CN108492190A - Intelligent security Portfolio Selection Based method, apparatus and equipment - Google Patents
Intelligent security Portfolio Selection Based method, apparatus and equipment Download PDFInfo
- Publication number
- CN108492190A CN108492190A CN201810242821.4A CN201810242821A CN108492190A CN 108492190 A CN108492190 A CN 108492190A CN 201810242821 A CN201810242821 A CN 201810242821A CN 108492190 A CN108492190 A CN 108492190A
- Authority
- CN
- China
- Prior art keywords
- data
- investment combination
- investment
- machine learning
- portfolio
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention provides a kind of intelligent security Portfolio Selection Based method, apparatus and equipment, it is related to the technical field of investment securities, stock portfolio data by obtaining external system production, which are realized, treats being continuously replenished and updating for ticket data splitting of selecting stocks, to obtain all obtainable stock portfolio data informations, realization selects optimal Portfolio from all possible investment combination.It obtains target investment combination data and distributed pretreatment and calculating is carried out to it, invalid data can be removed and improve the quality of data, be the form for being suitable for machine learning by data normalization.By being trained to historical data, optimizes machine learning model, be supplied to user to select with current-period data prediction next period earning rate, the general common people is made also can to customize one's own optimal Portfolio according to the demand of oneself.
Description
Technical field
The present invention relates to investment securities technical fields, more particularly, to a kind of intelligent security Portfolio Selection Based method, dress
It sets and equipment.
Background technology
It is well known that the investment target in financial market is numerous.With economic and finance development, investment target more they tends to more
Sample.For A share market, a total of more than 2800 stock of A share market at present, then being demonstrate,proved caused by this 2800 stocks
Just there are about as many as 10 33 powers for certificate investment combination quantity.The investment combination of our this 10 33 power is referred to as universe investment group
It closes, English abbreviation is UPID (Universal Portfolio Identity), and the space where them is known as universe investment group
Close space.And conventional investment theory is due to the limitation of computing capability at that time, the investment combination amount of space analyzed well below
This quantity.Therefore we need to utilize state-of-the-art big data technology, from above-mentioned universe investment combination space, with more
Analysis A share market is gone at wide visual angle, establishes the machine learning system for being directed to super flood tide Portfolio Selection, and make its efficiency
It can be considerably beyond previous traditional quantization investment analysis with income.
In novel quantization investment field, quantitative analysis teacher be typically from personal share, by statistical method or
Machine learning is predicted, selects stock.These methods to a certain extent customer service conventional investment field the shortcomings that, but this side
The combination gone out selected by method is not global optimum's combination.
Invention content
In view of this, the purpose of the present invention is to provide a kind of intelligent security Portfolio Selection Based method, apparatus and equipment,
Using investment combination as research object, is selected from all possible investment combination and generate optimal Portfolio.
In a first aspect, an embodiment of the present invention provides a kind of intelligent security Portfolio Selection Based methods, including：Data update
Partly, pretreatment and calculating section, machine learning part, the rapid of method are：Obtain the stock portfolio data of external system production；
According to stock portfolio data update investment combination database；Investment combination database data is screened, to obtain target investment combination
Data；Target investment combination data are pre-processed and calculated, to obtain history feature value；According to history feature value to machine
Learning model is trained, to obtain optimization machine learning model；Latest features value is obtained, and by optimizing machine learning model
Investment portfolio yield is predicted, the next period indicated yield to obtain investment combination is selected for user.
With reference to first aspect, an embodiment of the present invention provides the first possible embodiments of first aspect, wherein sieve
Investment combination database data is selected, the step of to obtain target investment combination data, including：Root according to the rule pre-set, from
Investment combination is selected in investment combination database；The lower target investment combination of personal share correlation is chosen from investment combination.
With reference to first aspect, an embodiment of the present invention provides second of possible embodiments of first aspect, wherein root
Machine learning model is trained according to history feature value, with obtain optimization machine learning model the step of, including：Pass through history
Characteristic value predicts the recent earning rate of corresponding investment combination, to obtain recent earning rate predicted value；History feature value is obtained to correspond to
Investment combination in recent earning rate actual value；Using the difference of recent earning rate predicted value and recent earning rate actual value as
Object function obtains corresponding model parameter when target function value minimum.
