CN112418833A - Method and device for predicting price change trend of digital currency - Google Patents
Method and device for predicting price change trend of digital currency Download PDFInfo
- Publication number
- CN112418833A CN112418833A CN201910772405.XA CN201910772405A CN112418833A CN 112418833 A CN112418833 A CN 112418833A CN 201910772405 A CN201910772405 A CN 201910772405A CN 112418833 A CN112418833 A CN 112418833A
- Authority
- CN
- China
- Prior art keywords
- digital currency
- price
- trend
- prediction
- target digital
- 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
Classifications
-
- 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
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/04—Payment circuits
- G06Q20/06—Private payment circuits, e.g. involving electronic currency used among participants of a common payment scheme
- G06Q20/065—Private payment circuits, e.g. involving electronic currency used among participants of a common payment scheme using e-cash
- G06Q20/0652—Private payment circuits, e.g. involving electronic currency used among participants of a common payment scheme using e-cash e-cash with decreasing value according to a parameter, e.g. time
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention provides a method and a device for predicting a price change trend of digital currency. The method comprises the following steps: acquiring the price of the target digital currency and the value of each predictive variable in the lag phase; inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting a prediction result of the price change trend of the current target digital currency; the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical value of each prediction variable; the trend prediction model is a model constructed based on vector autoregression. According to the method and the device for predicting the price change trend of the digital currency, provided by the embodiment of the invention, the trend prediction model is constructed on the basis of vector autoregression, and the trend prediction model outputs the prediction result of the price change trend of the current-period target digital currency according to the price of the target digital currency in the lag period and the values of all prediction variables, so that the price change trend prediction result with higher precision can be obtained.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a method and a device for predicting price change trend of digital currency.
Background
At present, for the prediction of the price change trend of digital currency, a prediction model is generally constructed by combining methods such as neural network and multiple linear regression on the basis of the theory of economics or finance, and a prediction result is obtained.
However, because the digital currency price has a plurality of influence factors, the influence mechanism is very complex, and the influence factors have correlation influence, the prediction accuracy of the prediction model constructed based on the existing economics or finance theory is low.
Disclosure of Invention
The embodiment of the invention provides a method and a device for predicting a price change trend of digital currency, which are used for solving or at least partially solving the defect of low prediction precision in the prior art.
In a first aspect, an embodiment of the present invention provides a method for predicting a price variation trend of digital currency, including:
acquiring the price of the target digital currency and the value of each predictive variable in the lag phase;
inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting a prediction result of the price change trend of the target digital currency in the current period;
the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical values of the prediction variables; the trend prediction model is a model constructed based on vector autoregression.
Preferably, the specific step of inputting the price of the target digital currency and the value of each prediction variable in the lag period into a trend prediction model and outputting the prediction result of the price change trend of the target digital currency in the current period includes:
inputting the price of the target digital currency in the lag period and the value of each prediction variable into a vector autoregressive submodel in the trend prediction model, and outputting the price prediction value of the target digital currency in the current period;
and acquiring a prediction result of the price change trend of the current period of the target digital currency according to the price prediction value of the current period of the target digital currency and the price of the last period of the target digital currency.
Preferably, before inputting the price of the target digital currency and the value of each predictive variable in the lag period into the trend prediction model, the method further comprises:
according to the historical value of each influence factor of the target digital currency price change, performing stability inspection and collaborative analysis on each influence factor to obtain each prediction variable;
and training according to the historical price of the target digital currency and the historical values of the forecasting variables to obtain the trend forecasting model.
Preferably, before inputting the price of the target digital currency and the value of each predictive variable in the lag period into the trend prediction model, the method further comprises:
according to the historical value of each influence factor of the price change of the target digital currency, performing stability inspection and collaborative analysis on each influence factor to obtain each candidate variable;
training according to the historical price of the target digital currency and the historical value of each candidate variable to obtain a first prediction model;
the first prediction model is checked, and each prediction variable is screened out from each candidate variable according to a check result;
and training according to the historical price of the target digital currency and the historical values of the forecasting variables to obtain the trend forecasting model.
Preferably, the specific step of testing the first prediction model includes:
the first predictive model is tested according to at least one of a granger causal relationship test, an impulse response function test, and a variance decomposition test.
Preferably, obtaining the trend prediction model further comprises:
the trend prediction model is tested according to at least one of a granger causal relationship test, an impulse response function test, and a variance decomposition test.
Preferably, the influencing factors include supply and demand factors, investment attraction factors, macro-economic and financial factors, and self-factors of the target digital currency.
