CN108734335A - Electronic device, finance data processing method and computer readable storage medium - Google Patents
Electronic device, finance data processing method and computer readable storage medium Download PDFInfo
- Publication number
- CN108734335A CN108734335A CN201810298032.2A CN201810298032A CN108734335A CN 108734335 A CN108734335 A CN 108734335A CN 201810298032 A CN201810298032 A CN 201810298032A CN 108734335 A CN108734335 A CN 108734335A
- Authority
- CN
- China
- Prior art keywords
- factor data
- data
- pending
- factor
- model
- 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
- 238000003672 processing method Methods 0.000 title claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 61
- 238000000034 method Methods 0.000 claims description 30
- 239000011159 matrix material Substances 0.000 claims description 28
- 238000013507 mapping Methods 0.000 claims description 18
- 230000015654 memory Effects 0.000 claims description 18
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 230000007787 long-term memory Effects 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 abstract description 7
- 238000012549 training Methods 0.000 description 34
- 238000012360 testing method Methods 0.000 description 18
- 238000003062 neural network model Methods 0.000 description 14
- 230000008859 change Effects 0.000 description 8
- 238000002790 cross-validation Methods 0.000 description 8
- 238000012795 verification Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000011478 gradient descent method Methods 0.000 description 4
- 241001269238 Data Species 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000013499 data model Methods 0.000 description 3
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 2
- 230000029142 excretion Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Technology Law (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of electronic device, finance data processing method and computer readable storage medium.Present invention determine that the corresponding factor data prediction model of pending factor data, the predicted value of pending factor data is calculated according to the factor data prediction model, the observation of factor data in first preset time section and predicted value are input to data processing model, to obtain removing the factor data after making an uproar.Compared to the prior art, this invention removes the noises in factor data observation, improve the accuracy of factor data, and good basis has been established for the follow-up analysis for factor data.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of electronic device, finance data processing method and calculating
Machine readable storage medium storing program for executing.
Background technology
Finance data includes stock factor data, bond factor data, futures factor data, fund factor data etc..Example
Such as, the factorial analysis based on stock factor data is usually used to as the important evidence for judging the stock future trend, because
The type of subdata is varied, different types of factor data, and corresponding calculation is also different.Due to collection mode
Or spatiotemporal reason, the factor data that can be observed are considered the data of unclean (being full of noise) substantially.For example, by
When the factor of momentum there are time delay, observed at a certain moment is positive value, does not represent the stock price and will continue to increase upwards
It is long, it is also possible to which that there are momentum reversal developments.Therefore the analysis result accuracy analyzed based on the data for being full of noise
It is not high.
Invention content
The main object of the present invention is to solve the factor data that observes to be full of noise, based on the factor data carry out because
The not high problem of sub- analysis result accuracy.
To achieve the above object, electronic device proposed by the present invention, the electronic device include memory and processor, institute
State the finance data processing system that is stored with and can run on the processor on memory, the finance data processing system quilt
The processor realizes following steps when executing：
S10 obtains the observation of each pending factor data in the first preset time section；
S20, in the factor data forecasting model database pre-established, according to predetermined each factor data and because of subnumber
It is predicted that the mapping relations between model, the corresponding factor data prediction model of each pending factor data of inquiry；
It is pre- to obtain described first according to the corresponding factor data prediction model of each pending factor data by S30
If the predicted value of each pending factor data in time interval；
S40, by the prediction of the observation of each pending factor data of acquisition and each pending factor data
Value inputs predetermined data processing model, to obtain revised factor data.
Preferably, before the step S20, the processor is additionally operable to execute the finance data processing system, with
Realize following steps：
S50 carries out classification processing according to predetermined classifying rules to each factor data；
S60 establishes the corresponding factor data prediction of each factor data respectively according to the type of each factor data
Model；
S70, by the mapping relations between the corresponding factor data of each factor data prediction model of foundation into
Row storage is handled.
Preferably, the step S60 includes：
When the type of Graph One factor data is the first data class, then established based on following operation expression described because of subnumber
According to corresponding factor data prediction model：
X (K+1)=Z (K)
Wherein, X (K+1) is the predicted value of the factor data at K+1 moment, and Z (K) is the observation of the factor data at K moment.
Preferably, the step S60 further includes：
When the type of Graph One factor data is the second data class, the corresponding factor data prediction model of the factor data
Method for building up be：
Acquire the factor data in the second preset time section；
The factor data of acquisition is pre-processed；
The factor data pair is built using the pretreated factor data, and based on shot and long term Memory Neural Networks
The factor data prediction model answered.
