CN110232593A - The total retail sales of consumer goods add up the prediction technique and system of amplification - Google Patents
The total retail sales of consumer goods add up the prediction technique and system of amplification Download PDFInfo
- Publication number
- CN110232593A CN110232593A CN201910492416.2A CN201910492416A CN110232593A CN 110232593 A CN110232593 A CN 110232593A CN 201910492416 A CN201910492416 A CN 201910492416A CN 110232593 A CN110232593 A CN 110232593A
- Authority
- CN
- China
- Prior art keywords
- prediction
- amplification
- index
- consumer goods
- retail sales
- 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
Classifications
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
Abstract
The present invention provides the prediction technique and system of a kind of accumulative amplification of the total retail sales of consumer goods, is related to data prediction field.The following steps are included: obtaining total retail sales of consumer goods related data, history of forming data;Prediction index is determined based on the historical data;Add up amplification forecast system based on the prediction index building total retail sales of consumer goods;Vector Autoression Models and shot and long term Memory Neural Networks model based on pre-training add up amplification forecast system to the total retail sales of consumer goods respectively and handle, and obtain the first prediction result and the second prediction result;Initial predicted result is obtained based on first prediction result and second prediction result;The initial predicted result is modified based on emergency event, obtains the prediction result that the total retail sales of consumer goods add up amplification.The present invention can be with the accumulative amplification of the Accurate Prediction total retail sales of consumer goods.
Description
Technical field
The present invention relates to data to predict field, and in particular to a kind of total retail sales of consumer goods add up the prediction of amplification
Method and system.
Background technique
With being skyrocketed through for China's economic, it is further important to consume occupied effect.The total volume of retail sales of social consumer goods
The main conditions of a regional level of consumption are able to reflect, and the accumulation amplification of the total volume of retail sales can embody the change of the level of consumption
Change trend.Therefore, the accumulative amplification that reasonable prediction goes out the total retail sales of consumer goods has great importance.
In the prior art, the accumulative amplification that the total retail sales of consumer goods are generally predicted with ARIMA model, by searching
Collect the historical data of the total retail sales of consumer goods, and carries out the fitting of the Multiplicative Seasonality Model in ARIMA model, analysis to it
The change conditions of the accumulative amplification of the total retail sales of consumer goods out.
However, the method for the prior art only passes through the income of residents, size of population etc. prediction, not in view of influencing
Influence of the industry of regional economy to the accumulative amplification of the total retail sales of consumer goods.Therefore the analysis level of the prior art is not
It is enough comprehensively, can not Accurate Prediction go out the accumulative amplification of the total retail sales of consumer goods.
Summary of the invention
(1) the technical issues of solving
In view of the deficiencies of the prior art, the present invention provides the predictions that a kind of total retail sales of consumer goods add up amplification
Method and system, solve the prior art can not Accurate Prediction go out the total retail sales of consumer goods add up amplification the technical issues of.
(2) technical solution
In order to achieve the above object, the present invention is achieved by the following technical programs:
The present invention solves a kind of prediction side of the accumulative amplification of the total retail sales of consumer goods provided by its technical problem
Method, the prediction technique are executed by computer, comprising the following steps:
Obtain total retail sales of consumer goods related data, history of forming data;
Determine that the total retail sales of consumer goods add up the prediction index of amplification based on the historical data, the prediction refers to
Mark includes: industry index;
Add up amplification forecast system based on the prediction index building total retail sales of consumer goods;
Based on the Vector Autoression Models of pre-training to the total retail sales of consumer goods add up amplification forecast system into
Row processing, obtains the first prediction result;Shot and long term Memory Neural Networks model based on pre-training is to the social consumer goods zero
It sells the accumulative amplification forecast system of total value to be handled, obtains the second prediction result;
Initial predicted result is obtained based on first prediction result and second prediction result;
The initial predicted result is modified based on emergency event, obtains the accumulative increasing of the total retail sales of consumer goods
The prediction result of width.
Preferably, it includes: the total retail sales of consumer goods that the total retail sales of consumer goods, which add up amplification forecast system,
The accumulative monthly forecast system of amplification and the total retail sales of consumer goods add up amplification Seasonal prediction system.
Preferably, the prediction index that the total retail sales of consumer goods add up the monthly forecast system of amplification includes: price
With level of consumption index, income level index, industry index, finance index, macro-performance indicator, industry index, opening
Spend index, taxation target and other indexs.
Preferably, the prediction index that the total retail sales of consumer goods add up amplification Seasonal prediction system includes: price
With level of consumption index, income level index, macro-performance indicator, industry index, industry index, consumer confidence index index and its
His index.
Preferably, the industry index includes: national real estate fixed investment, national merchandise building face
Growth is counted in accumulation, province's automobile adds up to increase than same period last year, saves the accumulative amplification of wholesale and retail business electricity consumption.
Preferably, the acquisition methods of the initial predicted result are as follows:
The weight of default first prediction result and the second prediction result, in conjunction with weight to first prediction result and institute
It states the second prediction result to be handled, obtains the initial predicted result.
Preferably, the emergency event includes: political event, international environment variation and social environment variation.