In a second aspect, an embodiment of the present invention provides an apparatus for predicting a price variation trend of digital currency, including:
the acquisition module is used for acquiring the price of the target digital currency and the value of each predictive variable in the lag phase;
the prediction module is used for inputting the price of the target digital currency and the values of the prediction variables in the lag period into a trend prediction model and outputting the prediction result of the price change trend of the target digital currency in the current period;
the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical values of the prediction variables; the trend prediction model is a model constructed based on vector autoregression.
In a third aspect, an embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when executing the computer program, the method for predicting a price change trend of digital currency provided in any one of the various possible implementations of the first aspect is implemented.
In a fourth aspect, embodiments of the present invention provide a non-transitory computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the steps of the method for predicting a trend of a price change of digital money as provided by any one of the various possible implementations of the first aspect.
According to the method and the device for predicting the price change trend of the digital currency, provided by the embodiment of the invention, the trend prediction model is constructed on the basis of vector autoregression, and the trend prediction model outputs the prediction result of the price change trend of the current-period target digital currency according to the price of the target digital currency in the lag period and the values of all prediction variables, so that the price change trend prediction result with higher precision can be obtained.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flow chart illustrating a method for predicting a price variation trend of digital currency according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a device for predicting a price change trend of digital money according to an embodiment of the present invention;
fig. 3 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In order to overcome the above problems in the prior art, an embodiment of the present invention provides a method and an apparatus for predicting a price change trend of digital money, wherein a prediction model is constructed based on a vector autoregressive method, and since the vector autoregressive model does not have any pre-constraint condition, and is not based on a finance or economics theory, a plurality of price influencing factors can be covered, and a time series having a correlation is predicted by capturing linear correlations among a plurality of time series, so that a prediction result of a price change trend of digital money with higher accuracy can be obtained.
Fig. 1 is a flowchart illustrating a method for predicting a price change trend of digital currency according to an embodiment of the present invention. As shown in fig. 1, the method includes: and step S101, acquiring the price of the target digital currency in the lag period and the value of each forecasting variable.
It should be noted that the lag period refers to a plurality of consecutive periods before the current period. For example, it is today (day 10 of A month), and the lag phase may be one week before today (day 3 to 9 of A month).
The number of phases included in the lag phase is predetermined.
The predictive variable is a variable related to a price change of the target digital currency. The predictive variable may be a variable that is positively or negatively correlated with the price change of the target digital currency. For example, in general, the total stock of the target digital currency is a variable that is negatively related to the price change of the target digital currency, and the economic scale of the target digital currency is a variable that is positively related to the price change of the target digital currency; the positive message of the target digital currency is a variable positively correlated with the price change of the target digital currency, and the negative message is a variable negatively correlated with the price change of the target digital currency.
For each phase in the lag phase, the price of the target digital currency and the values of the predictive variables in the phase are acquired, so that the price of the target digital currency and the values of the predictive variables in the lag phase can be acquired.
And S102, inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting the prediction result of the price change trend of the current-period target digital currency.
The trend prediction model is obtained by training according to the historical price of the target digital currency and the historical value of each prediction variable; the trend prediction model is a model constructed based on vector autoregression.
Before step S101, a sample data set is constructed according to the price of the target digital currency and the values of the predictive variables in the historical period, and the trend prediction model is trained according to the sample data set to obtain a trained trend prediction model. The trained trend prediction model can be used for predicting the price change trend of the target digital currency.
Vector Autoregression (VAR) is used to capture linear correlations between multiple time series. Vector autoregression is commonly used to predict time series systems of correlations and to analyze the dynamic effects of random disturbance terms on variable systems. All variables in the VAR enter the model in the same way: each variable has an equation that accounts for its evolution in terms of its own lag value, the lag values of the other model variables, and an error term.
The price of the target digital currency and each prediction variable are used as variables in the trend prediction model, and through training, an error term and a coefficient (parameter) to be determined in the trend prediction model can be determined, so that the trained trend prediction model is obtained.
Vector autoregressive has at least the following advantages: 1. the estimation of the parameters is easier; 2. the universality of the VAR model form is that the VAR model is not based on financial economic theory, so that other explanatory variables, such as variables with one-way causal relationship, can be added to the VAR model to a great extent optionally, and can also be used as exogenous variables to enhance the explanatory power of the model on dependent variables; 3. the method has no initial constraint condition, is very accurate in recent prediction, can predict the variation trend when performing long-term prediction, and has great reference value.
Specifically, the price of the target digital currency in the lag period and the value of each predictive variable are the lag value of each variable in the trend prediction model and are input into the trend prediction model, and the trend prediction model can output the prediction result of the price change trend of the current-period target digital currency.
The price variation trend of the current-stage target digital currency refers to the variation of the current-stage price of the target digital currency relative to the previous-stage price, and specifically includes rising, falling or unchanging.
