Mercurial > hg > medcouple
annotate pymedcouple @ 24:cd3094a08e03
Remove unused numpy import
author | Jordi Gutiérrez Hermoso <jordigh@octave.org> |
---|---|
date | Sat, 17 Jan 2015 20:05:24 -0500 |
parents | 29a178b23219 |
children | 1a2bb52722c2 |
rev | line source |
---|---|
5
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
1 #!/usr/bin/python |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
2 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
3 import random |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
4 |
7 | 5 from itertools import tee, izip |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
6 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
7 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
8 def partsort(L, n): |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
9 """This is a partial sort of the weighted list L = [(a,w) for a in A, |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
10 w in W], so that the element at position n is in the correct spot. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
11 Also known as quickselect. The elements of W are ignored for the |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
12 purposes of this partial sort. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
13 """ |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
14 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
15 beg = 0 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
16 end = len(L) |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
17 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
18 while True: |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
19 pivot = random.randint(beg, end - 1) |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
20 pivotval = L[pivot][0] |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
21 L[pivot], L[end - 1] = L[end - 1], L[pivot] |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
22 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
23 idx = beg |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
24 for i in xrange(beg, end): |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
25 if L[i][0] < pivotval: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
26 L[idx], L[i] = L[i], L[idx] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
27 idx += 1 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
28 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
29 L[idx], L[end-1] = L[end-1], L[idx] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
30 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
31 if idx == n: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
32 return L[idx] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
33 elif idx < n: |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
34 beg = idx + 1 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
35 else: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
36 end = idx |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
37 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
38 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
39 def wmedian(A, W): |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
40 """This computes the weighted median of array A with corresponding |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
41 weights W. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
42 """ |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
43 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
44 AW = zip(A, W) |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
45 n = len(AW) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
46 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
47 wtot = sum(W) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
48 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
49 beg = 0 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
50 end = n - 1 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
51 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
52 while True: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
53 mid = (beg + end)//2 |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
54 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
55 partsort(AW, mid) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
56 trial = AW[mid][0] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
57 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
58 wleft = wright = 0 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
59 for (a, w) in AW: |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
60 if a < trial: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
61 wleft += w |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
62 else: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
63 # This also includes a == trial, i.e. the "middle" |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
64 # weight. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
65 wright += w |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
66 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
67 if 2*wleft > wtot: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
68 end = mid |
21
d63e763f8ac1
wmedian: don't infinite loop if wleft == wright
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
20
diff
changeset
|
69 elif 2*wright < wtot: |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
70 beg = mid |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
71 else: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
72 return trial |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
73 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
74 |
10
3d145bcc8694
Remove more debug code
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
9
diff
changeset
|
75 def medcouple_1d(X, eps1 = 2**-52, eps2 = 2**-1022): |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
76 """Calculates the medcouple robust measure of skewness. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
77 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
78 Parameters |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
79 ---------- |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
80 y : array-like, 1-d |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
81 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
82 Returns |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
83 ------- |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
84 mc : float |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
85 The medcouple statistic |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
86 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
87 .. [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
|
88 Skewness." Journal of Computational and Graphical Statistics, Vol. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
89 13, No. 4 (Dec., 2004), pp. 996- 1017 |
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 .. [2] D. B. Johnson and T. Mizoguchi "Selecting the Kth Element |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
92 in $X + Y$ and $X_1 + X_2 + \cdots X_m$". SIAM Journal of |
9 | 93 Computing, Vol. 7, No. 2 (May 1978), pp. 147-53. |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
94 |
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 # FIXME: Figure out what to do about NaNs. |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
97 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
98 n = len(X) |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
99 n2 = (n - 1)//2 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
100 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
101 if n < 3: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
102 return 0 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
103 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
104 Z = sorted(X, reverse=True) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
105 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
106 if n % 2 == 1: |
7 | 107 Zmed = Z[n2] |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
108 else: |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
109 Zmed = (Z[n2] + Z[n2 + 1])/2 |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
110 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
111 #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
|
112 if abs(Z[0] - Zmed) < eps1*(eps1 + abs(Zmed)): |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
113 return -1.0 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