The present invention solves a kind of prediction system of the accumulative amplification of the total retail sales of consumer goods provided by its technical problem
System, the system comprises computer, the computer includes:
At least one storage unit;
At least one processing unit;
Wherein, at least one instruction is stored at least one described storage unit, at least one instruction is by described
At least one processing unit is loaded and is executed to perform the steps of
Obtain total retail sales of consumer goods related data, history of forming data;
Determine that the total retail sales of consumer goods add up the prediction index of amplification based on the historical data, the prediction refers to
Mark includes: industry index;
Add up amplification forecast system based on the prediction index building total retail sales of consumer goods;
Based on the Vector Autoression Models of pre-training to the total retail sales of consumer goods add up amplification forecast system into
Row processing, obtains the first prediction result;Shot and long term Memory Neural Networks based on pre-training are sold the social consumer goods total
Volume adds up amplification forecast system and is handled, and obtains the second prediction result;
Initial predicted result is obtained based on first prediction result and second prediction result;
The initial predicted result is modified based on emergency event, obtains the accumulative increasing of the total retail sales of consumer goods
The prediction result of width.
Preferably, the industry index includes: national real estate fixed investment, national merchandise building face
Growth is counted in accumulation, province's automobile adds up to increase than same period last year, saves the accumulative amplification of wholesale and retail business electricity consumption.
Preferably, the acquisition methods of the initial predicted result are as follows:
The weight of default first prediction result and the second prediction result, in conjunction with weight to first prediction result and institute
It states the second prediction result to be handled, obtains the initial predicted result.
(3) beneficial effect
The present invention provides prediction techniques and system that a kind of total retail sales of consumer goods add up amplification.With existing skill
Art is compared, have it is following the utility model has the advantages that
The present invention passes through the prediction index for determining that the total retail sales of consumer goods add up amplification, and is based on prediction index structure
It builds out the total retail sales of consumer goods and adds up amplification forecast system, wherein prediction index includes industry index;Based on pre-training
Vector Autoression Models and shot and long term Memory Neural Networks model respectively to the total retail sales of consumer goods add up amplification prediction
System is handled, and the first prediction result and the second prediction result are obtained;Based on the first prediction result and the second prediction result
Obtain initial predicted result;Initial predicted result is modified based on emergency event, it is tired to obtain the total retail sales of consumer goods
Count the prediction result of amplification.The present invention considers influence of the industry index factor to regional economy, comprehensively analyzes area
Economic situation of change, can go out the accumulative amplification of the total retail sales of consumer goods with Accurate Prediction, meanwhile, by vector auto regression
The accumulative amplification that the prediction total retail sales of consumer goods are combined with shot and long term Memory Neural Networks, so that result is more quasi-
Really.Support can be provided for departments of government decision based on prediction result, enterprise provides information for consumer goods production and consumption
Service.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will to embodiment or
Attached drawing needed to be used in the description of the prior art is briefly described, it should be apparent that, the accompanying drawings in the following description is only
Some embodiments of the present invention, for those of ordinary skill in the art, without creative efforts, also
Other drawings may be obtained according to these drawings without any creative labor.
Fig. 1 is the overall flow figure for the prediction technique that the total retail sales of consumer goods add up amplification in the embodiment of the present invention;
Fig. 2 is the functional image of activation primitive in the embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, to the technology in the embodiment of the present invention
Scheme is clearly and completely described, it is clear that and described embodiments are some of the embodiments of the present invention, rather than whole
Embodiment.Based on the embodiments of the present invention, those of ordinary skill in the art are obtained without creative efforts
The every other embodiment obtained, shall fall within the protection scope of the present invention.
The prediction technique and system that the embodiment of the present application adds up amplification by providing a kind of total retail sales of consumer goods,
Solve the problems, such as the prior art can not Accurate Prediction go out social consumer goods the total volume of retail sales add up amplification, realize social consumption
The product total volume of retail sales adds up the Accurate Prediction of amplification.
Technical solution in the embodiment of the present application is in order to solve the above technical problems, general thought is as follows:
The embodiment of the present invention passes through the prediction index for determining that the total retail sales of consumer goods add up amplification, and based on prediction
Index constructs the total retail sales of consumer goods and adds up amplification forecast system, wherein prediction index includes industry index;It is based on
The Vector Autoression Models and shot and long term Memory Neural Networks model of pre-training are accumulative to the total retail sales of consumer goods respectively to be increased
Width forecast system is handled, and the first prediction result and the second prediction result are obtained;It is predicted based on the first prediction result and second
As a result initial predicted result is obtained;Initial predicted result is modified based on emergency event, obtains social consumer goods retail
Total value adds up the prediction result of amplification.The embodiment of the present invention considers influence of the industry index factor to regional economy, comprehensively
Ground analyzes the situation of change of regional economy, and the accumulative amplification of the total retail sales of consumer goods can be gone out with Accurate Prediction, meanwhile,
The accumulative amplification that vector auto regression and shot and long term Memory Neural Networks are combined to the prediction total retail sales of consumer goods, makes
It is more accurate to obtain result.Support can be provided for departments of government decision based on prediction result, be consumer goods production and consumption
Enterprise provides information service.
In order to better understand the above technical scheme, right in conjunction with appended figures and specific embodiments
Above-mentioned technical proposal is described in detail.
The embodiment of the invention provides the prediction techniques that a kind of total retail sales of consumer goods add up amplification, such as Fig. 1 institute
Show.Above-mentioned prediction technique is executed by computer, comprising the following steps:
S1, total retail sales of consumer goods related data, history of forming data are obtained;
Determine that the total retail sales of consumer goods add up the prediction index of amplification based on above-mentioned historical data, above-mentioned prediction refers to
Mark includes: industry index;
S2, add up amplification forecast system based on the above-mentioned prediction index building total retail sales of consumer goods;
S3, amplification predictor is added up to the above-mentioned total retail sales of consumer goods based on the Vector Autoression Models of pre-training
System is handled, and the first prediction result is obtained;Shot and long term Memory Neural Networks model based on pre-training is to above-mentioned social consumption
The product total volume of retail sales adds up amplification forecast system and is handled, and obtains the second prediction result;
S4, initial predicted result is obtained based on above-mentioned first prediction result and above-mentioned second prediction result;
S5, above-mentioned initial predicted result is modified based on emergency event, it is accumulative obtains the total retail sales of consumer goods
The prediction result of amplification.
