CN109615106A - Stock yield method for pushing, device, storage medium and server - Google Patents
Stock yield method for pushing, device, storage medium and server Download PDFInfo
- Publication number
- CN109615106A CN109615106A CN201811205222.1A CN201811205222A CN109615106A CN 109615106 A CN109615106 A CN 109615106A CN 201811205222 A CN201811205222 A CN 201811205222A CN 109615106 A CN109615106 A CN 109615106A
- Authority
- CN
- China
- Prior art keywords
- stock
- prediction model
- data
- prediction
- stability coefficient
- 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
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Human Resources & Organizations (AREA)
- Finance (AREA)
- General Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Technology Law (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
It includes obtaining the historical stock data of default industry within a preset period of time that the present invention, which provides a kind of stock yield method for pushing, device, storage medium and server, the stock yield method for pushing, and the historical stock data are divided into training set and test set；It determines at least two classifiers, obtains corresponding at least two prediction model；Each prediction model is tested respectively, obtains the first stability coefficient of each test model；According to the stock certificate data after preset time period, the second stability coefficient and head combination information ratio are obtained；According to the first stability coefficient, the second stability coefficient and head combination information ratio, an optimum prediction model is determined, to generate stock yield data；If having the income of stock in stock yield data is more than preset threshold, the stock more than preset threshold is pushed into user.The present invention can determine an optimum prediction model according to multiple classifiers, improve the accuracy of stock yield prediction, reduce customer investment loss.
Description
Technical field
The present invention relates to field of computer technology, specifically, the present invention relates to a kind of stock yield method for pushing, dress
It sets, storage medium and server.
Background technique
Existing stock analysis or financial product analysis are much to be realized by professional by its professional standing and experience
, this analysis mode is easy the subjective impact by professional, and accuracy is lower.It is existing it is some utilize historical data carry out
In stock or the method for financial analysis, generally use static method, qualitative description it is more, quantitative description lacks, it is also difficult to by shadow
The factors for ringing stock market compare and analyze simultaneously, therefore, have biggish limitation；Moreover, existing static method master
Will use multiple-factor linear model, but the influence for having many factor pair next period incomes be not it is linear, so as to cause existing more
The prediction result of factor linear model differs greatly with actual result；Moreover, for different industries, the impact factor of stock market
Difference, therefore, it is difficult to relatively accurately predict that the stock market of all industries moves towards only with a kind of prediction model, and in part
In the case of, the prediction result in the short time is also not sufficiently stable, to influence the investment judgement of user.
Summary of the invention
The present invention is directed to the shortcomings that existing way, proposes a kind of stock yield method for pushing, device, storage medium and service
Device, to solve the problems, such as it is existing in the prior art at least one.
The prediction technique of stock yield proposed by the present invention includes:
The default historical stock data of industry within a preset period of time are obtained, the historical stock data are divided into training set
And test set；
It determines at least two classifiers, every kind of classifier is respectively trained according to the data of the training set, obtains corresponding
At least two prediction models；
According to the historical stock data of the test set, each prediction model is tested respectively, obtains each test model
First stability coefficient；
According to stock certificate data and each prediction model of the stock in the test set after preset time period,
The future profits for predicting stock in the test set respectively obtains the second stability coefficient and head combination information ratio；
According to first stability coefficient, the second stability coefficient and head combination information ratio, from described at least two
The optimum prediction model of a default industry is determined in a prediction model；
According to the optimum prediction model, stock yield data are generated；
Judging whether to have in the stock yield data income of stock is more than preset threshold, if so, then will be more than default
The stock of threshold value pushes to user.
Further, the historical stock data include each sample stock multiple preset time points stock price and
The basic side factor；The basic side factor includes: net profit, net profit speedup, estimates net profit speedup, net profit margin, net assets
Earning rate is full of and receives speedup, estimates to be full of and receive speedup, ratio, Leveraged rate, the p/e ratio phase of operational cash flow net amount and net profit
To profit rate of increase (PEG index), it is full of receipts scale, net profit scale.
Further, the historical stock data according to the test set, test each prediction model respectively, obtain every
First stability coefficient of a test model, comprising:
The input parameter that each prediction model is obtained from the historical stock data of the test set obtains each prediction mould
First probability value of type output；
According to the size of first probability value, the stock in the test set is divided into several grade intervals；
Judge the corresponding effective yield of all stocks in optimal grade interval whether than any one other grade area
Between in the corresponding effective yield of all stocks it is high；
If so, test result is denoted as 1；
If it is not, test result is then denoted as 0；
According to the test result of each preset time point, the average test number in the preset time period is obtained
Value, using the average test numerical value as the first stability coefficient.
Further, stock certificate data of the stock according in the test set after preset time period, and it is each
The prediction model predicts the future profits of stock in the test set respectively, obtains the second stability coefficient and head combination
Information ratio, comprising:
Determine multiple times timing points after the preset time period；
The stock in the test set is obtained in the corresponding stock certificate data of each described time timing points, from the stock certificate data
The middle input parameter for obtaining each prediction model obtains each time timing points correspond to the output of each prediction model second
Probability value；
According to the size of second probability value, the stock in the test set is divided into several grade intervals；Sentence
Whether the corresponding effective yield of all stocks broken in optimal grade interval is than all in any one other grade interval
The corresponding effective yield of stock is high；If so, prediction result is denoted as 1；If it is not, prediction result is then denoted as 0；According to each
The prediction result of described time timing points obtains the consensus forecast numerical value after the preset time period, by the average survey
Numerical value is tried as the second stability coefficient；
Using current time as the last one time timing points, according to the corresponding stock certificate data of current time and the prediction mould
Type obtains corresponding second probability value of each prediction model, and second probability value is suitable according to from big to small
Sequence is ranked up, using the stock of preset quantity in the top as head dynamic combined；Calculate separately each prediction mould
Head combination information ratio IR:IR=α/ω of the corresponding head dynamic combined of type, wherein α is the head dynamic combined
Excess earnings, ω be the head dynamic combined active risk.
