1mlpack_approx_kfn(1) General Commands Manual mlpack_approx_kfn(1)
2
6 mlpack_approx_kfn - approximate furthest neighbor search
7
9 mlpack_approx_kfn [-h] [-v]
10
12 This program implements two strategies for furthest neighbor search.
13 These strategies are:
14 · The 'qdafn' algorithm from 'Approximate Furthest Neighbor in
16 High Dimensions' by R. Pagh, F. Silvestri, J. Sivertsen, and
17 M. Skala, in Similarity Search and Applications 2015 (SISAP).
18 · The 'DrusillaSelect' algorithm from 'Fast approximate fur‐
20 thest neighbors with data-dependent candidate selection, by
21 R.R. Curtin and A.B. Gardner, in Similarity Search and Appli‐
22 cations 2016 (SISAP).
23 These two strategies give approximate results for the furthest neighbor
25 search problem and can be used as fast replacements for other furthest
26 neighbor techniques such as those found in the mlpack_kfn program. Note
27 that typically, the 'ds' algorithm requires far fewer tables and pro‐
28 jections than the 'qdafn' algorithm.
29 Specify a reference set (set to search in) with --reference_file, spec‐
31 ify a query set with --query_file, and specify algorithm parameters
32 with --num_tables (-t) and --num_projections (-p) (or don't and
33 defaults will be used). The algorithm to be used (either 'ds'---the
34 default---or 'qdafn') may be specified with --algorithm. Also specify
35 the number of neighbors to search for with --k. Each of those options
36 also has short names; see the detailed parameter documentation below.
37 If no query file is specified, the reference set will be used as the
39 query set. A model may be saved with --output_model_file (-M), and an
40 input model may be loaded instead of specifying a reference set with
41 --input_model_file (-m).
42 Results for each query point are stored in the files specified by
44 --neighbors_file and --distances_file. This is in the same format as
45 the mlpack_kfn and mlpack_knn programs: each row holds the k distances
46 or neighbor indices for each query point.
47
49 --algorithm (-a) [string]
50 Algorithm to use: 'ds' or 'qdafn'. Default value 'ds'.
51 --calculate_error (-e)
53 If set, calculate the average distance error for the first fur‐
54 thest neighbor only. --distances_file (-d) [string] File to
55 save furthest neighbor distances to. Default value ''.
56 --exact_distances_file (-x) [string] File containing exact dis‐
57 tances to furthest neighbors; this can be used to avoid explicit
58 calculation when --calculate_error is set. Default value ''.
59 --help (-h)
61 Default help info.
62 --info [string]
64 Get help on a specific module or option. Default value ''.
65 --input_model_file (-m) [string] File containing input model.
66 Default value ''.
67 --k (-k) [int]
69 Number of furthest neighbors to search for. Default value 0.
70 --neighbors_file (-n) [string] File to save furthest neighbor
71 indices to. Default value ''. --num_projections (-p) [int]
72 Number of projections to use in each hash table. Default value
73 5.
74 --num_tables (-t) [int]
76 Number of hash tables to use. Default value 5.
77 --query_file (-q) [string]
79 File containing query points. Default value ’'. --refer‐
80 ence_file (-r) [string] File containing reference points.
81 Default value ’'.
82 --verbose (-v)
84 Display informational messages and the full list of parameters
85 and timers at the end of execution.
86 --version (-V)
88 Display the version of mlpack.
89
91 --output_model_file (-M) [string] File to save output model to. Default
92 value ’'.
93
96 For further information, including relevant papers, citations, and the‐
97 ory, For further information, including relevant papers, citations, and
98 theory, consult the documentation found at http://www.mlpack.org or
99 included with your consult the documentation found at
100 http://www.mlpack.org or included with your DISTRIBUTION OF MLPACK.
101 DISTRIBUTION OF MLPACK.
102 mlpack_approx_kfn(1)