Specific embodiment
In order to make those skilled in the art more fully understand the technical solution in this specification, below in conjunction with this explanation
Attached drawing in book embodiment is clearly and completely described the technical solution in this specification embodiment, it is clear that described
Embodiment be only this specification a part of the embodiment, instead of all the embodiments.The embodiment of base in this manual,
Every other embodiment obtained by those of ordinary skill in the art without making creative efforts, all should belong to
The range of this specification protection.
Due to the fast development paid face to face under overseas line and overseas receive single business etc., payment application, such as Alipay,
As a kind of one of means of payment that its people generally use, in the world, support under different currency buyer and trade company it
Between Zhi Fuyu collect money relationship.And in the foreign exchange business of the means of payment such as Alipay, it needs next by purchasing in advance
Each dealing amount of foreign exchange of a billing cycle of buying foreign exchange carries out profit and loss control to reduce potential exchange rate opening fluctuation risk.In turn,
In order to carry out profit and loss control, the dealing amount of foreign exchange to each billing cycle of buying foreign exchange is needed to carry out more accurate prediction.It is general and
Speech, multiple business days when billing cycle of buying foreign exchange is usually a business day and festivals or holidays differ.Current daily prediction of buying foreign exchange
Work faces following difficult point and problem: 1, data available dimension is insufficient: dealing amount of foreign exchange is by emergency event or promoting activities
Etc. factors be affected, but effectively acquire the difficult of relevant information in advance;2, data cleansing difficulty is big, washes out
The reliability of characteristic need a large amount of verifyings.Especially when festivals or holidays being related to the always amount of the buying foreign exchange prediction of multiple business days, more
It is to be exaggerated these problems.And when the period span of buying foreign exchange for needing to predict is multiple business days, the difficulty of prediction work rises sharply,
Most models are specifically referred to for the performance situation of missing periodicity T+n prediction.For example, during the Spring Festival, overseas under scene
Longest billing cycle of buying foreign exchange is 5 business days (such as 15 to 19 2 months 2018, because of the Spring Festival after No. 15
Factor cannot buy foreign exchange).Such as the strategy and in this scenario for " foreign exchange is competing to optimal " that supports Alipay to pay face to face
The profit and loss is controlled, the coverage accuracy rate bought foreign exchange in advance is required very high.Simultaneously because the relationship of service competition, during the entire Spring Festival
Unpredictable factor it is very more.So current, prediction work importance is strong on the whole in the industry, and difficulty is big.
There is also some prediction models for trade company's trading volume in currently available technology, such as integrate rolling average and return certainly
Return model (Autoregressive Integrated Moving Average, ARIMA), it can be by the data mould of Parametric drive
Type;Or Support vector regression (Support Vector Regression, SVR) model etc..Due to the time sequence under this scene
Column data is difficult the mathematical formulae description parsed with parameterized model, causes ARIMA predictablity rate not ideal enough.Other moulds
Type such as SVR or artificial neural network (Artificial Neural NetWork, ANN), but the input of these models require for
The characteristics of shortage of data is serious under static and the historical data and this scene of fixed length and data difficult description forms opposition, causes
It cannot export effective reliable result.
Shot and long term memory network (Long Short-Term Memory, LSTM), is a kind of time recurrent neural network, it
Can Chief Learning Officer, CLO observe value sequence and can effectively predicted time sequence, be suitable for being spaced and prolonging in processing and predicted time sequence
Critical event relatively long late.LSTM may be considered a kind of special RNN (Recurrent Neural Network,
RNN, Recognition with Recurrent Neural Network), it can be disappeared and gradient explosion issues with the gradient in effective solution long sequence data training process.
In the internal structure of LSTM, transmission state can be controlled by gating state, remember to need memory for a long time, forget not
Important information;A kind of memory stacked system compared to common RNN is more suitably applied to need the operation of " long-term memory "
Task, as shown in Figure 1.Fig. 1 is a kind of structural schematic diagram of LSTM model, and there are three doors: input gate 1 forgets door 3 and out gate
2.These decide whether to allow new input (input gate), delete information, because its inessential (forgetting door) or allowing it current
Time (out gate) does not influence output.Therefore, this specification embodiment provide a kind of dealing amount of foreign exchange prediction technique, device and
System can use LSTM model and be able to maintain that the key characteristic of good internal state is more efficient, accurately predict trading volume.
Certainly, it can also be passed using the model of modification or the improvement of LSTM, or other class times in this specification other embodiments
Return neural network, such as GRU (Gated Recurrent Unit, GRU).In some embodiments of this specification, GRU generation is being used
When for LSTM, can by GRU by the input gate of LSTM, forget door, out gate and become updating door and reset door, and by cell-like
A state is merged into state and output.
Fig. 2 is a kind of structural schematic diagram that dealing amount of foreign exchange forecasting system is carried out based on LSTM that this specification provides.Such as
System shown in Fig. 2 may include:
Public sentiment module 201: corresponding public feelings information can be obtained by crawler technology.The public feelings information is often referred to can
Directly or indirectly to influence the information of trading volume, such as common may include that the payments such as trade company's action message, Unionpay, wechat are flat
The action message of platform, political impact (such as: government policy, vital emergent event).These public feelings informations can be to normal state
Border agiotage trading volume has an impact or impacts.
Real time data module 202: can be with real-time synchronization trade company trading volume data.It can be special by accessing in concrete implementation
Data platform or service platform synchronize trade company's trading volume.
Business data module 203: it can be handed over the relevant business datum of synchronous service information, such as the country information of transaction
Easy stroke count, crowd's label of trading, merchant type label etc..Synchronizing for business datum can realize by data warehouse table, can
With from different data warehouse synchronous service data.Data warehouse (also referred to as counting storehouse) can be for store transaction data, be
(the Subject Oriented) of one subject-oriented, integrated (Integrate), metastable (Non-Volatile),
Reflecting history changes the data acquisition system of (Time Variant), can be used for supporting administrative decision.Data warehouse is a storage
The warehouse of data, its inside house various data, these data are next according to some structures, rule generally according to needing
Tissue and storage, realizing can be constructed by transaction platform or be constructed by other data platforms, and this specification embodiment is not made to have
Body limits.
