CN103310286A - Product order prediction method and device with time series characteristics - Google Patents
Product order prediction method and device with time series characteristics Download PDFInfo
- Publication number
- CN103310286A CN103310286A CN2013102612648A CN201310261264A CN103310286A CN 103310286 A CN103310286 A CN 103310286A CN 2013102612648 A CN2013102612648 A CN 2013102612648A CN 201310261264 A CN201310261264 A CN 201310261264A CN 103310286 A CN103310286 A CN 103310286A
- Authority
- CN
- China
- Prior art keywords
- prediction
- order
- data
- time series
- neural network
- 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
Images
Abstract
The invention discloses a product order prediction method and device with time series characteristics. The product order prediction method comprises obtaining statistics of order data of enterprises at every time point according to stored historical order data; selecting a prediction model according to the time series characteristics of the order data and determining a prediction output equation of the prediction model; enabling the statistics of the historical order data to be processed as a prediction input table according to the requirements of the prediction model and training a corresponding prediction network model; and utilizing the prediction input table of the prediction order quantity to calculate to obtain the prediction order quantity of orders according to the prediction model which is well-trained through prediction orders and the prediction output equation. The invention also provides the product order prediction device according to the product order prediction method. The product order prediction device mainly comprises a data acquisition module, a data preprocessing module, a time series modeling module and an order prediction module. The product order prediction method and device with the time series characteristics have the advantages of solving the nonlinear problem of product order prediction, meeting the requirements of system real-time performance and improving the prediction accuracy.
Description
Technical field
The invention belongs to the enterprise production management field, be specifically related to a kind of product order forecast method and device with time series characteristic.
Background technology
Order forecasting is one of important content of modern enterprise production management.Modern economy is fierce market economy, and client's demand is tending towards personalized and diversified in the market, and this has just proposed Secretary to production and operation pattern and the management method of enterprise.Therefore, the order form mode that represents customer demand has just become source and the terminal point of enterprise operation and management and supply chain management.Can modern market be changed to the buyer's market by the seller's market, therefore, catch client as much as possible to become the key of Business survival and maintenance competitive power.And the key of catching the client just is whether product can satisfy client's demand to greatest extent, whether can in time tackle the variation of market and customer demand.This just requires enterprise should organize according to market and client's demand order volume a series of business activities such as buying, production, logistics distribution of all departments of enterprise, could transfer to greatest extent the enthusiasm for production of each department of enterprise like this, improve greatly the production efficiency of enterprise and whole piece supply chain, cut operating costs.Just under such background, predict that exactly the product order just seems especially important, and become the correct guide of the business activities such as enterprise procurement, producing and selling, for the equilibrium of stock plan of next production cycle provides guidance, improve the sensitivity of supply chain, make the production division can the reasonable arrangement production capacity and prepare raw materials for production, adapt to that steel industry is intelligent, the needs of modernization development.
Time prediction is the field of a multidisciplinary intersection, also is one of important means of dynamic data analyzing and processing.In the face of present increasing mass data, no matter be on the upper or space of time, traditional data analysis is processed means and has been difficult to competently, has in a large number the information of important actual production meaning and is wasted because can't obtain timely, correct processing.So, for the dynamic data with time response is carried out analyzing and processing, must from analysis time sequence data variation characteristic start with, set up suitable forecast model, again according to the inertia principle, suppose that forecasting object variation tendency in the past can be extended in the future, thereby make the comparatively prediction of science.By trend analysis, can in more rational situation, make the prediction of long-term or short-term.In fact trend analysis is exactly the time series prediction, all has suitable applied research at numerous areas such as science, economy, engineerings at present and is worth.The present invention introduces time series forecasting with Wavelet Neural Network Method on the basis of original Dynamic Time Series predicted application research, and this is carried out comparatively deep applied research.
At present, the method of order forecasting is a lot, method from current research, to carry out with the method for data mining mostly, there is the limitation of algorithm complexity in these methods in the production application process, and shortage to the visual exploration of data, was applied to dynamic data analysis advantage not obvious simultaneously before data prediction, can not embody the development trend of dynamic data.
Studies show that, the wavelet neural network theory is introduced in the time series predicting model, and the dynamic time sequence data are carried out forecast analysis, have very strong using value.Because wavelet neural network is a kind of artificial intelligence neural networks, it can imitate and extend the functions such as human brain intelligence, thinking, consciousness, has very strong self study and adaptive ability, and can carry out nonlinear dynamic processing, solve nonlinear problem, satisfy the requirement of system real time.
Summary of the invention
The present invention is directed to the characteristics that the product order forecasting has time series models, a kind of product order forecast method and device with time series characteristic is provided, to improve the order forecasting precision, for the equilibrium of stock plan of next production cycle provides guidance, make the production division can the reasonable arrangement production capacity and prepare raw materials for production, adapt to the needs of enterprise intelligent, modernization development.
In order to achieve the above object, the present invention adopts following technical scheme:
A kind of product order forecast method with time series characteristic comprises:
According to the time series characteristic of order data, choose wavelet neural network as time series predicting model, be x in input History Order amount sequence samples
i(i=1,2 ..., in the time of k), hidden layer output computing formula is
In the formula, h (j) is j node output valve of hidden layer; ω
IjFor input layer is connected the connection weights with hidden layer; b
jBe wavelet basis function h
jShift factor; a
jBe wavelet basis function h
jContraction-expansion factor; h
jBe wavelet basis function.
