1mlpack_pca(1) User Commands mlpack_pca(1)
2
3
4
6 mlpack_pca - principal components analysis
7
9 mlpack_pca -i string [-c string] [-d int] [-s bool] [-r double] [-V bool] [-o string] [-h -v]
10
12 This program performs principal components analysis on the given
13 dataset using the exact, randomized, randomized block Krylov, or QUIC
14 SVD method. It will transform the data onto its principal components,
15 optionally performing dimensionality reduction by ignoring the princi‐
16 pal components with the smallest eigenvalues.
17
18 Use the '--input_file (-i)' parameter to specify the dataset to perform
19 PCA on. A desired new dimensionality can be specified with the
20 ’--new_dimensionality (-d)' parameter, or the desired variance to
21 retain can be specified with the '--var_to_retain (-r)' parameter. If
22 desired, the dataset can be scaled before running PCA with the '--scale
23 (-s)' parameter.
24
25 Multiple different decomposition techniques can be used. The method to
26 use can be specified with the '--decomposition_method (-c)' parameter,
27 and it may take the values 'exact', 'randomized', or 'quic'.
28
29 For example, to reduce the dimensionality of the matrix 'data.csv' to 5
30 dimensions using randomized SVD for the decomposition, storing the out‐
31 put matrix to 'data_mod.csv', the following command can be used:
32
33 $ pca --input_file data.csv --new_dimensionality 5 --decomposi‐
34 tion_method randomized --output_file data_mod.csv
35
37 --input_file (-i) [string]
38 Input dataset to perform PCA on.
39
41 --decomposition_method (-c) [string]
42 Method used for the principal components analysis: 'exact',
43 'randomized', 'randomized-block-krylov', 'quic'. Default value
44 'exact'.
45
46 --help (-h) [bool]
47 Default help info.
48
49 --info [string]
50 Get help on a specific module or option. Default value ''.
51
52 --new_dimensionality (-d) [int]
53 Desired dimensionality of output dataset. If 0, no dimensional‐
54 ity reduction is performed. Default value 0.
55
56 --scale (-s) [bool]
57 If set, the data will be scaled before running PCA, such that
58 the variance of each feature is 1.
59
60 --var_to_retain (-r) [double]
61 Amount of variance to retain; should be between 0 and 1. If 1,
62 all variance is retained. Overrides -d. Default value 0.
63
64 --verbose (-v) [bool]
65 Display informational messages and the full list of parameters
66 and timers at the end of execution.
67
68 --version (-V) [bool]
69 Display the version of mlpack.
70
72 --output_file (-o) [string]
73 Matrix to save modified dataset to. Default value ''.
74
76 For further information, including relevant papers, citations, and the‐
77 ory, consult the documentation found at http://www.mlpack.org or
78 included with your distribution of mlpack.
79
80
81
82mlpack-3.0.4 21 February 2019 mlpack_pca(1)