Data prediction and sorting module 204: can be by real-time data synchronization template, real-time data synchronization module, business
The data that data simultaneous module obtains are converted into available numerical information by the rule set.It simultaneously can also be by above-mentioned carriage
Two in feelings information, real-time deal amount data, transaction business information etc. are multinomial or whole progress data Layer fusions.Then lead to
Cross arrangement, convert the data into meet prediction model enter ginseng standard (it is required that) data format.
By data prediction and sorting module 204, the data that can be will acquire (may include historical trading data or reality
When data) be processed into required for information dimension and meet prediction model and enter to join the data format of standard.
Correlating module 205: can to by data prediction and sorting module obtain it is each enter ginseng data
Feature is analyzed, the correlation for parsing and determining it with regressor.Judged into ginseng according to the correlation if appropriate for mould
Type uses.
The method of correlation analysis can be selected according to actual needs, such as can pass through drawing, mathematical statistics side
Method analyzes the correlation between initial transaction data and regressor.
In this specification one embodiment, Pearson correlation coefficient method PCC (Pearson correlation can be passed through
) or Spearman rank correlation method SRCC (Spearman's rank correlation coefficient) coefficient
Calculate the correlation between initial transaction information and regressor.PCC can indicate degree of correlation between a kind of two variables of measurement
Method, it is a value between 1 and -1, wherein 1 indicates variable perfect positive correlation, and 0 indicates unrelated, and -1 indicates complete
It is negatively correlated.SRCC is mainly used for solving the problems, such as to claim name data related to alphabetic data, suitable for two column variables, and have etc.
Grade variable property has the data of linear relationship.The accuracy of data screening can be improved using PCC or SRCC, be subsequent transaction
Amount prediction provides accurate data basis.
Prediction model 206 based on LSTM: may include training and two stages of prediction in model prediction processing.This
Embodiment is in the processing stage being trained using LSTM model to model, it can be determined that whether reaches target number of repetition n;
If not up to n, judge whether to reach the frequency of training m for keeping learning rate;Learning rate is kept if not up to, to learning rate
Decay multiplied by decay factor.Certainly, in addition to the model training index of above-mentioned setting, other model trainings can also be set
It is required that.Then data input model can be obtained in batches from training set queue to be trained.Collect team from verifying after the completion of training
Data input model is obtained in column in batches to be verified.
The data acquisition system that uses of prediction model training can become training dataset, may include above-mentioned public sentiment module,
The data of business data module etc..Corresponding label has can be set in training data.In model training stage, can be used predetermined
The data of number window are trained, and one of window can be the data of one day or multiple days.It, can be with such as in an example
The data that one window is 7 days are set, selects the data of 300 windows of time point to be recently trained model, obtains pre-
Survey model.When carrying out trading volume prediction using prediction model, it can choose a nearest time window and predicted, obtained
Prediction result.
Certainly, training dataset can be updated over time, can be to training number when obtaining new data
It is updated according to collection, new data can be incorporated into training data concentration.The training dataset of update can be used to prediction mould
Type is trained, and updates prediction model.When being predicted using prediction model, updated prediction model can be used and carry out in advance
It surveys.In practical application, such as by the unit time of day count trading volume because every day all can once data output, training
Can all there be new data input in data set daily, therefore prediction model is trained using new training dataset, after training
Prediction model may be will be updated, such as the variation of transaction stream can cause the Parameters variation of prediction model, so that prediction model
The variation of transaction stream is followed in time.Therefore in some embodiments, the prediction model actually used when being predicted every time is
Newest prediction model based on updates such as newest transaction data, public feelings informations.
Fig. 3 is a kind of schematic diagram that model training and prediction Processing Example are carried out using LSTM that this specification provides.
The present embodiment can take 7 days as a time window, and nearest time window is on January -23 on the 17th, it is corresponding as shown in figure 3,
A upper time window is on January -22 on the 16th, then a upper time window is on January -21 on the 15th, and so on.Model training
Stage can choose 300 time windows or 400 time windows etc. and be trained.Some embodiments that this specification provides
In, specifically without limitation to the number of training window, it can be adjusted according to the actual situation in the training process.And one
It must be identical time cycle, such as 10 nearest time window periods that a little embodiments, which do not limit each time window, yet
It can be 10 days data, remaining window period is then still 7 days.
Assuming that prediction result is y=f (a with the functional relation for entering ginseng1,a2,a3,…,a100), one, which shares 100, enters ginseng.anIt is the parameters (entering ginseng) for participating in model training chosen, for an, such as a1, can be in corresponding time window
The average value in each day, summing value, or the value obtained by other processing transformation/conversions.Entering ginseng can be by above-mentioned data
Pretreatment and sorting module 204 carry out data fusion and arrangement and correlating module 205 carries out correlation analysis screening
What is determined afterwards enters ginseng.After the training data for obtaining on January -23 on the 17th, each value for entering ginseng of this time window is determined, pass through
Cross the predicted value y0 (the prediction trading volume on January 24) that function handles this proper time window.Similarly, January -22 days on the 16th
The available predicted value y1 of data, the available predicted value y2 of the data in -21 days on the 15th January, and so on.January 24 is served as
Available after 24 points of day to obtain real trade amount on January 24, this real trade amount can be used as above-mentioned correlation analysis
Regressor.Such as real trade amount is 100 three days ago, day before yesterday real trade amount is 80, and real trade amount yesterday is 85, in model
Training stage can learn the relationship of real data and regressor to the regressor updated every time, and can be corresponding based on this
Adjustment model parameter, update/correction model.In this way, this specification embodiment can be by accessing real-time prediction data, it can
To capture the trend currently traded automatically, and predicted value is adjusted, so that prediction result is more accurate, reliable.
In some embodiments, business datum or other real time datas can also update mould by the real time data being specially arranged
Block 207 is passed to LSTM prediction model, accelerates model training or prediction processing.