The wavelet basis function that the embodiment of the invention adopts is the female wavelet basis function of Morlet, and mathematical formulae is:
Wavelet neural network output layer computing formula is
In the formula, ω
JkFor hidden layer arrives the output layer weights; H (i) is the output of i node; L is the hidden layer node number; M is the output layer nodes.
Step 3, the statistical value of History Order amount data is the prediction input table, the corresponding prediction of training network model according to the requirement arrangement of forecast model;
Initialization wavelet function contraction-expansion factor a
k, shift factor b
kAnd network connection weights omega
Ij, ω
Jk, e-learning speed η is set;
In order to eliminate the difference of the order of magnitude between each dimension data, avoid causing the neural network forecast error larger because inputoutput data order of magnitude difference is large, then all data are carried out normalized.Choose the minimax method in the present embodiment and carry out normalized, functional form is
x
k=(x
k-x
min)/(x
max-x
min)
X wherein
MaxBe maximal value in the data sequence, x
MinBe minimum value in the data sequence.
The input parameter of wavelet neural network is { X
1, X
2..., X
k, input History Order amount burst sample is x
i(i=1,2 ..., k), all samples being divided into training sample and test sample book, training sample is used for training network, and test sample book is used for the test network precision of prediction, and prediction of output result;
With the training sample input neural network, the prediction that obtains wavelet neural network is output as y
j(j=1,2 ..., m), and the error e of computational grid output y (k) and desired output yn (k), wherein error e is:
According to error e roll-off network weights
And wavelet parameter
Make the neural network forecast value approach expectation value;
Revise wavelet neural network weights and wavelet basis function coefficient according to predicated error e:
In the formula, η is learning rate.
The above step that circulates is trained wavelet neural network, and training finishes in the allowed band until predicated error is stable at.
Step 4, the forecast model that has trained according to described prediction order and prediction output equation, and the prediction input table of described prediction order calculates the prediction order volume of described order, and the predicted data that draws is carried out renormalization process, functional form is
y=y*(x
max-x
min)+x
min
Thus, can draw the once order volume of prediction.
Predicting the outcome for fear of wavelet neural network only has one uncertainty, the present invention adopts repeatedly cyclic wavelet neural network prediction model, draw a plurality of predicting the outcome, and these are predicted the outcome ask the method for its average, the average that draws is as the final output that predicts the outcome.
A kind of product order forecasting device with time series characteristic comprises:
Data acquisition module, be used for to gather this product within each time period (as day or month) actual order volume, and its arrangement become the prediction input table;
Data preprocessing module is used for the actual order volume of described prediction order is done normalized;
The time series modeling module is used for determining time series models corresponding to described order forecasting and prediction output equation according to described order data with time series characteristic;
The order forecasting module is used for obtaining the prediction order volume of described product according to the historical data of described prediction product and forecast model and the prediction output equation of described prediction product order.
Beneficial effect of the present invention: a kind of product order forecasting device and method with time series characteristic of proposition, be based on the time series predicting model of wavelet neural network, this model takes full advantage of wavelet neural network and has good Nonlinear Processing analysis ability and very strong self study and adaptive characteristics, solve this nonlinear problem of product order forecasting, satisfied the requirement of system real time.And, improve traditional prediction model based on wavelet neural network and only produced a uncertainty that predicts the outcome, prediction model based on wavelet neural network circular flow is repeatedly asked for average as last predicting the outcome, improved the accuracy of prediction.
Description of drawings
Fig. 1 is a kind of process flow diagram with order forecast method of time series characteristic.
Fig. 2 is the algorithm flow chart of wavelet neural network.
Fig. 3 is the topological structure of wavelet neural network.
Fig. 4 is certain iron and steel enterprise high line of in March, 2013 figure that predicts the outcome of 60 times of circulation that predicts the outcome.
Fig. 5 is that certain iron and steel enterprise predicts order volume and actual order volume contrast broken line graph year March in June, 2012 to 2013.
Fig. 6 is that certain iron and steel enterprise predicts order volume and actual order volume error broken line graph year March in June, 2012 to 2013.
Fig. 7 is a kind of formation schematic diagram with order forecasting device of time series characteristic that the embodiment of the invention provides.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is described in detail.Obviously, described embodiment only is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.
In order to predict exactly the product order volume, satisfy the requirement of product form ordering system real-time, the embodiment of the invention provides a kind of seasonal effect in time series product order volume Forecasting Methodology that has, and as shown in Figure 1, comprising:
The production of present most of enterprise is based on a moon order, and the present invention is take the high line order volume statistics (seeing Table 1) of certain iron and steel enterprise as the order data statistical value.Here, in order to protect enterprise's privacy, the embodiment of the invention has been done respective handling to the order data of this enterprise.
Certain iron and steel enterprise of table 1 high line forecast model order data (actual order data being processed)
According to the time series characteristic of order data, choose wavelet neural network as time series predicting model, the algorithm flow chart of wavelet neural network is as shown in Figure 2.
