Glossary · Phase 3 · Machine learning
Sketch
A small, mergeable data structure that answers a question approximately, such as HyperLogLog for distinct counts or t-digest for quantiles.
For the statistics that won't merge, there are sketches: small, mergeable structures that give approximate answers with known error.
Try it now
Response times with the server that handled each request. Group by server and compare the mean with the count.
Learn it properly
Data in pieces: partitions and mergeable summaries · Split, summarize and combine; which statistics merge exactly, merging variances, sketches for the rest, and partitioning well. About 30 minutes.
Other terms from the same lesson
- HyperLogLog: A sketch that estimates how many distinct values a huge dataset has, within a few percent, using a few kilobytes.
- mergeable summary: A summary of a partition that can be combined with another partition's to give the summary of both, such as a count, a sum, or a mean kept with its count.
Never trained an AI?
Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account.