CN109410041A - A kind of high-dimensional deal maker's appraisal procedure and system by data-driven - Google Patents
A kind of high-dimensional deal maker's appraisal procedure and system by data-driven Download PDFInfo
- Publication number
- CN109410041A CN109410041A CN201710709808.0A CN201710709808A CN109410041A CN 109410041 A CN109410041 A CN 109410041A CN 201710709808 A CN201710709808 A CN 201710709808A CN 109410041 A CN109410041 A CN 109410041A
- Authority
- CN
- China
- Prior art keywords
- deal maker
- training
- target
- characteristic
- deal
- 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
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- 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
Abstract
The present invention provides a kind of high-dimensional deal maker's appraisal procedures by data-driven, which comprises establishes deal maker's analysis foundation feature database；Based on the foundation characteristic library, the feature in the foundation characteristic library is combined, to obtain a comprehensive character library；Based on a specific assessment dimension target, the transaction record of deal maker is handled, generates learning objective for each deal maker's sample；According to the trading processing is obtained training, target data set, be trained to obtain evaluation model M using nonlinear model；The characteristic value for obtaining the deal maker obtains the predicted target values of the deal maker in conjunction with the characteristic value and evaluation model M.This method and system can carry out comprehensive analysis to multiple dimensions, and on different analytic angles, provide corresponding deal maker's ability, risk assessment by the transaction record of deal maker.
Description
Technical field
The present invention relates to data analysis field more particularly to a kind of high-dimensional deal maker's appraisal procedure by data-driven and
System.
Background technique
Security practitioner needs to establish the complete investment capacity and trend data library to investor (and deal maker of broad sense),
And long-term accurately tracking is carried out to its trading capacity, risk partiality, to provide the information such as corresponding investment ability to bear grading.
The often static, low-dimensional data that conventional approach is relied on, i.e., present total assets, flowing implementations, risk
Then questionnaire etc. carries out the analysis of single dimension, the accuracy evaluated is low, subjectivity ambiguity is relatively strong, can not comprehensively consider
The information of deal maker's various dimensions.
Summary of the invention
Based on this, it is necessary to provide a kind of high-dimensional deal maker's appraisal procedures and system by data-driven by the present invention.
A kind of high-dimensional deal maker's appraisal procedure by data-driven, which comprises
Establish deal maker's analysis foundation feature database；
Based on the foundation characteristic library, the feature in the foundation characteristic library is combined, to obtain a comprehensive character library；
Based on a specific assessment dimension target, the transaction record of deal maker is handled, is each deal maker's sample
Generate learning objective；
According to the trading processing is obtained training, target data set, be trained and commented using nonlinear model
Valence model M；
The characteristic value for obtaining the deal maker obtains the prediction mesh of the deal maker in conjunction with the characteristic value and evaluation model M
Scale value.
The characteristic value for obtaining the deal maker in the step in one of the embodiments, in conjunction with the characteristic value and is commented
Valence model M, after obtaining the predicted target values of the deal maker, the method also includes:
The predicted target values are ranked up；
And the ranking results are normalized to target zone, to obtain the assessment result of deal maker.
It is described based on a specific assessment dimension target in one of the embodiments, to transaction
The step of transaction record of member is handled, and generates learning objective for each deal maker's sample
Specifically:
According to timeslice cutting training set and object set；
Giving a certain specific deal maker T, wherein the first stroke exchange hour point is A, and last transaction time point is S,
Future time piece is N；
The period for intercepting (A, S~N) is training data section, and (S~N, S) is target data segment, carries out cutting；
According to the training set of cutting, feature extraction is carried out to the training set；
Different object sets is respectively obtained according to different assessment targets for obtained object set.
The basis obtains the trading processing in one of the embodiments, training, target data set, use are non-
Linear model is trained the step of obtaining evaluation model M and includes:
According to being handled the transaction to obtain training dataset, target data set；
It takes gradient to integrate tree-model to be trained to obtain model M.
The feature in the foundation characteristic library includes: in one of the embodiments,
Performance characteristic, the performance characteristic include that Sharpe Ratio and maximum are withdrawn；
Behavioural characteristic, the behavioural characteristic include opening close a position Annual distribution and position in storehouse distribution；
Psychological characteristics, the psychological characteristics include misalignment effects and Herd Behavior.