114 if abs(Z[-1] - Zmed) < eps1*(eps1 + abs(Zmed)): |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
115 return 1.0 |
7 | 116 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
117 # Centre Z wrt median, so that median(Z) = 0. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
118 Z = [z - Zmed for z in Z] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
119 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
120 # Scale inside [-0.5, 0.5], for greater numerical stability. |
7 | 121 Zden = 2*max(Z[0], -Z[-1]) |
5
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
122 Z = [z/Zden for z in Z] |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
123 Zmed /= Zden |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
124 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
125 Zeps = eps1*(eps1 + abs(Zmed)) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
126 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
127 # These overlap on the entries that are tied with the median |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
128 Zplus = [z for z in Z if z >= -Zeps] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
129 Zminus = [z for z in Z if Zeps >= z] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
130 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
131 n_plus = len(Zplus) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
132 n_minus = len(Zminus) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
133 |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
134 |
10
3d145bcc8694
Remove more debug code
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
9
diff
changeset
|
135 def h_kern(i, j): |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
136 """Kernel function h for the medcouple, closing over the values of |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
137 Zplus and Zminus just defined above. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
138 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
139 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
|
140 is the signum of their position. |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
141 |
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 a = Zplus[i] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
144 b = Zminus[j] |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
145 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
146 if abs(a - b) <= 2*eps2: |
7 | 147 h = signum(n_plus - 1 - i - j) |
148 else: | |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
149 h = (a + b)/(a - b) |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
150 |
7 | 151 return h |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
152 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
153 # Init left and right borders |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
154 L = [0]*n_plus |
7 | 155 R = [n_minus - 1]*n_plus |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
156 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
157 Ltot = 0 |
7 | 158 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
|
159 medc_idx = Rtot//2 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
160 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
161 # kth pair algorithm (Johnson & Mizoguchi) |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
162 while Rtot - Ltot > n_plus: |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
163 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
164 # First, compute the median inside the given bounds |
6 | 165 # (Be stingy, reuse same generator) |
166 [I1, I2] = tee(i for i in xrange(0, n_plus) if L[i] <= R[i]) | |
167 | |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
168 A = [h_kern(i, (L[i] + R[i])//2) for i in I1] |
7 | 169 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
|
170 Am = wmedian(A, W) |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
171 |
7 | 172 Am_eps = eps1*(eps1 + abs(Am)) |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
173 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
174 # Compute new left and right boundaries, based on the weighted |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
175 # median |
7 | 176 P = [] |
177 Q = [] | |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
178 |
7 | 179 j = 0 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
180 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
|
181 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
|
182 j += 1 |
12
c81f6d263897
Spaces around plus and minus
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
11
diff
changeset
|
183 P.append(j - 1) |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
184 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
185 P.reverse() |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
186 |
7 | 187 j = n_minus - 1 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
188 for i in xrange(0, n_plus): |
7 | 189 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
|
190 j -= 1 |
12
c81f6d263897
Spaces around plus and minus
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
11
diff
changeset
|
191 Q.append(j + 1) |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
192 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
193 # Check on which side of those bounds the desired median of |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
194 # the whole matrix may be. |
7 | 195 sumP = sum(P) + len(P) |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
196 sumQ = sum(Q) |
11
4c7d7ff28a0e
Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
10
diff
changeset
|
197 |
4c7d7ff28a0e
Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
10
diff
changeset
|
198 if medc_idx <= sumP - 1: |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
199 R = P |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
200 Rtot = sumP |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
201 else: |
11
4c7d7ff28a0e
Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
10
diff
changeset
|
202 if medc_idx > sumQ - 1: |
7 | 203 L = Q |
204 Ltot = sumQ | |
205 else: | |
206 return Am | |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
207 |
7 | 208 # Didn't find the median, but now we have a very small search |
209 # space to find it in, just between the left and right boundaries. | |
210 # This space is of size Rtot - Ltot which is <= n_plus | |
211 A = [] | |
212 for (i, (l, r)) in enumerate(izip(L, R)): | |
213 for j in xrange(l, r + 1): | |
10
3d145bcc8694
Remove more debug code
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
9
diff
changeset
|
214 A.append(h_kern(i, j)) |
7 | 215 A.sort() |
216 A.reverse() | |
217 | |
23
29a178b23219
Whitespace style fixes
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
22
diff
changeset
|
218 Am = A[medc_idx - Ltot] |
11
4c7d7ff28a0e
Rename mid_idx to medc_idx
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
10
diff
changeset
|
219 |
7 | 220 return Am |
13
077261db7a58
Whitespace cleanup
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
12
diff
changeset
|
221 |
4
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
222 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
223 def signum(x): |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
224 if x > 0: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
225 return 1 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
226 if x < 0: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
227 return -1 |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
228 else: |
74d0d08dbc95
Add Python implementation
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
diff
changeset
|
229 return 0 |
5
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
230 |
8
b3e878bb793d
Remove debug statements
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
7
diff
changeset
|
231 |
5
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
232 def main(): |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
233 import sys |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
234 fname = sys.argv[1] |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
235 with open(fname) as f: |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
236 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
|
237 |
7 | 238 print "%.16g" % medcouple_1d(data) |
5
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
239 |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
240 if __name__ == "__main__": |
cbe17f888c79
Make it executable
Jordi Gutiérrez Hermoso <jordigh@octave.org>
parents:
4
diff
changeset
|
241 main() |