In another embodiment that this specification provides, predicted value and other parameters can be compared, and can tie
Public feelings information, such as trade company's activity, policy change etc. are closed, judges whether predicted value is reliable.And it can be adopted according to the result of judgement
It is adjusted with corresponding scheme, in this way, with the predicted value that effective guarantee model exports mentioned significantly in reliable range
The high reliability of prediction result.Therefore, in another embodiment, exception monitoring module 208 has been can be set, in concrete application
May include:
The comparison parameter of the selections such as the real trade amount of the same period in upper period, this Periodic Mean, trading volume yesterday can be passed through
This predicted value is detected, confirms whether this predicted value is abnormal.It is abnormal if it exists, then it can be based on presetting correlation
Coefficient carries out corresponding adjustment to predicted value.In a specific example, if this predicted value is equal higher than the same period in upper period, this week
The certain ratios such as value, trading volume yesterday, and the carriage for not being found for example that trade company's activity, policy change etc. trading volume being caused to rise sharply
Feelings information, then at this time can be by predicted value multiplied by correlation coefficient r 1, to ensure that output valve is in confidence band;Likewise, such as
Fruit predicted value lower than the certain ratios such as the same period in upper period, this week mean value, trading volume yesterday, and be not found for example that trade company's activity,
Policy change etc. leads to the public feelings information of trading volume rapid drawdown, then at this time can be defeated to ensure by predicted value multiplied by correlation coefficient r 2
Value is in confidence band out.The correlation coefficient r 1, correlation coefficient r 2 can be configured according to business scenario demand.Example
Such as, if detection the result is that predicted value is higher, r1 can be set less than 1;Correspondingly, if detection the result is that predicted value
It is relatively low, then correlation coefficient r 2 can be set greater than 1.The value of specific above-mentioned related coefficient rule of thumb or can compare parameter
It is configured with the difference of predicted value.
It should be noted that except the comparisons such as the same period in upper period of upper embodiment description, this Periodic Mean, trading volume yesterday ginseng
It measures, other comparison parameters can also be chosen in other embodiments to detect predicted value with the presence or absence of exception, such as nearest three
Mean value, this week intermediate value of all same periods etc..On the other hand, exception is determined whether there is with comparing parameter and be compared in predicted value
When, it can choose predicted value in some embodiments and compare parameter with one of them and be compared, such as only compared with this week mean value.
In the other embodiments that this specification provides, predicted value can be compared with multiple parameters that compare, as described above simultaneously
It was compared with the upper same period in period, this week mean value, trading volume yesterday.
During comparing processing, can judge as described in above-described embodiment choose comparison parameter whether be greater than or
Less than the certain ratio of predicted value, such as 25%.Directly comparison prediction value and parameter can also be compared in other embodiments
Absolute difference judges predicted value with the presence or absence of abnormal according to the size of absolute difference.
In one embodiment that this specification provides, predicted value with it is multiple compare parameter compared in processing, predicted value with it is right
Determine the predicted value with the presence or absence of abnormal as a result, different judgment conditions therefore can be set there are a variety of than what data compared.
Such as in one embodiment, it can be set that same period last week, this week mean value, that three comparison parameters of trading volume yesterday are below this is pre-
When measured value 25%, then it can be determined that today predicted value is higher, there are exceptions;Or above three compares parameter to be above this pre-
Determine that predicted value today is relatively low when measured value 30%, 1 related coefficient can be greater than multiplied by one.Certainly, in other embodiments
Can also part correlation data abnormal, such as this predicted value in an example is determined when being higher or lower than predicted value certain proportion
Although being lower than same period last week and this week mean value 25%, it is only below trading volume yesterday 5%, then can be set and determine that result is this
Predicted value does not occur exception.
In other embodiments, different correlation datas also can be set different ratio or difference, such as this
Predicted value exists 30%, compared to this week mean value 20%, compared to trading volume yesterday compared to same period last week fluctuation range
20% and without that the warning message for may cause trading volume and rising sharply or die-offing (public feelings information) is detected, then can determine this
Predicted value is within confidence band.
The exception monitoring module of multiple embodiments can be with the robustness of effective guarantee prediction model output valve among the above.
One or more embodiments that this specification provides, the prediction model based on LSTM simultaneously access real-time prediction data,
Current trading volume variation tendency can be captured automatically, and output predicted value is adjusted according to variation tendency.Secondly, what this specification provided
Some schemes can be monitored trade company's trading volume, and external public sentiment inclination can be acquired by public sentiment module, knows that trade company is living
The emotionally information of condition, policy change or other influences trading volume, and the information is automatically fed to prediction model to adjust prediction
Value.
A kind of dealing amount of foreign exchange prediction technique is provided in this specification one embodiment, by obtaining transaction letter in real time
Breath is trained adjustment to prediction model using real-time deal information, exports trading volume predicted value, realize the accurate of trading volume
Prediction, reduces the risk of loss abroad paid.
Further, the system can also include decision-making module 209, and the decision-making module can be according to prediction mould
Real-time deal trend and historical trend that type obtains, business KPI (Key Performance Indicator, KPI, Key Performance
Index) etc. information, predicted value is adjusted by above-mentioned related coefficient, and can be according to predicted value output optimal execution strategy.Pass through
Prediction model obtains reliable and accurate predicted value, under normal circumstances, this predicted value be generate trading strategies important references according to
According to.It therefore, can be according to the different designs of business and scene using the predicted value that the example scheme that this specification provides obtains
Obtain different trading strategies.Further can be in conjunction with external policy, such as the bottom-line plan (alternative) of design etc., it can
Optimal execution strategy is selected with finishing screen.Therefore, system described in another embodiment can also include plan outside bottom-line plan
Slightly module 210, can be used for storing preset trading strategies.The preset trading strategies of the storage may include
Trading strategies based on the predicted value formulation that default value-LSTM prediction module obtains, also may include being not suitable for the predicted value
Obtained trading strategies.Such as foreign exchange transaction strategy can be collected, such as this pre- only 60% position of covering of buying foreign exchange, it is whole to trade
Amount need to be lower than the trading strategies in the case of true sale amount.Bottom-line plan external policy module 210 can be adjusted for decision-making module 209
With making the comprehensive multi-exchange strategy of the decision-making module 209 or further combined with the optimal friendship of the associated informations output such as business KPI
Easily strategy.The system can also include output and display module 211, can be used for carrying out the predicted value of prediction model output
It writes table output and/or is shown prediction result, the content for exporting and showing can submit to business side and use, such as can be with
It is realized by modes such as online report, timing mails.
