CN110223105B - Transaction strategy generation method and engine based on artificial intelligence model - Google Patents
Transaction strategy generation method and engine based on artificial intelligence model Download PDFInfo
- Publication number
- CN110223105B CN110223105B CN201910414985.5A CN201910414985A CN110223105B CN 110223105 B CN110223105 B CN 110223105B CN 201910414985 A CN201910414985 A CN 201910414985A CN 110223105 B CN110223105 B CN 110223105B
- Authority
- CN
- China
- Prior art keywords
- strategy
- trading
- transaction
- precision
- test
- 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.)
- Expired - Fee Related
Links
Images
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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
Abstract
The invention discloses a trading strategy generation method and engine based on an artificial intelligence model, which comprises a trading index layer, an AI model layer, a trading logic layer, a target combination layer and a macro filter layer from bottom to top, wherein a structured trading signal is mined by using an AI and machine learning method, the trading signal is expanded into a complete trading strategy, the trading strategy is delivered to a trading system to run in a fully automatic way, an AI algorithm is used for replacing a fund manager to carry out asset allocation and stock trading, market trading opportunities are automatically mined, and a trading strategy of a dynamic self-adaptive market is generated.
Description
Technical Field
The invention relates to a transaction strategy generation method and engine based on an artificial intelligence model.
Background
In recent years, the rise of financial science and technology enables enterprises to use financial services more efficiently, and meanwhile, the threshold of using the financial services by consumers is lowered, so that the establishment of a fairer social environment is facilitated. The asset management industry is an important component of a mature financial market, and the investable assets of operating customers are managed in a centralized manner by professional investment institutions, so that the value preservation and the value increment of the assets can be realized, the resource allocation is optimized, the capital market efficiency is improved, and the financial market structure is stabilized.
Under the current large background of strong supervision of the resource management industry, remodeling of an industry chain and external opening of a financial market, the domestic resource management industry faces unprecedented challenges and pressure, but a strategy development method of a professional resource management institution has a plurality of problems:
(1) the product return is low, and the goal of value-added value-keeping for the customer is difficult to realize. According to the Wind data, the average earning rate of the fund management industry in 2018 is less than 5%, wherein the highest value of the average expected earning rate of the bank financing earning rate is 4.91%, the average expected earning rate of the trust is 8.24%, the average annual earning rate of the insurance fund is 4.3%, the average annual earning rate of the fund policeman is-1.98%, the average annual earning rate of the public fund is-6%, and the average annual earning rate of the private fund is-12.71%. The reduction of the profit scale of the industry also leads to the slowing of the capital inflow speed and the insufficient afterward force of the large-scale increase of the capital;
(2) the mechanism has poor performance stability, and is difficult to obtain long-term stable and high-consistency return. The head capital management organization shows strong performance in the cattle market and shows weak and even negative return in other periods, so that the net capital value curve is unsmooth, the fluctuation rate is high, the correlation degree with the market is too high in long periods of three years, five years and ten years, and the long-term value-keeping increment of large assets is not facilitated;
(3) investment strategies are small in capacity and difficult to achieve large-scale growth of assets and profits. The investment strategy has the advantages that the amount of capital which can be accommodated by the strategy is limited by few asset combination targets, weak Alpha factors, too short investment period, high investment cost, large market impact and the like, so that the capital management mechanism is difficult to obtain scale advantage, and a large amount of resources are forced to be invested to research and develop corresponding amount of strategies along with the increase of the capital amount;
(4) high investment and operation cost, and difficult accumulation of industrial profits and improvement of industrial efficiency. Investment institutions excessively depend on investment managers, researchers and traders to develop and execute strategies, labor, development, operation and maintenance costs are high, most profits are consumed, repeated research and development and product development among different institutions cause a great amount of waste of industrial manpower, and the efficiency of resource allocation is reduced.
Disclosure of Invention
The invention provides a trading strategy generation method and engine based on an artificial intelligence model, wherein an AI algorithm is used for replacing a fund manager to carry out asset allocation and security trading, market trading opportunities are automatically mined, and a trading strategy of a dynamic self-adaptive market is generated.
