Abstract :
In this paper, we focus on the problem of massive data retrieval, which is of great importance in data management and searching. To enhance the effectiveness of massive data retrieval, we introduce the Top-k query technology in this work. Top-k denotes to the method which only returns the top K most important objects according to a given ranking function. To tackle the limitations of the existing Top-k query, we proposed a modified Top-k query algorithm. In this algorithm, we select the data elements which have higher ranking scores on each attribute, and then run a threshold controlling scheme on these data elements. Finally, to make performance evaluation, we collect a dataset from US census dataset. Experimental results demonstrate that compared with PDG method, our algorithm can achieve better performance both in retrieval effectiveness and retrieval accuracy.