It is further, described according to first stability coefficient, the second stability coefficient and head combination information ratio,
The optimum prediction model of a default industry is determined from least two prediction model, comprising:
By corresponding first stability coefficient of each prediction model, the second stability coefficient and head combination information ratio
Rate is multiplied, and obtains the corresponding product of each prediction model；
Using the maximum prediction model of product as optimum prediction model.
Further, described according to the optimum prediction model, after generating stock yield data, further includes:
According to the stock yield data, several highest stocks of income are determined in multiple industries, by different industries
The stock be combined, form stock portfolio, and push to user.
Further, the classifier include Logic Regression Models, support vector machines, Gauss model-naive Bayesian, with
One or more of machine forest model.
The present invention also proposes that a kind of prediction meanss of stock yield, described device include:
Data acquisition module, for obtaining the historical stock data of default industry within a preset period of time, by the history
Stock certificate data is divided into training set and test set；
Prediction model module is respectively trained every kind according to the data of the training set for determining at least two classifiers
Classifier obtains corresponding at least two prediction model；
Model measurement module is tested each prediction model respectively, is obtained for the historical stock data according to the test set
To the first stability coefficient of each test model；
Model prediction module, for the stock certificate data according to the stock in the test set after preset time period, and
Each prediction model, predicts the future profits of stock in the test set respectively, obtains the second stability coefficient and head
Combined information ratio；
Model determining module, for according to first stability coefficient, the second stability coefficient and head combination information
Ratio determines the optimum prediction model of a default industry from least two prediction model；
Earnings forecast module, for according to the optimum prediction model prediction stock yield；
Pushing module is more than preset threshold for judging whether to have in the stock yield data income of stock, if so,
The stock more than preset threshold is then pushed into user.
The present invention also proposes a kind of computer readable storage medium, is stored thereon with computer program, and the program is processed
Device realizes stock yield method for pushing described in aforementioned any one when executing.
The present invention also proposes that a kind of server, the server include:
One or more processors；
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes aforementioned described in any item stock yield method for pushing.
The invention has the following advantages:
1, the classifier that the present invention can be different using historical stock data training, to obtain different prediction models,
And according to first stability coefficient, the second stability coefficient and head combination information ratio, an optimum prediction mould is determined
Type improves the accuracy of stock yield prediction, reduces customer investment loss.
2, stock can be divided into several grades respectively according to the size of the first probability value and the second probability value by the present invention
Section with the stock of the optimal grade interval of determination, and the stock of optimal grade interval is compared with effective yield, can be further
It filters out and predicts more quasi- prediction model, be conducive to the higher stock of further screening income.
The additional aspect of the present invention and advantage will be set forth in part in the description, these will become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect and advantage of the invention will become from the following description of the accompanying drawings of embodiments
Obviously and it is readily appreciated that, in which:
Fig. 1 is the flow diagram of stock yield method for pushing first embodiment of the present invention；
Fig. 2 is the flow diagram of stock yield method for pushing second embodiment of the present invention；
Fig. 3 is the flow diagram of stock yield method for pushing second embodiment of the present invention；
Fig. 4 is the structural schematic diagram of server example of the present invention.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, and for explaining only the invention, and is not construed as limiting the claims.
Those skilled in the art of the present technique are appreciated that unless expressly stated, singular " one " used herein, " one
It is a ", " described " and "the" may also comprise plural form, " first " used herein, " second " are only used for distinguishing same technology special
Sign, is not limited the sequence of the technical characteristic and quantity etc..It is to be further understood that in specification of the invention
The wording " comprising " used refers to that there are the feature, integer, step, operation, element and/or component, but it is not excluded that depositing
Or add other one or more features, integer, step, operation, element, component and/or their group.
Those skilled in the art of the present technique are appreciated that unless otherwise defined, all terms used herein (including technology art
Language and scientific term), there is meaning identical with the general understanding of those of ordinary skill in fields of the present invention.Should also
Understand, those terms such as defined in the general dictionary, it should be understood that have in the context of the prior art
The consistent meaning of meaning, and unless idealization or meaning too formal otherwise will not be used by specific definitions as here
To explain.
