EfficientApproximateSearchonStringCollectionsPartII.pptVIP

Efficient Approximate Search on String CollectionsPart II;Overview;Selection Queries Using Sketch Based Algorithms;What is a Sketch;Using Sketches for Selection Queries;Known sketches;Prefix Filter ;Example;Example continued;Example with weighted vectors w(s?t) ? θ ( w(s?t) = Σq?s?tw(q) ) Sort by weights (not lexicographically anymore) Keep prefix pf(s) s.t. w[pf(s)] ? w(s) - α ;Continued;Properties;How do I Choose α?;Extend to Jaccard;Technicality;Extend to Edit Distance;Edit Distance Continued;Edit Distance Candidates;Constructing the Prefix;Choosing α;Pros/Cons;Mismatch Filter;Mismatch Filter Continued;Mismatch Condition;Pros/Cons;Minhash;How to use minhash;Pros/Cons;PartEnum;Example;Pros/Cons;Compression(BJL+09);A Global Approach;Simple strategies;Combining Lists;General Observation;Selectivity Estimationfor Selection Queries;The Problem;Flavors;Edit Distance;Clustering - Sepia;Minhash - VSol;VSol Estimator;Selectivity Estimation;Example;The m-β Similarity ;OptEQ – wild-card q-grams;Example;Assuming Replacements Only;Replacement Intersection Lattice;Replacement Lattice;General Formulas;Hashed Sampling;Visual Example;Construction;Example;Selectivity Estimation;Count Distinct;The Bottom-k Sketch;Transformations/Synonyms(ACGK08);Transformations;Observations;Augmented Generative Grammar;Efficiency;Conclusion;Conclusion;Thank you!;References;References;References