With reference to first aspect, an embodiment of the present invention provides the third possible embodiments of first aspect, wherein root
The step of according to stock portfolio data update investment combination database, including：By in stock portfolio data investment combination data with
Available data is compared in investment combination database；If the investment combination data in stock portfolio data do not appear in existing throwing
It provides in combined data base, then assigns the investment combination data new mark, and be added in investment combination database.
With reference to first aspect, an embodiment of the present invention provides the 4th kind of possible embodiment of first aspect, method is also
Including client, the screening conditions for receiving and feedback user inputs, to provide finally selected investment combination data.
With reference to first aspect, an embodiment of the present invention provides the 5th kind of possible embodiment of first aspect, method is also
Including slave part, for other data in replacement investment combined data base.
Second aspect, the embodiment of the present invention also provide a kind of intelligent security Portfolio Selection Based device, including：Data update
Partly, pretreatment and calculating section, machine learning part, device include：Data acquisition module, for obtaining external system production
Stock portfolio data；Data update module, for according to stock portfolio data update investment combination database；Data screening mould
Block, for screening investment combination database data, to obtain target investment combination data；Data processing computing module, for pair
Target investment combination data are pre-processed and are calculated, to obtain history feature value；Training module, for according to history feature value
Machine learning model is trained, to obtain optimization machine learning model；
Prediction module, for obtaining latest features value, and by optimizing machine learning model to investment portfolio yields
Rate is predicted that the next period indicated yield to obtain investment combination is selected for user.
In conjunction with second aspect, an embodiment of the present invention provides the first possible embodiments of second aspect, and device is also
Including interactive module, the screening conditions for receiving and feedback user inputs, to provide finally selected investment combination data.
In conjunction with second aspect, an embodiment of the present invention provides second of possible embodiments of second aspect, and device is also
Including supplementary module, for other data in replacement investment combined data base.
The third aspect, the embodiment of the present invention also provide a kind of electronic equipment, including memory, processor, are deposited in memory
Contain the computer program that can be run on a processor, processor realized when executing computer program above-mentioned first aspect and
Its any one possible embodiment.
The embodiment of the present invention brings following advantageous effect：Selecting party is combined in intellectual investment provided in an embodiment of the present invention
The continuous of ticket data splitting of selecting stocks is treated in method, device and equipment, the stock portfolio data realization by obtaining external system production
Supplement and update, to obtain all obtainable stock portfolio data informations, realization is selected from all possible investment combination
Go out optimal Portfolio.It obtains target investment combination data and distributed pretreatment and calculating is carried out to it, it is invalid to remove
Data improve the quality of data, are the form for being suitable for machine learning by data normalization.It is excellent by being trained to historical data
Change machine learning model, is supplied to user to select with current-period data prediction next period earning rate, makes the general common people also can be according to certainly
Oneself demand customizes one's own optimal Portfolio.
Other feature and advantage of the disclosure will illustrate in the following description, alternatively, Partial Feature and advantage can be with
Deduce from specification or unambiguously determine, or by implement the disclosure above-mentioned technology it can be learnt that.
To enable the above objects, features, and advantages of the disclosure to be clearer and more comprehensible, preferred embodiment cited below particularly, and coordinate
Appended attached drawing, is described in detail below.
Description of the drawings
It, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical solution in the prior art
Embodiment or attached drawing needed to be used in the description of the prior art are briefly described, it should be apparent that, in being described below
Attached drawing is some embodiments of the present invention, for those of ordinary skill in the art, before not making the creative labor
It puts, other drawings may also be obtained based on these drawings.