A kind of high-dimensional deal maker's assessment system by data-driven, the system comprises:
Module is established, for establishing deal maker's analysis foundation feature database；
Composite module is based on the foundation characteristic library, is combined to the feature in the foundation characteristic library, complete to obtain one
Region feature library；
Target generation module handles the transaction record of deal maker, is every based on a specific assessment dimension target
A deal maker's sample generates learning objective；
Training module, for according to the trading processing is obtained training, target data set, using nonlinear model into
Row training obtains evaluation model M；
Target value obtains module and, in conjunction with the characteristic value and evaluation model M, obtains for obtaining the characteristic value of the deal maker
To the predicted target values of the deal maker.
In one of the embodiments, the system also includes:
Sorting module, for being ranked up to the predicted target values；
Module is normalized, for normalizing the ranking results to target zone, to obtain the assessment result of deal maker.
The target generation module includes: in one of the embodiments,
Cutting unit, for according to timeslice cutting training set and object set；
Given unit, for giving a certain specific deal maker T, wherein the first stroke exchange hour point is A, finishing touch
Exchange hour point is S, and future time piece is N；
Interception unit, the period for intercepting (A, S~N) is training data section, and (S~N, S) is target data segment,
Carry out cutting；
Extracting unit carries out feature extraction to the training set for the training set according to cutting；
Object set unit is obtained, for respectively obtaining different for obtained object set according to different assessment targets
Object set.
The training module includes: in one of the embodiments,
Processing unit handles the transaction to obtain training dataset, target data set for basis；
Training unit is trained to obtain model M for taking gradient to integrate tree-model.
The feature in the foundation characteristic library includes: in one of the embodiments,
Performance characteristic, the performance characteristic include that Sharpe Ratio and maximum are withdrawn；
Behavioural characteristic, the behavioural characteristic include opening close a position Annual distribution and position in storehouse distribution；
Psychological characteristics, the psychological characteristics include misalignment effects and Herd Behavior.
The utility model has the advantages that
The present invention provides a kind of high-dimensional deal maker's appraisal procedures by data-driven, which comprises establishes and hands over
Easy member's analysis foundation feature database；Based on the foundation characteristic library, the feature in the foundation characteristic library is combined, to obtain one
Comprehensive character library；Based on a specific assessment dimension target, the transaction record of deal maker is handled, is each deal maker's sample
This generation learning objective；According to the trading processing is obtained training, target data set, be trained using nonlinear model
Obtain evaluation model M；The characteristic value for obtaining the deal maker obtains the deal maker's in conjunction with the characteristic value and evaluation model M
Predicted target values.This method and system can carry out comprehensive analysis to multiple dimensions by the transaction record of deal maker, and
On different analytic angles, corresponding deal maker's ability, risk assessment are provided.
Detailed description of the invention
Fig. 1 is a kind of high-dimensional deal maker's appraisal procedure flow chart by data-driven of the invention.
Fig. 2 is a kind of high-dimensional deal maker's assessment system block diagram by data-driven of the invention.
Specific embodiment
To make those skilled in the art that the technical problems to be solved by the invention, technical side be more clearly understood
Case and advantageous effects, below in conjunction with drawings and examples, the present invention is further elaborated.
Please refer to Fig. 1, a kind of high-dimensional deal maker's appraisal procedure by data-driven, which comprises
S100: deal maker's analysis foundation feature database is established.
It should be noted that the foundation characteristic library is to analyze base from the deal maker of deal maker's record foundation of most original
Plinth feature database, feature database includes multiple class another characteristics, including performance characteristic, behavioural characteristic, psychological characteristics etc..Wherein,
Parameters, the maximums such as performance characteristic includes Sharpe Ratio and maximum is withdrawn withdraw rate: any history time point is backward within the selected period
It pushes away, earning rate when product net value goes to minimum point withdraws the maximum value of amplitude.Maximum is withdrawn can after describing to buy in product
The case where worst that can occur.It is an important risk indicator that maximum, which is withdrawn, and hedge fund sum number quantization strategy is handed over
Easily, the index is also more important than stability bandwidth.Sharpe Ratio (Sharpe Ratio), be otherwise known as Sharp Ratio, fund valuation
Standardized index.Sharpe Ratio is in modern investment theory studies have shown that the size of risk has on determining combined performance
Basic effect.In addition, behavioural characteristic includes opening close a position Annual distribution and position in storehouse distribution；Psychological characteristics include misalignment effects and
Herd Behavior etc..
S200: being based on the foundation characteristic library, be combined to the feature in the foundation characteristic library, comprehensively special to obtain one
Levy library.