The embodiment of the present invention passes through the prediction index for determining that the total retail sales of consumer goods add up amplification, and based on prediction
Index constructs the total retail sales of consumer goods and adds up amplification forecast system, wherein prediction index includes industry index;It is based on
The Vector Autoression Models and shot and long term Memory Neural Networks model of pre-training are accumulative to the total retail sales of consumer goods respectively to be increased
Width forecast system is handled, and the first prediction result and the second prediction result are obtained;It is predicted based on the first prediction result and second
As a result initial predicted result is obtained;Initial predicted result is modified based on emergency event, obtains social consumer goods retail
Total value adds up the prediction result of amplification.The embodiment of the present invention considers influence of the industry index factor to regional economy, comprehensively
Ground analyzes the situation of change of regional economy, and the accumulative amplification of the total retail sales of consumer goods can be gone out with Accurate Prediction, meanwhile,
The accumulative amplification that vector auto regression and shot and long term Memory Neural Networks are combined to the prediction total retail sales of consumer goods, makes
It is more accurate to obtain result.Support can be provided for departments of government decision based on prediction result, be consumer goods production and consumption
Enterprise provides information service.
Each step is described in detail below.
In step sl, total retail sales of consumer goods related data, history of forming data are obtained;Based on above-mentioned history
Data determine that the total retail sales of consumer goods add up the prediction index of amplification.
Specifically, the embodiment of the present invention obtains the dependency number for influencing the total retail sales of consumer goods based on internet platform
According to history of forming data.
The prediction index that the total retail sales of consumer goods add up amplification is extracted from historical data.Specifically, of the invention
Embodiment predicts the accumulative amplification of the total retail sales of consumer goods in terms of monthly and season two.
Specifically, the total retail sales of consumer goods, which add up prediction index of the amplification on monthly, shares 9, it is respectively as follows:
Price and level of consumption index, income level index, industry index, finance index, macro-performance indicator,
Industry index, openness index, taxation target and other indexs.
Wherein, price and level of consumption index include following two-level index:
It saves Consumer Prices index CPI, save retail sales index, national enterprise commodity price combined index, country
Enterprise's commodity price combined index-agricultural product, national enterprise commodity price combined index-mineral products, national enterprise commodity price always refer to
Number-kerosene electricity, national consumer satisfaction index, the national index of consumer confidence, saves expenditure approach at national consumer anticipation index
Final consumption.
Income level index includes following two-level index:
Province's urban residents' disposable income per capita, province's town dweller's per capita consumption expenditure, national town dweller can prop up per capita
It is rural per-capita with income, national town dweller's per capita consumption expenditure, the rural per-capita disposable income of country, country
The consumption expenditure.
Industry index includes following two-level index:
The national secondary industry sum of investments in fixed assets used, saves second at the national tertiary industry sum of investments in fixed assets used
Industrial investment completes volume, saves secondary industry electricity consumption, save tertiary industry finished value of investment, save tertiary industry electricity consumption.
Finance index includes following two-level index:
State revenue, state financial spending (repaying principal without debt), national currency and quasi-money (M2) supply,
National currency (M1) supply increases by a year-on-year basis.
Macro-performance indicator includes following two-level index:
National special drawing rights unit converts into RMB, national dollar converts into RMB (end of term number), national dollar folding
It closes RMB (average), the accumulative growth of national total import and export value, save inlet and outlet amplification.
Industry index includes following two-level index:
National real estate fixed investment, the accumulative growth of national Marketable Housing Area Sold, province's automobile are accumulative than upper
The same period in year increases, saves the accumulative amplification of wholesale and retail business electricity consumption.
Openness index includes following two-level index:
It saves foreign trader and Chinese Hong Kong, Macau and Taiwan businessman enterprise rate of production and marketing, save foreign trader and Chinese Hong Kong, Macau and Taiwan businessman's loss of enterprise enterprise number.
Taxation target includes following two-level index:
Save foreign trader and Chinese Hong Kong, Macau and Taiwan businessman enterprise value added tax payable, national various kinds of taxes.
Other indexs include following two-level index:
The national sum of investments in fixed assets used is accumulative to be increased, the accumulative amplification of loss-making enterprise, large enterprise, province number, saves population certainly
Right growth rate.
Specifically, the total retail sales of consumer goods, which add up prediction index of the amplification on season, shares 7, it is respectively as follows:
Price and level of consumption index, income level index, macro-performance indicator, industry index, industry index, boom
Index index and other indexs.
Wherein, price and level of consumption index include following two-level index:
Consumer Prices index CPI, country's Consumer Prices index-city, national Consumer Prices is saved to refer to
Number-rural area saves retail sales index, national consumer anticipation index, national consumer satisfaction index, national consumer
Confidence index saves expenditure approach final consumption.
Income level index includes following two-level index:
It saves the accumulative growth of urban residents' disposable income per capita, save town dweller's per capita consumption expenditure accumulative amplification, country
Urban residents' disposable income per capita adds up amplification, national town dweller's per capita consumption expenditure adds up amplification, national rural resident
Per capita disposable income adds up amplification, the rural per-capita consumption expenditure of country adds up amplification, saves staff average salary.
