Skip to content

Latest commit

 

History

History
17 lines (13 loc) · 798 Bytes

File metadata and controls

17 lines (13 loc) · 798 Bytes

Approximate Nearest Neighbours LSH

University Assignment: Implement LSH algorithm for d-dimensional images in C++, using Lloyd's algorithm and Hypercube & L-Hashtables structures. L1 metric: Manhattan.

We can run it like this, from outside of the src/ folder make && ./lsh -d Datasets/dataset -q Datasets/queryset -k 4 -L 1 -o test.txt -N 2 -R 10000

Or like this to run the cube algorithm make && ./cube -d Datasets/dataset -q Datasets/queryset -k 14 -M 10 -probes 5 -o results2.txt -N 5 -R 10000

Cluster: make && ./cluster -i Datasets/dataset -c cluster.conf -o testOutputCluster.txt -complete -m Classic

File paths have to be without "" Provided that we have the dataset in the folder ./Datasets/dataset Provided that we have the dataset in the folder ./Datasets/queryset