ABSTRACT:

High-dimensional databases pose a challenge withrespect to efficient access. High-dimensional indexes do notwork as a result of of the oft-cited "curse of dimensionality'. However, users are typically curious about querying data over a relativelysmall subset of the complete attribute set at a time. A potential answer is to use lower dimensional indexes that accurately represent the user access patterns. Query response using physical database style developed based on a static snapshot of the question workload might significantly degrade if the question patterns amendment.To address these issues, we tend to introduce a parameterizable technique to advocate indexes based on index sorts frequently used forhigh-dimensional knowledge sets and to dynamically regulate indexesas the underlying query workload changes. We tend to incorporate aquery pattern amendment detection mechanism to work out when the access patterns have changed enough to warrant modification inthe physical database style. By adjusting analysis parameters,we have a tendency to trade off analysis speed against analysis resolution. We perform experiments with a number of information sets, question sets, and parameters to point out the effect that varying these characteristics has on analysis results.