Fig. 1 is intelligent security Portfolio Selection Based method flow diagram provided in an embodiment of the present invention；
Fig. 2 is the stream of model parameter preparation method in intelligent security Portfolio Selection Based method provided in an embodiment of the present invention
Cheng Tu；
Fig. 3 is the structure diagram of intelligent security Portfolio Selection Based device provided in an embodiment of the present invention；
Fig. 4 is the structural frames for including interactive module of intelligent security Portfolio Selection Based device provided in an embodiment of the present invention
Figure；
Fig. 5 is the structural frames for including supplementary module of intelligent security Portfolio Selection Based device provided in an embodiment of the present invention
Figure；
Fig. 6 is the structure diagram of the sub- equipment of intelligent security Portfolio Selection Based provided in an embodiment of the present invention.
Icon：
31- data acquisition modules；32- data update modules；33- data screening modules；34- data processing computing modules；
35- training modules；36- prediction modules；37- interactive modules；38- supplementary modules；61- memories；62- processors.
Specific implementation mode
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with attached drawing to the present invention
Technical solution be clearly and completely described, it is clear that described embodiments are some of the embodiments of the present invention, rather than
Whole embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not making creative work premise
Lower obtained every other embodiment, shall fall within the protection scope of the present invention.
It is well known that the investment target in financial market is numerous.With economic and finance development, investment target more they tends to more
Sample.For A share market, a total of more than 2800 stock of A share market at present, then being demonstrate,proved caused by this 2800 stocks
Just there are about as many as 10 33 powers for certificate investment combination quantity.The investment combination of our this 10 33 power is referred to as universe investment group
It closes, English abbreviation is UPID (Universal Portfolio Identity), and the space where them is known as universe investment group
Close space.And conventional investment theory is due to the limitation of computing capability at that time, the investment combination amount of space analyzed well below
This quantity.Therefore we need to utilize state-of-the-art big data technology, from above-mentioned universe investment combination space, with more
Analysis A share market is gone at wide visual angle, establishes the machine learning system for being directed to super flood tide Portfolio Selection, and make its efficiency
It can be considerably beyond previous traditional quantization investment analysis with income.
In novel quantization investment field, quantitative analysis teacher be typically from personal share, by statistical method or
Machine learning is predicted, selects stock.These methods to a certain extent customer service conventional investment field the shortcomings that, but this side
The combination gone out selected by method is not global optimum's combination.Based on this, a kind of intelligent investment securities provided in an embodiment of the present invention
Combination selection method, device and equipment are selected and are generated from all possible investment combination using investment combination as research object
Optimal Portfolio.
For ease of understanding the present embodiment, first to a kind of intelligent investment securities group disclosed in the embodiment of the present invention
Selection method is closed to describe in detail.
Embodiment 1
The embodiment of the present invention 1 provides a kind of intelligent security Portfolio Selection Based method, intelligent security shown in Figure 1
Model parameter preparation method in Portfolio Selection Based method flow diagram and intelligent security Portfolio Selection Based method shown in Fig. 2
Flow chart, including data update part, pretreatment and calculating section, machine learning part, the rapid of this method are：
Step S102 obtains the stock portfolio data of external system production.
Stock portfolio data are generated by external system, include the throwing for having the fixed investment beginning and end date of more stock
Money combination further includes the number of different stock portfolio, and time started publicity phase of contained stock, raises at publicity end time phase
Time started phase raises end time phase, running time started phase, running end time phase, the maximum carrying amount of money, minimum investment
Share, accumulative return rate, expected yield, risk class, date created, the slice day of trade in period, risk numerical value, plate number
In amount, plate under most stock quantity, plate id (the trade classification id in asset table information table), accounting, the accounting upper limit, accounting
The essential informations such as limit, minimum share, security Internal Code, security id.After daily closing quotation, when same day data acquisition finishes, pass through
Timing plan target timing operation is arranged in crontab, obtains stock portfolio data automatically from external system.
Step S104, according to stock portfolio data update investment combination database.
The data that the condition of satisfaction is extracted from stock portfolio data are added in investment combination database.Investment
Data in combined data base are on the increase with the production of stock portfolio data, and the final size of investment combination database is
Universe investment combination space.
In step S104, the step of according to stock portfolio data update investment combination database, including, by stock portfolio
Investment combination data in data are compared with available data in investment combination database；If the investment in stock portfolio data
Data splitting does not appear in existing investment combined data base, then assigns the investment combination data new mark, and be added to throwing
It provides in combined data base.