Preferably, the step S40 includes：
The observation of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the observing matrix of pending factor data；
The predicted value of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the prediction matrix of pending factor data；
The prediction matrix of the observing matrix of the pending factor data and the pending factor data is input to pre-
First determining Kalman filter model, to obtain the revised factor data.
In addition, to achieve the above object, the present invention provides a kind of finance data processing method, and the method comprising the steps of：
S10 obtains the observation of each pending factor data in the first preset time section；
S20, in the factor data forecasting model database pre-established, according to predetermined each factor data and because of subnumber
It is predicted that the mapping relations between model, the corresponding factor data prediction model of each pending factor data of inquiry；
It is pre- to obtain described first according to the corresponding factor data prediction model of each pending factor data by S30
If the predicted value of each pending factor data in time interval；
S40, by the prediction of the observation of each pending factor data of acquisition and each pending factor data
Value inputs predetermined data processing model, to obtain revised factor data.
Preferably, before the step S20, the finance data processing method further includes：
S50 carries out classification processing according to predetermined classifying rules to each factor data；
S60 establishes the corresponding factor data prediction of each factor data respectively according to the type of each factor data
Model；
S70, by the mapping relations between the corresponding factor data of each factor data prediction model of foundation into
Row storage is handled.
Preferably, the step S60 includes：
When the type of Graph One factor data is the first data class, then established based on following operation expression described because of subnumber
According to corresponding factor data prediction model：
X (K+1)=Z (K)
Wherein, X (K+1) is the predicted value of the factor data at K+1 moment, and Z (K) is the observation of the factor data at K moment；
When the type of Graph One factor data is the second data class, the corresponding factor data prediction model of the factor data
Method for building up be：
Acquire the factor data in the second preset time section；
The factor data of acquisition is pre-processed；
The factor data pair is built using the pretreated factor data, and based on shot and long term Memory Neural Networks
The factor data prediction model answered.
Preferably, the step S40 includes：
The observation of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the observing matrix of pending factor data；
The predicted value of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the prediction matrix of pending factor data；
The prediction matrix of the observing matrix of the pending factor data and the pending factor data is input to pre-
First determining Kalman filter model, to obtain the revised factor data.
In addition, to achieve the above object, the present invention provides a kind of computer readable storage medium, which is characterized in that described
Computer-readable recording medium storage has finance data processing system, the finance data processing system can be by least one processing
Device executes, so that at least one processor executes finance data processing method as described in any one of the above embodiments.
Present invention determine that the corresponding factor data prediction model of pending factor data, according to the factor data prediction model
The predicted value for calculating pending factor data, the observation of the factor data in the first preset time section and predicted value is defeated
Enter to data processing model, to obtain removing the factor data after making an uproar；Compared to the prior art, the present invention fully considered it is each because
Characteristic is different between subdata, used for it different factor data prediction models can make the predicted value of factor data more subject to
Really, more meet truth, in addition, the factor data removed after making an uproar through data processing model eliminates in factor data observation
Noise, improve the accuracy of factor data, good basis established for the follow-up analysis for factor data.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
The structure shown according to these attached drawings obtains other attached drawings.
Fig. 1 is the flow diagram of finance data processing method first embodiment of the present invention；
Fig. 2 is the flow diagram of finance data processing method second embodiment of the present invention；
Fig. 3 is the running environment schematic diagram of financial data system first embodiment of the present invention；
Fig. 4 is the Program modual graph of financial data system first embodiment of the present invention；
Fig. 5 is the Program modual graph of financial data system second embodiment of the present invention.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific implementation mode
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the present invention.
As shown in FIG. 1, FIG. 1 is the flow diagrams of finance data processing method first embodiment of the present invention.
In the present embodiment, this method includes：
S10 obtains the observation of each pending factor data in the first preset time section；
If by taking the factor data of stock as an example, above-mentioned factor data may include stock same day amount of increase, rolling average valence
(moving average), stock trading volume etc..
User can be arranged as required to the first preset time section, for example, the first preset time section of setting is 2018 1
On January 22nd, No. 21 moon.
The frequency of the observation of each pending factor data can be set as needed in the first preset time of above-mentioned acquisition section
It sets, for example, in real time, fixed time interval (for example, 1 day) is either set or in the acquisition instruction for receiving user and sending out
Shi Zhihang obtains the step of observation of each pending factor data in the first preset time section.It should be noted that
Factor data (for example, net profit increases, operating profit increases etc.) is if desired obtained by calculation, then can first carry out calculating step in advance
Suddenly, and factor data that the calculating step obtains is stored to preset memory space, when executing step S10, in memory space
Factor data needed for reading, in some embodiments, above-mentioned steps S10 also may include the calculating step of factor data.