The topological structure of wavelet neural network as shown in Figure 3, wherein, X
1, X
2..., X
kBe the input parameter of wavelet neural network, Y
1, Y
2..., Y
mBe the prediction output of wavelet neural network, ω
IjAnd ω
JkWeights for wavelet neural network.Be x in input History Order amount sequence samples
i(i=1,2 ..., in the time of k), hidden layer output computing formula is
In the formula, h (j) is j node output valve of hidden layer; ω
IjFor input layer is connected the connection weights with hidden layer; b
jBe wavelet basis function h
jShift factor; a
jBe wavelet basis function h
jContraction-expansion factor; h
jBe wavelet basis function.
The wavelet basis function that the embodiment of the invention adopts is the female wavelet basis function of Morlet, and mathematical formulae is:
Wavelet neural network output layer computing formula is
In the formula, ω
JkFor hidden layer arrives the output layer weights; H (i) is the output of i node; L is the hidden layer node number; M is the output layer nodes.
Step 3, the statistical value of History Order amount data is the prediction input table, the corresponding prediction of training network model according to the requirement arrangement of forecast model;
Initialization wavelet function contraction-expansion factor a
k, shift factor b
kAnd network connection weights omega
Ij, ω
Jk, e-learning speed η is set;
In order to eliminate the difference of the order of magnitude between each dimension data, avoid causing the neural network forecast error larger because inputoutput data order of magnitude difference is large, then all data are carried out normalized.Choose the minimax method in the present embodiment and carry out normalized, functional form is
x
k=(x
k-x
min)/(x
max-x
min)
X wherein
MaxBe maximal value in the data sequence, x
MinBe minimum value in the data sequence.
The input parameter of wavelet neural network is { X
1, X
2..., X
k, input History Order amount burst sample is x
i(i=1,2 ..., k), all samples being divided into training sample and test sample book, training sample is used for training network, and test sample book is used for the test network precision of prediction, and prediction of output result;
With the training sample input neural network, the prediction that obtains wavelet neural network is output as y
j(j=1,2 ..., m), and the error e of computational grid output y (k) and desired output yn (k), wherein error e is:
According to error e roll-off network weights
And wavelet parameter
Make the neural network forecast value approach expectation value;
Revise wavelet neural network weights and wavelet basis function coefficient according to predicated error e:
In the formula, η is learning rate.
The above step that circulates is trained wavelet neural network, and training finishes in the allowed band until predicated error is stable at.
What choose in the embodiment of the invention is 4-6-1 based on the wavelet neural network structure, and namely the input layer number is 4, and the hidden layer node number is 6, and the output layer nodes is 1.
Learn according to historical experience, when round-off error e when at every turn predicting simultaneously about 12 times, the cumulative errors value is minimum, is 12 row so the prediction input table of the present embodiment is set line number.
After setting wavelet neural network training circulation 100 times according to historical experience in the present embodiment, the acquiescence error e has fallen in the ideal range and has been stable in this scope, obtain network weight and wavelet basis function coefficient according to the round-off error e that obtains, and the fixed value during as subsequent prediction input data.
Can be determined by above-mentioned steps, in the embodiment of the invention, if will predict certain iron and steel enterprise of iron and steel enterprise high dose of X-ray in March, 2013, then its prediction input table is the form of 12 row, 9 row, and is as shown in table 2 below.
Table 2 prediction input data table
Wherein, front 5 classify the training sample of neural network as, 1-4 classifies training input sample as, the training input sample that wavelet neural network is listed as according to 1-4 draws training and predicts the outcome, compare with the 5th row training output sample, draw predicated error e, according to error e weights and the wavelet coefficient of network are revised repeatedly; After the circuit training 100 times, draw round-off error at this moment and network weight, wavelet basis function coefficient, the value when inputting as the subsequent prediction data; 5-9 is listed as test sample book, is used for the wavelet neural network that has trained is tested, and draws to predict the outcome.
Step 4, the forecast model that has trained according to described prediction order and prediction output equation, and the prediction input table of described prediction product order calculates the prediction order volume of described product.
In the present embodiment, according to the described forecast model that has trained and output equation, utilize the test input data of 5-9 row in the prediction input table, draw the prediction Output rusults, and the predicted data that draws is carried out renormalization process, functional form is
y=y*(x
max-x
min)+x
min
Thus, can draw the once order volume of prediction.
Predicting the outcome for fear of wavelet neural network only has one uncertainty, the present invention adopts repeatedly cyclic wavelet neural network prediction model, draw a plurality of predicting the outcome, and these are predicted the outcome ask the method for its average, the average that draws is as the final output that predicts the outcome.
In the embodiment of the invention, setting cycle index is 60 times, and then 60 each the predicting the outcome as shown in Figure 4 that circulate average to 60 times predict the outcome, and the high line order volume that can draw in March, 2013 prediction is prediction=0.6970.And actual value can be got by table 1, be 0.68831, both absolute errors are 0.0126, according to the requirement (≤30%) of iron and steel enterprise to the predicated error scope, error in allowed band and prediction effect better, illustrate that the used Wavelet Neural Network Method of the present invention can be used for carrying out order forecasting.