It should be noted that the comprehensive character library includes compound characteristics library and foundation characteristic library, it, should after being combined
Comprehensive character library possesses a character engine.
S300: based on a specific assessment dimension target, the transaction record of deal maker is handled, is each deal maker
Sample generates learning objective；
It should be noted that the assessment target of the learning objective collection includes earnings target.Air control target etc..
S400: according to the trading processing is obtained training, target data set, be trained using nonlinear model
To evaluation model M；
It should be noted that in the present embodiment, the decision tree in the nonlinear model is classical efficient machine learning
Sorting algorithm is highly suitable for linear model effect and is unable to meet demand, and rule description is distributed proper scene, and decision
The methods of tree is combined together with traditional thought, just formation Assembled tree model method, including.It is pushed away in Baidu search keyword search
It recommends in system strategy, experiments have shown that integrated tree-model estimates classification accuracy with very high.
S500: obtaining the characteristic value of the deal maker, in conjunction with the characteristic value and evaluation model M, obtains the deal maker's
Predicted target values.
The characteristic value for obtaining the deal maker in the step in one of the embodiments, in conjunction with the characteristic value and is commented
Valence model M, after obtaining the predicted target values of the deal maker, the method also includes:
The predicted target values are ranked up；
And the ranking results are normalized to target zone, to obtain the assessment result of deal maker.
It is described based on a specific assessment dimension target in one of the embodiments, to the transaction record of deal maker into
Row processing is the step of each deal maker's sample generates learning objective specifically:
According to timeslice cutting training set and object set；
Giving a certain specific deal maker T, wherein the first stroke exchange hour point is A, and last transaction time point is S,
Future time piece is N；
The period for intercepting (A, S~N) is training data section, and (S~N, S) is target data segment, carries out cutting；
According to the training set of cutting, feature extraction is carried out to the training set；
Different object sets is respectively obtained according to different assessment targets for obtained object set.
The basis obtains the trading processing in one of the embodiments, training, target data set, use are non-
Linear model is trained the step of obtaining evaluation model M and includes:
According to being handled the transaction to obtain training dataset, target data set；
It takes gradient to integrate tree-model to be trained to obtain model M.
The feature in the foundation characteristic library includes: in one of the embodiments,
Performance characteristic, the performance characteristic include that Sharpe Ratio and maximum are withdrawn；
Behavioural characteristic, the behavioural characteristic include opening close a position Annual distribution and position in storehouse distribution；
Psychological characteristics, the psychological characteristics include misalignment effects and Herd Behavior.
It should be noted that opening Annual distribution of closing a position is a concept in financial field, in general, forward price meeting
Go up, just do more (buy and open a position), rising (selling) is closed a position, and earn: price difference=exit price-opens a position valence.If forward price can under
Fall, just does empty (sell and open a position), fall and (buy) and close a position, earn: the price difference=valence of opening a position-exit price.It opens a position and exactly signs (futures)
Deal contract；Close a position is exactly to fulfil (futures) deal contract.In addition, Herd Behavior refers in the market do not have in the presence of those in economics
Have to form the investor being expected or there is no proficiency information of oneself, they will change certainly according to the behavior of other investors
Oneself behavior.
The present invention provides a kind of high-dimensional deal maker's appraisal procedures by data-driven, which comprises establishes and hands over
Easy member's analysis foundation feature database；Based on the foundation characteristic library, the feature in the foundation characteristic library is combined, to obtain one
Comprehensive character library；Based on a specific assessment dimension target, the transaction record of deal maker is handled, is each deal maker's sample
This generation learning objective；According to the trading processing is obtained training, target data set, be trained using nonlinear model
Obtain evaluation model M；The characteristic value for obtaining the deal maker obtains the deal maker's in conjunction with the characteristic value and evaluation model M
Predicted target values.This method and system can carry out comprehensive analysis to multiple dimensions by the transaction record of deal maker, and
On different analytic angles, corresponding deal maker's ability, risk assessment are provided.
