Created
April 8, 2017 01:39
-
-
Save vznvzn/47d5fbd792204a0883db72c4130926bf to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
require 'statsample' | |
def f2(n) | |
n = (n * 3 + 1) / 2 while (n.odd?) | |
n /= 2 while (n.even?) | |
return n | |
end | |
def setx(x, mn, mx) | |
x['x'] = (x['r'] - mn) / (mx - mn) * 9.99 if (!mn.nil? && mn != mx) | |
end | |
def adv(x, seen, l1, mn = nil, mx = nil) | |
x['n'] = x['nb'].to_i(2) | |
n1 = n = x['n'] | |
return if (seen.member?(n)) | |
seen[n] = x | |
l = [n] | |
c = 10 | |
while (n >= n1 && n != 1 && l.size < c) | |
n = f2(n) | |
l << n | |
end | |
x['n_2'] = n | |
x['ls'] = l.size | |
x['ns'] = x['n'].to_s(2).length | |
x['r'] = nil | |
if (l.size == c) then | |
x['r'] = n.to_s(2).length.to_f / x['ns'] | |
setx(x, mn, mx) if (!mn.nil?) | |
end | |
x['.'] = x['x'].nil? ? '?' : x['x'].to_i.to_s | |
l1 << x | |
end | |
def stat(l) | |
l = [0] if (l.empty?) | |
t = t2 = 0 | |
l.each \ | |
{ | |
|x| | |
t += x | |
t2 += x ** 2 | |
} | |
c = l.size | |
a = t.to_f / c | |
z = t2.to_f / c - a ** 2 | |
sd = Math.sqrt(z < 0 ? 0 : z) | |
return a, sd, l.max.to_f, l.min.to_f | |
end | |
def init(w, f) | |
s = (0...w).map { rand(2).to_s }.join | |
s[0, 1] = '1' if (f) | |
return s | |
end | |
def freq(l, h = {}) | |
l1 = l.map { |x| x['x'].nil? ? nil : x['x'].to_i }.compact | |
h.replace(Hash[hist(l1)]) | |
# $stderr.puts(h.inspect) | |
l1 = h.sort_by { |k, v| v }.reverse.map { |x| x[0] } | |
h2 = Hash[[l1, (0...l1.size).to_a].transpose] | |
$stderr.puts(h.sort.map { |k, v| [k, v, h2.fetch(k, 0) ] }.inspect) | |
return h2 | |
end | |
def minmax(l) | |
l = l.map { |x| x['r'] }.compact | |
return l.min, l.max | |
end | |
def rescale(l, mn, mx) | |
l.each \ | |
{ | |
|x| | |
setx(x, mn, mx) if (!x['r'].nil?) | |
} | |
end | |
def hist(l) | |
h = {} | |
l.each \ | |
{ | |
|x| | |
h[x] = h.fetch(x, 0) + 1 | |
} | |
return h.sort | |
end | |
def dist(w, c) | |
l1 = [] | |
f0 = f1 = true | |
seen = {} | |
20.times { adv({'nb' => init(w, f0)}, seen, l1) } | |
i = n = z = 0 | |
mn = mx = nil | |
h1 = {} | |
h = {} | |
while (n < c) | |
i += 1 | |
if (i % 100 == 0) then | |
mn, mx = minmax(seen.values) | |
rescale(seen.values.select { |x| x.member?('x') }, mn, mx) | |
h1 = freq(l1.select { |x| !x.member?('*') && x.member?('x') }, h) | |
h2 = freq(seen.values.select { |x| x.member?('*') }) | |
$stderr.puts([h1.size, h2.size].inspect) | |
$stderr.puts | |
end | |
l1 = l1.sort_by { |x| [x['ls'], x.member?('x') ? h2.fetch(x['x'].to_i, 9) : 0] }.reverse | |
t = f1 ? l1.size / 4 : [l1.size, 400].min | |
case i % 3 | |
when 0 | |
x = l1[rand(t)] | |
s = x['nb'].dup | |
r = f0 ? (rand(s.length - 1) + 1) : rand(s.length) | |
s[r, 1] = (s[r, 1].to_i ^ 1).to_s | |
a = x['.'] | |
when 1 | |
x = l1[rand(t)] | |
y = l1[rand(t)] | |
