Precise recommendation of followers helps in improving the user experienceand maintaining the prosperity of twitter and microblog platforms. In thispaper, we design a hybrid recommender system of microblog as a solution of KDDCup 2012, track 1 task, which requires predicting users a user might follow inTencent Microblog. We describe the background of the problem and present thealgorithm consisting of keyword analysis, user taxonomy, (potential)interestsextraction and item recommendation. Experimental result shows the highperformance of our algorithm. Some possible improvements are discussed, whichleads to further study.
展开▼