Multiple embodiments of the present embodiment are retouched above by a complete detailed forecasting system structure
It states.It should be noted that including other supplemental functionalities or can also may not necessarily want complete in some embodiments
All modular structures, or the deformation based on system structure shown in Fig. 2, transformation including system shown in Figure 2 etc..Utilize this theory
Bright book example scheme is trained prediction model by accessing real-time prediction data, captures current transaction stream automatically simultaneously
The model that timely updates adjusts predicted value, so that trading volume predicted value is more accurate, reliable.It is existing to be also based on predicted value generation
Trading volume is inclined to the optimal execution strategy under (trend), reduces potential exchange rate opening fluctuation risk, preferably carries out profit and loss control
System.
Description based on the embodiment above, this specification provide a kind of dealing amount of foreign exchange prediction technique, and Fig. 4 is this explanation
The flow diagram of trading volume prediction technique in book one embodiment.As shown in figure 4, the method may include:
S0: obtaining the prediction data in the prediction window phase, and the prediction data includes at least public feelings information, customer transaction
One of amount, business datum;
S2: the prediction data is pre-processed and is converted into meeting the data format that prediction model enters ginseng requirement, is obtained
To prediction input data;
S4: being handled the prediction input data using the trading volume prediction model of building, obtains trading volume prediction
Value, the trading volume prediction model include: constructed based on time recurrent neural network, and, based on the newest day of trade
Output data updates the training set of the trading volume prediction model, and utilizes the training data pair in updated training set
The trading volume prediction model is trained.
The present embodiment can train obtained prediction model to carry out trading volume prediction by time recurrent neural network, when
Between recurrent neural network being capable of learning time longer observation value sequence and can effectively predicted time sequence.This programme is by it
It is able to maintain that the key characteristic of good internal state, can be realized effective, reliable trading volume prediction in foreign exchange business.
As previously mentioned, the present embodiment can choose a variety of time recurrent neural networks, such as GRU.In preferred embodiment,
It can choose LSTM (shot and long term memory network), the friendship that can be preferably applicable in foreign exchange business in prediction model modeling
Easily amount prediction.Therefore, in another embodiment of the method, the trading volume prediction model includes based on LSTM (shot and long term note
Recall network) model construction obtains.In this specification embodiment, prediction model can be shot and long term memory network model LSTM
(Long Short-Term Memory Network).Certainly, GRU (Gated can also be used according to actual needs
Recurrent Unit), it can be understood as a type time recurrent neural network.GRU can be by the input gate of LSTM, forgetting
Door, out gate become updating door and reset door, and location mode and output are merged into a state, and speed training can reach
To theoretic promotion.Increasingly complex feature is extracted by self-characteristic using LSTM model, to reach better fitting
Effect.The predictablity rate for being better than original system is obtained in having the prediction under trade company's moving obstacle, and is generated existing transaction and inclined
Downward optimal execution strategy.
It, can be according to the transaction data got daily such as in this specification embodiment: real-time deal amount data, public sentiment letter
Breath, transaction business information carry out new training to the previous day trained prediction model and update, and obtain updated prediction model.
The training update method of model can be with reference to the process of model training in above-described embodiment, and details are not described herein again.
The prediction model training stage can be with after obtaining public feelings information, historical trading volume, the training datas such as business datum
Data are pre-processed, are merged, are converted etc. to be processed into and meeting prediction model and enter the data format that ginseng requires.Business datum, carriage
The data information of the polymorphic types such as feelings data, which may generate, largely enters ginseng, also will use when prediction model is trained and predicts multiple
Enter ginseng.Enter in ginseng from what the various data informations that disparate databases obtain and arrange were formed, it is understood that there may be some enters ginseng
Data characteristics and without using prediction model or some enter the reflected information of ginseng and predicted with trading volume in this implementation and uncorrelated
Or correlation very little (influence to trading volume can be ignored).Therefore, the trading volume prediction technique that this specification provides
Another embodiment in, as shown in figure 5, Fig. 5 be this specification provide another embodiment in trading volume prediction technique
Flow diagram, training data is by the pretreatment and is converted into after meeting the data format that prediction model enters ginseng requirement,
The method can also include:
S6: regressor corresponding in the data characteristics and the trading volume prediction model for entering ginseng is carried out at correlation analysis
Reason determines and enters ginseng suitable for the trading volume prediction model;
Correspondingly, being suitable for the trading volume using the determination in the training managing of the trading volume prediction model
The ginseng that enters of prediction model is trained.
In the present embodiment can by PCC or SRCC to it is each enter ginseng data characteristics analyze, parse itself and recurrence
The correlation of amount, and can judge that this enters the data characteristics of ginseng if appropriate for using in current trading volume prediction model accordingly
In.This embodiment scheme to it is a variety of enter ginseng carry out data screening, can choose correlation it is higher enter ginseng use it is pre- in trading volume
In the model of survey, so that prediction model is relatively reliable, prediction result is more accurate.
This specification is provided in another embodiment of the method, and predicted value and choosing can will be obtained using prediction data
The comparison parameter taken is compared, detects, and whether the predicted value of judgment models prediction is abnormal.If abnormal, phase can be carried out
The adjustment answered can be in this way in confidence band with the predicted value of effective guarantee trading volume prediction model final output.Specifically
, in another embodiment of the method, the method can also include:
S802: detecting the trading volume predicted value using the comparison parameter of selection, confirms the trading volume prediction
Value is with the presence or absence of abnormal;
S804: if it is determined that the trading volume predicted value exist it is abnormal, then according to the result of the detection by the trading volume
Predicted value is adjusted rear predicted value multiplied by corresponding related coefficient.
Fig. 6 is the method flow schematic diagram of another embodiment of the method that this specification provides, as shown in fig. 6,
The method can also include:
S806: the corresponding public feelings information of the predicted time window phase is obtained;
If correspondingly, the confirmation trading volume predicted value with the presence or absence of it is abnormal include: the trading volume predicted value with
The absolute difference for comparing parameter is more than or less than preset difference value or preset percentage, and is not had based on public feelings information determination
There is the information that trading volume can be caused to rise sharply or die-off, it is determined that the trading volume predicted value exists abnormal.