Referring to figure 2., a kind of high-dimensional deal maker's assessment system by data-driven, the system comprises:
Module 10 is established, for establishing deal maker's analysis foundation feature database；
Composite module 20 is based on the foundation characteristic library, is combined to the feature in the foundation characteristic library, to obtain one
Comprehensive character library；
Target generation module 30 handles the transaction record of deal maker, is based on a specific assessment dimension target
Each deal maker's sample generates learning objective；
Training module 40, the training that the trading processing is obtained for basis, target data set, using nonlinear model
It is trained to obtain evaluation model M；
It should be noted that in the present embodiment, the decision tree in the nonlinear model is classical efficient machine learning
Sorting algorithm is highly suitable for linear model effect and is unable to meet demand, and rule description is distributed proper scene, and decision
The methods of tree is combined together with traditional thought, just formation Assembled tree model method, including.It is pushed away in Baidu search keyword search
It recommends in system strategy, experiments have shown that integrated tree-model estimates classification accuracy with very high.
Target value obtains module 50, for obtaining the characteristic value of the deal maker, in conjunction with the characteristic value and evaluation model M,
Obtain the predicted target values of the deal maker.
In one of the embodiments, the system also includes:
Sorting module, for being ranked up to the predicted target values；
Module is normalized, for normalizing the ranking results to target zone, to obtain the assessment result of deal maker.
The target generation module includes: in one of the embodiments,
Cutting unit, for according to timeslice cutting training set and object set；
Given unit, for giving a certain specific deal maker T, wherein the first stroke exchange hour point is A, finishing touch
Exchange hour point is S, and future time piece is N；
Interception unit, the period for intercepting (A, S~N) is training data section, and (S~N, S) is target data segment,
Carry out cutting；
Extracting unit carries out feature extraction to the training set for the training set according to cutting；
Object set unit is obtained, for respectively obtaining different for obtained object set according to different assessment targets
Object set.
The training module includes: in one of the embodiments,
Processing unit handles the transaction to obtain training dataset, target data set for basis；
Training unit is trained to obtain model M for taking gradient to integrate tree-model.
The feature in the foundation characteristic library includes: in one of the embodiments,
Performance characteristic, the performance characteristic include that Sharpe Ratio and maximum are withdrawn；
Behavioural characteristic, the behavioural characteristic include opening close a position Annual distribution and position in storehouse distribution；
Psychological characteristics, the psychological characteristics include misalignment effects and Herd Behavior.
The present invention provides a kind of high-dimensional deal maker's assessment system by data-driven, the system comprises: establish mould
Block, for establishing deal maker's analysis foundation feature database；Composite module is based on the foundation characteristic library, to the foundation characteristic library
Feature be combined, to obtain a comprehensive character library；Target generation module, based on a specific assessment dimension target, to friendship
The transaction record of Yi Yuan is handled, and generates learning objective for each deal maker's sample；Training module, for according to the friendship
Easy to handle obtained training, target data set are trained to obtain evaluation model M using nonlinear model；Target value obtains mould
Block, in conjunction with the characteristic value and evaluation model M, obtains the prediction mesh of the deal maker for obtaining the characteristic value of the deal maker
Scale value.The system can carry out comprehensive analysis to multiple dimensions, and at different analysis angles by the transaction record of deal maker
On degree, corresponding deal maker's ability, risk assessment are provided.
The above description is only a preferred embodiment of the present invention, rather than does limitation in any form to the present invention.This field
Technical staff can impose various equivalent changes and improvement, all institutes within the scope of the claims on the basis of the above embodiments
The equivalent variations or modification done, should all fall under the scope of the present invention.
Claims (10)
1. a kind of high-dimensional deal maker's appraisal procedure by data-driven, which is characterized in that the described method includes:
Establish deal maker's analysis foundation feature database；
Based on the foundation characteristic library, the feature in the foundation characteristic library is combined, to obtain a comprehensive character library；
Based on a specific assessment dimension target, the transaction record of deal maker is handled, is generated for each deal maker's sample
Learning objective；
According to the trading processing is obtained training, target data set, be trained to obtain evaluation mould using nonlinear model
Type M；
The characteristic value for obtaining the deal maker obtains the prediction target of the deal maker in conjunction with the characteristic value and evaluation model M
Value.
2. the method according to claim 1, wherein obtain the characteristic value of the deal maker in the step, in conjunction with
The characteristic value and evaluation model M, after obtaining the predicted target values of the deal maker, the method also includes:
The predicted target values are ranked up；
And the ranking results are normalized to target zone, to obtain the assessment result of deal maker.
3. the method according to claim 1, wherein described be based on a specific assessment dimension target, to transaction
The step of transaction record of member is handled, and generates learning objective for each deal maker's sample specifically:
According to timeslice cutting training set and object set；
Give a certain specific deal maker T, wherein the first stroke exchange hour point is A, and last transaction time point is S, future
Timeslice is N；
The period for intercepting (A, S~N) is training data section, and (S~N, S) is target data segment, carries out cutting；
According to the training set of cutting, feature extraction is carried out to the training set；
Different object sets is respectively obtained according to different assessment targets for obtained object set.
