http://blog.csdn.net/lili72
Background: Need to be more realistic understanding of the user's use of the product, take daily water data, statistical analysis of a time period of user behavior characteristics:
User region attributes: where The most recent occurrences of the day have occurred, the user is presumed to be in the place where the user appears most and where the user recently appeared.
User preferences: The most recent user's menu, based on the user's preference for the product, more in-depth knowledge of the user's age level, directed to the user push similar products.
User session Properties: The most recent user time period statistics (such as: morning, morning, afternoon, evening, etc.), according to speculate users use the product period, each time period of stay time, presumably speculate the user's working hours.
User Login times: The most recent user logon times, speculated about the user's frequency of use, whether loyal users, the possibility of user churn. ( more than 7 consecutive days not logged in, may have been lost)
Users use the network: users use the network type for the most recent days, thus speculating on the network preferences used by users.
User clicks on the number of ads: The most recent users click on the type of ad and the number of ads, speculate about the user's personal preferences, age level.
User Search content: recent user Search content situation, speculation user preferences, personality characteristics.
User Click Track: The latest user Click on the track situation, from landing to log out of this period of time a continuous trajectory movement
User Label design