As described above, the trading volume predicted value that can choose different comparison parameter and model prediction compares, inspection
Trading volume predicted value is surveyed with the presence or absence of abnormal.Different comparison parameters usually have different testing results, therefore, the method
Another embodiment in provide it is a kind of compare parameter selection embodiment.In specific one embodiment, the selection
It may include at least one of following for comparing parameter:
S8020: the same period in upper period trading volume corresponding with forecast date, this Periodic Mean, the friendship of a last day of trade
Yi Liang.
The same period in the upper period trading volume corresponding with forecast date, the same period in upper period typically refer to and this prediction
The predicted time window corresponding upper time window phase on the corresponding date trading volume.Under some cases, described upper one
The trading volume of a day of trade and the of even date trading volume that the yesterday, trading volume referred to.Specifically show for example, current pre-
Surveying time window is 17-23 days January (7 days are a time window phase or period), and the trading volume for predicting January 24 is P, then institute
The corresponding same period in the upper period trading volume of the forecast date stated (same period in upper period, on January -22 on the 16th prediction January 17) is January
Trading volume P1 on the 17th, this week mean value were the mean value P2 of daily real trade amount in this 7 days -23 days on the 17th January, a upper day of trade
Trading volume be then January 23 trading volume P3.
When whether the trading volume P in the January 24 of detection prediction is abnormal, P1, P2, P3 can be compared with P respectively,
Such as P1/P is taken, P2/P, P3/P.If P1/P, P2/P, P3/P be above it is certain very, such as 25%, and in real time obtain
There is no the warning messages that trade company's activity rises sharply in public feelings information, it may be considered that 24 daily trading volume in January of this prediction is higher,
Belong to exception.Prediction can be obtained into 24 daily trading volume P in January multiplied by a related coefficient a2 less than 1, such as a2=at this time
0.8.Judging that trading volume predicted value deposits when abnormal according to testing result multiplied by corresponding related coefficient, can effectively control defeated
Value is in confidence band out.Certainly, other comparison parameters can also be chosen in other embodiments, or with the comparison of selection
One or partial testing result in parameter as judge trading volume predicted value whether Yi Chang foundation.In addition, for difference
Comparison parameter also can choose the different very absolute differences compared with trading volume predicted value.
After obtaining trading volume predicted value, the trading volume predicted value of acquisition being sent to business side, (such as: table is opened up online
Show or the modes such as timing mail is sent be sent to business side), business side can buy foreign exchange according to the trading volume predicted, be state
It converts remittance business and data basis is provided in border.
Business side directly can be shown or be sent to by the trading volume that prediction obtains in some embodiments to use, it is another
Optimal execution strategy can also be determined based on the trading volume predicted value in a little embodiments, be subsequently transmitted to corresponding industry
Business side uses.Therefore, in this specification one embodiment, as shown in fig. 7, the method can also include:
S10: optimal execution strategy is determined based on the trading volume predicted value.
The trading strategies that the trading volume predicted value that the optimal execution strategy can be exported based on prediction model generates,
It may include the optimal transaction plan determined under the aid decisions scheme such as external transaction strategy, bottom-line plan external policy of combination
Slightly.Trading strategies either under some predetermined cases, such as this pre- buy foreign exchange only 60% position of covering or whole transaction
Amount can need to provide the optimal execution strategy of similar expert system in conjunction with prediction scheme of revealing all the details (at this time lower than true sale amount etc.
It can be the optimal case in candidate scheme).
This specification embodiment, by access real-time transaction data, (daily business datum, trade company's trading volume, public sentiment are believed
The update of the training datas set such as breath), current transaction stream can be captured automatically adjusts predicted value.Some implementations of this specification
Example can be monitored trade company's trading volume, once the situation for having transaction abnormal can the adjusting transaction of automatic feedback prediction model
Measure predicted value.By acquiring external public feelings information, to know trade company's activity condition, and the information is passed to prediction model tune automatically
Save trading volume predicted value.By the robustness of abnormality detection assurance model output valve, external transaction is then combined by integrated decision-making
Strategy can obtain optimal execution strategy instantly.This specification embodiment has accessed real time data, expansion factor of payment application etc.
The full dose factor carries out data fusion, expanded target signature space and can training information, model part then can use LSTM mould
Type extracts increasingly complex feature by self-characteristic, to reach better fitting effect.In the case where there is trade company's moving obstacle
The predictablity rate for being better than original system is obtained in prediction, and generates the existing vergens optimal execution strategy of transaction.
Various embodiments are described in a progressive manner for the above method in this specification, identical between each embodiment
Similar part may refer to each other, and each embodiment focuses on the differences from other embodiments.Correlation
Place illustrates referring to the part of embodiment of the method.
Embodiment of the method provided by this specification embodiment can be in mobile terminal, terminal, server, block
It is executed in chain network, distributed network or similar arithmetic unit.For running on the server, Fig. 8 is using this hair
A kind of hardware block diagram of dealing amount of foreign exchange predictive server of bright embodiment.As shown in figure 8, server 10 may include one
(processor 100 can include but is not limited to Micro-processor MCV or can compile a or multiple (one is only shown in figure) processor 100
The processing unit of journey logical device FPGA etc.), memory 200 for storing data and the transmission mould for communication function
Block 300.This neighborhood those of ordinary skill is appreciated that structure shown in Fig. 8 is only to illustrate, not to above-mentioned electronic device
Structure causes to limit.For example, server 10 may also include the more or less component than shown in Fig. 8, such as can also wrap
Other processing hardware are included, such as database or multi-level buffer, GPU, or with the configuration different from shown in Fig. 8.
