Prefetching XML data with abstract query mining

Web applications using the XML technologies are increasingly popular. Besides caching, another query performance improvement approach is prefetching. Similar to other database queries, XML queries often have their own query patterns. We present a preprocessing algorithm that extracts abstract queries from similar query constructs but with different constraints values. These abstract queries help to produce useful access patterns by applying association mining on them. Through our preliminary experiments, we found that our prefetching approach improved both the cache hit ratio and precision of the XML cache system.