Those skilled in the art of the present technique are appreciated that " server " used herein above, " server apparatus " both include nothing
The equipment of line signal receiver only has the equipment of the wireless signal receiver of non-emissive ability, and including receiving and emitting
The equipment of hardware has the reception that on bidirectional communication link, can execute two-way communication and emits the equipment of hardware.It is this
Equipment may include: honeycomb or other communication equipments, with single line display or multi-line display or no multi-line
The honeycomb of display or other communication equipments；PCS (Personal Communications Service, PCS Personal Communications System),
It can be with combine voice, data processing, fax and/or communication ability；PDA (Personal Digital Assistant,
Personal digital assistant), it may include radio frequency receiver, pager, the Internet/intranet access, web browser, account
Originally, calendar and/or GPS (Global Positioning System, global positioning system) receiver；Conventional laptop and/or
Palmtop computer or other equipment, have and/or conventional laptop and/or palmtop computer including radio frequency receiver
Or other equipment." server " used herein above, " server apparatus " can be it is portable, can transport, be mounted on traffic work
Have in (aviation, sea-freight and/or land), or be suitable for and/or be configured in local runtime, and/or with distribution form, fortune
Row is run in any other of the earth and/or space position." server " used herein above, " server apparatus " can also be
The equipment such as the communication server, service on net device.
The present invention proposes a kind of stock yield method for pushing, for improving the income of customer investment stock, as shown in Figure 1
First embodiment includes the following steps:
Step S10: obtaining the default historical stock data of industry within a preset period of time, by the historical stock data point
For training set and test set；
Step S20: determining at least two classifiers, and every kind of classifier is respectively trained according to the data of the training set, obtains
To corresponding at least two prediction model；
Step S30: according to the historical stock data of the test set, each prediction model is tested respectively, obtains each survey
First stability coefficient of die trial type；
Step S40: according to stock certificate data of the stock in the test set after preset time period and each described pre-
Model is surveyed, the future profits of stock in the test set is predicted respectively, obtains the second stability coefficient and head combination information ratio
Rate；
Step S50: according to first stability coefficient, the second stability coefficient and head combination information ratio, from institute
State the optimum prediction model that a default industry is determined at least two prediction models；
Step S60: according to the optimum prediction model, stock yield data are generated；
Step S70: judging whether to have in the stock yield data income of stock is more than preset threshold, if so, then will
Stock more than preset threshold pushes to user.
Wherein, each step is specific as follows:
Step S10: obtaining the default historical stock data of industry within a preset period of time, by the historical stock data point
For training set and test set.
The division of the default industry can be divided according to universal classifications standards such as industrial sectors of national economy classification, can also be tied
The industry characteristic self-defining for combining listed company in city divides.It is accurate that the preset time period can be predicted according to stock yield
Property and server operation determine, server license in the case where, the long period can be set by the preset time period, with
The prediction model is set to train more accurate model parameter according to great amount of samples data；The preset time period can also be according to row
Industry feature determines, such as the period where the stock certificate data of Software Industry listed company is shorter, and the listing of biological medicine class is public
Period where the stock certificate data of department is relatively long, then it is pre- can to set the preset time period of Software Industry to shorter first
If the period, and longer second preset time period is set by the preset time period of biological medicine class listed company, can be improved
Establish the speed of every profession and trade prediction model.
The historical stock data may include a variety of factors for influencing stock price such as the basic side factor and the funds factor.
The basic side factor may include book market value ratio (book value/market value), earnings yield rate (earn per
Share), dynamic profit rate (present price/whole year estimated per share earning ratio) price/earn, (present price/last year is estimated for static p/e ratio
Per share earning ratio) price/earn, ROE (return on equity) Return on Net Assets/common stock holder's equity earning rate,
Manage the factors such as rate of gross profit, net profit margin (P/R) mainly.The funds factor may include asset-liabilities (liability/asset), consolidate
Determine assets ratio (faculty assset/asset), circulation value (the circulating stock number of share of stock * circulating stock share price that can be traded at that time)
Etc. the factors.The historical stock data, which may also include, influences variation factor of intangible assets in enterprises corresponding to stock etc..It is described each
The definition of the factor and calculation can be found in the definition and calculation of existing financial industry, and details are not described herein.
Such as: historical stock data of the agricultural between 2013.1.1-2018.1.1 in listed company are obtained, and should
80% sample data in historical stock data is as training set, and 20% sample data is as test set.The training set
For establishing the prediction model, the test set is used to test the accuracy rate for the prediction model established.The training set
It should be strictly separated with the data in the test set, avoid repeating, to guarantee the prediction model test result trained
The accuracy of accuracy and the subsequent recommendation stock obtained according to each coefficient.
Step S20: determining at least two classifiers, and every kind of classifier is respectively trained according to the data of the training set, obtains
To corresponding at least two prediction model.
In one embodiment of the invention, the classifier includes Logic Regression Models, support vector machines, Gauss simplicity
One or more of Bayesian model, Random Forest model；It can also be derived according to the parameter of each classifier more different
Classifier.According to the historical stock data in the training set, different classifications device can be respectively trained out to different prediction models；
The prediction result of each different prediction model and actual stock certificate data are compared again, from different prediction models
The determining and immediate optimum prediction model of actual conditions.This step can be improved to obtain by determining multiple and different classifiers
The prediction model accuracy.
Step S30: according to the historical stock data of the test set, each prediction model is tested respectively, obtains each survey
First stability coefficient of die trial type.