Memory 200 can be used for storing the software program and module of application software, such as the friendship in this specification embodiment
The easily corresponding program instruction/module of amount prediction technique, the software program that processor 100 is stored in memory 200 by operation
And module, thereby executing various function application and data processing.Memory 200 may include high speed random access memory, may be used also
It is stored including nonvolatile memory, such as one or more magnetic storage device, flash memory or other nonvolatile solid states
Device.In some instances, memory 200 can further comprise the memory remotely located relative to processor 100, these are long-range
Memory can pass through network connection to terminal 10.The example of above-mentioned network includes but is not limited to internet, in enterprise
Portion's net, local area network, mobile radio communication and combinations thereof.
Transmission module 300 is used to that data to be received or sent via a network.Above-mentioned network specific example may include
The wireless network that the communication providers of terminal 10 provide.In an example, transmission module 300 includes that a network is suitable
Orchestration (Network Interface Controller, NIC), can be connected by base station with other network equipments so as to
Internet is communicated.In an example, transmission module 300 can be radio frequency (Radio Frequency, RF) module,
For wirelessly being communicated with internet.
Based on the description of trading volume prediction technique and system related embodiment described above, this specification is one or more
Embodiment also provides a kind of dealing amount of foreign exchange prediction meanss.The device may include having used described in this specification embodiment
The system (including distributed system) of method, software (application), module, component, server, client etc. simultaneously combine necessary reality
Apply the device of hardware.Based on same innovation thinking, the device in one or more embodiments that this specification embodiment provides is such as
Described in the following examples.Since the implementation that device solves the problems, such as is similar to method, this specification embodiment is specific
The implementation of device may refer to the implementation of preceding method, overlaps will not be repeated.It is used below, term " unit " or
The combination of the software and/or hardware of predetermined function may be implemented in person's " module ".Although device described in following embodiment is preferable
Ground is realized with software, but the realization of the combination of hardware or software and hardware is also that may and be contemplated.
Specifically, Fig. 9 is the modular structure signal of dealing amount of foreign exchange prediction meanss one embodiment that this specification provides
Figure, as shown in figure 9, the trading volume prediction meanss provided in this specification may include:
Data acquisition module 20, can be used for obtaining the prediction data in the prediction window phase, and the prediction data is at least wrapped
Include one of public feelings information, customer transaction amount, business datum;
Preprocessing module 21 can be used for pre-processing the prediction data and being converted into meeting prediction model and enter ginseng
It is required that data format, obtain prediction input data;
Prediction module 22 can be used for using the trading volume prediction model of building to the prediction input data
Reason, obtains trading volume predicted value, the trading volume prediction model include: constructed based on time recurrent neural network, and,
Output data based on the newest day of trade updates the training set of the trading volume prediction model, and utilizes updated training set
Training data in conjunction is trained the trading volume prediction model.
The trading volume prediction meanss that this specification embodiment provides, can construct the time recurrent neural network of such as LSTM
Prediction model, can Chief Learning Officer, CLO observe value sequence and can effectively predicted time sequence, by the transaction data that gets to pre-
It surveys model and is trained update, trading volume predicted value is predicted according to updated prediction model.It is predicted in this embodiment scheme
The trading volume predicted value of model output is the predicted value exported according to historical trading trend, can effectively improve trading volume prediction
Accuracy provides accurate data basis for foreign exchange transaction, reduces transaction risk.
In one embodiment of described device, trading volume prediction model used in the prediction module 22 is based on shot and long term
Memory network model construction obtains.It is of course also possible to select other time recurrent neural networks, such as GRU.
Figure 10 is the structural schematic diagram of trading volume prediction meanss in another embodiment of this specification, as shown in Figure 10, institute
Stating device can also include:
Correlation processing module 23 can be used for training data by the pretreatment and be converted into meeting prediction model and
After joining desired data format, phase is carried out to regressor corresponding in the data characteristics and the trading volume prediction model for entering ginseng
The analysis processing of closing property, determines and enters ginseng suitable for the trading volume prediction model;
Correspondingly, the prediction module 22 is in the training managing of the trading volume prediction model, it is suitable using the determination
The ginseng that enters for the trading volume prediction model is trained.
On the basis of the above embodiments, it can use Pearson correlation coefficient method or Spearman rank correlation method calculate
Correlation between the mode input data and corresponding regressor.
Figure 11 is the structural schematic diagram in another embodiment of this specification described device, as shown in figure 11, described device
Can also include:
Monitoring adjustment module 24, can be used for detecting the trading volume predicted value using the comparison parameter of selection,
Confirm the trading volume predicted value with the presence or absence of abnormal;
And determining that the trading volume predicted value deposits when abnormal, according to the result of the detection by the trading volume
Predicted value is adjusted rear predicted value multiplied by corresponding related coefficient.
The trading volume prediction meanss that this specification embodiment provides, can be by the comparison parameter of the trading volume of prediction and selection
It is compared, the trading volume of judgment models prediction is with the presence or absence of abnormal.Once the situation for having transaction abnormal can automatic feedback
To prediction model, predicted value is adjusted, improves the accuracy of trading volume prediction.
In another embodiment of this specification, described device can also include:
Public feelings information module 25 can be used for obtaining the corresponding public feelings information of the predicted time window phase;
If correspondingly, monitoring adjustment module 24 confirms that the trading volume predicted value may include: with the presence or absence of abnormal
The trading volume predicted value and the absolute difference for comparing parameter are more than or less than preset difference value or preset percentage, and are based on
The public feelings information determines the information that trading volume can not caused to rise sharply or die-off, it is determined that there are different for the trading volume predicted value
Often.
Comparing parameter can be by multiple choices, in one embodiment that this specification provides, the comparison parameter of the selection
May include at least one of following:
The same period in upper period trading volume corresponding with forecast date, this Periodic Mean, the trading volume of a last day of trade.
The above three that the present embodiment is chosen compares parameter, can effectively, accurately reflect the variation of Recent Activity amount
Whether trend, being combined according to above-mentioned variation tendency has the public feelings information for leading to trading volume larger fluctuation, can effectively ensure
The predicted value of model final output is within the scope of reasonable, reliable.
Figure 12 is the structural schematic diagram of trading volume prediction meanss in another embodiment of this specification, as shown in fig. 7, described
Device can also include:
As a result output module 26 can be used for determining optimal execution strategy based on the trading volume predicted value.