S20, in the factor data forecasting model database pre-established, according to predetermined each factor data and because of subnumber
It is predicted that the mapping relations between model, the corresponding factor data prediction model of each pending factor data of inquiry；
Establish corresponding factor data prediction model for each factor data in advance, due to each factor data characteristic not
Together, the method that corresponding factor data prediction model is established for it is also different, and each factor data prediction model of foundation is deposited
In storage to factor data forecasting model database, while preserving the mapping relations (example between factor data and factor data prediction model
Such as, reflect the mapping table of factor data and factor data prediction model mapping relations).
It is pre- to obtain described first according to the corresponding factor data prediction model of each pending factor data by S30
If the predicted value of each pending factor data in time interval；
For example, the predicted value of the pending factor data to obtain the K moment, then pending because of subnumber by the K-1 moment
According to observation be input in factor data prediction model, factor data prediction model export result be the pending of K moment
The predicted value of factor data.
S40, by the prediction of the observation of each pending factor data of acquisition and each pending factor data
Value inputs predetermined data processing model, to obtain revised factor data.
The present embodiment determines the corresponding factor data prediction model of pending factor data, and mould is predicted according to the factor data
Type calculates the predicted value of pending factor data, by the observation and predicted value of the factor data in the first preset time section
It is input to data processing model, to obtain removing the factor data after making an uproar；Compared to the prior art, the present embodiment fully considers
Characteristic is different between each factor data, uses different factor data prediction models that can make the predicted value of factor data more for it
For it is accurate, more meet truth, in addition, removing the factor data after making an uproar through data processing model eliminates factor data observation
Noise in value improves the accuracy of factor data, and good basis has been established for the follow-up analysis for factor data.
Preferably, in the present embodiment, above-mentioned steps S40 can be specially：
The observation of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the observing matrix of pending factor data；
The predicted value of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the prediction matrix of pending factor data；
The prediction matrix of the observing matrix of the pending factor data and the pending factor data is input to pre-
First determining Kalman filter model, to obtain the revised factor data.
In the present embodiment, the step of observing matrix of the above-mentioned pending factor data of generation, can also execute the step S10
After be immediately performed；Equally, the step of prediction matrix of the pending factor data of above-mentioned generation can be after executing the step S30 immediately
It executes.
Data processing model is Kalman filter model in the present embodiment, and Kalman filter model can efficient excretion factor
The noise of data makes factor data more close to its actual value.
As shown in Fig. 2, Fig. 2 is the flow diagram of finance data processing method second embodiment of the present invention.
In finance data processing method second embodiment of the present invention, the present embodiment is described on the basis of first embodiment
Before step S20, this method further includes：
S50 carries out classification processing according to predetermined classifying rules to each factor data；
S60 establishes the corresponding factor data prediction of each factor data respectively according to the type of each factor data
Model；
S70, by the mapping relations between the corresponding factor data of each factor data prediction model of foundation into
Row storage is handled.
Preferably, in the present embodiment, above-mentioned steps S60 includes the following steps：
When the type of Graph One factor data is the first data class, then established based on following operation expression described because of subnumber
According to corresponding factor data prediction model：
X (K+1)=Z (K)
Wherein, X (K+1) is the predicted value of the factor data at K+1 moment, and Z (K) is the observation of the factor data at K moment.
The method for building up of the corresponding factor data prediction model of factor data of above-mentioned first data class is suitable for low frequency
Factor data, so-called low frequency factor data are with price change, slowly varying factor data.
Preferably, in the present embodiment, above-mentioned steps S60 is further comprising the steps of：
When the type of Graph One factor data is the second data class, the corresponding factor data prediction model of the factor data
Method for building up be：
Acquire the factor data in the second preset time section；
The factor data of acquisition is pre-processed (for example, normalized)；
The factor data pair is built using the pretreated factor data, and based on shot and long term Memory Neural Networks
The factor data prediction model answered.
It is above-mentioned described because of subnumber based on shot and long term memory (LSTM, Long Short-Term Memory) neural network structure
It is specially according to the step of corresponding factor data prediction model：
Based on cross-validation method (cross-validation), pretreated factor data is divided into training set, is commented
Estimate collection and test set (for example, the factor data of 70% quantity, as training set, the factor data of 10% quantity collects as assessment,
The factor data of 20% quantity is as test set).Training set is input to LSTM neural network models, and is based on gradient descent method
(for example, stochastic gradient descent method) is trained.Assessment collection in the training process to LSTM neural network models into
Row verification, the LSTM neural network models that test set input training is obtained, with the LSTM obtained to training
Neural network model is verified, when the LSTM neural network models that training obtains meet the first default verification condition (example
Such as, it is less than predetermined threshold value with verification result difference), then training is completed, and sets the LSTM neural network models that training is completed to
The factor data prediction model of the factor data.