By that analogy, measurable a plurality of months high line order volume.For the accuracy that shows that the inventive method predicts the outcome, the present invention according to year February in January, 2011 to 2013 in the table 1 high line order data, predicted certain iron and steel enterprise from the high line order volume in June, 2012 to 2013 year March, and contrasted with actual order volume.Order volume predicts the outcome with actual order volume comparison diagram as shown in Figure 5, and error result as shown in Figure 6.
As can be seen from Figure, predict the outcome less with the error of actual order volume, according to the condition of iron and steel enterprise to predicated error scope (≤30%), show the predicated error of this method in the error range that allows, so the method for the used wavelet neural network of the present invention prediction order volume is feasible.
With said method accordingly, the embodiment of the invention also provides a kind of order forecasting device with time series characteristic, as shown in Figure 7, comprising:
Data acquisition module, be used for to gather this product within each time period (as day or month) actual order volume, and its arrangement become the prediction input table;
Data preprocessing module is used for the actual order volume of described prediction order is done normalized;
The time series modeling module is used for determining time series models corresponding to described order forecasting and prediction output equation according to described order data with time series characteristic;
The order forecasting module is used for obtaining the prediction order volume of described product according to the historical data of described prediction product and forecast model and the prediction output equation of described prediction product order.
Further, described data preprocessing module specifically comprises:
Historical data gathers submodule, is used for obtaining the History Order data of storage;
The historical data reorganization submodule is for the data entry form that the historical data of obtaining is organized into the capable m row of n according to the requirement of Forecasting Methodology.
Further, described time series modeling module specifically comprises:
Time series models are determined submodule, are used for having the time series characteristic according to the order of described product, determine that used forecast model is the time series predicting model of wavelet neural network;
Parameter is determined submodule, is used for the correlation parameter according to the described time series models of described time series models initialization;
The error correction submodule is used for the feature request according to described prediction model based on wavelet neural network, and its network error is revised, and further determines the correlation parameter of forecast model according to the error of revising, and is more accurate to guarantee prediction output;
The prediction output sub-module is used for the Parameters in Mathematical Model according to described time series predicting model and process error correction, determines its output equation, as the prediction output equation of described time series predicting model.
Further, described order forecasting module specifically comprises:
Predicted data is obtained submodule, for the data that predict the outcome that obtain via the predictive equation generation of previous step forecast model;
Data subsequent treatment submodule is used for that predicting the outcome of obtaining carried out renormalization and processes, and in order to improve prediction accuracy, with the circular flow of time series modeling module repeatedly, obtains predicting the outcome and ask its average repeatedly, obtains last prediction order volume.
Further, described data subsequent treatment submodule specifically comprises:
Data renormalization processing unit is used for carrying out renormalization by predicting the outcome of predictive equation output and processes;
The data mean value processing unit is used for forecast model is carried out repeatedly circular flow, and a plurality of operation result averaged to obtaining, and obtains last prediction order volume.
A kind of product order forecasting device with time series characteristic that the embodiment of the invention proposes, be based on the time series predicting model of wavelet neural network, this device takes full advantage of wavelet neural network and has good Nonlinear Processing analysis ability and very strong self study and adaptive characteristics, solve this nonlinear problem of product order forecasting, satisfied the requirement of system real time.
One of ordinary skill in the art will appreciate that all or part of flow process that realizes in above-described embodiment method, can come the relevant hardware of instruction to finish described method and device by computer program.
The above; be the specific embodiment of the present invention only, but protection scope of the present invention is not limited to this, anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; can expect easily changing or replacing, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (10)
1. the product order forecast method with time series characteristic is characterized in that, comprising:
Step 1, according to the statistical value of order data of each time point of History Order data acquisition enterprise of storage;
Step 2, choose forecast model according to the time series characteristic of order data, determine the prediction output equation of forecast model;
According to the time series characteristic of order data, choose wavelet neural network as time series predicting model, be x in input History Order amount sequence samples
i(i=1,2 ..., in the time of k), hidden layer output computing formula is
In the formula, h (j) is j node output valve of hidden layer; ω
IjFor input layer is connected the connection weights with hidden layer; b
jBe wavelet basis function h
jShift factor; a
jBe wavelet basis function h
jContraction-expansion factor; h
jBe wavelet basis function;
Described wavelet basis function is the female wavelet basis function of Morlet, and mathematical formulae is:
Wavelet neural network output layer computing formula is
In the formula, ω
JkFor hidden layer arrives the output layer weights; H (i) is the output of i node; L is the hidden layer node number; M is the output layer nodes;
Step 3, the statistical value of History Order amount data is the prediction input table, the corresponding prediction of training network model according to the requirement arrangement of forecast model;
Initialization wavelet function contraction-expansion factor a
k, shift factor b
kAnd network connection weights omega
Ij, ω
Jk, e-learning speed η is set;
In order to eliminate the difference of the order of magnitude between each dimension data, avoid causing the neural network forecast error larger because inputoutput data order of magnitude difference is large, then all data are carried out normalized; Choose the minimax method and carry out normalized, functional form is
x
k=(x
k-x
min)/(x
max-x
min)
X wherein
MaxBe maximal value in the data sequence, x
MinBe minimum value in the data sequence;
The input parameter of wavelet neural network is { X
1, X
2..., X
k, input History Order amount burst sample is x
i(i=1,2 ..., k), all samples being divided into training sample and test sample book, training sample is used for training network, and test sample book is used for the test network precision of prediction, and prediction of output result;
With the training sample input neural network, the prediction that obtains wavelet neural network is output as y
j(j=1,2 ..., m), and the error e of computational grid output y (k) and desired output yn (k), wherein error e is:
According to error e roll-off network weights
And wavelet parameter
Make the neural network forecast value approach expectation value;
Revise wavelet neural network weights and wavelet basis function coefficient according to predicated error e:
In the formula, η is learning rate;
The above step that circulates is trained wavelet neural network, and training finishes in the allowed band until predicated error is stable at;
Step 4, the forecast model that trains according to described prediction order and prediction output equation utilize the prediction input table of described prediction order volume to calculate the prediction order volume of described order.