4. the method according to claim 1, wherein the basis trading processing is obtained training, mesh
Data set is marked, being trained the step of obtaining evaluation model M using nonlinear model includes:
According to being handled the transaction to obtain training dataset, target data set；
It takes gradient to integrate tree-model to be trained to obtain model M.
5. the method according to claim 1, wherein the feature in the foundation characteristic library includes:
Performance characteristic, the performance characteristic include that Sharpe Ratio and maximum are withdrawn；
Behavioural characteristic, the behavioural characteristic include opening close a position Annual distribution and position in storehouse distribution；
Psychological characteristics, the psychological characteristics include misalignment effects and Herd Behavior.
6. a kind of high-dimensional deal maker's assessment system by data-driven, which is characterized in that the system comprises:
Module is established, for establishing deal maker's analysis foundation feature database；
Composite module is based on the foundation characteristic library, is combined to the feature in the foundation characteristic library, comprehensively special to obtain one
Levy library；
Target generation module handles the transaction record of deal maker, is each friendship based on a specific assessment dimension target
Easy member's sample generates learning objective；
Training module, the training obtained for basis to the trading processing, target data set, is instructed using nonlinear model
Get evaluation model M；
Target value obtains module and, in conjunction with the characteristic value and evaluation model M, obtains institute for obtaining the characteristic value of the deal maker
State the predicted target values of deal maker.
7. system according to claim 6, which is characterized in that the system also includes:
Sorting module, for being ranked up to the predicted target values；
Module is normalized, for normalizing the ranking results to target zone, to obtain the assessment result of deal maker.
8. system according to claim 6, which is characterized in that the target generation module includes:
Cutting unit, for according to timeslice cutting training set and object set；
Given unit, for giving a certain specific deal maker T, wherein the first stroke exchange hour point is A, last transaction
Time point is S, and future time piece is N；
Interception unit, the period for intercepting (A, S~N) is training data section, and (S~N, S) is target data segment, is carried out
Cutting；
Extracting unit carries out feature extraction to the training set for the training set according to cutting；
Object set unit is obtained, for respectively obtaining different targets according to different assessment targets for obtained object set
Collection.
9. system according to claim 6, which is characterized in that the training module includes:
Processing unit handles the transaction to obtain training dataset, target data set for basis；
Training unit is trained to obtain model M for taking gradient to integrate tree-model.
10. system according to claim 6, which is characterized in that the feature in the foundation characteristic library includes:
Performance characteristic, the performance characteristic include that Sharpe Ratio and maximum are withdrawn；
Behavioural characteristic, the behavioural characteristic include opening close a position Annual distribution and position in storehouse distribution；
Psychological characteristics, the psychological characteristics include misalignment effects and Herd Behavior.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710709808.0A CN109410041A (en) | 2017-08-17 | 2017-08-17 | A kind of high-dimensional deal maker's appraisal procedure and system by data-driven |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710709808.0A CN109410041A (en) | 2017-08-17 | 2017-08-17 | A kind of high-dimensional deal maker's appraisal procedure and system by data-driven |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109410041A true CN109410041A (en) | 2019-03-01 |
Family
ID=65455182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710709808.0A Pending CN109410041A (en) | 2017-08-17 | 2017-08-17 | A kind of high-dimensional deal maker's appraisal procedure and system by data-driven |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109410041A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101564328A (en) * | 2009-05-07 | 2009-10-28 | 杭州电子科技大学 | Laptop artificial limb multi-movement-mode identifying method based on support vector data description |
| CN102214320A (en) * | 2010-04-12 | 2011-10-12 | 宋威 | Neural network training method and junk mail filtering method using same |
| WO2014111540A1 (en) * | 2013-01-21 | 2014-07-24 | Ides Technologies Sa | System and method for characterizing financial messages |