Macro-performance indicator includes following two-level index:
It is tired that national GDP adds up amplification, province regional GDP country DP, national gross domestic product index number
Evaluation.
Industry index includes following two-level index:
National real estate fixed investment, the accumulative growth of national Marketable Housing Area Sold, province's automobile are accumulative than upper
The same period in year increases, saves the accumulative amplification of wholesale and retail business electricity consumption.
Industry index includes following two-level index:
It saves secondary industry value, save tertiary industry value, national value-added of the tertiary industry index aggregate-value, national secondary industry
Increase value index number aggregate-value, national GDP contribution rate aggregate-value, national tertiary industry contribution rate aggregate-value, country
Secondary industry contribution rate aggregate-value.
Consumer confidence index index includes following two-level index:
The current income of country experiences index, the following income confidence index of country, the following price forward index of country, country
Entrepreneur confidence exponent, national business condition index, man, state bank confidence index, state bank's industry consumer confidence index.
Other indexs include following two-level index:
It saves total output of building industry, national fixed assets devaluation preparation aggregate-value, national various kinds of taxes (hundred million yuan), save
The natural growth rate of population.
In step s 2, add up amplification forecast system based on the prediction index building total retail sales of consumer goods.
Specifically, some prediction index only have season since some prediction index only have monthly data without season data
Degree is according to without monthly data.Therefore the embodiment of the present invention is in the specific implementation, while establishing of both monthly and season
The total retail sales of consumer goods add up amplification forecast system.That is: the total retail sales of consumer goods add up the monthly forecast system of amplification
Add up amplification Seasonal prediction system with the total retail sales of consumer goods.
Add up the monthly forecast system of amplification based on the monthly prediction index building total retail sales of consumer goods, such as 1 institute of table
Show:
Table 1
Add up amplification Seasonal prediction system based on the Seasonal prediction index building total retail sales of consumer goods, such as 2 institute of table
Show:
Table 2
In step s3, the Vector Autoression Models based on pre-training (VAR) are tired to the above-mentioned total retail sales of consumer goods
Meter amplification forecast system is handled, and the first prediction result is obtained;Shot and long term Memory Neural Networks model based on pre-training
(LSTM) add up amplification forecast system to the above-mentioned total retail sales of consumer goods to handle, obtain the second prediction result.
It is corresponding that the prediction index in amplification forecast system is added up to the above-mentioned total retail sales of consumer goods based on two kinds of models
Data be respectively processed calculating, obtain two kinds of prediction results, specific processing method is as follows:
It should be noted that Vector Autoression Models are a kind of common econometric models, it is the prior art.
Specifically, will save unit Retail sales of consumer goods above norm in the building process of monthly model and add up amplification,
Save Consumer Prices index CPI, save inlet and outlet amplification, save foreign trader and Chinese Hong Kong, Macau and Taiwan businessman's enterprise's value added tax payable be used as to
Measure the endogenous variable of autoregression model;National currency (M1) supply is increased by a year-on-year basis simultaneously, national dollar converts into RMB
(average), exogenous variable of the state revenue as Vector Autoression Models, constructs the simultaneous equations of vector auto regression
Group, and predicted.
Since January does not have data in monthly data, so according to routine, selection 11 phases of lag are predicted, obtained connection
In vertical equation group, it can be predicted using following equation:
X1=C (1,1) * X1 (- 1)+C (1,2) * X1 (- 2)+C (1,3) * X1 (- 3)+C (1,4) * X1 (- 4)+C (1,5) *
X1(-5)+C(1,6)*X1(-6)+C(1,7)*X1(-7)+C(1,8)*X1(-8)+ C(1,9)*X1(-9)+C(1,10)*X1(-
10)+C(1,11)*X1(-11)+C(1,12)*X2(-1)+ C(1,13)*X2(-2)+C(1,14)*X2(-3)+C(1,15)*X2
(-4)+C(1,16)*X2(-5)+ C(1,17)*X2(-6)+C(1,18)*X2(-7)+C(1,19)*X2(-8)+C(1,20)*X2
(-9)+ C(1,21)*X2(-10)+C(1,22)*X2(-11)+C(1,23)*X3(-1)+C(1,24)*X3(-2)+ C(1,25)*
X3(-3)+C(1,26)*X3(-4)+C(1,27)*X3(-5)+C(1,28)*X3(-6)+ C(1,29)*X3(-7)+C(1,30)*
X3(-8)+C(1,31)*X3(-9)+C(1,32)*X3(-10)+ C(1,33)*X3(-11)+C(1,34)*X4(-1)+C(1,
35)*X4(-2)+C(1,36)*X4(-3)+ C(1,37)*X4(-4)+C(1,38)*X4(-5)+C(1,39)*X4(-6)+C(1,
40)*X4(-7)+ C(1,41)*X4(-8)+C(1,42)*X4(-9)+C(1,43)*X4(-10)+C(1,44)*X4(-11)+ C
(1,45)+C(1,46)*X5+C(1,47)*X6+C(1,48)*X7。
Wherein:
X1 indicates that saving unit Retail sales of consumer goods above norm adds up amplification;
X2 indicates to save Consumer Prices index CPI;
X3 indicates to save foreign trader and Chinese Hong Kong, Macau and Taiwan businessman enterprise value added tax payable;
X4 indicates to save inlet and outlet amplification;
X5 indicates that national currency (M1) supply increases by a year-on-year basis;
X6 indicates state revenue;
X7 indicates that national dollar converts into RMB (average);
Symbol such as C (1,1) before each variable indicates coefficient to be solved;
Symbol after each variable indicates the lag period, as X1 (- 1) indicates the X1 variable of one phase of lag.