Data in investment combination database are exactly unduplicated part in the stock portfolio data produced, i.e., for packet
Two groups of stock portfolio containing same stock can be regarded as two groups of different stock portfolio data, still if their date of manufacture is different
It can be regarded as the data in identical investment combination database.By the stock portfolio data of new production and existing investment combination database
Middle data compare, if being already present in existing investment combination database, are not processed, if stock portfolio does not appear in
In existing investment combination database, then assign the stock portfolio new mark, and be added in investment combination database.Investment
Each group of investment combination data have unique mark in combined data base.
Step S106 screens investment combination database data, to obtain target investment combination data.
The relatively high investment combination of historical yield is chosen from investment combination database, then screens wherein personal share correlation ratio
Lower investment combination is as target investment combination.Pre-set rule can be above specific earning rate.
In step s 106, investment combination database data is screened, the step of to obtain target investment combination data, packet
It includes：Root according to the rule pre-set, investment combination is selected from investment combination database；It is related that personal share is chosen from investment combination
The lower target investment combination of property.
For example, pre-set rule can be：Investment combination of the earning rate 5% or more, then from investment combination data
Investment combination of the earning rate 5% or more is chosen in library.Income is highly relevant before interaction shows as stock between stock
Property, i.e., same before stock rise with falling, we screen can be calculated when investment combination combine in correlation between personal share, specifically
Computational methods are as follows：
Calculate the related coefficient of all personal share yield volatilities between any two in investment combination；Calculate all related coefficients
Average value, the correlation as investment combination；The lower investment combination of correlation is chosen as target investment combination.
The related coefficient of all personal share yield volatilities between any two in investment combination is calculated, i.e.,：
Wherein, X is the yield volatility of stock 1, and Y is the yield volatility of stock 2.Calculate the phase of all investment combinations
Guan Xinghou chooses the lower investment combination of correlation as target investment combination.
After the correlation data for obtaining all investment combinations, we choose the lower combination of correlation between some personal shares,
The influence of the interaction between stock is eliminated by disperseing personal share risk
Step S108 is pre-processed and is calculated to target investment combination data, to obtain history feature value.
Characteristic value data is the 1*m vectors by being obtained in target investment combination, and m is characterized number, such as m=8+5+8+7
=28, the as summation of 8 technical parameters and 5 basic side indexs and 8 statistical parameters and 7 spectrum signatures.Technology is joined
Number includes simple Moving Average, weighted moving average line, index Moving Average, smooth unusual fluctuation Moving Average, the rate of change
Index, inertia index, relative strength index, basic side index include scale beta, market beta, value/growth beta, the country
Total output value factor is open, price index factor is open, and statistical parameter includes standard deviation, the degree of bias, kurtosis, Jarque-Bera inspections
Test, Jarque-Bera is examined, KPSS is examined, Dickey-Fuller GLS are examined, Hurst indexes, spectrum signature includes energy
Mutual information, kurtosis, harmonic wave index, frequency domain center of gravity, frequency variance, time domain center of gravity, comentropy.
It is that single machine is stored by MYSQL database to obtain after stock portfolio data above-mentioned, is dumped to by SQOOP
After HDFS file system, become distributed storage, therefore the processing and calculating to characteristic value data are all based on point of HDFS systems
Cloth stores, and the quick processing of large-scale data may be implemented.
The corresponding data of suspended stock in target investment combination are removed, to be cleaned to target investment combination data；
To multiple power processing before being carried out except power and/or ex dividend stock data in target investment combination, with to target investment combination data into
Row removes noise；Characteristic value data is based on distributed storage, and calculating is normalized based on large-scale data processing frame, with
Characteristic value is obtained, by the characteristic value of normalized, distributional pattern is partial to normal distribution.
Step S110 is trained machine learning model according to history feature value, to obtain optimization machine learning model.
In step s 110, machine learning model is trained according to history feature value, to obtain optimization machine learning
The step of model, including：
Step S202 predicts the recent earning rate of corresponding investment combination by history feature value, to obtain recent earning rate
Predicted value.