After obtaining at least two classifiers, each classifier is tested in the test set, it is every to test
Whether the result of a different classifier output is consistent with actual conditions.The output result of the classifier can be according to demand
It is set as ups and downs or the income rate score etc. of stock.In some embodiments, the output of the prediction model can be income
Rate, so that the income correlation of the result of output and equity investment, is conducive to user and intuitively judges.For from all stocks
More good stock is filtered out in ticket data, this step can preset several numerical intervals, and each numerical intervals are one corresponding
Grade interval；For example, default five earning rate numerical intervals corresponding to five grade intervals, according to each different prediction mould
Each stock correspondence in the test set is divided in preset five grade intervals by the income rate score of type output, wherein
The first estate section is optimal grade, corresponds to the highest numerical intervals of earning rate；If described in five grade intervals
The effective yield of optimal grade interval is highest in all grade intervals really, then test result is remembered 1, on the contrary then be denoted as
0；If by the preset time period monthly No. 1 be used as testing time point, in the testing time point repeat this operation, i.e.,
One group of array being made of 1 and 0 can will be obtained, the mean value of the array is sought, the first stability coefficient C1 can be obtained, this is first steady
Qualitative coefficient C1 can represent prediction degree of stability of the corresponding prediction model in the preset time period.Therefore the present invention also proposes
Another embodiment as shown in Figure 2:
The historical stock data according to the test set, test each prediction model respectively, obtain each test mould
First stability coefficient of type, comprising:
Step S31: obtaining the input parameter of each prediction model from the historical stock data of the test set, obtains every
First probability value of a prediction model output；
Step S32: according to the size of first probability value, the stock in the test set is divided into several grades
Section；
Step S33: judge whether the corresponding effective yield of all stocks in optimal grade interval is more any one than other
The corresponding effective yield of all stocks in a grade interval is high；If so, thening follow the steps S34；If it is not, thening follow the steps
S35；
Step S34: test result is denoted as 1；
Step S35: test result is then denoted as 0；
Step S36: according to the test result of each preset time point, the survey in the preset time period is obtained
The average test numerical value of test result, using the average test numerical value as the first stability coefficient C1.
The probability value and earning rate correlation of each classifier output, the obtained probability value is bigger, indicates
The future profits rate of prediction is bigger.The input parameter of each different prediction model is identical, but can select the base according to demand
One or more factors in this face factor are as one or more factors in input parameter, or the selection funds factor
As input parameter and other a variety of factors for influencing stock price as input parameter.
In some embodiments, the first probability value of each different prediction model output, can be to receive with the stock
The coefficient that benefit is positively correlated, in order to which user directly judges whether to buy in the stock of prediction.
In order to further increase the accuracy of the information pushed to user, the present embodiment can also be according to first probability value
Size, the stock in the test set is divided into several grade intervals, with from it is all participate in test stocks in determine
Optimal stock；It is more connect when each grade interval corresponds to a probability value section, or by first probability value
When close stock is divided into a grade interval, then multiple stocks can be corresponded in each grade interval；If by the probability value area
Between highest section as optimal grade interval, then can also correspond to multiple stocks in the optimal grade interval；The present embodiment into
Whether one step judges the corresponding effective yield of all stocks in the optimal grade interval than any one other grade area again
Between effective yield it is high；If the effective yield than any one other grade interval is all high, illustrate according to described pre-
The stock and the higher stock of effective yield surveyed in the corresponding optimal grade interval of the first probability value that model obtains have height
The consistency of degree is, it can be achieved that screen the purpose of stock.
Since the stock of one group of optimal grade interval can be obtained in each predicted time point；And work as the prediction mould
When type has multiple, the stock of one group of optimal grade interval can be obtained in each predicted time point for each prediction model.For
Compare the stability of each prediction model prediction, can by each prediction model each predicted time point test result into
Row compares, to obtain the output most stable of prediction model of result.The present embodiment is according to the survey of each preset time point
The average test numerical value in the preset time period is calculated in test result, and using the average test numerical value as described first
Stability coefficient C1 can determine whether each different prediction model stability lasting in each predicted time point.
Step S40: according to stock certificate data of the stock in the test set after preset time period and each described pre-
Model is surveyed, the future profits of stock in the test set is predicted respectively, obtains the second stability coefficient and head combination information ratio
Rate.
Since the data of the test set are generally the historical data in a period, test result has certain
Limitation.In order to improve the prediction model in the follow-up time accuracy still with higher of the period, the present invention is also
The accuracy of the prediction model is further judged by the second stability coefficient C2 and head combination information ratio IR.
In one embodiment of the invention, as shown in figure 3, the stock according in the test set is in preset time
Stock certificate data and each prediction model after section, predict the future profits of stock in the test set respectively, obtain the
Two stability coefficient C2 and head combination information ratio IR, comprising:
Step S41: multiple times timing points after the preset time period are determined.Described time timing points can for the period compared with
For regular time point, such as it regard back No. 1, No. 15 monthly as timing points；In another example if the section of the preset time period
For 2013.1.1-2018.1.1, then the timing points that return can be No. 15 of the every month after on January 1st, 2018, with basis
Previous No. 15 stock certificate datas of the moon predict next month No. 15 second probability values；Again by second probability value of prediction with
The effective yield on next month 15 is compared, and obtains prediction result；And so on, the prediction knot of every month 15 can be obtained
Fruit, to obtain consensus forecast numerical value as second stability coefficient.Over time, the number of described time timing points
Amount can also increase, therefore can regularly update the second stability coefficient C2.
Step S42: obtaining stock in the test set in the corresponding stock certificate data of each described time timing points, from described
It is defeated corresponding to each prediction model to obtain each time timing points for the input parameter that each prediction model is obtained in stock certificate data
The second probability value out.