In season model construction process, amount of social consumption product retail, province is added up into amplification, saving town dweller can prop up per capita
With the accumulative growth of income, the rural per-capita disposable income of country adds up amplification as endogenous variable, by national domestic production
Total value adds up amplification, national value-added of the tertiary industry index aggregate-value, and state bank's industry consumer confidence index is constructed as exogenous variable
Model.
Thus model will be included in the variable that the total retail sales of consumer goods are affected, on the one hand can preferably into
Row prediction, another convenience are also convenient for analyzing these principal elements.The lag period was selected as 4 phases in season model, obtained
Simultaneous Equations in, predicted using following equation:
Y1=D (1,1) * Y1 (- 1)+D (1,2) * Y1 (- 2)+D (1,3) * Y1 (- 3)+D (1,4) * Y1 (- 4)+D (1,5) *
Y2(-1)+D(1,6)*Y2(-2)+D(1,7)*Y2(-3)+D(1,8)* Y2(-4)+D(1,9)*Y3(-1)+D(1,10)*Y3(-
2)+D(1,11)*Y3(-3)+D (1,12)*Y3(-4)+D(1,13)+D(1,14)*Y4+D(1,15)*Y5+D(1,16)*Y6。
Wherein:
Y1 indicates that amount of social consumption product retail, province adds up amplification;
Y2 indicates that the rural per-capita disposable income of country adds up amplification;
Y3 indicates to save the accumulative growth of urban residents' disposable income per capita;
Y4 indicates that national GDP adds up amplification;
Y5 indicates national value-added of the tertiary industry index aggregate-value;
Y6 indicates state bank's industry consumer confidence index.
Symbol such as D (1,1) before each variable indicates coefficient to be solved;
Symbol after each variable indicates the lag period, as Y1 (- 1) indicates the Y1 variable of one phase of lag.
It should be noted that neural network model is a kind of imitation animal nerve network behavior feature, distribution is carried out simultaneously
The algorithm mathematics model of row information processing is the prior art.
Specifically, using python language, being taken under Linux-CentOS operating system during neural net model establishing
The deep learning frame based on Keras is built out, the prediction model building based on LSTM is then carried out.
Model construction groundwork is to adjust ginseng, and the quality of parameter adjustment directly affects prediction effect and the prediction of model
Precision, in this experiment, the parameter that we are related to adjustment mainly has:
(1) activation primitive: being a kind of special function that input is mapped to specific output space.What this was related to swashs
Function living has sigmoid, tanh etc..The functional image property of different activation primitives is as shown in Figure 2.
(2) loss function: loss function is the index for measuring precision of prediction.Loss function value is smaller, represents prediction result
Be really that error between result is smaller, the effect of model is also better.Model mainly uses Mean_squared_error
(MSE) with Mean_absolute_error (MAE) two kinds of loss functions;
(3) neural net layer structure: the model of this research and establishment, in addition to input layer, output layer, hidden layer has three layers, packet
Include one LSTM layers and two full articulamentums (Dense layers).It is generally believed that LSTM layers are no more than 3 layers, then it is more when training can ratio
It is difficult to converge, while next layer is added dimensionality reduction of common (Dense layers) of neural net layer for LSTM output result.
(4) node rejection rate: over-fitting occurs in order to prevent, and the rejection rate that node is arranged is 0.05 or 0.1;
(5) optimizer function: representing the iteration optimization mode of parameter, selects RMSprop, Adam, SGD, additionally wants
The floating number that learning efficiency is 0-1 is set;And epsilon number (floating number more than or equal to 0);
(6) it the number of iterations: is set as 1000 times.
In step s 4, initial predicted result is obtained based on above-mentioned first prediction result and above-mentioned second prediction result.
Specifically, weight can be assigned to the first prediction result and the second prediction result in advance, according to two prediction results and
Weight calculation shared by it obtains initial predicted result.
In embodiments of the present invention, the first prediction result is identical with the weight of the second prediction result, therefore two are predicted
Results are averaged obtains initial predicted result.
In step s 5, above-mentioned initial predicted result is modified based on emergency event, obtains social consumer goods retail
Total value adds up the prediction result of amplification.
Specifically, emergency event includes following several situations.
Political class event, such as: national monetary policy, the variation of financial policy.The variation of social environment, such as: it is natural
Disaster, festivals or holidays etc..The variation of international environment, such as: buterfly effect.And more close state is contacted with Chinese trade
Family economically generates the case where large variation.
Based on emergency event of that month or when season generation, prediction result is modified, obtains social consumer goods retail
Total value adds up the prediction result of amplification.
For example, the Spring Festival in 2018, then 2 months 2018 total retail sales of consumer goods were tired 2 months posterior segments
Counting speedup can be slightly lower, and the shut-down because long holidays in the Spring Festival usually close a business and produced along with sale causes economic data this period to go out
Now certain deviation, and then can be higher than 2 months in subsequent March, because March is not influenced by Spring Festival Effect, so correcting
When, to consider this point, the month influenced by Spring Festival Effect is modified.
Specifically, can be to be adjusted based on expert's historical experience.For example, on the 2 months data influenced by Spring Festival Effect
It slightly turns down, turns down 0.2 percentage point.
Below with the accuracy of the specific example verifying embodiment of the present invention.
By taking Anhui Province as an example, calculates separately out monthly and season the total retail sales of consumer goods and add up amplification.