The embodiment of the invention constructs the trend prediction model based on vector autoregression, and the trend prediction model outputs the prediction result of the price change trend of the current-period target digital currency according to the price of the target digital currency in the lag period and the value of each prediction variable, so that the price change trend prediction result with higher precision can be obtained.
Based on the content of the above embodiments, the specific steps of inputting the price of the target digital currency in the lag period and the value of each prediction variable into the trend prediction model, and outputting the prediction result of the price change trend of the current-period target digital currency include: and inputting the price of the target digital currency in the lag period and the value of each prediction variable into a vector autoregressive submodel in the trend prediction model, and outputting the price prediction value of the current-period target digital currency.
Specifically, the trend prediction model comprises a vector autoregression submodel, and the vector autoregression submodel is used for obtaining the price predicted value of the current-period target digital currency and the predicted value of each predicted variable of the current period.
The expression of the vector autoregressive submodel with N variable lag periods is
Yt＝c+П1Yt-1+П2Yt-2+…+ПpYt-p+ut
Wherein, Yt＝(y1，t，y2，t，…，yN，t)T；c＝(c1，c2，…，cN)T；
In general, y1Representing the price, y, of the target digital currency2，…，yNRespectively represent the values of the respective predictor variables, y1，t，y2，t，…，yN，tRespectively representing the price predicted value of the current-period target digital currency and the predicted value of each predicted variable of the current period.
Through training, pij(j ═ 1, 2, …, p), c, and utCan be determined so that Y can be obtained based on the price of the target digital currency in the lag phase and the value of each predictive variablet-1，Yt-2，…Yt-pIs a reaction of Yt-1，Yt-2，…Yt-pSubstituting the above-mentioned expression of vector autoregressive submodel to obtain YtThe price predicted value of the current-period target digital currency and the predicted value of each predicted variable of the current period can be obtained.
And acquiring a prediction result of the price change trend of the current-period target digital currency according to the price prediction value of the current-period target digital currency and the price of the previous-period target digital currency.
Specifically, the predicted value of the price of the current-stage target digital currency is compared with the price of the last-stage target digital currency, and the predicted result of the price change trend of the current-stage target digital currency can be obtained.
The price prediction value of the current-stage target digital currency and the prediction values of the current-stage prediction variables are replaced by the price of the first-stage target digital currency and the values of the prediction variables in the lag stage, the price of the new lag-stage target digital currency and the values of the prediction variables are obtained, the price of the new lag-stage target digital currency and the values of the prediction variables are input into the trend prediction model, and the price prediction value of the next-stage target digital currency, the prediction values of the next-stage prediction variables and the prediction result of the price change trend of the next-stage target digital currency can be obtained. And by analogy, the price change trend of the target digital currency in the current period and the continuous multiple periods after the current period can be obtained according to the price of the target digital currency in the lag period and the value of each predictive variable.
The continuous periods after the period are long, the accuracy of the long-term prediction result of the price change trend of the common prediction model is low, and the embodiment can obtain the long-term prediction result with higher accuracy.
For example, the current date is today (day 10 of a month), the lag phase may be one week before today (day 3 of a month to day 9), and although a general prediction model can predict the price change trend from day 10 of a month to day 25 of a month according to the data from day 3 of a month to day 9 of a month, the prediction result from day 10 of a month to day 12 may have higher accuracy, and the prediction result from day 13 of a month may have lower accuracy; according to the method provided by the embodiment of the invention, the prediction results from 10 days to 25 days of the A month have high precision, and the prediction precision of the long-term prediction results is greatly improved.
The specific number of multiple periods in the consecutive multiple periods after the current period may be set according to needs, and this is not particularly limited in the embodiment of the present invention.
According to the embodiment of the invention, the price prediction value of the current-period target digital currency is obtained according to the vector autoregressive submodel, and a more accurate price prediction result of the current-period target digital currency can be obtained, so that a price change trend prediction result with higher precision can be obtained.
Based on the above embodiments, before inputting the price of the target digital currency and the value of each predictive variable in the lag period into the trend prediction model, the method further includes: and according to the historical value of each influence factor of the price change of the target digital currency, performing stability inspection and collaborative analysis on each influence factor to obtain each prediction variable.
Specifically, because there are many influencing factors of the digital currency price change, before the trend prediction model is trained, each influencing factor can be screened through stationarity check and co-integration analysis, and a part of each influencing factor and/or a plurality of linear combinations formed by two influencing factors are/is selected as a prediction variable.
For vector autoregression, regression analysis is based on data stationarity. The vector autoregressive requirement is that the variable sequence is a stationary time sequence, and stationarity is defined as: the statistical regularity of the time series does not change with the time, that is, the statistical characteristics do not change with the time.