The optimal execution strategy exported by embodiment device, especially obtains in having the prediction under trade company's moving obstacle
It is substantially better than the predictablity rate of original system, and generates the existing vergens optimal execution strategy of transaction.
It should be noted that device described above can also include other embodiment party according to the description of embodiment of the method
Formula.Concrete implementation mode is referred to the description of related method embodiment, does not repeat one by one herein.
Various embodiments are described in a progressive manner for the above method in this specification, identical between each embodiment
Similar part may refer to each other, and each embodiment focuses on the differences from other embodiments.Correlation
Place illustrates referring to the part of embodiment of the method.
This specification embodiment also provides a kind of dealing amount of foreign exchange prediction processing equipment, comprising: at least one processor with
And the memory for storage processor executable instruction, the processor realize to include following step when executing described instruction
It is rapid:
The prediction data in the prediction window phase is obtained, the prediction data includes at least public feelings information, customer transaction amount, industry
One of data of being engaged in;
The prediction data is pre-processed to and is converted into meeting the data format that prediction model enters ginseng requirement, is obtained pre-
Survey input data;
The prediction input data is handled using the trading volume prediction model of building, obtains trading volume predicted value,
The trading volume prediction model include: constructed based on time recurrent neural network, and, the output based on the newest day of trade
Data update the training set of the trading volume prediction model, and using the training data in updated training set to described
Trading volume prediction model is trained.
Certainly, according to preceding method or Installation practice, the processing equipment can also include other embodiments,
Such as the trading volume prediction model includes the embodiment obtained based on shot and long term memory network model construction, based on comparison parameter
Determine whether abnormal and corresponding adjustment be adjusted the embodiment of rear predicted value, comparisons parameter for the same period in upper period trading volume,
This Periodic Mean, embodiment of trading volume of a last day of trade etc..According to preceding method or device or it can specifically show
What example described accordingly obtains, and all should belong within the practical range that the application is protected, and it is real not do embodiment one by one herein
Existing scheme is repeated.
This specification also provides a kind of dealing amount of foreign exchange forecasting system, and the system can be individual trading volume prediction system
System, can also apply in a variety of Data Analysis Services systems.The system can be individual server, also may include
One or more the methods of this specification or the server cluster of one or more embodiment devices, system (packet are used
Include distributed system, the privately owned chain in block chain, alliance's chain etc.), software (application), practical operation device, logic gates dress
It sets, quantum computer etc. and combine the necessary terminal installation for implementing hardware.The trading volume forecasting system may include at least
One processor and the memory for storing computer executable instructions, the processor realize above-mentioned when executing described instruction
The step of method described in one or more embodiment of anticipating or device.
Method or apparatus described in above-described embodiment that this specification provides or system can be realized by computer program
Service logic simultaneously records on a storage medium, and the storage medium can be read and be executed with computer, realizes that this specification is real
Apply the effect of scheme described by example.
This specification embodiment provide above-mentioned trading volume prediction technique or device or system can in a computer by
Reason device executes corresponding program instruction to realize, such as using the c++ language of windows operating system in the realization of the end PC, linux system
System is realized or other are for example realized using android, iOS system programming language in intelligent terminal, and based on quantum
Processing logic realization of calculation machine etc..
The storage medium of the memory may include the physical unit for storing information among the above, usually by information
It is stored again by the media in the way of electricity, magnetic or optics etc. after digitlization.It may include: to utilize that the storage medium, which has,
Electric energy mode stores the device of information such as, various memory, such as RAM, ROM;The device of information is stored in the way of magnetic energy such as,
Hard disk, floppy disk, tape, core memory, magnetic bubble memory, USB flash disk;Using optical mode storage information device such as, CD or
DVD.Certainly, there are also readable storage medium storing program for executing of other modes, such as quantum memory, graphene memory etc..
It is above-mentioned that this specification specific embodiment is described.Other embodiments are in the scope of the appended claims
It is interior.In some cases, the movement recorded in detail in the claims or step can be come according to the sequence being different from embodiment
It executes and desired result still may be implemented.In addition, process depicted in the drawing not necessarily require show it is specific suitable
Sequence or consecutive order are just able to achieve desired result.In some embodiments, multitasking and parallel processing be also can
With or may be advantageous.
As previously mentioned, system embodiment can also include other embodiments according to the description of related method embodiment,
Concrete implementation mode is referred to the description of corresponding method embodiment, does not repeat one by one herein.
All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for hardware+
For program class embodiment, since it is substantially similar to the method embodiment, so being described relatively simple, related place is referring to side
The part of method embodiment illustrates.
This specification embodiment is not limited to meet industry communication standard, standard programming language and data storage
Situation described in rule or this specification one or more embodiment.Certain professional standards use customized mode or reality
Apply example description practice processes on embodiment modified slightly also may be implemented above-described embodiment it is identical, it is equivalent or close,
Or the implementation result being anticipated that after deformation.Using these modifications or deformed data acquisition, storage, judgement, processing mode etc.
The embodiment of acquisition still may belong within the scope of the optional embodiment of this specification embodiment.
In the 1990s, the improvement of a technology can be distinguished clearly be on hardware improvement (for example,
Improvement to circuit structures such as diode, transistor, switches) or software on improvement (improvement for method flow).So
And with the development of technology, the improvement of current many method flows can be considered as directly improving for hardware circuit.
Designer nearly all obtains corresponding hardware circuit by the way that improved method flow to be programmed into hardware circuit.Cause
This, it cannot be said that the improvement of a method flow cannot be realized with hardware entities module.For example, programmable logic device
(Programmable Logic Device, PLD) (such as field programmable gate array (Field Programmable Gate
Array, FPGA)) it is exactly such a integrated circuit, logic function determines device programming by user.By designer
Voluntarily programming comes a digital display circuit " integrated " on a piece of PLD, designs and makes without asking chip maker
Dedicated IC chip.Moreover, nowadays, substitution manually makes IC chip, this programming is also used instead mostly " is patrolled
Volume compiler (logic compiler) " software realizes that software compiler used is similar when it writes with program development,
And the source code before compiling also write by handy specific programming language, this is referred to as hardware description language
(Hardware Description Language, HDL), and HDL is also not only a kind of, but there are many kind, such as ABEL
(Advanced Boolean Expression Language)、AHDL(Altera Hardware Description
Language)、Confluence、CUPL(Cornell University Programming Language)、HDCal、JHDL
(Java Hardware Description Language)、Lava、Lola、MyHDL、PALASM、RHDL(Ruby
Hardware Description Language) etc., VHDL (Very-High-Speed is most generally used at present
Integrated Circuit Hardware Description Language) and Verilog.Those skilled in the art also answer
This understands, it is only necessary to method flow slightly programming in logic and is programmed into integrated circuit with above-mentioned several hardware description languages,
The hardware circuit for realizing the logical method process can be readily available.