In monthly calculating, 2 months in December, 2018 in 2012 is had chosen, annual 11,77 datas are used for structure altogether
The data of selection are substituted into Vector Autoression Models and solve coefficient by established model, are obtained following equation and are used to predict:
X1=0.133995015034*X1 (- 1)+0.060016090106*X1 (- 2)+0.0978166981793*X1 (-
3)-0.0238814056106*X1(-4)- 0.121274622693*X1(-5)+0.0129273671249*X1(-6)-
0.151727622773*X1(-7)-0.0575730399457*X1(-8)+ 0.0971350322117*X1(-9)-
0.181977972025*X1(-10)+ 0.481760121398*X1(-11)-0.138944553201*X2(-1)+
0.359929237865*X2(-2)-0.289457667103*X2(-3)+ 0.170026177445*X2(-4)+
0.661977400198*X2(-5)- 0.584831493548*X2(-6)+1.13522156247*X2(-7)+
0.144255517779*X2(-8)-0.226116032734*X2(-9)+ 1.12156193824*X2(-10)-
1.12784871659*X2(-11)+ 0.00144348418934*X3(-1)+0.00450417759919*X3(-2)+
0.0166994523403*X3(-3)+0.00406697866786*X3(-4)+ 0.00973352046645*X3(-5)+
0.00608518908198*X3(-6)+ 0.00406244443386*X3(-7)-0.00849983651043*X3(-8)-
0.0204040876124*X3(-9)-0.009942458413*X3(-10)- 0.026702579068*X3(-11)+
0.00557058821666*X4(-1)+ 0.0262278452699*X4(-2)+0.0288876536536*X4(-3)-
0.0156407662872*X4(-4)-0.00499849845508*X4(-5)+ 0.0122241205706*X4(-6)-
0.00838275063912*X4(-7)+ 0.00161159630757*X4(-8)+0.012903191558*X4(-9)-
0.00233036984986*X4(-10)-0.00265814549668*X4(-11)-134.10159802 +
0.0137622402002*X5+0.0469025882258*X6+2.44084311106*X7
Wherein:
Coefficient to be solved such as C (1,1)=0.133995015034, and so on.
It obtains partial results and error is as shown in table 3:
Table 3
In season calculates, the first quarter in 2012 to the fourth quater in 2018 is had chosen, annual four season, 28 altogether
The data of selection are substituted into Vector Autoression Models for constructing model and solve coefficient, obtained following equation and be used to by data
Prediction:
Y1=-0.0495735195232*Y1 (- 1) -0.263237414751*Y1 (- 2)+0.319030296378*Y1
(-3)+0.120643799918*Y1(-4)-0.141255765482* Y2(-1)+0.0463146651022*Y2(-2)+
0.0075351578163*Y2(-3)+ 0.0327421414129*Y2(-4)+0.406054359527*Y3(-1)-
0.270545186528* Y3(-2)+0.424735876986*Y3(-3)-0.302541892945*Y3(-4)-
64.4301992795-0.00521234505865*Y4+0.695795738938*Y5- 0.0105579506688*Y6。
Wherein:
Coefficient to be solved such as D (1,1)=- 0.0495735195232, and so on.
It obtains partial results and error is as shown in table 4:
Table 4
Based on the above results it is found that the prediction result error very little of the embodiment of the present invention, therefore obtained social consumer goods
It is as accurately that the total volume of retail sales, which adds up amplification,.
The embodiment of the invention also provides the forecasting system that a kind of total retail sales of consumer goods add up amplification, above-mentioned systems
System includes computer, and above-mentioned computer includes:
At least one storage unit;
At least one processing unit;
Wherein, at least one instruction is stored at least one above-mentioned storage unit, above-mentioned at least one instruction is by above-mentioned
At least one processing unit is loaded and is executed to perform the steps of
Obtain total retail sales of consumer goods related data, history of forming data;
Determine that the total retail sales of consumer goods add up the prediction index of amplification based on above-mentioned historical data, above-mentioned prediction refers to
Mark includes: industry index;
Add up amplification forecast system based on the above-mentioned prediction index building total retail sales of consumer goods;
Based on the Vector Autoression Models of pre-training to the above-mentioned total retail sales of consumer goods add up amplification forecast system into
Row processing, obtains the first prediction result;Shot and long term Memory Neural Networks based on pre-training are sold above-mentioned social consumer goods total
Volume adds up amplification forecast system and is handled, and obtains the second prediction result;
Initial predicted result is obtained based on above-mentioned first prediction result and above-mentioned second prediction result;
Above-mentioned initial predicted result is modified based on emergency event, obtains the accumulative increasing of the total retail sales of consumer goods
The prediction result of width.
It will be appreciated that above-mentioned forecasting system provided in an embodiment of the present invention is corresponding with above-mentioned prediction technique, it is related
The part such as explanation, citing, beneficial effect of content can add up in the prediction technique of amplification with reference to the total retail sales of consumer goods
Corresponding contents, details are not described herein again.
In conclusion compared with prior art, have it is following the utility model has the advantages that
The embodiment of the present invention passes through the prediction index for determining that the total retail sales of consumer goods add up amplification, and based on prediction
Index constructs the total retail sales of consumer goods and adds up amplification forecast system, wherein prediction index includes industry index;It is based on
The Vector Autoression Models and shot and long term Memory Neural Networks model of pre-training are accumulative to the total retail sales of consumer goods respectively to be increased
Width forecast system is handled, and the first prediction result and the second prediction result are obtained;It is predicted based on the first prediction result and second
As a result initial predicted result is obtained;Initial predicted result is modified based on emergency event, obtains social consumer goods retail
Total value adds up the prediction result of amplification.The embodiment of the present invention considers influence of the industry index factor to regional economy, comprehensively
Ground analyzes the situation of change of regional economy, and the accumulative amplification of the total retail sales of consumer goods can be gone out with Accurate Prediction, meanwhile,
The accumulative amplification that vector auto regression and shot and long term Memory Neural Networks are combined to the prediction total retail sales of consumer goods, makes
It is more accurate to obtain result.Support can be provided for departments of government decision based on prediction result, be consumer goods production and consumption
Enterprise provides information service.
