Application of data mining in the field of financial securities trading

On the evening of July 28, the "Data mining in the field of financial securities trading application Experience sharing", organized by the CTO Club of CSDN's senior technical Management headquarters, was held successfully at the Haidian Bridge Garage Café, the first line of the CTO Club's financial industry software Professional Committee since its inception.

The event invited to Phoenix Senior technical manager Wang, and Rui Network CTO Gang Jianhua, Sohu Financial Division senior Engineer Zhaoschang, finance technical director Wu Yu and other guests, sharing in the securities trading market data analysis field involved in the experience and technology, And the application of data mining in the Internet-related fields is deeply communicated and discussed. The activity scene is hot, has nearly hundred industry technical superintendent to the scene.

The activity scene is hot

and Rui Network CTO Gang Jianhua introduced the data characteristics of the domestic securities trading industry: All dynamic time series data, new data per second thousands; 21 years historical data; a number of trading markets, a variety of securities categories; The original data amount is 2G per day in the Shanghai and Shenzhen cities; the financial statements and other data are unstructured data The historical data will never change. On data cleaning and denoising, he talked about several basic mechanisms-coding standardization, Unit standardization, warehousing inspection rules making, missing data processing mechanism. In the construction of the Data Warehouse--Custom file database, Gang Jianhua analyzes why the NoSQL is currently used: because the efficiency of the disk is still low, so the use of MongoDB, memory to store the data, 4 servers with 128G of memory, to achieve 27,000 per second write speed.

and Rui Network CTO Gang Jianhua

In research and development process planning, he focuses on: Product planning (product manager)-Model design (financial engineer)-Model development (development team)-Regression testing (test Department)-Solid disk verification (product manager)-Model monitoring (product manager) mode.

Phoenix Senior Technical Manager Wang

Phoenix Senior Technical manager Wang shared his future prospects for data mining in the field of financial securities trading-can you give users a hint based on the metric value? Can you prompt the user to appear on the line chart, can be from the mass data automatic analysis to get operation tips; trend forecast; Arbitrage Opportunity (Cross a| b| h| warrants | fund | commodities; analyze financial data according to specific models; Can the theory of Zhouyi and astrology be used in stock analysis.

Interactive questions and answers (from left to right: Wang, Zhaoschang, Wu Yu, Gang Jianhua)

On-site audience questions, mainly focused on several points: Financial industry data analysis how to integrate with their own business, BS/CS (client) architecture differences, the advantages and disadvantages of data security, how to obtain better quality securities trading data, data transmission, cleaning methods and so on. When it comes to data analysis solutions bs/cs architectural differences, financial technology director Wu Yu that BS is slower than CS and may be less timely when data volume is large, and may be two seconds slow. Ordinary customers feel unimportant, but half a second is important for high-end users. The second is the securities trading data acquisition and use of compliance, the Shanghai Shenzhen regulations have different differences, many Web sites because they do not get a license to do, some can obtain data analysis but can not trade. Again, data security is another area that cannot be overlooked.

In this regard, Sohu Financial Department senior engineer Zhaoschang not agree, he thought that BS architecture can reach the CS data level, Sohu Technical department has done testing, speed difference will not be too much, a level. CS structure, because there are clients, can store a large number of local data to do analysis, BS architecture may be limited by the Internet, but can be data to the client processing, the browser is only a display of processing results can be, technically have been very mature.

Overall, the future of the financial securities Industry data Services, the next technical focus will be on how to provide personalized services to individuals, on-site interaction is active, the participants put forward a lot of good needs and suggestions.

For more highlights and next-line events at the CSDN CTO Club, please keep an eye on Csdn CTO Club.

CTO Club is the CSDN for the vast number of technical managers of the online stronghold, adopted a strict invitation to audit system, only the company's senior technical managers, such as CTO, technical director, development Director, Engineering Director can apply to join, now more than 10,000 members. The CTO club regularly conducts offline activities to discuss related topics and facilitate the exchange of members ' interactions.

(Responsible editor: admin)