annotate medcouple.py @ 74:305b7361a5bd default tip @

showalgo: save a snapshot instead of waiting for keyboard input
author Jordi Gutiérrez Hermoso <jordigh@octave.org>
date Sun, 29 May 2016 19:05:01 -0400
parents 242afe8021b4
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
5
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
1 #!/usr/bin/python
40
2493b21d7d8b medcouple.py: add utf-8 source encoding line
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 37
diff changeset
2 # -*- coding: utf-8 -*-
5
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
3
37
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
4 # medcouple.py ---
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
5
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
6 # Copyright © 2015 Jordi Gutiérrez Hermoso <jordigh@octave.org>
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
7
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
8 # Author: Jordi Gutiérrez Hermoso
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
9
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
10 # This program is free software; you can redistribute it and/or
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
11 # modify it under the terms of the GNU General Public License
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
12 # as published by the Free Software Foundation; either version 3
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
13 # of the License, or (at your option) any later version.
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
14
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
15 # This program is distributed in the hope that it will be useful,
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
16 # but WITHOUT ANY WARRANTY; without even the implied warranty of
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
17 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
18 # GNU General Public License for more details.
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
19
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
20 # You should have received a copy of the GNU General Public License
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
21 # along with this program. If not, see <http://www.gnu.org/licenses/>.
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
22
4fb3b87b8610 Add GPL'ed header
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 33
diff changeset
23
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
24 import random
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
25
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
26 from itertools import tee, izip
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
27
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
28 def wmedian(A, W):
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
29 """This computes the weighted median of array A with corresponding
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
30 weights W.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
31 """
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
32
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
33 AW = zip(A, W)
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
34 n = len(AW)
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
35
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
36 wtot = sum(W)
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
37
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
38 beg = 0
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
39 end = n - 1
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
40
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
41 while True:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
42 mid = (beg + end)//2
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
43
31
1a2bb52722c2 pymedcouple: replace slow Python partsort with faster built-in full sort
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 24
diff changeset
44 AW = sorted(AW, key = lambda x: x[0]) # A partial sort would suffice here
1a2bb52722c2 pymedcouple: replace slow Python partsort with faster built-in full sort
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 24
diff changeset
45
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
46 trial = AW[mid][0]
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
47
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
48 wleft = wright = 0
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
49 for (a, w) in AW:
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
50 if a < trial:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
51 wleft += w
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
52 else:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
53 # This also includes a == trial, i.e. the "middle"
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
54 # weight.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
55 wright += w
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
56
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
57 if 2*wleft > wtot:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
58 end = mid
21
d63e763f8ac1 wmedian: don't infinite loop if wleft == wright
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 20
diff changeset
59 elif 2*wright < wtot:
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
60 beg = mid
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
61 else:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
62 return trial
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
63
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
64
10
3d145bcc8694 Remove more debug code
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 9
diff changeset
65 def medcouple_1d(X, eps1 = 2**-52, eps2 = 2**-1022):
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
66 """Calculates the medcouple robust measure of skewness.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
67
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
68 Parameters
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
69 ----------
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
70 y : array-like, 1-d
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
71
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
72 Returns
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
73 -------
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
74 mc : float
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
75 The medcouple statistic
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
76
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
77 .. [1] G. Brys, M. Hubert, and A. Struyf "A Robust Measure of
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
78 Skewness." Journal of Computational and Graphical Statistics, Vol.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
79 13, No. 4 (Dec., 2004), pp. 996- 1017
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
80
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
81 .. [2] D. B. Johnson and T. Mizoguchi "Selecting the Kth Element
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
82 in $X + Y$ and $X_1 + X_2 + \cdots X_m$". SIAM Journal of
9
e6bcaf38edaa Fix citation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 8
diff changeset
83 Computing, Vol. 7, No. 2 (May 1978), pp. 147-53.
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
84
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
85 """
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
86 # FIXME: Figure out what to do about NaNs.
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
87
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
88 n = len(X)
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
89 n2 = (n - 1)//2
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
90
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
91 if n < 3:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
92 return 0
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
93
52
242afe8021b4 medcouple.py: make sure data is floats
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 40
diff changeset
94 Z = sorted([float(x) for x in X], reverse=True)
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
95
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
96 if n % 2 == 1:
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
97 Zmed = Z[n2]
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
98 else:
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
99 Zmed = (Z[n2] + Z[n2 + 1])/2
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
100
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
101 #Check if the median is at the edges up to relative epsilon
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
102 if abs(Z[0] - Zmed) < eps1*(eps1 + abs(Zmed)):
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
103 return -1.0
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
104 if abs(Z[-1] - Zmed) < eps1*(eps1 + abs(Zmed)):
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
105 return 1.0
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
106
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
107 # Centre Z wrt median, so that median(Z) = 0.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
108 Z = [z - Zmed for z in Z]
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
109
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
110 # Scale inside [-0.5, 0.5], for greater numerical stability.