For each influencing factor, a stationarity test is performed on a time series formed by historical values of the influencing factor.
The stationarity test can be performed by a unit root test method.
The unit root test is to test whether a unit root exists in the sequence, because the existence of the unit root is a non-stationary time sequence.
Vector autoregressive requires that the time sequence is smooth, but because most time sequences are non-smooth in practical application, a difference method is usually adopted to eliminate non-smooth trends contained in the sequences, so that the sequences are modeled after smoothing. While some scalar are themselves non-stationary sequences, their linear combination is likely to be stationary. This smooth linear combination is called a covariance equation and can be interpreted as a long-term stable equilibrium relationship between variables. The reason for the synergy is that this set of time series has some common tendency, which can be counteracted by some linear combination, with the effect of becoming non-stationary to stationary.
If a sequence can become a stable sequence after once difference, the sequence is called as a first-order single integer; if a sequence can become a stable sequence after d differences, the sequence is called d-order single integer.
If both time series are d order singlets and the linear combination of the two time series is (d-b) order singlets, then the two time series are referred to as (d-b) order co-integrals. The coefficient vector that constitutes a linear combination of two variables is called the "co-integral vector".
By performing stationarity check on each influence factor, the order of the single integer of the influence factor can be obtained.
And performing co-integration check on the two influence factors according to the single-integration order of every two influence factors, and judging whether the two influence factors are co-integrated or not.
According to the results of the stationarity test and the co-integration analysis, a plurality of influence factors with stationarity and/or a plurality of linear combinations of two co-integrated influence factors can be selected from the influence factors to serve as prediction variables.
The number of the prediction variables can be selected according to actual conditions, for example, 5-9 prediction variables can be selected, but is not limited to the above.
And training according to the historical price of the target digital currency and the historical values of the forecasting variables to obtain a trend forecasting model.
It should be noted that, for the lag period number p in the vector autoregressive, the value of p is too small, and the residual error may have autocorrelation and cause inconsistency of parameter estimation; by increasing the p value appropriately, the autocorrelation present in the residual can be eliminated. But the p value cannot be too large, otherwise, the parameters to be estimated are too excessive, the degree of freedom is seriously reduced, and the effectiveness of model parameter estimation is directly influenced.
The value of the number of lag periods p may be preselected or determined according to the Akage Information Criterion (AIC) and/or the Schwarz Criterion (SC).
After the lag period number p and each predictive variable are determined, training can be performed according to the historical price of the target digital currency and the historical value of each predictive variable, and the undetermined coefficient (parameter) in the trend prediction model is determined to obtain the trained trend prediction model.
The embodiment of the invention carries out stability inspection and collaborative analysis on each influence factor to obtain each prediction variable, can obtain a time sequence with stability to predict the price change trend, and can obtain a price change trend prediction result with higher precision.
Based on the above embodiments, before inputting the price of the target digital currency and the value of each predictive variable in the lag period into the trend prediction model, the method further includes: and performing stability inspection and collaborative analysis on each influence factor according to the historical value of each influence factor of the price change of the target digital currency to obtain each candidate variable.
And training according to the historical price of the target digital currency and the historical values of the candidate variables to obtain a first prediction model.
Specifically, according to the results of stability test and co-integration analysis on each influence factor, a plurality of influence factors with stability and/or a plurality of linear combinations of two influence factors with co-integration are selected and are not directly used as prediction variables but candidate variables.
And the first prediction model obtained by training according to the historical price of the target digital currency and the historical value of each candidate variable is not used as a final trend prediction model but used as a preliminarily obtained prediction model.
The first predictive model is also a model constructed based on vector autoregressive.
And testing the first prediction model, and screening out each prediction variable from each candidate variable according to a test result.
Specifically, after the first prediction model is obtained, the first prediction model is checked for the historical price of the target digital currency and the historical values of the candidate variables.
Through the examination, the influence trend and the degree of the fluctuation of each candidate variable on the price trend can be obtained.
According to the influence degree of the fluctuation of each candidate variable on the price trend, screening the candidate variables to remove the candidate variables with the influence degree smaller than a preset threshold value, reserving the candidate variables with the influence degree larger than the preset threshold value as prediction variables, and selecting the candidate variables with the influence degree larger than the preset threshold value as the prediction variables as far as possible.
And training according to the historical price of the target digital currency and the historical values of the forecasting variables to obtain a trend forecasting model.
Specifically, because the number of the predictive variables may be less than the number of the candidate variables, correspondingly, if the predictive variables are part of the candidate variables, that is, part of the candidate variables are removed, because the variables in the trend prediction model change, the trend prediction model is obtained by training again according to the historical price of the target digital currency and the historical values of the predictive variables; and if the predicted variables are the same as the candidate variables, namely any candidate variable is not removed, training is not carried out, and the first prediction model is used as a trend prediction model.