2. the product order forecast method with time series characteristic according to claim 1 is characterized in that, statistical value is put in order according to the requirement of prediction network be corresponding prediction input table, comprising:
The statistical value of order data obtains the History Order data of storage for the time interval (such as day or the moon) according to order;
According to the requirement of prediction network, the prediction input table is the form of the capable m row of n, and wherein, line number n and columns m are all determined by Forecasting Methodology.
3. the product order forecast method with time series characteristic according to claim 2 is characterized in that, Forecasting Methodology comprises:
Order data according to prediction has the time series characteristic, determines the time series models that described order forecasting is corresponding;
Determine the correlation parameter of this model according to described time series models;
Obtain the prediction output equation of this model according to the correlation parameter of this model.
4. the product order forecast method with time series characteristic according to claim 3 is characterized in that, the correlation parameter that described time series models are determined comprises:
Described time series models are the time series predicting model based on wavelet neural network;
When described time series models were time series predicting model based on wavelet neural network, definite correlation parameter comprised network structure, network weight and the wavelet parameter etc. of wavelet basis function, wavelet basis function.
5. any one described product order forecast method with time series characteristic according to claim 1-4, it is characterized in that, the historical data of the forecast model of described prediction product order and described prediction product order is obtained the prediction order volume of described prediction product order, comprising:
Obtain the historical real data of described product order;
Real data according to described product order is carried out normalized, and purpose is to eliminate the difference of the order of magnitude between each dimension data, obtains thus the data after the normalized;
Obtain the data that predict the outcome according to the real data of described product order and the predictive equation of described forecast model;
The data that predict the outcome are carried out renormalization process, and forecast model repeatedly moved obtain a plurality of predicting the outcome, to these averaged that predicts the outcome, obtain the final prediction order volume of described prediction product.
6. the product order forecasting device with time series characteristic is characterized in that, comprising:
Data acquisition module be used for to gather this product actual order volume of (day or month) within each time period, and its arrangement is become the prediction input table;
Data preprocessing module is used for the actual order volume of described prediction order is done normalized;
The time series modeling module is used for determining time series models corresponding to described order forecasting and prediction output equation according to described order data with time series characteristic;
The order forecasting module is used for obtaining the prediction order volume of described product according to the historical data of described prediction product and forecast model and the prediction output equation of described prediction product order.
7. the product order forecasting device with time series characteristic according to claim 6 is characterized in that, data preprocessing module comprises:
Historical data gathers submodule, is used for obtaining the History Order data of storage;
The historical data reorganization submodule is for the data entry form that the historical data of obtaining is organized into the capable m row of n according to the requirement of Forecasting Methodology.
8. the product order forecasting device with time series characteristic according to claim 6 is characterized in that, the time series modeling module comprises:
Time series models are determined submodule, are used for having the time series characteristic according to the order of described product, determine that used forecast model is the time series predicting model of wavelet neural network;
Parameter is determined submodule, is used for the correlation parameter according to the described time series models of described time series models initialization;
The error correction submodule is used for the feature request according to described prediction model based on wavelet neural network, and its network error is revised, and further determines the correlation parameter of forecast model according to the error of revising, and is more accurate to guarantee prediction output;
The prediction output sub-module is used for the Parameters in Mathematical Model according to described time series predicting model and process error correction, determines its output equation, as the prediction output equation of described time series predicting model.
9. the product order forecasting device with time series characteristic according to claim 6 is characterized in that, the order forecasting module comprises:
Predicted data is obtained submodule, for the data that predict the outcome that obtain via the predictive equation generation of previous step forecast model;
Data subsequent treatment submodule is used for that predicting the outcome of obtaining carried out renormalization and processes, and in order to improve prediction accuracy, with the circular flow of time series modeling module repeatedly, obtains predicting the outcome and ask its average repeatedly, obtains last prediction order volume.