The technical scheme of the invention is realized as follows:
a transaction strategy generation method based on an artificial intelligence model specifically comprises the following steps:
s1, acquiring all dimension indexes related to the trading strategy on the market;
s2, establishing a plurality of artificial intelligence models;
s3, the artificial intelligence model excavates a structured trading signal from all the dimension indexes, and the trading signal is added with a departure signal, bin management and risk management to be expanded into an initial trading strategy;
s4, judging the market form according to the strategy win rate, and switching reasonable and correct strategy logic according to the market form;
s5, screening out strategies according with the strategy logic from the initial trading strategies, calculating the strategy win rates one by one, and sequencing and selecting the optimal rate;
and S6, generating a trading strategy of a single target or a combined target, wherein the trading strategy of the single target is a trading strategy with a strategy win rate ranking and a priority ranking, and the trading strategy of the combined target is a trading strategy with a strategy optimization and a priority ranking on all standby targets.
As a preferred embodiment of the present invention, the structured transaction signal in step S3 specifically refers to a non-linear combination signal of all dimension indexes.
As a preferred embodiment of the present invention, the present invention further comprises the following steps;
s7, carrying out retest and optimization on the generated strategy;
the steps of back test and optimization are specifically
Randomly generating a test set according to all the dimension indexes, and testing whether the precision of the transaction strategy generated in the step S6 on the test set reaches a threshold value;
or is that
Performing a Walk-forward/backforward multi-round test, requiring that the transaction strategy generated in step S6 obtain MAR >1 and sharp >1 in at least 90% of the test rounds;
or is that
Different transaction paths were simulated using the monte carlo test to determine that MAR >1 and sharp >1 were obtained with 90% test runs.
A transaction strategy generation engine based on artificial intelligence model specifically comprises
(1) The trading index layer is responsible for the input of the engine and acquiring all dimension indexes related to the trading strategy on the market;
(2) the AI model layer is responsible for the logic drive of the engine and comprises a plurality of artificial intelligence models;
(3) the transaction logic layer is responsible for strategy logic and an optimization target of the engine on a single target, and the model optimization target is strategy win rate;
(4) the object combination layer is responsible for selecting trading objects in the asset combination when the asset combination strategy is generated, and simultaneously carrying out strategy optimization on all standby objects so as to sort and prefer;
(5) and the macro filter layer is responsible for switching of transactions.
As a preferred embodiment of the invention, the artificial intelligence model excavates a structured trading signal from all dimension indexes, and the trading signal is added into a field signal, a position management and a risk management to be expanded into a trading strategy.
As a preferred embodiment of the invention, the transaction logic layer is responsible for the strategy logic and the optimization target of the engine on a single target, and the model optimization target is the strategy win rate; specifically, the market form is judged according to the strategy win rate, and reasonable and correct strategy logic is switched according to the market form.
As a preferred embodiment of the present invention, the transaction logic layer is further configured to perform a backtest and optimization on the policy; the steps of back test and optimization are specifically
Randomly generating a test set according to all dimension indexes, and testing whether the precision of the strategy on the test set reaches a threshold value;
alternatively, a Walk-forward/Backward multi-run test is performed, requiring that the above strategy can obtain MAR >1, sharp >1 in at least 90% of the test runs;
alternatively, a Monte Carlo test was used to simulate different transaction paths to determine that the above strategy achieved MAR >1 and Sharpe >1 at 90% test runs.
As a preferred embodiment of the present invention, the engine's policy logic on a single target includes, but is not limited to, trend tracking, mean regression, event driven, statistical arbitrage.
The invention has the beneficial effects that: and compared with the traditional strategy development mode, the quantitative trading strategy with stable return and low risk can be developed more efficiently and at low cost.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of an embodiment of a transaction policy generation method based on an artificial intelligence model according to the present invention;
FIG. 2 is a schematic block diagram of an embodiment of a transaction policy generation method based on an artificial intelligence model according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, the present invention provides a transaction policy generation method based on an artificial intelligence model, which specifically includes the following steps:
s1, acquiring all dimension indexes related to the trading strategy on the market;
s2, establishing a plurality of artificial intelligence models;
s3, the artificial intelligence model excavates a structured trading signal from all the dimension indexes, and the trading signal is added with a departure signal, bin management and risk management to be expanded into an initial trading strategy; the structured transaction signal specifically refers to a non-linear combination signal of all dimension indexes, such as a tree structure, a graph structure, and the like.
Adding the trading signal into the departure signal such as the longest holding time of the stop-loss and the like, bin management such as fixed proportion pyramid bin adding and the like, risk management such as the maximum withdrawal punishment trading fund curve and the like to expand the trading signal into a complete trading strategy. After the complete transaction strategy is generated and determined, the system can automatically operate.