The prediction accuracy of the trend prediction model is higher because the influence degree of the prediction variables on the price trend is larger.
According to the embodiment of the invention, the first prediction model is checked, the prediction variables are further screened out, the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical values of the prediction variables, the trend prediction model with higher prediction precision can be obtained, and thus the price change trend prediction result with higher precision can be obtained.
Based on the content of the foregoing embodiments, the specific step of checking the first prediction model includes: the first predictive model is tested according to at least one of a granger causal relationship test, an impulse response function test, and a variance decomposition test.
Specifically, the first prediction model is tested, and at least one of a glange causal relationship test, an impulse response function test, and a variance decomposition test may be used.
Model stability means that when a pulse is applied to the information process of an equation in the first prediction model, the impact gradually disappears with time, and if the impact does not disappear, the model is unstable.
And the Glanker causal relationship test can be used for stability test. In particular, the granger causal test can test whether any of the predictive variables is logically causal to the price of the target digital currency.
The impulse response function describes the influence of the impact of each endogenous variable in the vector autoregression on the intrinsic variables and other endogenous variables, or the impulse response function is to observe the response of each variable in the model to the impact over time.
The impulse response function describes the response of an endogenous variable to residual impulses. In particular, it describes the dynamic effects on the current and future values of the endogenous variable after a standard deviation magnitude of the impulse is applied to the random error term, and this analysis method is called the impulse response function.
The variance decomposition is to further evaluate the contribution degree of each endogenous variable to the prediction variance.
The three inspection methods can analyze the influence trend and degree of the fluctuation of the prediction variable on the price trend, and greatly improve the interpretation capability of the trend prediction model.
According to the embodiment of the invention, the influence degree of fluctuation of each candidate variable on the price trend is obtained through at least one of the grand causal relationship test, the impulse response function test and the variance decomposition test, so that the candidate variable with larger influence degree on the price trend can be screened out as the prediction variable, the trend prediction model with higher prediction precision can be obtained, and the price change trend prediction result with higher precision can be obtained.
Based on the content of the foregoing embodiments, obtaining the trend prediction model further includes: the trend prediction model is tested according to at least one of a granger causal relationship test, an impulse response function test, and a variance decomposition test.
Specifically, the trend prediction model may be further checked according to at least one of a granger causal relationship check, an impulse response function check, and a variance decomposition check, so as to obtain an influence degree of fluctuation of each prediction variable on the price trend.
According to the embodiment of the invention, the trend prediction model is tested according to at least one of the granger causal relationship test, the impulse response function test and the variance decomposition test, so that the influence degree of the fluctuation of each prediction variable on the price trend can be obtained, and the price of the target digital currency can be conveniently regulated and controlled.
Based on the contents of the above embodiments, the respective influencing factors include supply and demand factors, investment attraction factors, macro-economic and financial factors, and self-factors of the target digital currency.
Specifically, one of the key factors in the price change of digital money is the supply-demand interaction of the market. Therefore, at least one of the total stock amount of the target digital money in circulation, the annual amount of the target digital money, the circulation speed of the target digital money, and the general price level of the digital money and the service, etc. may be selected as the supply and demand factors.
Digital currency has certain investment attributes and, therefore, investment appeal also has an impact on the price of the digital currency. In general, a demand for the target digital currency, which increases due to a higher attraction, may generate an upward trend in the price of the target digital currency, and a lower attraction may mean a decrease in the demand for the target digital currency and the price thereof. Variables in terms of investment attraction may be selected as investment attraction factors.
The role of the global or national macro-economic and financial development also affects the price of the target digital currency through the capital market such as stock market. Thus, macroscopic economic and financial factors may also be a factor.
The characteristics of the target digital currency itself may also affect its price, such as its time attribute, production cost, production efficiency, and the like.
The supply and demand factors, the investment attraction factors, the macro-economic and financial factors, and the factors of the target digital currency themselves may all include various.
According to the embodiment of the invention, the supply and demand factors, the investment attraction factors, the macroscopic economy and financial factors and the self factors of the target digital currency are taken as the influence factors, the coverage is higher, the vector autoregression is combined, the influence factors do not need to be excluded in advance, and the trend prediction model with higher prediction precision can be obtained by combining the vector autoregression, so that the price change trend prediction result with higher precision can be obtained.
In order to facilitate understanding of the embodiments of the present invention, the following description will be given by taking bitcoin as an example.