10. the product order forecasting device with time series characteristic according to claim 9 is characterized in that, data subsequent treatment submodule comprises:
Data renormalization processing unit is used for carrying out renormalization by predicting the outcome of predictive equation output and processes;
The data mean value processing unit is used for forecast model is carried out repeatedly circular flow, and a plurality of operation result averaged to obtaining, and obtains last prediction order volume.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013102612648A CN103310286A (en) | 2013-06-25 | 2013-06-25 | Product order prediction method and device with time series characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013102612648A CN103310286A (en) | 2013-06-25 | 2013-06-25 | Product order prediction method and device with time series characteristics |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103310286A true CN103310286A (en) | 2013-09-18 |
Family
ID=49135473
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2013102612648A Pending CN103310286A (en) | 2013-06-25 | 2013-06-25 | Product order prediction method and device with time series characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103310286A (en) |
Cited By (30)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104599002A (en) * | 2015-02-13 | 2015-05-06 | 北京嘀嘀无限科技发展有限公司 | Order value predicting method and equipment |
CN104766144A (en) * | 2015-04-22 | 2015-07-08 | 携程计算机技术(上海)有限公司 | Order forecasting method and system |
CN105373840A (en) * | 2015-10-14 | 2016-03-02 | 深圳市天行家科技有限公司 | Designated-driving order predicting method and designated-driving transport capacity scheduling method |
CN105574733A (en) * | 2015-04-23 | 2016-05-11 | 田小平 | Shared inventory offline and online selling method and system |
WO2016091173A1 (en) * | 2014-12-09 | 2016-06-16 | 北京嘀嘀无限科技发展有限公司 | User maintenance system and method |
CN105847080A (en) * | 2016-03-18 | 2016-08-10 | 上海珍岛信息技术有限公司 | Method and system for predicting network traffic |
TWI581197B (en) * | 2014-09-30 | 2017-05-01 | 三菱電機股份有限公司 | Method, device and program product for determining order quantity |
CN108459997A (en) * | 2018-02-07 | 2018-08-28 | 深圳市微埃智能科技有限公司 | High skewness data value probability forecasting method based on deep learning and neural network |
CN109284866A (en) * | 2018-09-06 | 2019-01-29 | 安吉汽车物流股份有限公司 | Goods orders prediction technique and device, storage medium, terminal |
CN109376180A (en) * | 2018-09-19 | 2019-02-22 | 蜜小蜂智慧(北京)科技有限公司 | A kind of method and device constructing variable |
CN109543924A (en) * | 2018-12-20 | 2019-03-29 | 上海德启信息科技有限公司 | Goods amount prediction technique, device and computer equipment |
CN109784524A (en) * | 2017-11-13 | 2019-05-21 | 顺丰科技有限公司 | Part amount prediction technique and device, equipment and storage medium |
CN109902847A (en) * | 2017-12-11 | 2019-06-18 | 北京京东尚科信息技术有限公司 | Prediction divides the method and apparatus of library order volume |
CN110084437A (en) * | 2019-05-09 | 2019-08-02 | 上汽安吉物流股份有限公司 | Prediction technique and device, the logistics system and computer-readable medium of order |
CN110097320A (en) * | 2019-05-09 | 2019-08-06 | 上汽安吉物流股份有限公司 | Order forecast method and device, logistics system and computer-readable medium |
CN110110931A (en) * | 2019-05-09 | 2019-08-09 | 上汽安吉物流股份有限公司 | Order forecast method and device, logistics system and computer-readable medium |
CN110110932A (en) * | 2019-05-09 | 2019-08-09 | 上汽安吉物流股份有限公司 | Order forecast method and device, logistics system and computer-readable medium |
CN110110950A (en) * | 2018-02-01 | 2019-08-09 | 北京京东振世信息技术有限公司 | Generate the method, apparatus and computer readable storage medium in dispatching road area |
CN110309947A (en) * | 2019-05-09 | 2019-10-08 | 上汽安吉物流股份有限公司 | Complete vehicle logistics order forecast method and device, logistics system and computer-readable medium |
CN110766184A (en) * | 2018-07-25 | 2020-02-07 | 北京京东尚科信息技术有限公司 | Order quantity prediction method and device |
CN110837909A (en) * | 2018-08-17 | 2020-02-25 | 北京京东尚科信息技术有限公司 | Method and device for predicting order quantity |
CN110837907A (en) * | 2018-08-17 | 2020-02-25 | 天津京东深拓机器人科技有限公司 | Method and device for predicting wave order quantity |
CN111242698A (en) * | 2020-01-20 | 2020-06-05 | 金华航大北斗应用技术有限公司 | Method and system for neural network model of commodity demand prediction |
CN109407654B (en) * | 2018-12-20 | 2020-08-04 | 浙江大学 | Industrial data nonlinear causal analysis method based on sparse deep neural network |
WO2020186380A1 (en) * | 2019-03-15 | 2020-09-24 | State Street Corporation | Techniques to forecast future orders using deep learning |
CN112990421A (en) * | 2019-12-02 | 2021-06-18 | 杭州海康威视数字技术股份有限公司 | Method, device and storage medium for optimizing operation process of deep learning network |
CN113535990A (en) * | 2020-11-10 | 2021-10-22 | 腾讯科技(深圳)有限公司 | Method, device, storage medium and electronic equipment for determining multimedia content |
CN113554483A (en) * | 2021-07-19 | 2021-10-26 | 浙江工企信息技术股份有限公司 | Order full life cycle management method and system |
