Shanda literature wins International data mining competition

Data mining technology who is the strongest? Cmki CUP (International Data Mining Competition), held in mainland China for the first time, has been closed recently, with outstanding international rivals such as ebay and the grand prize winning the title of the competition with excellent results. This not only shows the royal literature in the user demand mining strength, but also represents the "Big data" era, China's data mining and information retrieval in the international leading position.

CIKM's full name is Analysys conference on information and Knowledge Management, an internationally renowned academic conference in the field of information retrieval and knowledge management, called the "Olympic ", the first time this year in mainland China, the meeting was invited to Google's great God-class character Jeff Dean, Microsoft's executive vice president Dr. Lurkey and Germany Max Planck Cato for Informatics Gerhard Weikum Professor as the keynote speaker, Unprecedented。

The "Topdata" team, led by Grand Literature Chief data Officer Dr. Chen Yunwen, occupies the home and the world's data-mining scientists at a higher level-the goal of the competition is to use machine learning techniques to identify user query intent (query Intent detection,qid). The organizers combined Baidu Company, provided the Baidu line of more than 60 million users of the real query log and click Behavior Data, requires the team according to the user's behavior data, design algorithms to automatically interpret the user's demand intention. For example, when users search for "Beijing and Shanghai", while still looking for "Shanghai Weather", "high-speed train tickets" and other information, the computer should be able to intelligently judge the user's demand is "from Beijing to Shanghai Travel" and return the corresponding various kinds of high-value information. The system developed by the participating teams should be fully and accurately identified with the user's intention. The competition has attracted 520 teams from all over the world, academia and business circles, the Grand literary team designed the algorithm, through the careful analysis of user behavior characteristics, and a number of machine learning algorithm training and fusion technology, in the game constantly surpass, and finally with excellent results.

It is reported that the grand Literature data team who participated in the contest undertook the research and development of the whole company's large data-related system, and was responsible for mining high value content from the massive literature works and user data, not only to provide high-quality service for the readers, but also for the derivative of the copyright of literary works Game adaptation and other services provide a lot of data guidance and help. "Shanda literature is the world's largest community-driven network literature platform, the content created by our pioneering network literature model is inherently a complete mass of data covering content, community, traffic, user attributes, and user behavior, Chen Yunwen says:" Our team is in this way. " Data Ocean ' daily exercise, take world champion, who?! The