CN104715295A - Chinese chemical fertilizer price index prediction method based on BP neural network - Google Patents
Chinese chemical fertilizer price index prediction method based on BP neural network Download PDFInfo
- Publication number
- CN104715295A CN104715295A CN201510157489.8A CN201510157489A CN104715295A CN 104715295 A CN104715295 A CN 104715295A CN 201510157489 A CN201510157489 A CN 201510157489A CN 104715295 A CN104715295 A CN 104715295A
- Authority
- CN
- China
- Prior art keywords
- price index
- neural network
- index prediction
- chemical fertilizer
- prediction method
- 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 30
- 239000003337 fertilizer Substances 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 26
- 239000000126 substance Substances 0.000 title abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 15
- 210000002569 neuron Anatomy 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 10
- 238000012886 linear function Methods 0.000 claims description 5
- 238000012546 transfer Methods 0.000 claims description 4
- 239000003795 chemical substances by application Substances 0.000 claims 1
- 238000013277 forecasting method Methods 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 13
- 238000012545 processing Methods 0.000 description 4
- 238000013459 approach Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012271 agricultural production Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012892 rational function Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Fertilizers (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a prediction method and particularly relates to a Chinese chemical fertilizer price index prediction method based on a BP neural network, belonging to the technical field of time sequence prediction. According to the technical scheme provided by the invention, the Chinese chemical fertilizer price index prediction method based on a BP neural network comprises the following steps: (a) training the BP neural network by use of the historical data of the Chinese chemical fertilizer price to obtain a price index prediction model; and (b) predicting the price index by use of the price index prediction model. The method provided by the invention needs consideration of statistical property calculation and can be theoretically applied to the modeling of any non-linear time sequence; with a unique non-traditional expression way and inherent learning ability, the method has tremendous advantages in controlling highly non-linear and seriously uncertain systems; and the method realizes high accuracy of predicting the Chinese chemical fertilizer price index.
Description
Technical Field
The invention relates to a prediction method, in particular to a Chinese fertilizer price index prediction method based on a BP neural network, belonging to the technical field of time series prediction.
Background
The chemical fertilizer plays an important role in agriculture in China, can be used as a basis for production decision making of enterprises and vast farmers to master initiative and gain benefits for the enterprises and the farmers due to correct prediction of price indexes of the chemical fertilizer in China, and can provide scientific basis for relevant policies made by governments, so that effective allocation of agricultural resources is improved, and sustainable and healthy development of agriculture is promoted.
Time series prediction can be classified into a simple time-series average method, a weighted time-series average method, a simple moving average method, a weighted moving average method, an exponential smoothing method, a trend prediction method, a seasonal trend prediction method, a market life cycle prediction method, and the like.
The mechanism by which fertilizer prices are developed is a nonlinear system with a high degree of complexity. Factors affecting the price of the fertilizer are many, such as policy control, raw material price, market demand, inventory, sale month, sale market, variety specification and the like. The traditional time series prediction method is generally only suitable for linear models or some nonlinear models which can be linearized, or the parameter estimation is very difficult and not easy to use when in use, and the prediction effect is not ideal.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a Chinese fertilizer price index prediction method based on a BP (back propagation) neural network.
According to the technical scheme provided by the invention, the Chinese fertilizer price index prediction method based on the BP neural network comprises the following steps:
a. training a BP neural network by using historical data of Chinese fertilizer prices to obtain a price index prediction model;
b. and predicting the price index by using the price index prediction model.
In the step a, a linear function is adopted as a normalization function in the BP neural network; the number m of hidden layer neurons isWherein m is the number of neurons in the hidden layer, n is the number of input nodes, l is the number of output nodes, and a is an adjusting constant between 0 and 10; the transfer function is Sigmoid function, and f (x) isx is input data; the obtained price index prediction model F (x) is:
F(x)＝(f(x1),f(x2),…,f(xq))
wherein, represents a threshold value; wijAre hidden layer weights.
The BP neural network adopts a three-layer BP neural network and a threshold valueIs 0.92, the hidden layer weight WijIs (3.15, 40.06, 15.24, -8.29, 17.69, -16.53, -74.15, -23.48, 3.10, 2.73, -38.05, -17.35), and the input data is year, month and last fertilizer price index value.
The invention has the advantages that: the statistical characteristics need to be considered, and the method can be applied to any nonlinear time series modeling in theory; the unique non-traditional expression mode and the inherent learning ability have great advantages in the aspect of solving the system control with high nonlinearity and serious uncertainty; the accuracy of the Chinese fertilizer price index prediction is high.
Detailed Description
The present invention will be further described with reference to the following specific examples.
In order to improve the prediction precision of the Chinese fertilizer price index, the Chinese fertilizer price index prediction method comprises the following steps:
a. training a BP neural network by using historical data of Chinese fertilizer prices to obtain a price index prediction model;
in the embodiment of the invention, the data with accurate history (year, month and date index value) is selected as the sample, and the sample with accurate history can adopt the fertilizer price and sales data reported by member units distributed all over the country by the China Association for agricultural production data circulation.