Controller can be implemented in any suitable manner, for example, controller can take such as microprocessor or processing
The computer for the computer readable program code (such as software or firmware) that device and storage can be executed by (micro-) processor can
Read medium, logic gate, switch, specific integrated circuit (Application Specific Integrated Circuit,
ASIC), the form of programmable logic controller (PLC) and insertion microcontroller, the example of controller includes but is not limited to following microcontroller
Device: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicone Labs C8051F320 are deposited
Memory controller is also implemented as a part of the control logic of memory.It is also known in the art that in addition to
Pure computer readable program code mode is realized other than controller, can be made completely by the way that method and step is carried out programming in logic
Controller is obtained to come in fact in the form of logic gate, switch, specific integrated circuit, programmable logic controller (PLC) and insertion microcontroller etc.
Existing identical function.Therefore this controller is considered a kind of hardware component, and to including for realizing various in it
The device of function can also be considered as the structure in hardware component.Or even, it can will be regarded for realizing the device of various functions
For either the software module of implementation method can be the structure in hardware component again.
System, device, module or the unit that above-described embodiment illustrates can specifically realize by computer chip or entity,
Or it is realized by the product with certain function.It is a kind of typically to realize that equipment is server system.Certainly, the application does not arrange
Except the development with future computer technology, realize that the computer of above-described embodiment function for example can be personal computer, knee
Laptop computer, vehicle-mounted human-computer interaction device, cellular phone, camera phone, smart phone, personal digital assistant, media play
It is any in device, navigation equipment, electronic mail equipment, game console, tablet computer, wearable device or these equipment
The combination of equipment.
Although this specification one or more embodiment provides the method operating procedure as described in embodiment or flow chart,
It but may include more or less operating procedure based on conventional or without creativeness means.The step of being enumerated in embodiment
Sequence is only one of numerous step execution sequence mode, does not represent and unique executes sequence.Device in practice or
When end product executes, can be executed according to embodiment or the execution of method shown in the drawings sequence or parallel (such as it is parallel
The environment of processor or multiple threads, even distributed data processing environment).The terms "include", "comprise" or its
Any other variant is intended to non-exclusive inclusion so that include the process, methods of a series of elements, product or
Equipment not only includes those elements, but also including other elements that are not explicitly listed, or further include for this process,
Method, product or the intrinsic element of equipment.In the absence of more restrictions, being not precluded is including the element
There is also other identical or equivalent elements in process, method, product or equipment.If such as waiting words using to the first, the second
Language is used to indicate names, and is not indicated any particular order.
For convenience of description, it is divided into various modules when description apparatus above with function to describe respectively.Certainly, implementing this
The function of each module can be realized in the same or multiple software and or hardware when specification one or more, it can also be with
The module for realizing same function is realized by the combination of multiple submodule or subelement etc..Installation practice described above is only
It is only illustrative, for example, in addition the division of the unit, only a kind of logical function partition can have in actual implementation
Division mode, such as multiple units or components can be combined or can be integrated into another system or some features can be with
Ignore, or does not execute.Another point, shown or discussed mutual coupling, direct-coupling or communication connection can be logical
Some interfaces are crossed, the indirect coupling or communication connection of device or unit can be electrical property, mechanical or other forms.
The present invention be referring to according to the method for the embodiment of the present invention, the process of device (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
In a typical configuration, calculating equipment includes one or more processors (CPU), input/output interface, net
Network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage, graphene stores or other
Magnetic storage device or any other non-transmission medium, can be used for storage can be accessed by a computing device information.According to herein
In define, computer-readable medium does not include temporary computer readable media (transitory media), such as the data of modulation
Signal and carrier wave.
It will be understood by those skilled in the art that this specification one or more embodiment can provide as method, system or calculating
Machine program product.Therefore, this specification one or more embodiment can be used complete hardware embodiment, complete software embodiment or
The form of embodiment combining software and hardware aspects.Moreover, this specification one or more embodiment can be used at one or
It is multiple wherein include computer usable program code computer-usable storage medium (including but not limited to magnetic disk storage,
CD-ROM, optical memory etc.) on the form of computer program product implemented.
This specification one or more embodiment can computer executable instructions it is general on
It hereinafter describes, such as program module.Generally, program module includes executing particular task or realization particular abstract data type
Routine, programs, objects, component, data structure etc..This this specification one can also be practiced in a distributed computing environment
Or multiple embodiments, in these distributed computing environments, by being held by the connected remote processing devices of communication network
Row task.In a distributed computing environment, program module can be located at the local and remote computer including storage equipment
In storage medium.
All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for system reality
For applying example, since it is substantially similar to the method embodiment, so being described relatively simple, related place is referring to embodiment of the method
Part explanation.In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ",
The description of " specific example " or " some examples " etc. means specific features described in conjunction with this embodiment or example, structure, material
Or feature is contained at least one embodiment or example of this specification.In the present specification, to the signal of above-mentioned term
Property statement be necessarily directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described
It may be combined in any suitable manner in any one or more of the embodiments or examples.In addition, without conflicting with each other, this
The technical staff in field can be by the spy of different embodiments or examples described in this specification and different embodiments or examples
Sign is combined.
The foregoing is merely the embodiments of this specification one or more embodiment, are not limited to book explanation
Book one or more embodiment.To those skilled in the art, this specification one or more embodiment can have various
Change and variation.All any modification, equivalent replacement, improvement and so within the spirit and principle of this specification, should all wrap
It is contained within scope of the claims.