It should be noted that above-mentioned be divided into training set, assessment collection and test set based on cross-validation method by factor data
The step of can be replaced：Factor data is divided into training set and test set based on cross-validation method.And training set, assessment collection and
The quantity of factor data can be arranged as required in test set, however it is not limited to enumerated scheme.
The method for building up of the corresponding factor data prediction model of factor data of above-mentioned second data class is suitable for high frequency
Factor data, so-called high-frequency factor data are with price change, vertiginous factor data, this is because LSTM neural networks
There is more accurately predictive ability for high-frequency factor data.
Judge that Graph One factor data belong to low frequency factor data or the method for high-frequency factor data can refer to following example：
Variance method：
The factor data in a time interval is taken, and difference processing is made by variable of the time to the factor data；
The relative standard deviation of the factor data after difference processing is calculated as change frequency value；
If the change frequency value of Graph One factor data is greater than or equal to preset standard variation frequency values, by the factor data
It is determined as high-frequency factor data；
If the change frequency value of Graph One factor data is less than preset standard variation frequency values, which is determined as
Low frequency factor data；
Wherein, if by taking stock factor data as an example, the preset standard variation frequency values may be configured as closing price
Change frequency value.
Crosspoint method：
Take the factor data in a time interval, and using the time as variable difference processing is made to the factor data with obtain because
Subdata curve, the coordinate X-axis where the factor data curve are sampling instant, and Y-axis is the value of difference processing postfactor data；
In the time interval, the number of intersections of the factor data curve and mean value line is counted；Wherein, the mean value is
The average value of difference processing postfactor data；
If Graph One factor data and curves and the number of intersections of mean value line are greater than or equal to preset standard number of intersections, should
The corresponding factor data of factor data curve is determined as high-frequency factor data；
If Graph One factor data and curves and the number of intersections of mean value line are less than preset standard number of intersections, by this because of subnumber
It is determined as low frequency factor data according to the corresponding factor data of curve；
Wherein, if by taking stock factor data as an example, the preset standard number of intersections may be configured as being based on closing price
The standard number of intersections of closing price curve after difference processing.
In addition, in the present embodiment, above-mentioned steps S50, step S60 and step S70 can be replaced by following steps：
Acquire the factor data in third preset time section；
The factor data of acquisition is pre-processed；
By the pretreated factor data according to preset rules be divided into training set and test set (for example, according to when
Between sequencing, divide 80% quantity factor data be used as training set, divide 20% quantity factor data as test
Collection)；
Using training set, and based on multiple prediction models (for example, neural network model, Random Forest model, linear regression
Model, Logic Regression Models etc.) it is trained, the multiple prediction models for completing training are verified using test set, according to
It is predetermined that rule is selected to choose factor data of the prediction model as the factor data from above-mentioned multiple prediction models
Prediction model.
Wherein, it is above-mentioned it is predetermined select rule can be select with the most similar prediction model of verification result as because
Subdata prediction model.
The method for building up of above-mentioned factor data prediction model is suitable for all kinds of factor datas, and (this scheme is not required to factor data
Classification), based on Graph One factor data multiple prediction models of training simultaneously, find out optimal prediction model as the factor data because
Subdata prediction model.
In addition, the present invention also proposes a kind of finance data processing system.
Referring to Fig. 3, being the running environment schematic diagram of 10 preferred embodiment of finance data processing system of the present invention.
In the present embodiment, finance data processing system 10 is installed and is run in electronic device 1.Electronic device 1 can be with
It is the computing devices such as desktop PC, notebook, palm PC and server.The electronic device 1 may include, but not only limit
In memory 11, processor 12 and display 13.Fig. 3 illustrates only the electronic device 1 with component 11-13, it should be understood that
Be, it is not required that implement all components shown, the implementation that can be substituted is more or less component.
Memory 11 can be the internal storage unit of electronic device 1 in some embodiments, such as the electronic device 1
Hard disk or memory.Memory 11 can also be the External memory equipment of electronic device 1, such as electronics dress in further embodiments
Set the plug-in type hard disk being equipped on 1, intelligent memory card (Smart Media Card, SMC), secure digital (Secure
Digital, SD) card, flash card (Flash Card) etc..Further, memory 11 can also both include the interior of electronic device 1
Portion's storage unit also includes External memory equipment.Memory 11 is for storing the application software for being installed on electronic device 1 and all kinds of
Data, for example, finance data processing system 10 program code etc..Memory 11 can be also used for temporarily storing and export
Or the data that will be exported.