It should be noted that through the above description of the embodiments, those skilled in the art can be understood that
It can be realized by means of software and necessary general hardware platform to each embodiment.Based on this understanding, above-mentioned skill
Substantially the part that contributes to existing technology can be embodied in the form of software products art scheme in other words, the meter
Calculation machine software product may be stored in a computer readable storage medium, such as ROM/RAM, magnetic disk, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
Method described in certain parts of example or embodiment.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, knot is not been shown in detail
Structure and technology, so as not to obscure the understanding of this specification.
Herein, relational terms such as first and second and the like be used merely to by an entity or operation with
Another entity or operation distinguish, and without necessarily requiring or implying between these entities or operation, there are any this
Actual relationship or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to nonexcludability
Include so that include a series of elements process, method, article or equipment not only include those elements, but also
Further include other elements that are not explicitly listed, or further include for this process, method, article or equipment institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including
There is also other identical elements in the process, method, article or equipment of the element.
The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although with reference to the foregoing embodiments
Invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each implementation
Technical solution documented by example is modified or equivalent replacement of some of the technical features;And these modification or
Replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. the prediction technique that a kind of total retail sales of consumer goods add up amplification, which is characterized in that the prediction technique is by calculating
Machine executes, comprising the following steps:
Obtain total retail sales of consumer goods related data, history of forming data;
Determine that the total retail sales of consumer goods add up the prediction index of amplification, the prediction index packet based on the historical data
It includes: industry index;
Add up amplification forecast system based on the prediction index building total retail sales of consumer goods;
The total retail sales of consumer goods are added up at amplification forecast system based on the Vector Autoression Models of pre-training
Reason, obtains the first prediction result;Shot and long term Memory Neural Networks model based on pre-training is sold the social consumer goods total
Volume adds up amplification forecast system and is handled, and obtains the second prediction result;
Initial predicted result is obtained based on first prediction result and second prediction result;
The initial predicted result is modified based on emergency event, the total retail sales of consumer goods is obtained and adds up the pre- of amplification
Survey result.
2. prediction technique as described in claim 1, which is characterized in that the total retail sales of consumer goods add up amplification prediction
System includes: that the total retail sales of consumer goods add up the monthly forecast system of amplification and the total retail sales of consumer goods accumulative amplification season
Spend forecast system.
3. prediction technique as claimed in claim 2, which is characterized in that it is monthly that the total retail sales of consumer goods add up amplification
The prediction index of forecast system include: price and level of consumption index, income level index, industry index, finance index,
Macro-performance indicator, industry index, openness index, taxation target and other indexs.
4. prediction technique as claimed in claim 2, which is characterized in that the total retail sales of consumer goods add up amplification season
The prediction index of forecast system include: price and level of consumption index, income level index, macro-performance indicator, industry index,
Industry index, consumer confidence index index and other indexs.
5. prediction technique as described in claim 1, which is characterized in that the industry index includes: that national real estate is fixed
Assets investment volume, the accumulative growth of national Marketable Housing Area Sold, province's automobile add up to increase than same period last year, save wholesale and retail business
Electricity consumption adds up amplification.
6. prediction technique as described in claim 1, which is characterized in that the acquisition methods of the initial predicted result are as follows:
The weight of default first prediction result and the second prediction result, in conjunction with weight to first prediction result and described second
Prediction result is handled, and the initial predicted result is obtained.
7. prediction technique as described in claim 1, which is characterized in that the emergency event includes: political event, international environment
Variation and social environment variation.
8. the forecasting system that a kind of total retail sales of consumer goods add up amplification, which is characterized in that the system comprises computer,
The computer includes:
At least one storage unit;
At least one processing unit;
Wherein, be stored at least one instruction at least one described storage unit, at least one instruction by it is described at least
One processing unit is loaded and is executed to perform the steps of
Obtain total retail sales of consumer goods related data, history of forming data;
Determine that the total retail sales of consumer goods add up the prediction index of amplification, the prediction index packet based on the historical data
It includes: industry index;
Add up amplification forecast system based on the prediction index building total retail sales of consumer goods;
The total retail sales of consumer goods are added up at amplification forecast system based on the Vector Autoression Models of pre-training
Reason, obtains the first prediction result;Shot and long term Memory Neural Networks based on pre-training are tired to the total retail sales of consumer goods
Meter amplification forecast system is handled, and the second prediction result is obtained;
Initial predicted result is obtained based on first prediction result and second prediction result;
The initial predicted result is modified based on emergency event, the total retail sales of consumer goods is obtained and adds up the pre- of amplification
Survey result.
9. forecasting system as claimed in claim 8, which is characterized in that the industry index includes: that national real estate is fixed
Assets investment volume, the accumulative growth of national Marketable Housing Area Sold, province's automobile add up to increase than same period last year, save wholesale and retail business
Electricity consumption adds up amplification.