Each prediction model exports the calculating process of second probability value and obtains first probability value in this step
Process it is similar, the stock certificate data only obtained is different from the time interval where the historical stock data.For one
For a prediction model, can be obtained in multiple described time timing points include multiple second probability values one group of second probability
Value；When there are multiple and different prediction models, can correspond to obtain the second probability value of multiple groups.
Step S43: according to the size of second probability value, the stock in the test set is divided into several grades
Section；Judge the corresponding effective yield of all stocks in optimal grade interval whether than in any one other grade interval
The corresponding effective yield of all stocks it is high；If so, prediction result is denoted as 1；If it is not, prediction result is then denoted as 0；Root
According to the prediction result of each described time timing points, the consensus forecast numerical value after the preset time period is obtained, it will be described
Average test numerical value is as the second stability coefficient.
Second probability value is the numerical value of prediction being positively correlated with stock yield, this step can first preset several etc.
Grade section, the corresponding default value section of each grade interval, then second probability value of output is divided to pair
Answer the default value section of numerical value, and will wherein the maximum section of default value it is most excellent to obtain as optimal grade interval
The stock portfolio in grade section；Again by the effective yield of the stock in optimal grade interval and any one other grade interval
The effective yield of stock is compared, with determine prediction second probability value whether the stock with the optimal grade interval
The effective yield of ticket is positively correlated；Finally the prediction result that timing points are each returned in the optimal grade interval is averaged,
Obtain the second stability coefficient C2.The second stability coefficient C2 reflects the optimal grade that the prediction model obtains
The stability that each stock in section is predicted in each time timing points.
In another embodiment, this step also can divide again the stock in the step S33 in optimal grade interval
For several sub- grade intervals, it may be assumed that from the stock of optimal grade interval for corresponding to first probability value, again according to institute
It states the second probability value and is divided into the sub- grade interval more segmented, excellent middle select excellent purpose to realize.
Step S44: using current time as the last one time timing points, according to the corresponding stock certificate data of current time and institute
Prediction model is stated, corresponding second probability value of each prediction model is obtained；By second probability value according to from big
It is ranked up to small sequence, using the stock of preset quantity in the top as head dynamic combined；Calculate separately each institute
State head combination information ratio IR:IR=α/ω of the corresponding head dynamic combined of prediction model, wherein α is dynamic for the head
The excess earnings of state combination, ω are the active risk of the head dynamic combined.
The first stability coefficient C1, the second stability coefficient C2 were both needed to according to multiple time points before current time
Data calculated, timeliness is poor；This step, can be by calculating using current time as the last one time timing points
To corresponding second probability value of current time, to determine the stock of the preset quantity according to the size of second probability value
Ticket, and calculate head combination information ratio IR of the stock in current time of the preset quantity.The head combination information ratio
IR can reflect the prediction income of the stock of the preset quantity, to provide instant investment reference information for user.The head
The definition of dynamic combined, excess earnings, active risk can be found in the existing definition of existing financial industry, and details are not described herein.
Step S50: according to first stability coefficient, the second stability coefficient and head combination information ratio, from institute
State the optimum prediction model that a default industry is determined at least two prediction models.
First stability coefficient can reflect the stabilization for the optimal stock that each prediction model is selected from the test set
Property.Second stability coefficient can reflect in the time after the preset time period where each prediction model from the test set
The stability for the optimal stock selected, or, being segmented out again from the optimal stock portfolio selected in the test set optimal
When stock portfolio, for reflecting the stability of the optimal stock portfolio of son segmented out again, that is, reflects and excellent middle select excellent stock
The prediction stability of ticket.The head combination information ratio IR is for reflecting income of the corresponding stock on current point in time
Rate has stronger association with the current data of stock.The present invention can according to actual needs with above three parameter, from it is described to
The optimum prediction model that a default industry is determined in few two prediction models, is conducive to mention according to the actual situation for user
For more accurate prediction model and more accurate referential data.
In one embodiment of the invention, described according to first stability coefficient, the second stability coefficient and head
Portion's combined information ratio determines the optimum prediction model of a default industry, packet from least two prediction model
It includes:
By corresponding first stability coefficient of each prediction model, the second stability coefficient and head combination information ratio
Rate is multiplied, and obtains the corresponding product of each prediction model；
Using the maximum prediction model of product as optimum prediction model.
The present embodiment is directly by first stability coefficient, the second stability coefficient and head combination information ratio phase
Multiply, and optimum prediction model is determined according to resulting product, calculation is simple and effective.It certainly, in a practical situation, can also root
Weight system is respectively set according to the significance level of first stability coefficient, the second stability coefficient and head combination information ratio
Number, then each weight coefficient is believed with corresponding first stability coefficient, the second stability coefficient and head combination respectively
Breath ratio is summed after being multiplied, and determines optimum prediction model further according to the result of summation.
Step S60: according to the optimum prediction model, stock yield data are generated.
After obtaining the optimum prediction model, the subsequent income number that each stock is calculated only on the basis of the optimum prediction model
According to so that user obtains more accurate investment reference numerical value.It, can periodically more in order to improve the timeliness of the optimum prediction model
The new optimum prediction model, so that the parameter in the optimum prediction model can be redefined according to newest sample data,
To keep the timeliness of the optimum prediction model.
Step S70: judging whether to have in the stock yield data income of stock is more than preset threshold, if so, then will
Stock more than preset threshold pushes to user.