S4, judging the market form according to the strategy win rate, and switching reasonable and correct strategy logic according to the market form;
s5, screening out strategies according with the strategy logic from the initial trading strategies, calculating the strategy win rates one by one, and sequencing and selecting the optimal rate;
and S6, generating a trading strategy of a single target or a combined target, wherein the trading strategy of the single target is a trading strategy with a strategy win rate ranking and a priority ranking, and the trading strategy of the combined target is a trading strategy with a strategy optimization and a priority ranking on all standby targets.
The invention provides a mode of data layers from bottom to top, which can fully automate the generation process of the whole trading strategy, and when the covered markets are many, tens of thousands of signals and strategies can be generated in each round, and then the strategies are screened by adopting strict trading indexes and statistical indexes, and the best hundreds of strategies are selected for trading.
For example, the following steps are carried out:
training method for a quantitative strategy of trade constant index (HSI)
Training targets: high-precision binary trading classifier
Training step:
1) randomly dividing 20-year HSI data into a training set (1-1/e) and a testing set (1/e);
2) performing the back sampling on the training set data to form N sub-samples (each sample is at least 1/N of the data in size, and the sum of the sub-samples is to cover all the original training set data);
3) the following is performed for each subsample:
-performing Principal Component Analysis (PCA) to further reduce the metrics to M;
randomly selecting M/2 indexes from the M indexes to calculate node splitting according to the entropy value until a decision tree is generated;
-calculating the accuracy of the decision tree, stopping if higher than 80%, and if not, adding or subtracting an index, repeating step 3);
in addition, if the precision of the obtained decision tree is improved by less than 10% after the indexes are added than before the indexes are added, the indexes are removed from the index set of the next round of training;
4) determining a result according to the obtained P decision trees by a majority voting principle, namely a random forest, calculating the precision of the random forest on the whole training set, stopping if the precision is higher than 97%, and repeating the steps 2) -3) if the precision is not higher than 97%;
5) calculating the precision of the random forest on the test set, stopping if the precision is higher than 90%, and repeating the steps 2) -4) if the precision is not higher than 90%;
6) and repeating the steps 1) -5) for K times, stopping if the precision of the obtained random forest is higher than 90% in K times of tests, and otherwise declaring that the model training fails. N, M, K, P are all positive integers.
For further backtesting and optimization of the generated strategy. The objective of backtesting is to obtain a set of parameters that make the strategy robust and consistent. Also comprises the following steps;
s7, carrying out retest and optimization on the generated strategy;
the steps of back test and optimization are specifically
Randomly generating a test set according to all the dimension indexes, and testing whether the precision of the transaction strategy generated in the step S6 on the test set reaches a threshold value; testing a large number of parameter combinations, making results into an optimization space, and selecting the parameter combinations on a 'parameter plain';
or is that
Performing a Walk-forward/backforward multi-round test, requiring that the transaction strategy generated in step S6 obtain MAR >1 and sharp >1 in at least 90% of the test rounds;
or is that
Different transaction paths were simulated using the monte carlo test to determine that MAR >1 and sharp >1 were obtained with 90% test runs.
As shown in FIG. 2, the present invention further provides a transaction policy generation engine based on artificial intelligence model, which specifically includes
(1) The trading index layer is responsible for the input of the engine and acquiring all dimension indexes related to the trading strategy on the market; the dimension index covers more than 200 market indexes of basic surface, quantification, public opinion and finance;
(2) the AI model layer is responsible for the logic drive of the engine and comprises a plurality of artificial intelligence models;
(3) the transaction logic layer is responsible for strategy logic and optimization targets of the engine on a single target, the current main strategy logic framework comprises trend tracking, mean value regression, event driving, statistical arbitrage and the like, and the model optimization target is strategy win rate;
(4) the object combination layer is responsible for selecting trading objects in the asset combination when the asset combination strategy is generated, and simultaneously carrying out strategy optimization on all standby objects so as to sort and prefer;
(5) and the macro filter layer is responsible for switching of transactions.