Step S204 obtains history feature and is worth corresponding investment combination in recent earning rate actual value.
Step S206 is obtained using the difference of recent earning rate predicted value and recent earning rate actual value as object function
Corresponding model parameter when target function value minimum.
At the T moment, the historical juncture is divided into three points, t by we0、t1、t2, wherein t0<t1<t2<T.We obtain t0It arrives
t1The characteristic value of the investment combination of period, referred to as the history feature value of history investment combination.By t0To t1The history of period
Characteristic value inputs machine learning model, and decision function is generated according to the setting of machine mould itself, is predicted by decision function close
Phase, that is, t1To t2The earning rate of period：We are given input x, x, that is, history feature value, and model can export a predicted value fθ
(x), wherein θ is model parameter, decision function fθConcrete form determined by the setting of machine mould itself, such as in perceptron
In machine learning model,The decision function of different machines learning model is different.It obtains in the recent period i.e.
t1To t2The actual value y of period earning rate, by the predicted value f of recent earning rateθ(x) with the difference of actual value y as target letter
Number：J (θ)=(y-fθ(x))2, object function is minimized by gradient descent method, obtains corresponding model parameter.Model parameter packet
Include algorithm the convergence speed, the weight of characteristic value and model complexity punishment parameter.By model parameter input initial machine study
Model obtains optimization machine learning model.
Machine learning model can be modeled by nonlinear models such as random forests, it is contemplated that non-linear relation, so as to
Reduce information loss.
Step S112, obtain latest features value, and by optimize machine learning model to investment portfolio yield into
Row prediction, the next period indicated yield to obtain investment combination are selected for user.
Investment combination is obtained in current i.e. t2To the characteristic value of T time section income, this feature value is inputted into optimization engineering
Practise model, the prediction receipts of the machine learning model of optimization can provide the next period based on the model parameter after optimization i.e. T to T+c periods
Beneficial rate.By this predicted value, i.e. next period indicated yield is supplied to user, user that can be selected by brushing as the final output of model
The expected yield of each investment combination chooses the investment combination oneself admired.
Further include client, the screening conditions for receiving and feedback user inputs, to provide finally selected investment group
Close data.
Client can be app software interfaces or webpage, for being interacted for user and system.User can be in visitor
Family end inputs screening conditions, and screening conditions can be expected rate, risk partiality or custom parameter.Custom parameter can
Be the easy number of days of most short delivery, opposite deep bid winning rate, history average absolute income, history maximum withdraw, history Sharpe Ratio, combination
Included listed company is averaged annual net profit amplification etc., and user can be by inputting one or more screening in client
Condition, system can filter investment combination database according to the demand of client, obtain qualified simplifying investment combination.Such as with
Family can set in 250 day of trade, and using every 7 day of trade as sliding window, the earning rate of every 7 day of trade is above Shanghai and Shenzhen
300 investment combination.Programmer is given by all throwings after this condition filter by inputting the parameter in python scripts
Money combination.
Further include slave part, for other data in replacement investment combined data base.
Other data include stock information data and financial data, and slave part updates these data and data is supplied to locate in advance
Reason and calculating section carry out data cleansing, denoising and normalization, eventually for the expected yield for calculating investment combination.
Using single stock as object, investment is combined to predict, select more stocks by machine learning, there are following
Defect：Non- global optimum, as the maximum value in the combination not necessarily 2n peacekeepings space of the maximum value of two n-dimensional spaces,
Optimal solution in the combination not necessarily universe investment combination space of optimal solution in personal share dimension, it is also necessary to consider stock it
Between correlation and it is interactive；Price shock when large-scale financial institution is invested, can be bought in specific several on a large scale
Stock can pay impact cost, to reduce the income of investment combination if failing to strike a bargain according to predetermined price；Information loss,
Industry is now widely used or linear regression model (LRM) based on multi-sector model, has ignored the nonlinear dependence between variable
System does not use all obtainable information, has ignored the feature of many investment combinations in the most efficient manner.
