
8. the grain quality index prediction technique according to claim 7 based on BP neural network, it is characterised in that: according to
The grain quality achievement data got constructs BP neural network prediction model
From grain quality achievement data Xt—4It is middle to extract a part of data composing training collection XP t—4；
To the training set XP t—4X is obtained as normalizedR t—4；
According to XR t—4Construct BP neural network model.
9. the grain quality index prediction technique based on BP neural network characterized by comprising
Obtain grain prediction phase information t, wherein the grain prediction phase information t expression needs to predict that the quality of t phase grain refers to
Mark；
According to grain prediction phase information t, the grain quality achievement data for meeting time requirement is obtained；
BP neural network prediction model is constructed according to the grain quality achievement data got；
According to the quality index for the BP neural network prediction model prediction t phase grain that building is completed；
The index of quality of t phase grain is assessed according to the quality index of the t phase grain predicted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910750723.6A CN110458361A (en) | 2019-08-14 | 2019-08-14 | Grain quality index prediction technique based on BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910750723.6A CN110458361A (en) | 2019-08-14 | 2019-08-14 | Grain quality index prediction technique based on BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110458361A true CN110458361A (en) | 2019-11-15 |
Family
ID=68486688
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910750723.6A Pending CN110458361A (en) | 2019-08-14 | 2019-08-14 | Grain quality index prediction technique based on BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110458361A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111274538A (en) * | 2020-01-21 | 2020-06-12 | 中国农业大学 | Method and system for judging quality state of raw aquatic products in storage and transportation processes |
| CN112395753A (en) * | 2020-11-11 | 2021-02-23 | 中国农业大学 | Drying method, device, equipment and storage medium for directionally regulating and controlling quality of rice |
| CN112598116A (en) * | 2020-12-22 | 2021-04-02 | 王槐林 | Pet appetite evaluation method, device, equipment and storage medium |
| CN113763033A (en) * | 2021-08-04 | 2021-12-07 | 武汉轻工大学 | Method, device, equipment and storage medium for selecting fish seed putting in full-automatic fishing ground |
| CN116912244A (en) * | 2023-09-12 | 2023-10-20 | 中储粮成都储藏研究院有限公司 | Intelligent grain quality index detection method based on machine vision |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101655454A (en) * | 2009-09-15 | 2010-02-24 | 北京市农林科学院 | Rapid determination method for evaluation of storage quality of grain |
| US20110066579A1 (en) * | 2009-09-16 | 2011-03-17 | Oki Electric Industry Co., Ltd. | Neural network system for time series data prediction |
| US20110307228A1 (en) * | 2008-10-15 | 2011-12-15 | Nikola Kirilov Kasabov | Data analysis and predictive systems and related methodologies |
| US20140332994A1 (en) * | 2013-05-13 | 2014-11-13 | Jeffrey E. Danes | Detection in thermoplastics |
| CN105891430A (en) * | 2016-04-11 | 2016-08-24 | 上海大学 | Food quality monitoring device based on neural network |
| CN106227038A (en) * | 2016-07-29 | 2016-12-14 | 中国人民解放军信息工程大学 | Grain drying tower intelligent control method based on neutral net and fuzzy control |
| US20170154347A1 (en) * | 2013-09-18 | 2017-06-01 | Simpler Postage, Inc. | Method and system for generating delivery estimates |
| CN107358388A (en) * | 2016-11-03 | 2017-11-17 | 厦门嵘拓物联科技有限公司 | A kind of WMS based on Internet of Things and the storage quality risk appraisal procedure based on the system |
| CN108090628A (en) * | 2018-01-16 | 2018-05-29 | 浙江大学 | A kind of grain feelings security detection and analysis method based on PSO-LSSVM algorithms |
| CN109002917A (en) * | 2018-07-13 | 2018-12-14 | 山东省农业信息中心 | Total output of grain multidimensional time-series prediction technique based on LSTM neural network |
| CN109062296A (en) * | 2018-08-27 | 2018-12-21 | 华夏易通国际物流有限公司 | A kind of grain storage method based on temperature and humidity monitor |
| CN109840854A (en) * | 2019-03-04 | 2019-06-04 | 中国科学技术大学 | Grain mildew prediction technique and device |
| CN109991325A (en) * | 2019-03-07 | 2019-07-09 | 中国检验检疫科学研究院 | The evaluation method and its application of food quality based on metabolism group data fusion and artificial neural network |
| US20190228176A1 (en) * | 2018-01-19 | 2019-07-25 | Griffin Group Global, LLC | System and method for providing a prediction-based data structure having different-scheme-derived portions |
| US20190228855A1 (en) * | 2018-01-25 | 2019-07-25 | Kraft Foods Group Brands Llc | Method and system for improving food-related personalization |
-
2019
- 2019-08-14 CN CN201910750723.6A patent/CN110458361A/en active Pending
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110307228A1 (en) * | 2008-10-15 | 2011-12-15 | Nikola Kirilov Kasabov | Data analysis and predictive systems and related methodologies |
| CN101655454A (en) * | 2009-09-15 | 2010-02-24 | 北京市农林科学院 | Rapid determination method for evaluation of storage quality of grain |
| US20110066579A1 (en) * | 2009-09-16 | 2011-03-17 | Oki Electric Industry Co., Ltd. | Neural network system for time series data prediction |
| US20140332994A1 (en) * | 2013-05-13 | 2014-11-13 | Jeffrey E. Danes | Detection in thermoplastics |
| US20170154347A1 (en) * | 2013-09-18 | 2017-06-01 | Simpler Postage, Inc. | Method and system for generating delivery estimates |
| CN105891430A (en) * | 2016-04-11 | 2016-08-24 | 上海大学 | Food quality monitoring device based on neural network |