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
111 Zden = 2*max(Z[0], -Z[-1])
5
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
112 Z = [z/Zden for z in Z]
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
113 Zmed /= Zden
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
114
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
115 Zeps = eps1*(eps1 + abs(Zmed))
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
116
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
117 # These overlap on the entries that are tied with the median
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
118 Zplus = [z for z in Z if z >= -Zeps]
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
119 Zminus = [z for z in Z if Zeps >= z]
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
120
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
121 n_plus = len(Zplus)
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
122 n_minus = len(Zminus)
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
123
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
124
10
3d145bcc8694 Remove more debug code
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 9
diff changeset
125 def h_kern(i, j):
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
126 """Kernel function h for the medcouple, closing over the values of
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
127 Zplus and Zminus just defined above.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
128
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
129 In case a and be are within epsilon of the median, the kernel
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
130 is the signum of their position.
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
131
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
132 """
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
133 a = Zplus[i]
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
134 b = Zminus[j]
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
135
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
136 if abs(a - b) <= 2*eps2:
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
137 h = signum(n_plus - 1 - i - j)
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
138 else:
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
139 h = (a + b)/(a - b)
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
140
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
141 return h
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
142
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
143 # Init left and right borders
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
144 L = [0]*n_plus
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
145 R = [n_minus - 1]*n_plus
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
146
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
147 Ltot = 0
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
148 Rtot = n_minus*n_plus
22
f5b4a2ab6204 medcouple_1d: modify the definition of medc_idx to match R's definition
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 21
diff changeset
149 medc_idx = Rtot//2
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
150
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
151 # kth pair algorithm (Johnson & Mizoguchi)
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
152 while Rtot - Ltot > n_plus:
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
153
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
154 # First, compute the median inside the given bounds
6
e3b1dcc51e6a Fix tee usage
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 5
diff changeset
155 # (Be stingy, reuse same generator)
e3b1dcc51e6a Fix tee usage
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 5
diff changeset
156 [I1, I2] = tee(i for i in xrange(0, n_plus) if L[i] <= R[i])
e3b1dcc51e6a Fix tee usage
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 5
diff changeset
157
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
158 A = [h_kern(i, (L[i] + R[i])//2) for i in I1]
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
159 W = [R[i] - L[i] + 1 for i in I2]
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
160 Am = wmedian(A, W)
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
161
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
162 Am_eps = eps1*(eps1 + abs(Am))
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
163
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
164 # Compute new left and right boundaries, based on the weighted
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
165 # median
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
166 P = []
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
167 Q = []
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
168
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
169 j = 0
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
170 for i in xrange(n_plus - 1, -1, -1):
20
8e02346789b2 Fix off by one error in P computation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 13
diff changeset
171 while j < n_minus and h_kern(i, j) - Am > Am_eps:
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
172 j += 1
12
c81f6d263897 Spaces around plus and minus
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 11
diff changeset
173 P.append(j - 1)
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
174
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
175 P.reverse()
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
176
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
177 j = n_minus - 1
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
178 for i in xrange(0, n_plus):
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
179 while j >= 0 and h_kern(i, j) - Am < -Am_eps:
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
180 j -= 1
12
c81f6d263897 Spaces around plus and minus
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 11
diff changeset
181 Q.append(j + 1)
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
182
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
183 # Check on which side of those bounds the desired median of
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
184 # the whole matrix may be.
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
185 sumP = sum(P) + len(P)
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
186 sumQ = sum(Q)
11
4c7d7ff28a0e Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 10
diff changeset
187
4c7d7ff28a0e Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 10
diff changeset
188 if medc_idx <= sumP - 1:
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
189 R = P
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
190 Rtot = sumP
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
191 else:
11
4c7d7ff28a0e Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 10
diff changeset
192 if medc_idx > sumQ - 1:
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
193 L = Q
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
194 Ltot = sumQ
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
195 else:
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
196 return Am
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
197
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
198 # Didn't find the median, but now we have a very small search
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
199 # space to find it in, just between the left and right boundaries.
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
200 # This space is of size Rtot - Ltot which is <= n_plus
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
201 A = []
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
202 for (i, (l, r)) in enumerate(izip(L, R)):
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
203 for j in xrange(l, r + 1):
10
3d145bcc8694 Remove more debug code
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 9
diff changeset
204 A.append(h_kern(i, j))
31
1a2bb52722c2 pymedcouple: replace slow Python partsort with faster built-in full sort
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 24
diff changeset
205
1a2bb52722c2 pymedcouple: replace slow Python partsort with faster built-in full sort
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 24
diff changeset
206 A.sort() # A partial sort would suffice here
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
207 A.reverse()
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
208
23
29a178b23219 Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 22
diff changeset
209 Am = A[medc_idx - Ltot]
11
4c7d7ff28a0e Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 10
diff changeset
210
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
211 return Am
13
077261db7a58 Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 12
diff changeset
212
4
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
213
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
214 def signum(x):
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
215 if x > 0:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
216 return 1
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
217 if x < 0:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
218 return -1
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
219 else:
74d0d08dbc95 Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff changeset
220 return 0
5
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
221
8
b3e878bb793d Remove debug statements
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 7
diff changeset
222
5
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
223 def main():
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
224 import sys
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
225 fname = sys.argv[1]
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
226 with open(fname) as f:
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
227 data = [float(x) for x in f.readlines() if x.strip() != ""]
8
b3e878bb793d Remove debug statements
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 7
diff changeset
228
7
6c7c7dc9d8ef hairball
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 6
diff changeset
229 print "%.16g" % medcouple_1d(data)
5
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
230
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
231 if __name__ == "__main__":
cbe17f888c79 Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents: 4
diff changeset
232 main()