As a preferred embodiment of the invention, the artificial intelligence model excavates a structured trading signal from all dimension indexes, and the trading signal is added into a field signal, a position management and a risk management to be expanded into a trading strategy. The structured transaction signal specifically refers to a non-linear combination signal of all dimension indexes, such as a tree structure, a graph structure, and the like. Adding the trading signal into the departure signal such as the longest holding time of the stop-loss and the like, bin management such as fixed proportion pyramid bin adding and the like, risk management such as the maximum withdrawal punishment trading fund curve and the like to expand the trading signal into a complete trading strategy. After the complete transaction strategy is generated and determined, the system can automatically operate.
As a preferred embodiment of the invention, the transaction logic layer is responsible for the strategy logic and the optimization target of the engine on a single target, and the model optimization target is the strategy win rate; specifically, the market form is judged according to the strategy win rate, and reasonable and correct strategy logic is switched according to the market form. The strategy logic of the engine on a single target includes but is not limited to trend tracking, mean regression, event driving and statistical arbitrage.
The judgment of the strategy on the market form and the switching of the trading logic are determined by the optimization target, namely the win rate, of the artificial intelligence model, and the whole process is automatically completed by the engine. For example, two market patterns occurring continuously are divided into left and right sides, the left side has a steady growth and obvious trend, and the right side has a fluctuation and no obvious trend. In the left stage, the AI engine tests one by one from the alternative strategy framework, when a trend tracking framework is tested and a transaction signal combined by transaction indexes is mined by using an applicable artificial intelligence model, if the convergence rate is found to be extremely high and a transaction signal with a high win rate (for example, higher than 70 percent) can be obtained under few iteration times, the engine determines that the current transaction is a stable growth and an obvious trend, and the selected logic framework is reasonable and correct at the moment; when the market enters the right stage, the AI engine repeats the process, when a trend tracking frame is selected, the engine abandons the trend tracking frame and selects a mean regression frame from alternative frames instead of finding that the convergence rate is extremely low and a high-win-rate signal cannot be obtained by a small number of iteration times, and a machine learning model suitable for the engine is used for mining a transaction signal.
As a preferred embodiment of the present invention, the transaction logic layer is further configured to perform a backtest and optimization on the policy; the steps of back test and optimization are specifically
Randomly generating a test set according to all dimension indexes, and testing whether the precision of the strategy on the test set reaches a threshold value;
alternatively, a Walk-forward/Backward multi-run test is performed, requiring that the above strategy can obtain MAR >1, sharp >1 in at least 90% of the test runs;
alternatively, a Monte Carlo test was used to simulate different transaction paths to determine that the above strategy achieved MAR >1 and Sharpe >1 at 90% test runs.
The invention uses AI algorithm to replace fund manager to carry out asset allocation and security trading, fully automatically excavates market trading opportunities, and generates a trading strategy of a dynamic adaptive market.
Meanwhile, the engine can effectively help the asset management mechanisms such as banks, security traders, insurance, public and private fund raising and returning performance by combining with the practical requirements of the asset industry, so that the investment stability is increased, the asset risk is controlled, and the high-speed increase of the management scale and the volume is realized. On the basis of helping the management mechanism to realize stable and high return, the financial institution can actively go out to serve the management mechanisms at home and abroad, attract regional and even international fund inflow, and greatly improve the overall strength and the international competitiveness of the Chinese financial industry.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (4)
1. A transaction strategy generation method based on an artificial intelligence model is characterized by comprising the following steps:
s1, acquiring all dimension indexes related to the trading strategy on the market;
s2, establishing a plurality of artificial intelligence models;
training:
1) randomly dividing data into a training set and a testing set;
2) performing put-back sampling on the training set data to form N sub-samples, wherein the size of each sample is at least 1/N data, and the sum of the sub-samples is to cover all original training set data;
3) the following is performed for each subsample:
performing principal component analysis, and further reducing the dimension of the indexes into M;