This step can monitor whether occur the stock of high yield in the stock yield data, when the income of certain stock is super
When crossing the preset threshold, that is, judge the stock for the stock of high yield, can real-time recommendation buy in user.The user can be
The user of specific direction, such as it is bound to the user of particular terminal, or be directed toward the user of special handset number；It also can be not specific
The user of direction, such as push to the advertisement position of the client of APP and clicked to be shown to uncertain user, or for uncertain user
It checks.
The present invention can train different classifiers using the historical stock data, to obtain different prediction models, and
According to first stability coefficient, the second stability coefficient and head combination information ratio, an optimum prediction model is determined,
The accuracy for improving stock yield prediction reduces customer investment loss.
In one embodiment of the invention, the historical stock data include each sample stock in multiple preset times
The stock price and the basic side factor of point；The basic side factor include: net profit, net profit speedup, estimate net profit speedup,
Net profit margin, net assets income ratio are full of and receive speedup, estimate ratio, the lever for being full of and receiving speedup, operational cash flow net amount and net profit
Ratio, p/e ratio the relative profitability rate of increase (PEG index), the factors such as receipts scale, net profit scale that are full of；Wherein each factor
It is specifically defined the universal standard that can refer to existing financial industry with calculation, details are not described herein.
It is described according to the optimum prediction model in another embodiment of the present invention, after generating stock yield data,
Further include:
According to the stock yield data, several highest stocks of income are determined in multiple industries, by different industries
The stock be combined, form stock portfolio, and push to user.
Compared to single equity investment, multiple stock portfolio can be shared to the risk of single advance versus decline, advantageously reduced
The investment risk of user；Compared to the stock portfolio for recommending same industry, the present embodiment can reduce same industry and systematicness occurs
Investment risk when risk further reduced the investment risk of user.
The present invention also proposes a kind of stock yield driving means, and described device includes:
Data acquisition module, for obtaining the historical stock data of default industry within a preset period of time, by the history
Stock certificate data is divided into training set and test set；
Prediction model module is respectively trained every kind according to the data of the training set for determining at least two classifiers
Classifier obtains corresponding at least two prediction model；
Model measurement module is tested each prediction model respectively, is obtained for the historical stock data according to the test set
To the first stability coefficient of each test model；
Model prediction module, for the stock certificate data according to the stock in the test set after preset time period, and
Each prediction model, predicts the future profits of stock in the test set respectively, obtains the second stability coefficient and head
Combined information ratio；
Model determining module, for according to first stability coefficient, the second stability coefficient and head combination information
Ratio determines the optimum prediction model of a default industry from least two prediction model；
Earnings forecast module, for according to the optimum prediction model prediction stock yield；
Pushing module is more than preset threshold for judging whether to have in the stock yield data income of stock, if so,
The stock more than preset threshold is then pushed into user.
The embodiment of the present invention also provides a kind of computer readable storage medium, is stored thereon with computer program, the program
The step of stock yield method for pushing described in above-mentioned any one is realized when being executed by processor.Wherein, the storage medium
Including but not limited to any kind of disk (including floppy disk, hard disk, CD, CD-ROM and magneto-optic disk), ROM (Read-Only
Memory, read-only memory), RAM (Random AcceSS Memory, immediately memory), EPROM (EraSable
Programmable Read-Only Memory, Erarable Programmable Read only Memory), EEPROM (Electrically
EraSable Programmable Read-Only Memory, Electrically Erasable Programmable Read-Only Memory), flash memory, magnetic card
Or light card.It is, storage medium includes the form storage or transmission information by equipment (for example, computer) can read
Any medium.It can be read-only memory, disk or CD etc..
The embodiment of the present invention also provides a kind of server, and the server includes:
One or more processors；
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes the step of stock yield method for pushing described in above-mentioned any one.
Fig. 4 be server of the present invention structural schematic diagram, including processor 320, storage device 330, input unit 340 with
And the equal devices of display unit 350.It will be understood by those skilled in the art that the structure devices shown in Fig. 4 are not constituted to all clothes
The restriction of business device may include than illustrating more or fewer components, or the certain components of combination.Storage device 330 can be used for
Application program 310 and each functional module are stored, processor 320 runs the application program 310 for being stored in storage device 330, from
And execute the various function application and data processing of equipment.Storage device 330 can be built-in storage or external memory, or
Including both built-in storage and external memory.Built-in storage may include that read-only memory, programming ROM (PROM), electricity can be compiled
Journey ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory or random access memory.External memory can be with
Including hard disk, floppy disk, ZIP disk, USB flash disk, tape etc..Storage device disclosed in this invention includes but is not limited to depositing for these types
Storage device.Storage device 330 disclosed in this invention is only used as example rather than as restriction.
Input unit 340 is used to receive the input of signal, and receive the historical stock data etc..Input unit 340
It may include touch panel and other input equipments.Touch panel collect user on it or nearby touch operation (such as
User uses the operations of any suitable object or attachment on touch panel or near touch panel such as finger, stylus), and
Corresponding attachment device is driven according to a pre-set procedure；Other input equipments can include but is not limited to physical keyboard, function
One of energy key (such as broadcasting control button, switch key etc.), trace ball, mouse, operating stick etc. are a variety of.Display unit
350 can be used for showing the information of user's input or be supplied to the information of user and the various menus of computer equipment.Display is single
The forms such as liquid crystal display, Organic Light Emitting Diode can be used in member 350.Processor 320 is the control centre of computer equipment, benefit
With the various pieces of various interfaces and the entire computer of connection, by run or execute be stored in it is soft in storage device 330
Part program and/or module, and the data being stored in storage device are called, perform various functions and handle data.