TWI771615B (en) * | 2019-09-17 | 2022-07-21 | 英業達股份有限公司 | Product order prediction method |
CN116452243A (en) * | 2023-05-11 | 2023-07-18 | 长沙麦可思信息科技有限公司 | Enterprise order prediction method, system and medium based on big data |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2000010455A1 (en) * | 1998-08-24 | 2000-03-02 | Emory University | Method and apparatus for predicting the onset of seizures based on features derived from signals indicative of brain activity |
US20020054694A1 (en) * | 1999-03-26 | 2002-05-09 | George J. Vachtsevanos | Method and apparatus for analyzing an image to direct and identify patterns |
CN102056183A (en) * | 2010-12-10 | 2011-05-11 | 北京交通大学 | Network flow prediction method and device based on cognitive network |
CN102323970A (en) * | 2011-05-19 | 2012-01-18 | 河海大学 | Hydrological time series prediction method based on multiple-factor wavelet neural network model |
-
2013
- 2013-06-25 CN CN2013102612648A patent/CN103310286A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2000010455A1 (en) * | 1998-08-24 | 2000-03-02 | Emory University | Method and apparatus for predicting the onset of seizures based on features derived from signals indicative of brain activity |
US20020054694A1 (en) * | 1999-03-26 | 2002-05-09 | George J. Vachtsevanos | Method and apparatus for analyzing an image to direct and identify patterns |
CN102056183A (en) * | 2010-12-10 | 2011-05-11 | 北京交通大学 | Network flow prediction method and device based on cognitive network |
CN102323970A (en) * | 2011-05-19 | 2012-01-18 | 河海大学 | Hydrological time series prediction method based on multiple-factor wavelet neural network model |
Non-Patent Citations (3)
Title |
---|
ESEN H等: ""Artificial neural network and wavelet neural network approaches for modelling of a solar air heater"", 《EXPERT SYSTEMS WITH APPLICATIONS》 * |
余开华: ""小波神经网络模型在河道流量水位预测中的应用"", 《水资源与水工程学报》 * |
张清华: ""小波神经网络参数优化及其应用"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (40)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI581197B (en) * | 2014-09-30 | 2017-05-01 | 三菱電機股份有限公司 | Method, device and program product for determining order quantity |
WO2016091173A1 (en) * | 2014-12-09 | 2016-06-16 | 北京嘀嘀无限科技发展有限公司 | User maintenance system and method |
GB2547395A (en) * | 2014-12-09 | 2017-08-16 | Beijing Didi Infinity Tech And Dev Co Ltd | User maintenance system and method |
CN104599002A (en) * | 2015-02-13 | 2015-05-06 | 北京嘀嘀无限科技发展有限公司 | Order value predicting method and equipment |
CN104766144A (en) * | 2015-04-22 | 2015-07-08 | 携程计算机技术(上海)有限公司 | Order forecasting method and system |
CN105574733A (en) * | 2015-04-23 | 2016-05-11 | 田小平 | Shared inventory offline and online selling method and system |
CN105574733B (en) * | 2015-04-23 | 2021-02-09 | 田小平 | Sharing inventory offline and online sales method and system |
WO2017063356A1 (en) * | 2015-10-14 | 2017-04-20 | 深圳市天行家科技有限公司 | Designated-driving order predicting method and designated-driving transport capacity scheduling method |
CN105373840B (en) * | 2015-10-14 | 2018-12-11 | 深圳市天行家科技有限公司 | In generation, drives order forecast method and generation drives Transport capacity dispatching method |
CN105373840A (en) * | 2015-10-14 | 2016-03-02 | 深圳市天行家科技有限公司 | Designated-driving order predicting method and designated-driving transport capacity scheduling method |
CN105847080A (en) * | 2016-03-18 | 2016-08-10 | 上海珍岛信息技术有限公司 | Method and system for predicting network traffic |
CN109784524B (en) * | 2017-11-13 | 2023-06-16 | 顺丰科技有限公司 | Method, device, equipment and storage medium for predicting part quantity |
CN109784524A (en) * | 2017-11-13 | 2019-05-21 | 顺丰科技有限公司 | Part amount prediction technique and device, equipment and storage medium |
CN109902847A (en) * | 2017-12-11 | 2019-06-18 | 北京京东尚科信息技术有限公司 | Prediction divides the method and apparatus of library order volume |
CN109902847B (en) * | 2017-12-11 | 2022-04-12 | 北京京东尚科信息技术有限公司 | Method and device for predicting amount of orders in branch warehouse |
CN110110950A (en) * | 2018-02-01 | 2019-08-09 | 北京京东振世信息技术有限公司 | Generate the method, apparatus and computer readable storage medium in dispatching road area |
CN108459997A (en) * | 2018-02-07 | 2018-08-28 | 深圳市微埃智能科技有限公司 | High skewness data value probability forecasting method based on deep learning and neural network |
CN110766184A (en) * | 2018-07-25 | 2020-02-07 | 北京京东尚科信息技术有限公司 | Order quantity prediction method and device |
CN110837907A (en) * | 2018-08-17 | 2020-02-25 | 天津京东深拓机器人科技有限公司 | Method and device for predicting wave order quantity |
CN110837909A (en) * | 2018-08-17 | 2020-02-25 | 北京京东尚科信息技术有限公司 | Method and device for predicting order quantity |
CN109284866B (en) * | 2018-09-06 | 2021-01-29 | 安吉汽车物流股份有限公司 | Commodity order prediction method and device, storage medium and terminal |
CN109284866A (en) * | 2018-09-06 | 2019-01-29 | 安吉汽车物流股份有限公司 | Goods orders prediction technique and device, storage medium, terminal |
CN109376180A (en) * | 2018-09-19 | 2019-02-22 | 蜜小蜂智慧(北京)科技有限公司 | A kind of method and device constructing variable |
CN109407654B (en) * | 2018-12-20 | 2020-08-04 | 浙江大学 | Industrial data nonlinear causal analysis method based on sparse deep neural network |