randomly selecting M/2 indexes from the M indexes, calculating, and performing node splitting according to the entropy value until a decision tree is generated;
calculating the precision of the decision tree, stopping if the precision is higher than 80%, adding or reducing an index if the precision is not higher than 80%, and repeating the step 3); if the precision of the obtained decision tree is improved by less than 10 percent after the indexes are added compared with that before the indexes are added, the indexes are removed from the index set of the next round of training;
4) determining a result according to the obtained P decision trees by a majority voting principle, namely a random forest, calculating the precision of the random forest on the whole training set, stopping if the precision is higher than 97%, and repeating the steps 2) -3) if the precision is not higher than 97%;
5) calculating the precision of the random forest on the test set, stopping if the precision is higher than 90%, and repeating the steps 2) -4) if the precision is not higher than 90%;
6) repeating the steps 1) -5) for K times, stopping if the precision of the obtained random forest is higher than 90% in K times of tests, otherwise declaring that the model training fails, wherein N, M, K, P are positive integers;
s3, the artificial intelligence model excavates a structured trading signal from all the dimension indexes, and the trading signal is added with a departure signal, bin management and risk management to be expanded into an initial trading strategy;
s4, judging the market form according to the strategy win rate, and switching reasonable and correct strategy logic according to the market form;
s5, screening out strategies according with the strategy logic from the initial trading strategies, calculating the strategy win rates one by one, and sequencing and selecting the optimal rate;
s6, generating a trading strategy of a single target or a combined target, wherein the trading strategy of the single target is a trading strategy with a strategy win rate sequencing and a preferred sequence, and the trading strategy of the combined target is a trading strategy with a strategy optimization and a sequencing and a preferred sequence on all standby targets;
s7, carrying out retest and optimization on the generated strategy;
the steps of back test and optimization are specifically
Randomly generating a test set according to all the dimension indexes, and testing whether the precision of the transaction strategy generated in the step S6 on the test set reaches a threshold value;
or is that
Performing a Walk-forward/backforward multi-round test, requiring that the transaction strategy generated in step S6 obtain MAR >1 and sharp >1 in at least 90% of the test rounds;
or is that
Different transaction paths were simulated using the monte carlo test to determine that MAR >1 and sharp >1 were obtained with 90% test runs.
2. The method for generating transaction strategy based on artificial intelligence model as claimed in claim 1, wherein the structured transaction signal in step S3 is specifically a non-linear combined signal of all dimension indexes.
3. A transaction strategy generation engine based on an artificial intelligence model is characterized by specifically comprising
(1) The trading index layer is responsible for the input of the engine and acquiring all dimension indexes related to the trading strategy on the market;
(2) the AI model layer is responsible for the logic drive of the engine and comprises a plurality of artificial intelligence models; the artificial intelligence model excavates structured trading signals from all dimension indexes, the trading signals are added with a field signal, bin management and risk management to be expanded into a trading strategy, and the structured trading signals specifically refer to nonlinear combined signals of all dimension indexes;
(3) the transaction logic layer is responsible for strategy logic and an optimization target of the engine on a single target, and the model optimization target is strategy win rate;
(4) the object combination layer is responsible for selecting trading objects in the asset combination when the asset combination strategy is generated, and simultaneously carrying out strategy optimization on all standby objects so as to sort and prefer;
(5) a macro filter layer, a switch responsible for transactions;
the transaction logic layer is responsible for strategy logic and an optimization target of the engine on a single target, and the model optimization target is strategy win rate; judging the market form according to the strategy win rate, and switching a reasonable and correct strategy logic according to the market form;
the transaction logic layer is also used for carrying out retest and optimization on the strategy; the steps of back test and optimization are specifically
Randomly generating a test set according to all dimension indexes, and testing whether the precision of the strategy on the test set reaches a threshold value;
alternatively, a Walk-forward/Backward multi-run test is performed, requiring that the above strategy can obtain MAR >1, sharp >1 in at least 90% of the test runs;
alternatively, a Monte Carlo test was used to simulate different transaction paths to determine that the above strategy achieved MAR >1 and Sharpe >1 at 90% test runs.