| CN104657496A (en) * | 2015-03-09 | 2015-05-27 | 杭州朗和科技有限公司 | Method and equipment for calculating information hot value |
| CN106022907A (en) * | 2016-05-13 | 2016-10-12 | 清华大学 | Method and system for predicting trend of background core transaction event of large commercial bank |
| CN106445977A (en) * | 2015-08-12 | 2017-02-22 | 阿里巴巴集团控股有限公司 | Picture pushing method and device |
| CN106897918A (en) * | 2017-02-24 | 2017-06-27 | 上海易贷网金融信息服务有限公司 | A kind of hybrid machine learning credit scoring model construction method |
-
2017
- 2017-08-17 CN CN201710709808.0A patent/CN109410041A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101564328A (en) * | 2009-05-07 | 2009-10-28 | 杭州电子科技大学 | Laptop artificial limb multi-movement-mode identifying method based on support vector data description |
| CN102214320A (en) * | 2010-04-12 | 2011-10-12 | 宋威 | Neural network training method and junk mail filtering method using same |
| WO2014111540A1 (en) * | 2013-01-21 | 2014-07-24 | Ides Technologies Sa | System and method for characterizing financial messages |
| CN104657496A (en) * | 2015-03-09 | 2015-05-27 | 杭州朗和科技有限公司 | Method and equipment for calculating information hot value |
| CN106445977A (en) * | 2015-08-12 | 2017-02-22 | 阿里巴巴集团控股有限公司 | Picture pushing method and device |
| CN106022907A (en) * | 2016-05-13 | 2016-10-12 | 清华大学 | Method and system for predicting trend of background core transaction event of large commercial bank |
| CN106897918A (en) * | 2017-02-24 | 2017-06-27 | 上海易贷网金融信息服务有限公司 | A kind of hybrid machine learning credit scoring model construction method |
Non-Patent Citations (2)
| Title |
|---|
| 王红君 等: "基于能量诱导型PSO算法与LSSVM模型的钢铁企业高炉煤气消耗量预测", 《制造业自动化》 * |
| 蔡智勇: "基于典型度的相对熵相似度知识推荐", 《万方数据》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dunis | Artifi cial Intelligence in Financial Markets | |
| Gorgulho et al. | Applying a GA kernel on optimizing technical analysis rules for stock picking and portfolio composition | |
| US11645522B2 (en) | Method and system using machine learning for prediction of stocks and/or other market instruments price volatility, movements and future pricing by applying random forest based techniques | |
| US8930254B2 (en) | Financial methodology to valuate and predict the news impact of major events on financial instruments | |
| Xu et al. | A novel mobile personalized recommended method based on money flow model for stock exchange | |
| Money | Bank credits and agricultural development: Does it promote entrepreneurship performance | |
| Hamid et al. | Financial forecasting with neural networks | |
| Baranes et al. | Earning movement prediction using machine learning-support vector machines (SVM) | |
| Malla et al. | The crowding effects of basic and applied research: a theoretical and empirical analysis of an agricultural biotech industry | |
| Miciuła | The concept of FTS analysis in forecasting trends of exchange rate changes | |
| Wang et al. | Research on finance Credit Risk Quantification Model Based on Machine Learning Algorithm | |
| CN109086927A (en) | In conjunction with the multiple-factor method of commerce of big data the analysis of public opinion and Fusion Model | |
| Moedjahedy et al. | Stock price forecasting on telecommunication sector companies in Indonesia Stock Exchange using machine learning algorithms | |
| Degife et al. | Efficient predictive model for determining critical factors affecting commodity price: the case of coffee in Ethiopian Commodity Exchange (ECX) | |
| CN109410041A (en) | A kind of high-dimensional deal maker's appraisal procedure and system by data-driven | |
| Niknya et al. | Financial distress prediction of Tehran Stock Exchange companies using support vector machine | |
| Chen | Enhancement of stock market forecasting using a technical analysis-based approach | |
| Wani et al. | Unleashing Customer Insights: Segmentation Through Machine Learning | |
| Marti | Decoding the Quant Market: A Guide to Machine Learning in Trading | |
| Wang et al. | Research on Quantitative Investment Strategy Based on Random Forest Model and Risk Management | |
| Ghosh et al. | Panic Selling Analysis in Stock Market | |
| Traa et al. | Predicting takeover targets on Oslo stock exchange: an extension to the prediction literature on the Norwegian M&A market | |
| Yang et al. | Using Big Data Technology to Analyze the Fluctuation Trends in Stock Price–As a Practice of Taiwan Financial Stock Market | |
| Kiehl et al. | Evolving multi-variate time-series patterns for the discrimination of fraudulent financial filings | |
| Kushwaha et al. | Prospective Stock Analysis Model to improve the investment chances using Machine Learning |
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: 20190301 |
|
| RJ01 | Rejection of invention patent application after publication |