Video network new recommended engine online large data to help users recommend

April 7 News, recently, the new edition of the video network to launch the eight intimate experience, which the new recommended engine on-line, it will be through the shadow of the long, similar audience types, social platform data analysis and other "big data" for each user "tailor-made" favorite program.

It is understood that the user in the first landing page of the video site for the play page, the content of the recommended page on the richness and accuracy, determines whether the user continued to stay to watch the site of other films. At present, most of the video site content recommendation is single, and did not put the user really favorite programs recommended to them, resulting in the loss of users. Le Video network recommendation engine on the line, to achieve in the play page (landing page) to users including the same category, cross categories, hot Classic and other dimensions recommended. For example, the recent hot film "Ted", the following three different dimension recommendations appear below the playback interface:

"You might like to recommend the same kind of video that is similar to the movie you're watching on the user's page, such as the comedy" People are in the way "and" The Hundred Star Hotel ".

"Everyone is watching." It is recommended that other types of films be popular nowadays, such as "One Parting" and "1942", which are widely discussed recently.

"Classic Theater" is recommended in the past classic films, such as the recommended "Years of God Steal", "Sissi princess" and other enduring old movies.

Video network new recommendation engine using "collaborative filtering" as the main algorithm, using the same interest and similar experience of the http://www.aliyun.com/zixun/aggregation/8172.html "> User behavior Model, to recommend the content of interest to users, Recommended more accurate. For example, a woman watched the "Twilight City", b women also watched the "Twilight City", the system speculated that a woman and B is a shadow of the same, the system will be a woman like the "Harry Potter" also recommended to B women.

The recommendation engine is also effective in the "large data" analysis, the effective information provided to the recommendation engine as a basis for decision-making. For example, the length of the user is analyzed to determine how much it likes the movie. The shadow 5 minutes and the shadow 1 hours of users obviously have different preferences, the latter more sincere, through this effective information, the system will determine the real love of the film users who is, improve the recommended accuracy. In addition, the unlimited flow of information in the way, the number of films more, increased user recommended hit rate, users do not even have to go to the homepage, landing page (play page) to complete the selection work.

It is reported that the recent new edition of the network has introduced a number of user-friendly functions and experience, the new recommended engine for users to recommend accurate, sticky users, to achieve the maximum content of the video network. Analysis of the industry, the user experience of the continuous upgrade, will also promote the music network advertising marketing value of further promotion.