For supply and demand factors, the Total number of mined Bitcoins (totcbcs) can be selected as the Total stock of the circulating Bitcoins; representing Bitcoin economies of scale using the total Number of Unique Bitcoin Transactions per day (Bitcon Number of Transactions, NTRAN) and the Number of Unique Bitcoin Addresses Used per day (Bitcon Number of Unique Bitcoin Addresses Used, NADDU); using Bitcoin Transaction Confirmation Time (ATRCT) to indicate the currency circulation speed of Bitcoin circulation; to measure the general price level of digital currency and services for the global economy, the exchange rate between U.S. dollars and euros (USDEUR) is used because in practice the price of bitcoin is quoted in U.S. dollars. For example, if dollars are redeemed against euros to upgrade, it is likely that it will also upgrade, whereas an increase in euros to dollar exchange rates will result in a decrease in the amount of dollars that a bitcoin needs to pay, thereby reducing the price of the bitcoin.
For investment attraction factors, the information obtained from the bitcoin wallet MyWallet represents the degree of investment in bitcoin to represent the degree of attraction of digital currency to people. Specifically, the Number of registered Users of Bitcoin wallet (Bitcoin MyWallet of Users, MWNUS), the Number of transactions Per Day of Bitcoin wallet (Bitcoin mywall Number of Transaction Per Day, MWNTD), and the total Transaction amount of Bitcoin wallet (Bitcoin mywall Transaction Volume, MWTRV) may be selected as investment attraction factors.
For macroscopic economic and financial factors, bitcoin prices are affected by variables such as stock exchange indices, exchange rates, and oil price measurements. The effect of macroscopic economic and financial indicators on bitcoin price may play a role through a variety of channels. For example, a stock exchange index may reflect the overall macroscopic and financial development of the global economy. Advantageous macro-economic and financial developments may stimulate the use of bitcoin at trading and exchange, thereby enhancing its demand, which may have a positive impact on bitcoin prices. Specifically, the Price of petroleum (OPEC cloud Oil Price, OBR) and Doujones comprehensive index (DOWJONES, DJCI) can be selected. Gold prices (GoldPrice, GOLDP) and the calendar index (Nikkei 225Stock Price, N225) or other major Stock market indices may also be selected.
For the self factors of the target digital currency, the influence of the property of the bitcoin under the block chain basis can be considered, for example, as time goes on, the bitcoin ' digging ' speed is slower and the difficulty is larger, and the miner ' gains less and less while ' digging ', that is, the increase of the ' digging ' cost may have a negative influence on the bitcoin price to a certain extent. Specifically, the Hash Rate of a Block chain (bitjoint Hash Rate, HRATE), the Difficulty coefficient obtained by the Hash value of the Block chain (bitjoint Difficulty, DIFF), the Average Block Size of the Block chain (bitjoint Average Block Size, AVBLS), the miner profit of the Block chain (bitjoint Miners revived, MIREV), and the Transaction Number per Block in the Block chain (bitjoint Number of Transaction Perblock, NTRBL) are selected.
The TOTBC, NTRAN, NADDU, ATRCT, USDEUR, MWNUS, MWNTD, MWTRV, OBR, DJCI, GOLDP, N225, HRATE, DIFF, AVBLS, MIREV and NTRBL can be selected through stable type inspection and consistency analysis to obtain the prediction variable. Training according to the values of the prediction variables and the prices of the bitcoins from 12/1/2018 to 3/1/2019 to obtain a trend prediction model, predicting the change trend of the bitcoin prices from 3/2/2019 to 4/1/2019, wherein the prediction result is a continuous rising, comparing the prediction result with an actual bitcoin price change curve, and the actual bitcoin price is really a continuous rising in the time period.
Fig. 2 is a schematic structural diagram of a device for predicting a price change trend of digital money according to an embodiment of the present invention. Based on the content of the above embodiments, as shown in fig. 2, the apparatus includes an obtaining module 201 and a predicting module 202, wherein:
an obtaining module 201, configured to obtain a price of the target digital currency and values of the prediction variables in the lag phase;
the prediction module 202 is used for inputting the price of the target digital currency in the lag period and the value of each prediction variable into the trend prediction model and outputting the prediction result of the price change trend of the current-period target digital currency;
the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical value of each prediction variable; the trend prediction model is a model constructed based on vector autoregression.
Specifically, for each period in the lag period, the acquisition module 201 may acquire the price of the target digital currency and the values of the respective predictive variables in the period, and thus may acquire the price of the target digital currency and the values of the respective predictive variables in the lag period.
The prediction module 202 inputs the price of the target digital currency in the lag period and the values of the prediction variables into the lag values of the variables in the trend prediction model, and the trend prediction model can output the prediction result of the price change trend of the current-period target digital currency.