The embodiment of the present invention utilizes state-of-the-art big data technology, from above-mentioned universe investment combination space, with more
Analysis A share market is gone at wide visual angle, is established and is directed to the machine learning system of super flood tide Portfolio Selection, using HDFS and
HIVE realizes the distributed storage of large-scale data, passes through the Spark large-scale data processing blocks being deployed on a large amount of servers
Frame realizes the high speed data processing and machine learning task of TB ranks, makes its efficiency and income considerably beyond previous traditional
Quantify investment analysis.Research object is switched into investment combination from personal share, and builds universe investment combination space and is analyzed.Pass through
Using Spark and Hive big data technologies, quickly handle a large amount of dynamic markets data, excavate to conventional quantization system excavate less than
Information.Investment combination quantum of output is big, cost reduction, and investment consultant's service of profession can be enjoyed by making ordinary populace also.
Embodiment 2
The embodiment of the present invention 2 provides a kind of intelligent security Portfolio Selection Based device, intelligent security shown in Figure 3
The structure diagram of Portfolio Selection Based device, including：Data update part, pretreatment and calculating section, machine learning part, dress
Set including：Data acquisition module 31, the stock portfolio data for obtaining external system production；Data update module 32, is used for
According to stock portfolio data update investment combination database；Data screening module 33, for screening investment combination database data,
To obtain target investment combination data；Data processing computing module 34, for target investment combination data carry out pretreatment and
It calculates, to obtain history feature value；Training module 35, for being trained to machine learning model according to history feature value, with
Obtain optimization machine learning model；Prediction module 36, for obtaining latest features value, and by optimizing machine learning model to most
New investment portfolio yield is predicted that the next period indicated yield to obtain investment combination is selected for user.
Structure diagram shown in Figure 4, above-mentioned apparatus further include interactive module 37, for receiving simultaneously feedback user input
Screening conditions, to provide finally selected investment combination data.
Structure diagram shown in Figure 5, above-mentioned apparatus further include supplementary module 38, are used for replacement investment combined data base
In other data.
The technique effect and preceding method embodiment phase of the device that the embodiment of the present invention is provided, realization principle and generation
Together, to briefly describe, device embodiment part does not refer to place, can refer to corresponding contents in preceding method embodiment.
Device provided in an embodiment of the present invention, the method technical characteristic having the same provided with above-described embodiment, so
Also identical technical problem can be solved, identical technique effect is reached.
Embodiment 3
The embodiment of the present invention 3 provides a kind of electronic equipment, electronic devices structure block diagram shown in Figure 6, including deposits
Reservoir 61, processor 62 are stored with the computer program that can be run on the processor 62 in memory 61, and processor 62 executes meter
The method that arbitrary steps describe in embodiment 1 is realized when calculation machine program.
It is apparent to those skilled in the art that for convenience and simplicity of description, the equipment of foregoing description
Specific work process, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
Unless specifically stated otherwise, the opposite step of the component and step that otherwise illustrate in these embodiments, digital table
It is not limit the scope of the invention up to formula and numerical value.
In all examples being illustrated and described herein, any occurrence should be construed as merely illustrative, without
It is as limitation, therefore, other examples of exemplary embodiment can have different values.
Finally it should be noted that：Embodiment described above, only specific implementation mode of the invention, to illustrate the present invention
Technical solution, rather than its limitations, scope of protection of the present invention is not limited thereto, although with reference to the foregoing embodiments to this hair
It is bright to be described in detail, it will be understood by those of ordinary skill in the art that：Any one skilled in the art
In the technical scope disclosed by the present invention, it can still modify to the technical solution recorded in previous embodiment or can be light
It is readily conceivable that variation or equivalent replacement of some of the technical features；And these modifications, variation or replacement, do not make
The essence of corresponding technical solution is detached from the spirit and scope of technical solution of the embodiment of the present invention, should all cover the protection in the present invention
Within the scope of.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. a kind of intelligence security Portfolio Selection Based method, which is characterized in that including：Data update part, pretreatment and calculating
Partly, machine learning part, the rapid of the method are：
Obtain the stock portfolio data of external system production；
According to the stock portfolio data update investment combination database；
The investment combination database data is screened, to obtain target investment combination data；
The target investment combination data are pre-processed and calculated, to obtain history feature value；
Machine learning model is trained according to the history feature value, to obtain optimization machine learning model；
Latest features value is obtained, and investment portfolio yield is predicted by the optimization machine learning model, with
The next period indicated yield for obtaining the investment combination is selected for user.