In one embodiment, server includes one or more processors 320, and one or more storage devices
330, one or more application program 310, wherein one or more of application programs 310 are stored in storage device 330
And be configured as being executed by one or more of processors 320, one or more of application programs 310 are configured to carry out
Stock yield method for pushing described in above embodiments.
It should be understood that each functional unit in various embodiments of the present invention can be integrated in a processing module,
It can be physically existed alone, can also be integrated in two or more units in a module with each unit.It is above-mentioned integrated
Module both can take the form of hardware realization, can also be realized in the form of software function module.
The above is only some embodiments of the invention, it is noted that for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
It is considered as protection scope of the present invention.
Claims (10)
1. a kind of stock yield method for pushing, which is characterized in that comprising steps of
The default historical stock data of industry within a preset period of time are obtained, the historical stock data are divided into training set and survey
Examination collection；
Determine at least two classifiers, every kind of classifier be respectively trained according to the data of the training set, obtain it is corresponding at least
Two prediction models；
According to the historical stock data of the test set, each prediction model is tested respectively, obtains the first of each test model
Stability coefficient；
According to stock certificate data and each prediction model of the stock in the test set after preset time period, respectively
The future profits for predicting stock in the test set obtains the second stability coefficient and head combination information ratio；
It is pre- from described at least two according to first stability coefficient, the second stability coefficient and head combination information ratio
It surveys in model and determines the optimum prediction model of a default industry；
According to the optimum prediction model, stock yield data are generated；
Judging whether to have in the stock yield data income of stock is more than preset threshold, if so, then will be more than preset threshold
Stock push to user.
2. the method according to claim 1, wherein the historical stock data include each sample stock more
The stock price and the basic side factor of a preset time point；The basic side factor includes: net profit, net profit speedup, estimates
Net profit speedup, net assets income ratio, is full of and receives speedup, estimates to be full of and receive speedup, operational cash flow net amount and net profit net profit margin
Ratio, Leveraged rate, p/e ratio the relative profitability rate of increase (PEG index), be full of receipts scale, net profit scale.
3. the method according to claim 1, wherein the historical stock data according to the test set, divide
Each prediction model is not tested, obtains the first stability coefficient of each test model, comprising:
The input parameter that each prediction model is obtained from the historical stock data of the test set, it is defeated to obtain each prediction model
The first probability value out；
According to the size of first probability value, the stock in the test set is divided into several grade intervals；
Judge the corresponding effective yield of all stocks in optimal grade interval whether than in any one other grade interval
The corresponding effective yield of all stocks it is high；
If so, test result is denoted as 1；
If it is not, test result is then denoted as 0；
According to the test result of each preset time point, the average test numerical value in the preset time period is obtained,
Using the average test numerical value as the first stability coefficient.
4. the method according to claim 1, wherein the stock according in the test set is in preset time
Stock certificate data and each prediction model after section, predict the future profits of stock in the test set respectively, obtain the
Two stability coefficients and head combination information ratio, comprising:
Determine multiple times timing points after the preset time period；
The stock in the test set is obtained in the corresponding stock certificate data of each described time timing points, is obtained from the stock certificate data
The input parameter for taking each prediction model obtains the second probability that each time timing points correspond to the output of each prediction model
Value；
According to the size of second probability value, the stock in the test set is divided into several grade intervals；Judgement is most
Whether the corresponding effective yield of all stocks in excellent grade interval is than all stocks in any one other grade interval
Corresponding effective yield is high；If so, prediction result is denoted as 1；If it is not, prediction result is then denoted as 0；According to each described
The prediction result for returning timing points, obtains the consensus forecast numerical value after the preset time period, by the average test number
Value is used as the second stability coefficient；
Using current time as the last one time timing points, according to the corresponding stock certificate data of current time and the prediction model,
Obtain corresponding second probability value of each prediction model, by second probability value according to sequence from big to small into
Row sequence, using the stock of preset quantity in the top as head dynamic combined；Calculate separately each prediction model
Head combination information ratio IR:IR=α/ω of corresponding head dynamic combined, wherein α is the super of the head dynamic combined
Volume income, ω are the active risk of the head dynamic combined.
5. the method according to claim 1, wherein described stablize according to first stability coefficient, second
Property coefficient and head combination information ratio determine the best pre- of the default industry from least two prediction model
Survey model, comprising:
By corresponding first stability coefficient of each prediction model, the second stability coefficient and head combination information ratio phase
Multiply, obtains the corresponding product of each prediction model；
Using the maximum prediction model of product as optimum prediction model.
6. generation stock is received the method according to claim 1, wherein described according to the optimum prediction model
After beneficial data, further includes:
According to the stock yield data, several highest stocks of income are determined in multiple industries, by the institute of different industries
It states stock to be combined, forms stock portfolio, and push to user.
7. the method according to claim 1, wherein the classifier includes Logic Regression Models, supporting vector
One or more of machine, Gauss model-naive Bayesian, Random Forest model.