The specific method and process for implementing the corresponding functions by each module included in the digital currency price variation trend prediction device are described in the above digital currency price variation trend prediction method, and are not described herein again.
The device for predicting the price variation trend of the digital currency is used for the method for predicting the price variation trend of the digital currency in the embodiments. Therefore, the description and definition in the prediction method of the digital currency price change trend in the foregoing embodiments can be used for understanding of the execution modules in the embodiments of the present invention.
The embodiment of the invention constructs the trend prediction model based on vector autoregression, and the trend prediction model outputs the prediction result of the price change trend of the current-period target digital currency according to the price of the target digital currency in the lag period and the value of each prediction variable, so that the price change trend prediction result with higher precision can be obtained.
Fig. 3 is a block diagram of an electronic device according to an embodiment of the present invention. Based on the content of the above embodiment, as shown in fig. 3, the electronic device may include: a processor (processor)301, a memory (memory)302, and a bus 303; wherein, the processor 301 and the memory 302 complete the communication with each other through the bus 303; the processor 301 is configured to invoke computer program instructions stored in the memory 302 and executable on the processor 301 to perform the methods for predicting a trend of a digital currency price change provided by the above-described embodiments of the methods, including, for example: acquiring the price of the target digital currency and the value of each predictive variable in the lag phase; inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting a prediction result of the price change trend of the current target digital currency; the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical value of each prediction variable; the trend prediction model is a model constructed based on vector autoregression.
Another embodiment of the present invention discloses a computer program product, the computer program product includes a computer program stored on a non-transitory computer readable storage medium, the computer program includes program instructions, when the program instructions are executed by a computer, the computer can execute the method for predicting the digital currency price change trend provided by the above method embodiments, for example, the method includes: acquiring the price of the target digital currency and the value of each predictive variable in the lag phase; inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting a prediction result of the price change trend of the current target digital currency; the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical value of each prediction variable; the trend prediction model is a model constructed based on vector autoregression.
Furthermore, the logic instructions in the memory 302 may be implemented in software functional units and stored in a computer readable storage medium when sold or used as a stand-alone product. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or make a contribution to the prior art, or may be implemented in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the methods of the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Another embodiment of the present invention provides a non-transitory computer-readable storage medium, which stores computer instructions, the computer instructions causing a computer to execute the method for predicting a price change trend of digital currency, provided by the above method embodiments, for example, the method includes: acquiring the price of the target digital currency and the value of each predictive variable in the lag phase; inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting a prediction result of the price change trend of the current target digital currency; the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical value of each prediction variable; the trend prediction model is a model constructed based on vector autoregression.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. It is understood that the above-described technical solutions may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method of the above-described embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for predicting the price variation trend of digital currency is characterized by comprising the following steps:
acquiring the price of the target digital currency and the value of each predictive variable in the lag phase;
inputting the price of the target digital currency in the lag period and the value of each prediction variable into a trend prediction model, and outputting a prediction result of the price change trend of the target digital currency in the current period;
the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical values of the prediction variables; the trend prediction model is a model constructed based on vector autoregression.
2. The method for predicting the price change trend of digital currency according to claim 1, wherein the specific steps of inputting the price of the target digital currency and the values of the predictive variables in the lag period into a trend prediction model and outputting the prediction result of the price change trend of the target digital currency at the present time comprise:
inputting the price of the target digital currency in the lag period and the value of each prediction variable into a vector autoregressive submodel in the trend prediction model, and outputting the price prediction value of the target digital currency in the current period;
and acquiring a prediction result of the price change trend of the current period of the target digital currency according to the price prediction value of the current period of the target digital currency and the price of the last period of the target digital currency.
3. The method for predicting the price variation trend of digital currency according to claim 1, wherein before inputting the price of the target digital currency and the value of each predictive variable in the lag period to the trend prediction model, the method further comprises:
according to the historical value of each influence factor of the target digital currency price change, performing stability inspection and collaborative analysis on each influence factor to obtain each prediction variable;
and training according to the historical price of the target digital currency and the historical values of the forecasting variables to obtain the trend forecasting model.
4. The method for predicting the price variation trend of digital currency according to claim 1, wherein before inputting the price of the target digital currency and the value of each predictive variable in the lag period to the trend prediction model, the method further comprises:
according to the historical value of each influence factor of the price change of the target digital currency, performing stability inspection and collaborative analysis on each influence factor to obtain each candidate variable;
training according to the historical price of the target digital currency and the historical value of each candidate variable to obtain a first prediction model;
the first prediction model is checked, and each prediction variable is screened out from each candidate variable according to a check result;
and training according to the historical price of the target digital currency and the historical values of the forecasting variables to obtain the trend forecasting model.
