Personalized Recommendation Interface Design

Watercress Radio listening to more than 41,000 songs, guessed like the probability is not high, but through the radio found a lot of songs, which also embodies http://www.aliyun.com/zixun/aggregation/8086.html "> The recommendation system should balance diversity and accuracy. Unlike the way I spit before, I rely on playing songs I like, with precision but not enough diversity. Comparison of similar products shrimp, to play the most popular songs, the choice of a larger cost, the recommended long tail of the song's intensity is low.

Liking for content will change with time and mood, I tend to be popular in the early days, and when I listen to the hearts of my mobile phone, I choose to stop playing some songs, but I'm tired of the red channel, and the songs I recommend in my private MHz are similar to the one that you would like to change to a jazz or other user edited album.

The radio is constantly increasing the number of channels (MHz), "expert users" produced by the album and style, carefully to provide users with multi-dimensional recommendations, and the style of songs is based on the user's choice to generate.

Radio after the main broadcast, no songs related to the description, users need to browse the recommended interface needs to increase the reliability of the recommended reasons, such as scoring, based on a friend's recommendation or short comments, visual priority even higher than the content itself. Personally think God shakes the recommended reason copy is good enough, for this to recommend as the core products, copywriting vision is not outstanding.

Before the advent of the radio, Watercress has a music score and other user feedback, users submit more feedback, the system will be able to calculate the user's favorite. Feedback can be like, or can be the purchase of merchandise or, page stay time and video viewing time long. On the one hand the interface needs to analyze user behavior, on the other hand the interface itself to facilitate user feedback. Mobile version of the public comment on the interface of the operation cost is very high, need to score and write dozens of words of comment, very do not use the recommendation of the system, so the interface of the information is larger, not conducive to rapid selection.

The recommendation system needs consistency of content quality, otherwise the user feedback and content correspondence are inconsistent. Earlier, there was a concert version of the song, The voice is very noisy, recently did not appear again, may eliminate the low quality content or content to be recommended the probability has been greatly reduced. Content quality, rendering, interface layout, and interaction can all affect the user's choice.

Recommended system can improve the viscosity and enhance the key indicators, interaction can reduce operating costs, such as reducing user search, find the steps and interface between the number of jumps, the interface architecture can be used to avoid the homogenization of products.

Source: http://daichuanqing.com