Processor 12 can be in some embodiments a central processing unit (Central Processing Unit,
CPU), microprocessor or other data processing chips, the program code for being stored in run memory 11 or processing data, example
Such as execute finance data processing system 10.
Display 13 can be in some embodiments light-emitting diode display, liquid crystal display, touch-control liquid crystal display and
OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode) touches device etc..Display 13 is for being shown in
The information that is handled in electronic device 1 and for showing visual user interface.The component 11-13 of electronic device 1, which passes through, is
System bus is in communication with each other.
Referring to Fig. 4, being the Program modual graph of 10 preferred embodiment of finance data processing system of the present invention.In the present embodiment
In, finance data processing system 10 can be divided into one or more modules, one or more module is stored in storage
In device 11, and it is performed by one or more processors (the present embodiment is processor 12), to complete the present invention.For example, in Fig. 4
In, finance data processing system 10 can be divided into the first acquisition module 101, enquiry module 102, the second acquisition module 103
And data processing module 104.The so-called module of the present invention is the series of computation machine program instruction for referring to complete specific function
Section, the implementation procedure than program more suitable for description finance data processing system 10 in the electronic apparatus 1, wherein：
First acquisition module 101, the observation for obtaining each pending factor data in the first preset time section；
Enquiry module 102, in the factor data forecasting model database pre-established, according to predetermined each factor
Mapping relations between data and factor data prediction model, the corresponding factor data of each pending factor data of inquiry are pre-
Survey model；
Second acquisition module 103, for predicting mould according to the corresponding factor data of each pending factor data
Type obtains the predicted value of each pending factor data in first preset time section；
Data processing module 104, for by the observation of each pending factor data obtained and each described waiting locating
The predicted value for managing factor data inputs predetermined data processing model, to obtain revised factor data.
If by taking the factor data of stock as an example, above-mentioned factor data may include stock same day amount of increase, rolling average valence
(moving average), stock trading volume etc..
User can be arranged as required to the first preset time section, for example, the first preset time section of setting is 2018 1
On January 22nd, No. 21 moon.
The frequency of the observation of each pending factor data can be set as needed in the first preset time of above-mentioned acquisition section
It sets, for example, in real time, fixed time interval (for example, 1 day) is either set or in the acquisition instruction for receiving user and sending out
Shi Zhihang obtains the step of observation of each pending factor data in the first preset time section.It should be noted that
Factor data (for example, net profit increases, operating profit increases etc.) is if desired obtained by calculation, then can first carry out calculating step in advance
Suddenly, and factor data that the calculating step obtains is stored to preset memory space, when executing the first preset time of the acquisition
In section when the step of the observation of each pending factor data, the first acquisition module 101 memory space read needed for because
Subdata, in some embodiments, the first acquisition module 101 also can perform the calculating step of factor data.
Before 102 actuator query steps of enquiry module, finance data processing system 10 is directed to each factor data in advance
Corresponding factor data prediction model is established, since the characteristic of each factor data is different, corresponding factor data is established for it
The method of prediction model is also different, each factor data prediction model of foundation is stored into factor data forecasting model database, together
Mapping relations between Shi Baocun factor datas and factor data prediction model are (for example, reflection factor data and factor data are pre-
Survey the mapping table of model mapping relations).
Above-mentioned second acquisition module 103 obtains the pre- of each pending factor data in first preset time section
The step of measured value can concrete example be：Predicted value to the pending factor data for obtaining the K moment, then wait for the K-1 moment
The observation for the treatment of factors data is input in factor data prediction model, when factor data prediction model output result is K
The predicted value for the pending factor data carved.
The present embodiment determines the corresponding factor data prediction model of pending factor data, and mould is predicted according to the factor data
Type calculates the predicted value of pending factor data, by the observation and predicted value of the factor data in the first preset time section
It is input to data processing model, to obtain removing the factor data after making an uproar；Compared to the prior art, the present embodiment fully considers
Characteristic is different between each factor data, uses different factor data prediction models that can make the predicted value of factor data more for it
For it is accurate, more meet truth, in addition, removing the factor data after making an uproar through data processing model eliminates factor data observation
Noise in value improves the accuracy of factor data, and good basis has been established for the follow-up analysis for factor data.
Preferably, in the present embodiment, above-mentioned data processing module 104 is additionally operable to：
The observation of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the observing matrix of pending factor data；
The predicted value of each pending factor data of acquisition is ranked up according to chronological order, to generate
State the prediction matrix of pending factor data；
The prediction matrix of the observing matrix of the pending factor data and the pending factor data is input to pre-
First determining Kalman filter model, to obtain the revised factor data.
