中文摘要：

基于重启型随机游走模型和个人化PageRank箅法，提出一种新的图上关键字搜索算法。该算法将向量空间模型和随机游走模型进行有效的结合，使查询搜索得到的结果可以匹配查询关键字，通过充分挖掘利用图中隐含的结构信息，更好地提供搜索结果。实验结果证明了该算法的有效性。

英文摘要：

This paper presents a new keyword search on graphs algorithm based on random walk with restart model and personalized PageRank algorithm. By combining vector space model and random walk model effectively, it can make the results match the query keywords as well as the structural information implied in the graph, which provides better searching results. Experimental result proves that the algorithm is effective.