Prof. Chen is recognized as one of the experts in distributed/parallel query processing and data mining with strong research credentials. He has published more than 270 papers and edited two books, and more than 80 of his journal papers are published in major ACM/IEEE journals/transactions. According to Google Scholar, the publications of Prof. Chen have received more than 7,000 citations in total. He also filed 17 US patents and 7 ROC patents. More information for his research results can be found in http://www.ee.ntu.edu.tw/~mschen.

        One of Prof. Chen's works is on developing the framework and algorithms to improve the execution of distributed and parallel queries .His distributed query processing work goes beyond the traditional paradigm of only using semijoins as reducers for query cost reduction. Instead, he combined joins and semijoins as reducers and devised an innovative approach to interleaving a sequence of joins with properly-identified semijoins to minimize the query execution cost. Prof. Chen's work on parallel query processing exploited three levels of parallelism, namely intra-operator, inter-operator, and inter-query levels. The notion of using multiple partitioned hash tables he proposed has been validated to be a viable approach to significantly reducing false lock contention. To improve parallel transaction processing, he further devised a new hash apparatus for an important database product and this apparatus was shown to be able to reduce the locking overhead significantly. He was awarded an Outstanding Innovation Award by IBM Corp. for his contributions to parallel transaction and query processing.

        Prof. Chen also conducted pioneering research on data mining. Several association rule mining techniques he proposed have been widely referenced and adopted by subsequent mining works. Prof. Chen pioneered the work on exploring user moving pattern both in the Web and in a mobile computing environment, and also contributed to the areas of Web search and Web content mining. Explicitly, he was among the very first to explore path traversal pattern mining in the Web, which has later spawned a subsequent of studies. He devised a search method VIPAS which builds virtual hyperlinks in light of prior usage of search results to enable itself to render better ranked Web pages to users. His work on sequential data broadcasting has been widely cited by recent papers on mobile computing.