2. according to the method described in claim 1, it is characterized in that, described screen the investment combination database data, to obtain
The step of taking target investment combination data, including：
Root according to the rule pre-set, investment combination is selected from the investment combination database；
The lower target investment combination of personal share correlation is chosen from the investment combination.
3. according to the method described in claim 1, it is characterized in that, it is described according to the history feature value to machine learning model
Be trained, with obtain optimization machine learning model the step of, including：
The recent earning rate that corresponding investment combination is predicted by history feature value, to obtain recent earning rate predicted value；
It obtains the history feature and is worth corresponding investment combination in the recent earning rate actual value；
Using the difference of the recent earning rate predicted value and recent earning rate actual value as object function, the target letter is obtained
Corresponding model parameter when numerical value minimum.
4. according to the method described in claim 1, it is characterized in that, described according to the stock portfolio data update investment combination
The step of database, including：
Investment combination data in the stock portfolio data are compared with available data in investment combination database；
If the investment combination data in the stock portfolio data do not appear in existing investment combined data base, the throwing is assigned
The new mark of data splitting is provided, and is added in the investment combination database.
5. according to the method described in claim 1, it is characterized in that, further including client, for receiving simultaneously feedback user input
Screening conditions, to provide finally selected investment combination data.
6. according to the method described in claim 1, it is characterized in that, further including slave part, for updating the investment combination
Other data in database.
7. a kind of intelligence security Portfolio Selection Based device, which is characterized in that including：Data update part, pretreatment and calculating
Partly, machine learning part, described device include：
Data acquisition module, the stock portfolio data for obtaining external system production；
Data update module, for according to the stock portfolio data update investment combination database；
Data screening module, for screening the investment combination database data, to obtain target investment combination data；
Data processing computing module, for the target investment combination data to be pre-processed and calculated, to obtain history spy
Value indicative；
Training module, for being trained to machine learning model according to the history feature value, to obtain optimization machine learning
Model；
Prediction module, for obtaining latest features value, and by the optimization machine learning model to investment portfolio yields
Rate is predicted that the next period indicated yield to obtain the investment combination is selected for user.
8. device according to claim 7, which is characterized in that further include interactive module, it is defeated for receiving simultaneously feedback user
The screening conditions entered, to provide finally selected investment combination data.
9. device according to claim 7, which is characterized in that further include supplementary module, for updating the investment combination
Other data in database.
10. a kind of electronic equipment, including memory, processor, it is stored with and can runs on the processor in the memory
Computer program, which is characterized in that the processor realizes that the claims 1-6 is arbitrary when executing the computer program
The work step of system described in one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810242821.4A CN108492190A (en) | 2018-03-22 | 2018-03-22 | Intelligent security Portfolio Selection Based method, apparatus and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810242821.4A CN108492190A (en) | 2018-03-22 | 2018-03-22 | Intelligent security Portfolio Selection Based method, apparatus and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108492190A true CN108492190A (en) | 2018-09-04 |
Family
ID=63319540
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810242821.4A Pending CN108492190A (en) | 2018-03-22 | 2018-03-22 | Intelligent security Portfolio Selection Based method, apparatus and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108492190A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544348A (en) * | 2018-11-09 | 2019-03-29 | 深圳前海微众银行股份有限公司 | Assets security screening technique, equipment and computer readable storage medium |