10. forecasting system as claimed in claim 8, which is characterized in that the acquisition methods of the initial predicted result are as follows:
The weight of default first prediction result and the second prediction result, in conjunction with weight to first prediction result and described second
Prediction result is handled, and the initial predicted result is obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910492416.2A CN110232593A (en) | 2019-06-06 | 2019-06-06 | The total retail sales of consumer goods add up the prediction technique and system of amplification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910492416.2A CN110232593A (en) | 2019-06-06 | 2019-06-06 | The total retail sales of consumer goods add up the prediction technique and system of amplification |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110232593A true CN110232593A (en) | 2019-09-13 |
Family
ID=67859279
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910492416.2A Pending CN110232593A (en) | 2019-06-06 | 2019-06-06 | The total retail sales of consumer goods add up the prediction technique and system of amplification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110232593A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111353809A (en) * | 2019-12-12 | 2020-06-30 | 合肥工业大学 | Social consumer goods retail total quarterly accumulated amplification prediction method and system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030187770A1 (en) * | 2002-03-29 | 2003-10-02 | Hideyuki Kamiryo | Economic growth-rate forecasting program and a computer-readable recording media recorded with the same |
CN102385724A (en) * | 2010-08-27 | 2012-03-21 | 上海财经大学 | Spare part assembling demand forecasting information processing method applied to inventory management |
CN105550782A (en) * | 2015-12-31 | 2016-05-04 | 天津英福科技有限公司 | Interactive method for macroeconomic analysis system established based on macroeconomic data |
CN106022547A (en) * | 2016-07-04 | 2016-10-12 | 上海财经大学 | Macro-forecast system and method thereof for future development of regional economy |
CN107798465A (en) * | 2017-10-09 | 2018-03-13 | 中国电子科技集团公司第二十八研究所 | A kind of urban economy monitoring running state early warning analysis method and system |
-
2019
- 2019-06-06 CN CN201910492416.2A patent/CN110232593A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030187770A1 (en) * | 2002-03-29 | 2003-10-02 | Hideyuki Kamiryo | Economic growth-rate forecasting program and a computer-readable recording media recorded with the same |
CN102385724A (en) * | 2010-08-27 | 2012-03-21 | 上海财经大学 | Spare part assembling demand forecasting information processing method applied to inventory management |
CN105550782A (en) * | 2015-12-31 | 2016-05-04 | 天津英福科技有限公司 | Interactive method for macroeconomic analysis system established based on macroeconomic data |
CN106022547A (en) * | 2016-07-04 | 2016-10-12 | 上海财经大学 | Macro-forecast system and method thereof for future development of regional economy |
CN107798465A (en) * | 2017-10-09 | 2018-03-13 | 中国电子科技集团公司第二十八研究所 | A kind of urban economy monitoring running state early warning analysis method and system |
Non-Patent Citations (3)
Title |
---|
LYU XIAOZHONG等: "Sales Prediction by Integrating the Heat and Sentiments of Product Dimensions", 《SUSTAINABILITY》 * |
吕晓峰: "我国电子商务各细分市场规模与社会消费品零售总额之间的关系研究", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
李宇卿: "我国能源消费预测模型实证研究", 《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 (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111353809A (en) * | 2019-12-12 | 2020-06-30 | 合肥工业大学 | Social consumer goods retail total quarterly accumulated amplification prediction method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109377260A (en) | User behavior analysis system towards apparel industry | |
CN105809482A (en) | Big data analysis system and method for cross-border trade | |
Andriosopoulos et al. | Portfolio optimization and index tracking for the shipping stock and freight markets using evolutionary algorithms | |
Lin et al. | How to effectively stabilize China's commodity price fluctuations? | |
Yamova et al. | Integral valuation of an enterprise's competitiveness in the industrial economy | |
Liu et al. | The diverse impact of heterogeneous customer characteristics on supply chain finance: Empirical evidence from Chinese factoring | |
CN109242664A (en) | It is a kind of towards the tax risk prediction technique for newly setting up enterprise | |
Lu et al. | A clustering-based portfolio strategy incorporating momentum effect and market trend prediction | |
Liu et al. | Quantified moving average strategy of crude oil futures market based on fuzzy logic rules and genetic algorithms | |
Li et al. | Predicting business risks of commercial banks based on BP-GA optimized model | |
CN110175689A (en) | A kind of method of probabilistic forecasting, the method and device of model training | |
Yi et al. | Financial risk prediction in supply chain finance based on buyer transaction behavior | |
Tabak et al. | Modeling default probabilities: The case of Brazil | |
CN110390426A (en) | Prediction technique, system and the storage medium of industrial economy in district speedup | |
Chirichenko et al. | Marketing determinants of the development of the investment market: innovations in the assessment of demand and supply (case study for the construction industry) | |
CN110232593A (en) | The total retail sales of consumer goods add up the prediction technique and system of amplification | |
Meng | Modeling the impact of exchange rates using a multicurrency framework | |
Wang et al. | Identifying the factors of China's seasonal retail sales of consumer goods using a data grouping approach–based GRA method | |
Fedotova et al. | Quality of life of the population of Russia in the light of modern global challenges | |
Li et al. | Strengthen credit scoring system of small and micro businesses with soft information: Analysis and comparison based on neural network models | |
Srisawat | An application of association rule mining based on stock market | |
Su | Nonlinear ARIMA models with feedback SVR in financial market forecasting | |
Adedapo | Analysis of Non-Oil Exports–Economic Growth Relationship in Nigeria: The Role of Institutional Qualities | |
Shi et al. | An optimal decision assessment model based on the acceptable maximum LGD of commercial banks and its application | |
Wang et al. | Oil Price Shocks and Inflation Targeting in China |
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: 20190913 |
|
RJ01 | Rejection of invention patent application after publication |