| CN106227038A (en) * | 2016-07-29 | 2016-12-14 | 中国人民解放军信息工程大学 | Grain drying tower intelligent control method based on neutral net and fuzzy control |
| CN107358388A (en) * | 2016-11-03 | 2017-11-17 | 厦门嵘拓物联科技有限公司 | A kind of WMS based on Internet of Things and the storage quality risk appraisal procedure based on the system |
| CN108090628A (en) * | 2018-01-16 | 2018-05-29 | 浙江大学 | A kind of grain feelings security detection and analysis method based on PSO-LSSVM algorithms |
| US20190228176A1 (en) * | 2018-01-19 | 2019-07-25 | Griffin Group Global, LLC | System and method for providing a prediction-based data structure having different-scheme-derived portions |
| US20190228855A1 (en) * | 2018-01-25 | 2019-07-25 | Kraft Foods Group Brands Llc | Method and system for improving food-related personalization |
| CN109002917A (en) * | 2018-07-13 | 2018-12-14 | 山东省农业信息中心 | Total output of grain multidimensional time-series prediction technique based on LSTM neural network |
| CN109062296A (en) * | 2018-08-27 | 2018-12-21 | 华夏易通国际物流有限公司 | A kind of grain storage method based on temperature and humidity monitor |
| CN109840854A (en) * | 2019-03-04 | 2019-06-04 | 中国科学技术大学 | Grain mildew prediction technique and device |
| CN109991325A (en) * | 2019-03-07 | 2019-07-09 | 中国检验检疫科学研究院 | The evaluation method and its application of food quality based on metabolism group data fusion and artificial neural network |
Non-Patent Citations (3)
| Title |
|---|
| ANONYMOUS: "《IP.COM非专利全文库》", 15 December 2017, IP.COM * |
| 哈根: "《神经网络设计》", 30 September 2002, 机械工业出版社 * |
| 沈兵: "基于BP神经网络的储粮安全预测模型研究", 《交通信息与安全》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111274538A (en) * | 2020-01-21 | 2020-06-12 | 中国农业大学 | Method and system for judging quality state of raw aquatic products in storage and transportation processes |
| CN111274538B (en) * | 2020-01-21 | 2021-10-26 | 中国农业大学 | Method and system for judging quality state of raw aquatic products in storage and transportation processes |
| CN112395753A (en) * | 2020-11-11 | 2021-02-23 | 中国农业大学 | Drying method, device, equipment and storage medium for directionally regulating and controlling quality of rice |
| CN112395753B (en) * | 2020-11-11 | 2021-10-29 | 中国农业大学 | Drying method, device, equipment and storage medium for directionally regulating and controlling quality of rice |
| CN112598116A (en) * | 2020-12-22 | 2021-04-02 | 王槐林 | Pet appetite evaluation method, device, equipment and storage medium |
| CN113763033A (en) * | 2021-08-04 | 2021-12-07 | 武汉轻工大学 | Method, device, equipment and storage medium for selecting fish seed putting in full-automatic fishing ground |
| CN116912244A (en) * | 2023-09-12 | 2023-10-20 | 中储粮成都储藏研究院有限公司 | Intelligent grain quality index detection method based on machine vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110458361A (en) | Grain quality index prediction technique based on BP neural network | |
| CN109214948A (en) | A kind of method and apparatus of electric system heat load prediction | |
| CN105160400B (en) | The method of lifting convolutional neural networks generalization ability based on L21 norms | |
| CN106777874A (en) | The method that forecast model is built based on Recognition with Recurrent Neural Network | |
| CN106779867A (en) | Support vector regression based on context-aware recommends method and system | |
| CN109002917A (en) | Total output of grain multidimensional time-series prediction technique based on LSTM neural network | |
| CN110458362A (en) | Grain quality index prediction technique based on SVM supporting vector machine model | |
| CN106874355A (en) | The collaborative filtering method of social networks and user's similarity is incorporated simultaneously | |
| CN107239993A (en) | A kind of matrix decomposition recommendation method and system based on expansion label | |
| CN108537624A (en) | A kind of tourist service recommendation method based on deep learning | |
| Ahmed et al. | Relative efficacy of food and cash transfers in improving food security and livelihoods of the ultra-poor in Bangladesh | |
| CN108509573A (en) | Book recommendation method based on matrix decomposition collaborative filtering and system | |
| CN107609289A (en) | The building material cost management-control method and system of structural fortification based on BIM models | |
| CN106503438A (en) | A kind of H RFM user modeling method and system for pharmacy member analysis | |
| CN106469336B (en) | Two-dimension code anti-counterfeit prediction meanss and method based on BP neural network | |
| CN108038629A (en) | A kind of optimization method based on collaborative filtering | |
| CN107506845A (en) | A kind of electricity sales amount Forecasting Methodology and its system based on multi-model fusion | |
| CN106709758A (en) | Intelligent recommendation and distribution solution for service class commodities in electronic commerce products | |
| CN108334638A (en) | Collaborative Filtering method based on shot and long term Memory Neural Networks and interest migration | |
| CN111091241A (en) | BP neural network-based drug sales prediction and decision method and system | |
| CN107016122A (en) | Knowledge recommendation method based on time-shift | |
| CN109493151A (en) | Method for Sales Forecast method and system | |
| CN106202377A (en) | A kind of online collaborative sort method based on stochastic gradient descent | |
| CN109740734A (en) | A kind of method of neuron spatial arrangement in optimization convolutional neural networks | |
| CN109857939A (en) | Accurate method for pushing towards intelligence manufacture service |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191115 |
|
| RJ01 | Rejection of invention patent application after publication |