4. The artificial intelligence model based transaction policy generation engine of claim 3, wherein the policy logic of the engine on a single target includes but is not limited to trend tracking, mean regression, event driven, statistical arbitrage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910414985.5A CN110223105B (en) | 2019-05-17 | 2019-05-17 | Transaction strategy generation method and engine based on artificial intelligence model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910414985.5A CN110223105B (en) | 2019-05-17 | 2019-05-17 | Transaction strategy generation method and engine based on artificial intelligence model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110223105A CN110223105A (en) | 2019-09-10 |
| CN110223105B true CN110223105B (en) | 2020-12-01 |
Family
ID=67821416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910414985.5A Expired - Fee Related CN110223105B (en) | 2019-05-17 | 2019-05-17 | Transaction strategy generation method and engine based on artificial intelligence model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110223105B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111949271B (en) * | 2020-07-27 | 2024-03-12 | 东莞市龙兴基石智能科技有限公司 | Method, transaction system, device and storage medium for customizing transaction policy |
| WO2022064894A1 (en) * | 2020-09-28 | 2022-03-31 | ソニーグループ株式会社 | Information processing device, information processing method, and program |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105894379A (en) * | 2016-03-30 | 2016-08-24 | 上海坤士合生信息科技有限公司 | System and method for generating financial product transaction strategy |
| CN106650992A (en) * | 2016-10-10 | 2017-05-10 | 北京极派客科技有限公司 | Quantitative investment strategy generating method and apparatus |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104572786A (en) * | 2013-10-29 | 2015-04-29 | 华为技术有限公司 | Visualized optimization processing method and device for random forest classification model |
| CN104123592B (en) * | 2014-07-15 | 2017-10-24 | 清华大学 | Bank's backstage TPS transaction events trend forecasting method and system |
| CN105303262A (en) * | 2015-11-12 | 2016-02-03 | 河海大学 | Short period load prediction method based on kernel principle component analysis and random forest |
| CN107798609A (en) * | 2017-11-08 | 2018-03-13 | 上海宽全智能科技有限公司 | Quantify trading strategies generation method and device, equipment and storage medium |
| CN109615531A (en) * | 2018-12-18 | 2019-04-12 | 厦门依实信息科技有限公司 | Securities market quantifies precisely returning for investment tactics and surveys and assessment system and method |
| CN109711995A (en) * | 2019-01-03 | 2019-05-03 | 周鸣籁 | A method of automatically generating the trading strategies in time series |
-
2019
- 2019-05-17 CN CN201910414985.5A patent/CN110223105B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105894379A (en) * | 2016-03-30 | 2016-08-24 | 上海坤士合生信息科技有限公司 | System and method for generating financial product transaction strategy |
| CN106650992A (en) * | 2016-10-10 | 2017-05-10 | 北京极派客科技有限公司 | Quantitative investment strategy generating method and apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110223105A (en) | 2019-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111311401A (en) | Financial default probability prediction model based on LightGBM | |
| CN110738564A (en) | Post-loan risk assessment method and device and storage medium | |
| Adofu et al. | Manufacturing sector and economic growth in Nigeria (1990-2013) | |
| CN105354210A (en) | Mobile game payment account behavior data processing method and apparatus | |
| CN113537807B (en) | Intelligent wind control method and equipment for enterprises | |
| CN110378382A (en) | Novel quantization transaction system and its implementation based on deeply study | |
| Mason et al. | What is holding back UK productivity? Lessons from decades of measurement | |
| CN110223105B (en) | Transaction strategy generation method and engine based on artificial intelligence model | |
| CN107908536A (en) | To the performance estimating method and system of GPU applications in CPU GPU isomerous environments | |
| Gripaios et al. | Analysing the impact of Objective 1 funding in Europe: a review | |
| Pompe et al. | Bankruptcy prediction: The influence of the year prior to failure selected for model building and the effects in a period of economic decline | |
| Ratinger et al. | Are there any economic impacts of business R&D support? The case of the Czech Republic | |
| CN108122166A (en) | The method sorted out to stock and scored using intelligent quantization financial technology | |
| Mileris et al. | Data reduction influence on the accuracy of credit risk estimation models | |
| Doumpos et al. | Efficiency and performance evaluation of European cooperative banks | |
| Alfaro-Cid et al. | Minimising value-at-risk in a portfolio optimisation problem using a multi-objective genetic algorithm | |
| Urrutia et al. | Modelling and forecasting the exchange rate of the Philippines: A time series analysis | |
| CN116611911A (en) | Credit risk prediction method and device based on support vector machine | |
| CN115423499A (en) | Model training method, price prediction method, terminal device, and storage medium | |
| Chen et al. | A stable credit rating model based on learning vector quantization | |
| Soni et al. | A novel optimized classifier for the loan repayment capability prediction system | |
| Ünal et al. | A machine learning approach towards startup success prediction | |
| Kamal et al. | Modeling success factors for start-ups in western europe through a statistical learning approach | |
| Garcia et al. | FLBEIA: A toolbox to conduct Bio-Economic Impact Assessment of fisheries management strategies | |
| Ma et al. | Credit Risk Assessment of Chinese Listed Companies Based on SVM Improved by Shuffled Frog Leaping Algorithm |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201201 Termination date: 20210517 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |