1. national natural science fund committee, department of management science and department of information science interdisciplinary key projects: optimization and data mining
topic number: 70531040
research projects:

• further systemic discussion on the basis of existing results of optimization (including linear programming, multi-objective linear programming and svm) and data mining.
• establishment of the optimization method based on the theoretical framework for data mining.
• explore and develop high-level data mining technologies which has practical application value and can extract potential, innovative and useful knowledge from the massive data mining. and these include classification on basis of optimization, clustering, association, forecasts, and mode.
• research based on the new optimization data mining methods in modeling planning, algorithm implement, effectiveness in the process of data mining modeling and scalability issues in the massive data warehouse.
• unstructured, nonlinear, approximation and uncertainty issues based on the new optimization data mining methods.
• one or two true massive database with management sense of definitude or universality is combined to check up the practicality of the projects. comparative study on new method and the most popular non-optimized data mining methods.

2. national natural science fund committee, department of management science project: multi-objective nonlinear programming data mining method and application
topic number: 70472074
research projects:

• linear programming approach will be extended to nonlinear programming and develop into a series of nonlinear programming model.
• new data mining implementation model will be developed into application system based on c + + programming and linux platform.
• use the actual data warehouse we have to develop comparative study on nonlinear data mining methods, decision tree, neurons, fuzzy sets, rough sets and support vector machines, and other practical applications.

3. national natural science fund committee innovative research groups science foundation projects: data mining and intelligent knowledge management theory and application research
project approval: 70621001
research projects:
the project focuses on four directions in data mining and intelligent knowledge management theory and application:
    a. research on optimization theory and algorithms;
    b. data mining modeling, algorithm;
    c. intelligent knowledge management theory and model;
    d. demonstration research and application.
this group integrates the postulates of data technique and knowledge management to develop in-depth research, and make selective breakthrough in this cross-discipline.

4. 973 key basic research projects of state ministry of science and technology: molecular structure and function evolutionary research of antibodies
topic number: 2004cb720103
research projects:

• establishment of integrated structure and function of antibodies associated massive "virtual antibody data warehouse."
• simulation analysis of the reciprocity of antibody molecules and antigen epitope.
• data mining on the similarity of structure and activity of candidate antibodies.
• development of the original algorithm and software systems.

5. national natural science fund committee, foundation of yong researchers: research and application of commercial banks data mining system
research projects:

this project mainly concerns on the commercial banks’ management and data analysis platform which depends on the grid of data mining. this data mining system will be build on the next generation of commercial banks data mining grid, and it mainly solve the problems that how to effectively organize and manage the distributed and complex data sources. and this system will also resolve how to use the deduction and optimization method to the multi-sources, multi-directions analysis. it will extract the knowledge dynamically, this will serves the financial application well, especially supply the financial supervise department with precaution, prediction and decision support.

6. national natural science fund committee, foundation of yong researchers: the operation flow oriented knowledge management
research projects:

this project solves the selection of central flow, knowledge management audit, knowledge shifting, innovative system in the operation flow oriented knowledge management. it builds the reference models of the knowledge management, the implement scheme and the flow oriented knowledge management prototype….

7. national natural science fund committee, foundation of yong researchers: theory and method of convex programming in data mining
research projects:

this project will give a new view of the support vector machine (svm) from the convex optimization. it will build and solve the new formulation of svm use the linear programming, cone-programming and semi-definition programming method. and the new models will be applied to the regression , semi-supervised learning and kernel learning problems. at last, it will focus on the application area of the svms.

8. national natural science fund committee, foundation of yong researchers: technology and application of robust programming

(1) develop “personal credit scoring system” for the people’s bank of china. this system is the first such a system intended to provide the credit score for all 1.3 billion people in china
project introduction:
cooperation with the credit center of people's bank of china on the credit scoring system development project of 500 million chinese people’s personal data, currently the project is ongoing. the credit scoring project is to find hidden knowledge and pattern which reflects the customer risk characteristics and expected credit performance by using advanced data technology to excavate, analysis and extract information in depth from the records of the credit card and loan customer's credit history and operational activities, and by the way of scoring, the hidden knowledge can be summed up and be used as the scientific basis for management decisions. this credit scoring model is to predict the probability of the risk of defaults. one example is the forecast for probability of the bankrupt account in the next six months. from the point of view of modeling technology, we should firstly make it clear that what is "bad accounts", then the successive study includes original data analysis, integration, sample selection, feature variables extraction and derivation, modeling, model validation and explanation.

(2) developed “the framework for banking-to-banking commercial data exchanges” for the china’s foreign currency exchange center (this project was completed in december, 2006)
project introduction:
the inter-bank market data is an important part of china's money market operation data. it is also the important channel for the regulators, investors and other market participants to exchange the information and get the news on money market operations. moreover, it is the primary basis for the decision making on investment and monitoring implementation.the establishment of the unified inter-bank market business data exchange protocol aims at making the technical standards that can meet the demand of domestic money market mutual exchange of business data, so that the data exchange between china foreign exchange trading center and the market participants system can become convenient and consistent. this will increase market efficiency of the transaction and can effectively prevent and control the internal risk.

(3) developed “vip customers migration patterns and prediction” for netease co., a china’s major internet corporation (nasdaq ipo) (this project was completed in may 2006)
project introduction:
this project's research goals: 1, through data mining modeling analysis, provide a range of business analysis model, find the knowledge and pattern hidden in large amount of data and the provide operational decision support; 2, through the establishment of the specific model, form a set of data analysis methods, tools, procedures, which are available for later uses on other business analysis, and based on these model, develope a series of functional software, such as personalized search, advertising navigation etc. 3, the business analysis programs formed by this project can apply for national projects, the joint declaration of intellectual property rights. all these can help netease gain the core technical capabilities. 4, with the implementation of the project, cultivating the awareness of the value of their customers, the methodology of analyze problems with data and facts, and enhancing their execution.

(4) investigate “data mining approaches to petroleum and geological explorations” for bhp billiton, co, australia
project introduction:
due to the increased cost of oil and gas exploration and the difficulty to identify the location of oil and gas reservoirs, it is necessary to use new data mining technology such as mathematical methods to analyze the large amounts of data gained during the exploration process. in order to more accurately determine oil and gas structures and improve the success rate of the drill, we need to develop new mathematical method to analyze the exploration data.

the technologies used in this project include optimization data mining methods, statistics, decision trees, neural networks, fuzzy logic and computer programming technology. the project also tries to develop the practical data mining technology and software that can be applied to the oil and gas exploration. the new data mining software will be able to easily and commonly transmit data through some specific interface with the prevalent software in petroleum industry (such as geoqust / geoframe, landmark and geolog, etc.).

(5) develop “customers classification system” for the industrial and commercial bank of china (icbc).
project introduction:
this research aims at using data mining technology to find some hidden knowledge and pattern which can reflect customer risk characteristics and the performance of customer loyalty. the data mining activities includes the process and integration of original data of the credit history records and operational activity records related to the individual customer of icbc, the validation and amendment of the current classification rule on the individual customer data of icbc, and the excavation, analysis and extraction in depth of the data on the characteristics of individual customer activities and the distribution of customer contribution. and the research will provide the scientific basis to icbc about personal customer classification, risk management and marketing.

develop “auditing data warehouse and data mining” for the china’ national audit bureau. this system can be used to audit the financial activities of large –scale enterprises, especially the state–owned companies (under review)