The invention adopts three layers of BP neural networks, and the three layers of BP neural networks can realize multidimensional unit cube RtTo RdThe mapping of (2) can approach any rational function, increasing the number of layers can further reduce the error and improve the precision, but also complicates the network, thereby increasing the training time of the network weight.
Normalization function: normalization, also called normalization, means to limit the input and output data of the network to the [0,1] or [ -1,1] interval by transformation processing. The main reasons for normalization are two reasons: firstly, because each variable in the sample set has different physical meanings and different dimensions, normalization gives equal importance to each input component, and summarizes the statistical distribution of the unified sample, thereby providing convenience for the subsequent data processing. Secondly, a hidden layer of the BP neural network generally adopts a Sigmoid transfer function, and in order to improve training speed and sensitivity and effectively avoid a saturation region of the Sigmoid function, the value of input data is required to be between [0 and 1], so that data needs to be preprocessed, and different variables need to be preprocessed respectively.
Therefore, before neural network training, input samples must be normalized, and the normalized input value and output value uniformly fall within the [0,1] interval, and in practical application, there are generally three common normalization methods: normalizing the linear function; normalizing the logarithmic function; and normalizing the inverse cotangent function. The invention adopts a linear function to carry out normalization processing on input sample data.
Number of cryptic neurons: has certain influence on the performance of the neural network. When the number of the neurons in the hidden layer is too small, the learning capacity is limited, and the learning capacity is not enough to store all rules contained in a training sample, and when the number of the neurons in the hidden layer is too large, the network training time is increased, irregular contents such as interference and noise in the sample are stored, and the generalization capability is reduced. The number m of hidden layer neurons isWherein m is the number of neurons in the hidden layer, n is the number of input nodes, l is the number of output nodes, and a is an adjusting constant between 0 and 10; the transfer function is Sigmoid function, and f (x) isx is input data; in general, the prediction effect is best when the number m of the hidden layer neurons is 12.
Learning rate: the learning rate determines the weight variation generated in each round of training. A high learning rate may lead to instability of the system; but a low learning rate results in a longer training time, possibly slow convergence, but ensures that the network error values jump out of the valleys of the error surface and eventually approach the minimum error value. In general, a smaller learning rate tends to be selected to ensure system stability. The learning rate is selected to be in the range of 0.01-0.8. The learning rate employed in the present invention is 0.72.
The Sigmoid function was chosen as the output function because of its beneficial properties: non-linear, monotonic; infinite times can be micro; when the weight value is large, the threshold function is approximate, and when the weight value is small, the linear function can be approximate.
The method can be used for the actual application of Chinese fertilizer price index prediction after network training. And (3) inputting the sample data in the step (1) into a BP neural network after normalization processing, and training the network until the relative error meets the requirement. After the network training is completed, the important parameters of the network can be obtained: implicit layer weights and thresholds.
In the embodiment of the invention, the obtained price index prediction model F (x) is as follows:
F(x)＝(f(x1),f(x2),…,f(xq))
wherein, represents a threshold value; wijAre hidden layer weights.
The BP neural network adopts a three-layer BP neural network and a threshold valueIs 0.92, the hidden layer weight WijIs (3.15, 40.06, 15.24, -8.29, 17.69, -16.53, -74.15, -23.48, 3.10, 2.73, -38.05, -17.35), and the input data is year, month and last fertilizer price index value.
b. And predicting the price index by using the price index prediction model.
And (4) predicting the price index of the Chinese fertilizer by using the trained BP network. After the BP neural network training is finished, the price index of the Chinese fertilizer can be predicted, input data (the price index values of the fertilizer in the year, month and the previous period) are input into the network, and the neural network can calculate a predicted value.
The BP neural network algorithm can be realized by utilizing the existing NET software platform. The present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof, and it is therefore intended that all such changes and modifications as fall within the true spirit and scope of the invention be considered as within the following claims.
Claims (3)
1. A Chinese fertilizer price index prediction method based on a BP neural network is characterized by comprising the following steps:
(a) training the BP neural network by using historical data of Chinese fertilizer prices to obtain a price index prediction model;
(b) and predicting the price index by using the price index prediction model.
2. The BP-based agent of claim 1The Chinese fertilizer price index forecasting method through the network is characterized in that in the step (a), a linear function is adopted as a normalization function in a BP neural network; the number m of hidden layer neurons isWherein m is the number of neurons in the hidden layer, n is the number of input nodes, l is the number of output nodes, and a is an adjusting constant between 0 and 10; the transfer function is Sigmoid function, and f (x) isx is input data; the obtained price index prediction model F (x) is:
F(x)＝(f(x1),f(x2),…,f(xq))
wherein, represents a threshold value; wijAre hidden layer weights.