CN109543924A (en) * | 2018-12-20 | 2019-03-29 | 上海德启信息科技有限公司 | Goods amount prediction technique, device and computer equipment |
WO2020186380A1 (en) * | 2019-03-15 | 2020-09-24 | State Street Corporation | Techniques to forecast future orders using deep learning |
CN110110931A (en) * | 2019-05-09 | 2019-08-09 | 上汽安吉物流股份有限公司 | Order forecast method and device, logistics system and computer-readable medium |
CN110097320A (en) * | 2019-05-09 | 2019-08-06 | 上汽安吉物流股份有限公司 | Order forecast method and device, logistics system and computer-readable medium |
CN110084437A (en) * | 2019-05-09 | 2019-08-02 | 上汽安吉物流股份有限公司 | Prediction technique and device, the logistics system and computer-readable medium of order |
CN110309947A (en) * | 2019-05-09 | 2019-10-08 | 上汽安吉物流股份有限公司 | Complete vehicle logistics order forecast method and device, logistics system and computer-readable medium |
CN110110932A (en) * | 2019-05-09 | 2019-08-09 | 上汽安吉物流股份有限公司 | Order forecast method and device, logistics system and computer-readable medium |
TWI771615B (en) * | 2019-09-17 | 2022-07-21 | 英業達股份有限公司 | Product order prediction method |
CN112990421A (en) * | 2019-12-02 | 2021-06-18 | 杭州海康威视数字技术股份有限公司 | Method, device and storage medium for optimizing operation process of deep learning network |
CN112990421B (en) * | 2019-12-02 | 2023-09-05 | 杭州海康威视数字技术股份有限公司 | Method, device and storage medium for optimizing operation process of deep learning network |
CN111242698A (en) * | 2020-01-20 | 2020-06-05 | 金华航大北斗应用技术有限公司 | Method and system for neural network model of commodity demand prediction |
CN113535990A (en) * | 2020-11-10 | 2021-10-22 | 腾讯科技(深圳)有限公司 | Method, device, storage medium and electronic equipment for determining multimedia content |
CN113535990B (en) * | 2020-11-10 | 2023-12-15 | 腾讯科技(深圳)有限公司 | Method, device, storage medium and electronic equipment for determining multimedia content |
CN113554483A (en) * | 2021-07-19 | 2021-10-26 | 浙江工企信息技术股份有限公司 | Order full life cycle management method and system |
CN116452243A (en) * | 2023-05-11 | 2023-07-18 | 长沙麦可思信息科技有限公司 | Enterprise order prediction method, system and medium based on big data |
CN116452243B (en) * | 2023-05-11 | 2023-08-29 | 长沙麦可思信息科技有限公司 | Enterprise order prediction method, system and medium based on big data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103310286A (en) | Product order prediction method and device with time series characteristics | |
Wang et al. | A seasonal GM (1, 1) model for forecasting the electricity consumption of the primary economic sectors | |
Fan et al. | Short term load forecasting based on phase space reconstruction algorithm and bi-square kernel regression model | |
Andryushkevich et al. | Composition and application of power system digital twins based on ontological modeling | |
CN103577694B (en) | Aquaculture water quality short-time combination forecast method on basis of multi-scale analysis | |
CN103745273B (en) | Semiconductor fabrication process multi-performance prediction method | |
CN107563705A (en) | Household electrical appliances product safety stock and the system and method ordered goods again are analyzed using big data | |
Baidakov et al. | Methodical bases for developing predictive scenarios of agribusiness | |
Niu et al. | Research on short-term power load time series forecasting model based on BP neural network | |
CN107784397A (en) | A kind of power network material requirements forecasting system and its Forecasting Methodology | |
CN108596242A (en) | Power grid meteorology load forecasting method based on wavelet neural network and support vector machines | |
CN104537433A (en) | Sold electricity quantity prediction method based on inventory capacities and business expansion characteristics | |
CN109086930A (en) | A kind of user power utilization behavior analysis method based on electric power K line chart and depth network | |
Dyvak et al. | Modeling of dynamics of the company’s share in the solid fuel market | |
CN111738505A (en) | Bank branch workload prediction method and device, electronic equipment and storage medium | |
Shi | Application of back propagation (BP) neural network in marine regional economic forecast | |
CN114091776A (en) | K-means-based multi-branch AGCNN short-term power load prediction method | |
CN109993354A (en) | A method of it is predicted for energy consumption | |
CN104182854A (en) | Mixed energy consumption measuring method for enterprise energy management system | |
CN112053181A (en) | Agricultural product price prediction method and system | |
CN104504466B (en) | Consider the wind electric field power prediction method of atmospheric perturbation effect | |
CN111126827A (en) | Input-output accounting model construction method based on BP artificial neural network | |
CN107730025A (en) | A kind of power supply enterprise's Cash flow prospects model changed based on electricity | |
Takács et al. | Fuzzy cognitive map for student evaluation model | |
Wang et al. | Demand forecasting models of tourism based on ELM |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130918 |