In the present embodiment, the step of observing matrix of the above-mentioned pending factor data of generation, can also be by the first acquisition module
101 hold immediately after the step of having executed the observation of each pending factor data in the first preset time section of the acquisition
Row；Equally, the step of prediction matrix of the pending factor data of above-mentioned generation can also be by the second acquisition module 103 having executed
It is immediately performed after stating the step of obtaining the predicted value of each pending factor data in first preset time section.
Data processing model is Kalman filter model in the present embodiment, and Kalman filter model can efficient excretion factor
The noise of data makes factor data more close to its actual value.
As shown in figure 5, Fig. 5 is the Program modual graph of finance data processing system second embodiment of the present invention.
In finance data processing system second embodiment of the present invention, the present embodiment is described on the basis of first embodiment
Finance data processing system 10 further includes model construction module 105, and the model construction module 105 is used for：
According to predetermined classifying rules, classification processing is carried out to each factor data；According to each factor data
Type establishes the corresponding factor data prediction model of each factor data respectively；By each factor data prediction of foundation
Mapping relations between the corresponding factor data of model carry out storage processing.
Preferably, in the present embodiment, model construction module 105 described above is additionally operable to：
When the type of Graph One factor data is the first data class, then established based on following operation expression described because of subnumber
According to corresponding factor data prediction model：
X (K+1)=Z (K)
Wherein, X (K+1) is the predicted value of the factor data at K+1 moment, and Z (K) is the observation of the factor data at K moment.
The method for building up of the corresponding factor data prediction model of factor data of above-mentioned first data class is suitable for low frequency
Factor data, so-called low frequency factor data are with price change, slowly varying factor data.
Preferably, in the present embodiment, model construction module 105 described above is additionally operable to：
When the type of Graph One factor data is the second data class, the corresponding factor data prediction model of the factor data
Method for building up be：
Acquire the factor data in the second preset time section；
The factor data of acquisition is pre-processed (for example, normalized)；
The factor data pair is built using the pretreated factor data, and based on shot and long term Memory Neural Networks
The factor data prediction model answered.
It is above-mentioned described because of subnumber based on shot and long term memory (LSTM, Long Short-Term Memory) neural network structure
It is specially according to the step of corresponding factor data prediction model：
Based on cross-validation method (cross-validation), pretreated factor data is divided into training set, is commented
Estimate collection and test set (for example, the factor data of 70% quantity, as training set, the factor data of 10% quantity collects as assessment,
The factor data of 20% quantity is as test set).Training set is input to LSTM neural network models, and is based on gradient descent method
(for example, stochastic gradient descent method) is trained.Assessment collection in the training process to LSTM neural network models into
Row verification, the LSTM neural network models that test set input training is obtained, with the LSTM obtained to training
Neural network model is verified, when the LSTM neural network models that training obtains meet the first default verification condition (example
Such as, it is less than predetermined threshold value with verification result difference), then training is completed, and sets the LSTM neural network models that training is completed to
The factor data prediction model of the factor data.
It should be noted that above-mentioned be divided into training set, assessment collection and test set based on cross-validation method by factor data
The step of can be replaced：Factor data is divided into training set and test set based on cross-validation method.And training set, assessment collection and
The quantity of factor data can be arranged as required in test set, however it is not limited to enumerated scheme.
The method for building up of the corresponding factor data prediction model of factor data of above-mentioned second data class is suitable for high frequency
Factor data, so-called high-frequency factor data are with price change, vertiginous factor data, this is because LSTM neural networks
There is more accurately predictive ability for the factor data of sequential class.
In addition, in the present embodiment, model construction module 105 described above is additionally operable to：
Acquire the factor data in third preset time section；
The factor data of acquisition is pre-processed；
By the pretreated factor data according to preset rules be divided into training set and test set (for example, according to when
Between sequencing, divide 80% quantity factor data be used as training set, divide 20% quantity factor data as test
Collection)；
Using training set, and based on multiple prediction models (for example, neural network model, Random Forest model, linear regression
Model, Logic Regression Models etc.) it is trained, the multiple prediction models for completing training are verified using test set, according to
It is predetermined that rule is selected to choose factor data of the prediction model as the factor data from above-mentioned multiple prediction models
Prediction model.
Wherein, it is above-mentioned it is predetermined select rule can be select with the most similar prediction model of verification result as because
Subdata prediction model.
The method for building up of above-mentioned factor data prediction model is suitable for all kinds of factor datas, and (this scheme is not required to factor data
Classification), based on Graph One factor data multiple prediction models of training simultaneously, find out optimal prediction model as the factor data because
Subdata prediction model.
Further, the present invention also proposes that a kind of computer readable storage medium, the computer readable storage medium are deposited
Contain finance data processing system, the finance data processing system can be executed by least one processor so that it is described at least
One processor executes the finance data processing method in any of the above-described embodiment.
The foregoing is merely the preferred embodiment of the present invention, are not intended to limit the scope of the invention, every at this
Under the inventive concept of invention, using equivalent structure transformation made by description of the invention and accompanying drawing content, or directly/use indirectly
In the scope of patent protection that other related technical areas are included in the present invention.
Claims (10)
1. a kind of electronic device, which is characterized in that the electronic device includes memory and processor, is stored on the memory
There are the finance data processing system that can be run on the processor, the finance data processing system to be executed by the processor
Shi Shixian following steps：
S10 obtains the observation of each pending factor data in the first preset time section；
S20, it is pre- according to predetermined each factor data and factor data in the factor data forecasting model database pre-established
Survey the mapping relations between model, the corresponding factor data prediction model of each pending factor data of inquiry；
S30, according to the corresponding factor data prediction model of each pending factor data, obtain described first it is default when
Between in section each pending factor data predicted value；
S40, the predicted value of the observation of each pending factor data of acquisition and each pending factor data is defeated
Enter predetermined data processing model, to obtain revised factor data.
2. electronic device as described in claim 1, which is characterized in that before the step S20, the processor is additionally operable to
The finance data processing system is executed, to realize following steps：
S50 carries out classification processing according to predetermined classifying rules to each factor data；
S60 establishes the corresponding factor data prediction mould of each factor data respectively according to the type of each factor data
Type；
S70 deposits the mapping relations between the corresponding factor data of each factor data prediction model of foundation
Storage is handled.
3. electronic device as claimed in claim 2, which is characterized in that the step S60 includes：
When the type of Graph One factor data is the first data class, then the factor data pair is established based on following operation expression
The factor data prediction model answered：
X (K+1)=Z (K)
Wherein, X (K+1) is the predicted value of the factor data at K+1 moment, and Z (K) is the observation of the factor data at K moment.
4. electronic device as claimed in claim 2, which is characterized in that the step S60 further includes：
When the type of Graph One factor data is the second data class, the corresponding factor data prediction model of factor data is built
Cube method is：
Acquire the factor data in the second preset time section；
The factor data of acquisition is pre-processed；
Using the pretreated factor data, and it is corresponding based on the shot and long term Memory Neural Networks structure factor data
Factor data prediction model.
5. the electronic device as described in any one of Claims 1-4, which is characterized in that the step S40 includes：
The observation of each pending factor data of acquisition is ranked up according to chronological order, to be waited for described in generation
The observing matrix for the treatment of factors data；
The predicted value of each pending factor data of acquisition is ranked up according to chronological order, to be waited for described in generation
The prediction matrix for the treatment of factors data；
The prediction matrix of the observing matrix of the pending factor data and the pending factor data is input in advance really
Fixed Kalman filter model, to obtain the revised factor data.
6. a kind of finance data processing method, which is characterized in that the method comprising the steps of：
S10 obtains the observation of each pending factor data in the first preset time section；
S20, it is pre- according to predetermined each factor data and factor data in the factor data forecasting model database pre-established
Survey the mapping relations between model, the corresponding factor data prediction model of each pending factor data of inquiry；
S30, according to the corresponding factor data prediction model of each pending factor data, obtain described first it is default when
Between in section each pending factor data predicted value；
S40, the predicted value of the observation of each pending factor data of acquisition and each pending factor data is defeated
Enter predetermined data processing model, to obtain revised factor data.
7. finance data processing method as claimed in claim 6, which is characterized in that before the step S20, the finance
Data processing method further includes：
S50 carries out classification processing according to predetermined classifying rules to each factor data；
S60 establishes the corresponding factor data prediction mould of each factor data respectively according to the type of each factor data
Type；
S70 deposits the mapping relations between the corresponding factor data of each factor data prediction model of foundation
Storage is handled.
8. finance data processing method as claimed in claim 7, which is characterized in that the step S60 includes：
When the type of Graph One factor data is the first data class, then the factor data pair is established based on following operation expression
The factor data prediction model answered：
X (K+1)=Z (K)
Wherein, X (K+1) is the predicted value of the factor data at K+1 moment, and Z (K) is the observation of the factor data at K moment；
When the type of Graph One factor data is the second data class, the corresponding factor data prediction model of factor data is built
Cube method is：
Acquire the factor data in the second preset time section；
The factor data of acquisition is pre-processed；
Using the pretreated factor data, and it is corresponding based on the shot and long term Memory Neural Networks structure factor data
Factor data prediction model.
9. the finance data processing method as described in any one of claim 6 to 8, which is characterized in that the step S40 packets
It includes：
The observation of each pending factor data of acquisition is ranked up according to chronological order, to be waited for described in generation
The observing matrix for the treatment of factors data；
The predicted value of each pending factor data of acquisition is ranked up according to chronological order, to be waited for described in generation
The prediction matrix for the treatment of factors data；
The prediction matrix of the observing matrix of the pending factor data and the pending factor data is input in advance really
Fixed Kalman filter model, to obtain the revised factor data.
10. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has finance data
Processing system, the finance data processing system can be executed by least one processor, so that at least one processor is held
Finance data processing method of the row as described in any one of claim 6-9.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810298032.2A CN108734335A (en) | 2018-04-03 | 2018-04-03 | Electronic device, finance data processing method and computer readable storage medium |
| PCT/CN2018/102226 WO2019192136A1 (en) | 2018-04-03 | 2018-08-24 | Electronic device, financial data processing method and system, and computer-readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810298032.2A CN108734335A (en) | 2018-04-03 | 2018-04-03 | Electronic device, finance data processing method and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108734335A true CN108734335A (en) | 2018-11-02 |
Family
ID=63941290
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810298032.2A Pending CN108734335A (en) | 2018-04-03 | 2018-04-03 | Electronic device, finance data processing method and computer readable storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108734335A (en) |
| WO (1) | WO2019192136A1 (en) |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180053102A1 (en) * | 2016-08-16 | 2018-02-22 | Toyota Jidosha Kabushiki Kaisha | Individualized Adaptation of Driver Action Prediction Models |
| CN107730087A (en) * | 2017-09-20 | 2018-02-23 | 平安科技（深圳）有限公司 | Forecast model training method, data monitoring method, device, equipment and medium |
| CN107798604A (en) * | 2017-09-28 | 2018-03-13 | 平安科技（深圳）有限公司 | Become a shareholder when selecting method and terminal device based on machine learning |
| CN107766888A (en) * | 2017-10-24 | 2018-03-06 | 众安信息技术服务有限公司 | Data processing method and device |
-
2018
- 2018-04-03 CN CN201810298032.2A patent/CN108734335A/en active Pending
- 2018-08-24 WO PCT/CN2018/102226 patent/WO2019192136A1/en active Application Filing
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019192136A1 (en) | 2019-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107704946B (en) | Electronic device, Voice Navigation needing forecasting method and storage medium | |
| US8311860B2 (en) | Industry scenario mapping tool | |
| Chen et al. | Optimal variability sensitive condition-based maintenance with a Cox PH model | |
| CN109636212B (en) | Method for predicting actual running time of job | |
| CN108241900A (en) | Engineering project construction period prediction method, device and system | |
| CN109787958A (en) | Network flow real-time detection method and detection terminal, computer readable storage medium | |
| CN117235608B (en) | Risk detection method, risk detection device, electronic equipment and storage medium | |
| CN107025494A (en) | Data predication method, financing recommendation method, device and terminal device | |
| CN116883181B (en) | Financial service pushing method based on user portrait, storage medium and server | |
| CN110059083A (en) | A kind of data evaluation method, apparatus and electronic equipment | |
| CN108509259A (en) | Obtain the method and air control system in multiparty data source | |
| CN116823468A (en) | SAC-based high-frequency quantitative transaction control method, system and storage medium | |
| CN108734335A (en) | Electronic device, finance data processing method and computer readable storage medium | |
| CN108710994A (en) | Investment share-selecting method, device and storage medium based on the public sentiment factor | |
| CN109410046A (en) | Target stock selection method, device and storage medium | |
| CN115564410A (en) | State monitoring method and device for relay protection equipment | |
| US20220164659A1 (en) | Deep Learning Error Minimizing System for Real-Time Generation of Big Data Analysis Models for Mobile App Users and Controlling Method for the Same | |
| Ahmad | Designing accelerated life tests for generalised exponential distribution with log-linear model | |
| US11106344B2 (en) | Methods and devices for capturing heuristic information via a weighting tool | |
| CN113191540A (en) | Construction method and device of industrial link manufacturing resources | |
| CN110457304A (en) | Data cleaning method, device, electronic equipment and readable storage medium storing program for executing | |
| CN106301880A (en) | One determines that cyberrelationship degree of stability, Internet service recommend method and apparatus | |
| CN110020783A (en) | Electronic device, the metering method of assets expected loss and storage medium | |
| CN117971630B (en) | Heterogeneous computing platform, task simulation and time consumption prediction method, device and equipment thereof | |
| Lum et al. | The effects of data mining techniques on software cost estimation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181102 |