| CN109697205A (en) * | 2018-12-14 | 2019-04-30 | 北京向上一心科技有限公司 | Data reordering method, method for exhibiting data, device, equipment and storage medium |
| CN113781083A (en) * | 2021-01-06 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Commodity replenishment method, commodity replenishment device, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100010870A1 (en) * | 2000-12-20 | 2010-01-14 | Karl Millar | System and Method for Tuning Demand Coefficients |
| CN106251218A (en) * | 2016-07-11 | 2016-12-21 | 璇玑智能（北京）科技有限公司 | The generation method and apparatus of investment combination mode |
| CN106600369A (en) * | 2016-12-09 | 2017-04-26 | 广东奡风科技股份有限公司 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
| CN106709513A (en) * | 2016-12-10 | 2017-05-24 | 中泰证券股份有限公司 | Supervised machine learning-based security financing account identification method |
-
2018
- 2018-03-22 CN CN201810242821.4A patent/CN108492190A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100010870A1 (en) * | 2000-12-20 | 2010-01-14 | Karl Millar | System and Method for Tuning Demand Coefficients |
| CN106251218A (en) * | 2016-07-11 | 2016-12-21 | 璇玑智能（北京）科技有限公司 | The generation method and apparatus of investment combination mode |
| CN106600369A (en) * | 2016-12-09 | 2017-04-26 | 广东奡风科技股份有限公司 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
| CN106709513A (en) * | 2016-12-10 | 2017-05-24 | 中泰证券股份有限公司 | Supervised machine learning-based security financing account identification method |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544348A (en) * | 2018-11-09 | 2019-03-29 | 深圳前海微众银行股份有限公司 | Assets security screening technique, equipment and computer readable storage medium |
| CN109697205A (en) * | 2018-12-14 | 2019-04-30 | 北京向上一心科技有限公司 | Data reordering method, method for exhibiting data, device, equipment and storage medium |
| CN113781083A (en) * | 2021-01-06 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Commodity replenishment method, commodity replenishment device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kao | Efficiency decomposition and aggregation in network data envelopment analysis | |
| Yaseen et al. | Application of the hybrid artificial neural network coupled with rolling mechanism and grey model algorithms for streamflow forecasting over multiple time horizons | |
| Tabak et al. | Directed clustering coefficient as a measure of systemic risk in complex banking networks | |
| CN108492189A (en) | Portfolio Selection selection method, device and equipment | |
| CN108492190A (en) | Intelligent security Portfolio Selection Based method, apparatus and equipment | |
| Lia et al. | Cost Minimization is Essential for the Sustainable Development of an Industry: A Mathematical Economic Model Approach | |
| CN112700319A (en) | Enterprise credit line determination method and device based on government affair data | |
| Kulikova et al. | Planning of technological development of new products and its impact on the economic performance of the enterprise | |
| CN108090570A (en) | For selecting the method and system of the feature of machine learning sample | |
| CN110659937A (en) | Gradient-lifting-tree-based improved supplier quantitative scoring prediction algorithm | |
| CN108108820A (en) | For selecting the method and system of the feature of machine learning sample | |
| Maji et al. | Stock prediction and mutual fund portfolio management using curve fitting techniques | |
| CN108846695A (en) | The prediction technique and device of terminal replacement cycle | |
| CN106407305A (en) | Data mining system and method | |
| TWI696971B (en) | Financial correlation prediction system and the method thereof | |
| Wang et al. | Research on the construction of stock portfolios based on multiobjective water cycle algorithm and KMV algorithm | |
| CN108304975A (en) | A kind of data prediction system and method | |
| Lotfi et al. | A new two-stage data envelopment analysis (DEA) model for evaluating the branch performance of banks | |
| Panigrahi et al. | Normalize time series and forecast using evolutionary neural network | |
| Xu et al. | Dynamic portfolio management based on pair trading and deep reinforcement learning | |
| CN110750572A (en) | Adaptive method and device for heuristic evaluation of scientific and technological achievements | |
| CN107578130A (en) | A kind of method of the copper forecasting of futures prix of the TSK fuzzy systems of structure and parameter combination learning | |
| Rifa'i et al. | Optimized fuzzy backpropagation neural network using genetic algorithm for predicting Indonesian stock exchange composite index | |
| Zandi et al. | Research on stock portfolio based on time series prediction and multi-objective optimization | |
| CN114358519A (en) | Intelligent credit limit interest rate adjusting method and device |
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: 20180904 |