3. The BP neural network-based Chinese fertilizer price index prediction method as claimed in claim 2, wherein the BP neural network adopts a three-layer BP neural network and a threshold valueIs 0.92, the hidden layer weight WijIs (3.15, 40.06, 15.24, -8.29, 17.69, -16.53, -74.15, -23.48, 3.10, 2.73, -38.05, -17.35), and the input data is year, month and last fertilizer price index value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510157489.8A CN104715295A (en) | 2015-04-03 | 2015-04-03 | Chinese chemical fertilizer price index prediction method based on BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510157489.8A CN104715295A (en) | 2015-04-03 | 2015-04-03 | Chinese chemical fertilizer price index prediction method based on BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104715295A true CN104715295A (en) | 2015-06-17 |
Family
ID=53414603
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510157489.8A Pending CN104715295A (en) | 2015-04-03 | 2015-04-03 | Chinese chemical fertilizer price index prediction method based on BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104715295A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096558A (en) * | 2016-06-16 | 2016-11-09 | 长安大学 | A kind of binding reinforcing bars method based on neutral net |
| CN106296280A (en) * | 2016-08-05 | 2017-01-04 | 摩贝（上海）生物科技有限公司 | A kind of compound price index Estimation System and method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101276454A (en) * | 2007-12-05 | 2008-10-01 | 中原工学院 | Method for model building, forecasting and decision-making of stock market based on BP neural net |
| CN102982229A (en) * | 2012-09-06 | 2013-03-20 | 淮阴工学院 | Multi-assortment commodity price expectation data pre-processing method based on neural networks |
| CN104463359A (en) * | 2014-12-01 | 2015-03-25 | 河海大学常州校区 | Dredging operation yield prediction model analysis method based on BP neural network |
-
2015
- 2015-04-03 CN CN201510157489.8A patent/CN104715295A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101276454A (en) * | 2007-12-05 | 2008-10-01 | 中原工学院 | Method for model building, forecasting and decision-making of stock market based on BP neural net |
| CN102982229A (en) * | 2012-09-06 | 2013-03-20 | 淮阴工学院 | Multi-assortment commodity price expectation data pre-processing method based on neural networks |
| CN104463359A (en) * | 2014-12-01 | 2015-03-25 | 河海大学常州校区 | Dredging operation yield prediction model analysis method based on BP neural network |
Non-Patent Citations (1)
| Title |
|---|
| 李建华: "基于BP神经网络的二手船船价指数预测研究", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096558A (en) * | 2016-06-16 | 2016-11-09 | 长安大学 | A kind of binding reinforcing bars method based on neutral net |
| CN106096558B (en) * | 2016-06-16 | 2019-08-30 | 长安大学 | A kind of binding reinforcing bars method neural network based |
| CN106296280A (en) * | 2016-08-05 | 2017-01-04 | 摩贝（上海）生物科技有限公司 | A kind of compound price index Estimation System and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Prediction of stock prices based on LM-BP neural network and the estimation of overfitting point by RDCI | |
| CN106022521B (en) | Short-term load prediction method of distributed BP neural network based on Hadoop architecture | |
| CN108197773A (en) | Methods of electric load forecasting, load forecast device and terminal device | |
| CN107730006A (en) | A kind of nearly zero energy consumption controller of building based on regenerative resource big data deep learning | |
| CN102682345A (en) | Traffic flow prediction method based on quick learning neural network with double optimal learning rates | |
| CN107705556A (en) | A kind of traffic flow forecasting method combined based on SVMs and BP neural network | |
| Du et al. | Genetic algorithm combined with BP neural network in hospital drug inventory management system | |
| CN103886218A (en) | Lake and reservoir algal bloom predicating method based on multielement nonstationary time series analysis and neural network and support vector machine compensation | |
| CN108665322A (en) | The construction method of grain ration Consumption forecast model, Consumption forecast method and device | |
| CN110503251A (en) | A kind of non-festivals or holidays load forecasting method based on Stacking algorithm | |
| CN106650022A (en) | Method for predicting fault of complex electronic device | |
| Wu et al. | An ant colony algorithm for Master production scheduling optimization | |
| CN104715295A (en) | Chinese chemical fertilizer price index prediction method based on BP neural network | |
| CN107633294A (en) | A kind of improvement Sobol Sensitivity Analysis applied to RBF neutral nets | |
| CN110874635B (en) | Deep neural network model compression method and device | |
| Luo et al. | Stock market forecasting algorithm based on improved neural network | |
| Wang et al. | Prediction of water quality in South to North Water Transfer Project of China based on GA-optimized general regression neural network | |
| CN106250686A (en) | A kind of collective communication function modelling method of concurrent program | |
| Khan et al. | Stochastic computational approach for complex nonlinear ordinary differential equations | |
| CN115619449A (en) | Deep learning model-based fertilizer price prediction method, device and storage medium | |
| Lian et al. | Training algorithm for radial basis function neural network based on quantum-behaved particle swarm optimization | |
| CN115545490A (en) | Power station engineering cost analysis method and device based on small sample machine learning | |
| Luo et al. | Application of generalized regression neural network to the agricultural machinery demand forecasting | |
| Hu et al. | [Retracted] Support Vector Machine‐Based Nonlinear Model and Its Application in Regional Economic Forecasting | |
| Lv et al. | Application of an improved grey neural network in grain yield prediction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150617 |