8. a kind of prediction meanss of stock yield characterized by comprising
Data acquisition module, for obtaining the historical stock data of default industry within a preset period of time, by the historical stock
Data are divided into training set and test set；
Every kind of classification is respectively trained according to the data of the training set for determining at least two classifiers in prediction model module
Device obtains corresponding at least two prediction model；
Model measurement module tests each prediction model for the historical stock data according to the test set respectively, obtains every
First stability coefficient of a test model；
Model prediction module, for the stock certificate data according to the stock in the test set after preset time period, and it is each
The prediction model predicts the future profits of stock in the test set respectively, obtains the second stability coefficient and head combination
Information ratio；
Model determining module is used for according to first stability coefficient, the second stability coefficient and head combination information ratio,
The optimum prediction model of a default industry is determined from least two prediction model；
Earnings forecast module, for according to the optimum prediction model prediction stock yield；
Pushing module is more than preset threshold for judging whether to have in the stock yield data income of stock, if so, then will
Stock more than preset threshold pushes to user.
9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor
The prediction technique of stock yield as claimed in any of claims 1 to 7 in one of claims is realized when row.
10. a kind of server, which is characterized in that the server includes:
One or more processors；
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
The now prediction technique of stock yield as claimed in any of claims 1 to 7 in one of claims.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811205222.1A CN109615106A (en) | 2018-10-16 | 2018-10-16 | Stock yield method for pushing, device, storage medium and server |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811205222.1A CN109615106A (en) | 2018-10-16 | 2018-10-16 | Stock yield method for pushing, device, storage medium and server |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109615106A true CN109615106A (en) | 2019-04-12 |
Family
ID=66001684
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811205222.1A Pending CN109615106A (en) | 2018-10-16 | 2018-10-16 | Stock yield method for pushing, device, storage medium and server |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109615106A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110443374A (en) * | 2019-08-14 | 2019-11-12 | 腾讯科技（深圳）有限公司 | A kind of resource information processing method, device and equipment |
| TWI708202B (en) * | 2019-04-16 | 2020-10-21 | 元大證券投資信託股份有限公司 | Stock-to-bond ratio trend prediction system and stock-to-bond ratio trend prediction method |
| CN113781246A (en) * | 2021-09-14 | 2021-12-10 | 平安科技（深圳）有限公司 | Policy generation method and device based on preset label and storage medium |
-
2018
- 2018-10-16 CN CN201811205222.1A patent/CN109615106A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI708202B (en) * | 2019-04-16 | 2020-10-21 | 元大證券投資信託股份有限公司 | Stock-to-bond ratio trend prediction system and stock-to-bond ratio trend prediction method |
| CN110443374A (en) * | 2019-08-14 | 2019-11-12 | 腾讯科技（深圳）有限公司 | A kind of resource information processing method, device and equipment |
| CN113781246A (en) * | 2021-09-14 | 2021-12-10 | 平安科技（深圳）有限公司 | Policy generation method and device based on preset label and storage medium |
| CN113781246B (en) * | 2021-09-14 | 2024-03-05 | 平安科技（深圳）有限公司 | Strategy generation method and device based on preset label and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Banker | Maximum likelihood, consistency and data envelopment analysis: a statistical foundation | |
| Arrighetti et al. | Intangible assets and firm heterogeneity: Evidence from Italy | |
| Gonzálvez-Gallego et al. | Using integrated information systems in supply chain management | |
| Pál et al. | Financing constraints and firms’ cash policy in the euro area | |
| US20060004653A1 (en) | Apparatus, method and system for a designing and trading macroeconomic investment views | |
| US20210233177A1 (en) | Method and apparatus for determining inventor impact | |
| Firk et al. | Exploring Value‐Based Management sophistication: the role of potential economic benefits and institutional influence | |
| Guidolin et al. | Forecasting commodity futures returns with stepwise regressions: Do commodity-specific factors help? | |
| CN109615106A (en) | Stock yield method for pushing, device, storage medium and server | |
| WO2020150611A1 (en) | Systems and methods for entity performance and risk scoring | |
| Gupta-Mukherjee | When active fund managers deviate from their peers: Implications for fund performance | |
| Chuprinin et al. | Adjusting to the information environment: News tangibility and mutual fund performance | |
| Buhl et al. | Valuing customer portfolios under risk-return-aspects: a model-based approach and its application in the financial services industry | |
| Spelta et al. | Financial crises: Uncovering self-organized patterns and predicting stock markets instability | |
| Ma et al. | Deconstructing uncertainty | |
| KR100944117B1 (en) | Analysis method for tender using probability distribution of bid price | |
| Hoekman et al. | Aid for trade and trade in services | |
| Stuchfield et al. | Modeling the profitability of customer relationships: development and impact of Barclays de Zoete Wedd’s BEATRICE | |
| Celik Girgin et al. | Valuation mismatch and shipping q indicator for shipping asset management | |
| MacEachern | Measuring the added value of library and information services: the New Zealand approach | |
| Guidolin et al. | Chapter 16 The Economic and Statistical Value of Forecast Combinations Under Regime Switching: An Application to Predictable US Returns | |
| Son et al. | Supply chain information in analyst reports on publicly traded companies | |
| Masad et al. | Electronic Banking and its Impact on Job Security in Jordanian Banking Sector: Profitability is an Intermediate Variable | |
| Karim et al. | Marketing strategies and export of Iranian medicinal plants | |
| Gallagher et al. | Style factor timing: An application to the portfolio holdings of US fund managers |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190412 |
|
| RJ01 | Rejection of invention patent application after publication |