5. The method for predicting a trend of digital currency prices according to claim 4, wherein said step of testing said first predictive model comprises:
the first predictive model is tested according to at least one of a granger causal relationship test, an impulse response function test, and a variance decomposition test.
6. The method for predicting the trend of digital currency prices according to claim 3, wherein obtaining the trend prediction model further comprises:
the trend prediction model is tested according to at least one of a granger causal relationship test, an impulse response function test, and a variance decomposition test.
7. The method for predicting a price variation tendency of digital money according to any one of claims 3 to 6, wherein the respective influencing factors include supply and demand factors, investment attraction factors, macro-economic and financial factors, and self factors of the target digital money.
8. An apparatus for predicting a price change trend of digital money, comprising:
the acquisition module is used for acquiring the price of the target digital currency and the value of each predictive variable in the lag phase;
the prediction module is used for inputting the price of the target digital currency and the values of the prediction variables in the lag period into a trend prediction model and outputting the prediction result of the price change trend of the target digital currency in the current period;
the trend prediction model is obtained by training according to the historical price of the target digital currency and the historical values of the prediction variables; the trend prediction model is a model constructed based on vector autoregression.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the method for predicting a trend of digital currency prices as set forth in any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for predicting a trend of digital currency prices according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910772405.XA CN112418833A (en) | 2019-08-21 | 2019-08-21 | Method and device for predicting price change trend of digital currency |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910772405.XA CN112418833A (en) | 2019-08-21 | 2019-08-21 | Method and device for predicting price change trend of digital currency |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112418833A true CN112418833A (en) | 2021-02-26 |
Family
ID=74779721
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910772405.XA Pending CN112418833A (en) | 2019-08-21 | 2019-08-21 | Method and device for predicting price change trend of digital currency |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112418833A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024000152A1 (en) * | 2022-06-28 | 2024-01-04 | Chan Kin Kwan | A system and a method for analysing a market of exchangeable assets |
-
2019
- 2019-08-21 CN CN201910772405.XA patent/CN112418833A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024000152A1 (en) * | 2022-06-28 | 2024-01-04 | Chan Kin Kwan | A system and a method for analysing a market of exchangeable assets |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fritsch et al. | Business cycles, unemployment and entrepreneurial entry—evidence from Germany | |
| Castro et al. | Measuring and testing for the systemically important financial institutions | |
| Caglayan et al. | Firm productivity, exchange rate movements, sources of finance, and export orientation | |
| Hauner et al. | The interest group theory of financial development: Evidence from regulation | |
| Bellotti et al. | Retail credit stress testing using a discrete hazard model with macroeconomic factors | |
| Kasch et al. | Volatility threshold dynamic conditional correlations: An international analysis | |
| Nymoen et al. | Equilibrium unemployment dynamics in a panel of OECD countries | |
| Gulay et al. | Comparison of forecasting performances: Does normalization and variance stabilization method beat GARCH (1, 1)‐type models? Empirical Evidence from the Stock Markets | |
| Theodoridis et al. | News shocks and labour market dynamics in matching models | |
| Magaji et al. | Impact of international trade on economic growth: the granger causality test approach | |
| Nițescu et al. | The economic growth catalyzers at the European level, in the context of the 2008 financial crisis | |
| CN112418833A (en) | Method and device for predicting price change trend of digital currency | |
| Broer | Neoclassical theory and empirical models of aggregate firm behaviour | |
| Wiranto et al. | Analysis of LQ45 Index Stock Movements using the ARIMA Method during Uncertainty in Global Economic Conditions in 2023 | |
| Awe et al. | A time varying parameter state-space model for analyzing money supply-economic growth nexus | |
| Sefton et al. | Simulating household savings and labour supply: an application of dynamic programming | |
| Ventosa‐Santaulària et al. | Recessions and potential GDP: The case of Mexico | |
| Osmani et al. | Dynamic interactions between major macroeconomic aggregates in Albania. A vector autoregression approach | |
| US20220067460A1 (en) | Variance Characterization Based on Feature Contribution | |
| Kydland et al. | Contribution to dynamic macroeconomics: The time consistency of economic policy and the driving forces behind business cycles | |
| Humbatova | The impact of oil prices on state budget income and expenses: case of Azerbaijan | |
| Fioramanti et al. | The Econometrics of the EU Fiscal Governance: is the European Commission methodology still adequate? | |
| JP2021502653A (en) | Systems and methods for automated preparation of visible representations regarding the achievability of goals | |
| Wahyuningsih et al. | Effect of Macroeconomic Variables and Tax Rates on the Six ASEAN Countries Tax Ratio | |
| Sharpe | Lognormal model for stock prices |
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 |