sx = x['nb'] | |
sy = y['nb'] | |
s = '' | |
w.times \ | |
{ | |
|j| | |
s[j, 1] = (rand(2) == 0) ? sx[j, 1] : sy[j, 1] | |
} | |
a = x['.'] + ' x ' + y['.'] | |
when 2 | |
x = l1[rand(t)] | |
y = l1[rand(t)] | |
sx = x['nb'] | |
sy = y['nb'] | |
s = '' | |
r = rand(w + 1) | |
w.times \ | |
{ | |
|j| | |
s[j, 1] = (j < r) ? sx[j, 1] : sy[j, 1] | |
} | |
a = x['.'] + ' y ' + y['.'] | |
end | |
# p(a) | |
next if (s.to_i == 0) | |
l1.pop if (l1.size >= 500) | |
if (h1.size >= 5) then | |
x = l1.select { |x| x.member?('x') && !x.member?('*') && h.fetch(x['x'].to_i, 0) >= 3 }.max_by{ |x| h2.fetch(x['x'].to_i, 9) } | |
if (!x.nil?) then | |
x['*'] = nil | |
h[x['x'].to_i] -= 1 | |
end | |
end | |
adv({'nb' => s}, seen, l1, mn, mx) | |
n += 1 | |
end | |
mn, mx = minmax(l = seen.values.select { |x| x.member?('x') }) | |
rescale(l, mn, mx) | |
return l | |
end | |
def stat2(l, t, n) | |
return Hash[[["a#{n}", "sd#{n}", "mx#{n}"], stat(l)[0..2].map { |x| x / t }].transpose] | |
end | |
def d(s) | |
c = s.split('').select { |x| x == '1' }.size | |
d = c.to_f / s.length | |
return d | |
end | |
def data(n) | |
ns = n.to_s(2) | |
nl = ns.length | |
m = nl / 2 | |
nsh = ns[0..m] | |
nsl = ns[m..-1] | |
asdm1 = stat2(ns.split(/0+/).map { |x| x.length }, nl, 1) | |
l1 = ns.split(/1+/) | |
l1.shift | |
asdm0 = stat2(l1.map { |x| x.length }, nl, 0) | |
return {'d' => d(ns), 'dh' => d(nsh), 'dl' => d(nsl)}.merge(asdm1).merge(asdm0) | |
end | |
def fit(l, y, lx) | |
a = {} | |
(lx + [y]).each { |x| a[x] = l.map { |b| b[x] }.to_vector() } | |
ds = a.to_dataset() | |
r = Statsample::Regression.multiple(ds, y) | |
# $stderr.puts(r.summary) | |
return r.coeffs.merge({'c' => r.constant}) | |
end | |
def predict(l, y, z) | |
l.each \ | |
{ | |
|x| | |
t = z['c'] | |
(z.keys - ['c']).each { |k| t += z[k] * x[k] } | |
x["#{y}_p"] = t | |
} | |
end | |
def solve(l, y, x) | |
z = fit(l, y, x) | |
predict(l, y, z) | |
end | |
def sum(l) | |
t = 0.0 | |
l.each { |x| t += x } | |
return t | |
end | |
def av(l) | |
return nil if (l.empty?) | |
return sum(l) / l.size | |
end | |
def corr(l, y1, yp) | |
xav = av(l.map { |x| x[y1] }) | |
yav = av(l.map { |x| x[yp] }) | |
tx = ty = txy = e = 0.0 | |
l.each \ | |
{ | |
|z| | |
x = z[y1] | |
y = z[yp] | |
txy += (x - xav) * (y - yav) | |
tx += (x - xav) ** 2 | |
ty += (y - yav) ** 2 | |
e += (x - y) ** 2 | |
} | |
r = txy / (Math.sqrt(tx) * Math.sqrt(ty)) | |
e /= l.size | |
return r, e | |
end | |
def center(l, y) | |
a, sd = stat(l.map { |x| x[y] }) | |
l.each \ | |
{ | |
|x| | |
x["#{y}z"] = (x[y] - a) / sd | |
} | |
end | |
def balance2(l) | |
c = 20 | |
d = 10.0 / c | |
x0 = 0.0 | |
l2 = [] | |
seen = {} | |
while (x0 <= 10) | |
l.sort_by! { |x| (x['x'] - x0).abs } | |
x0 += d | |
l[0...10].each \ | |
{ | |
|x| | |
next if (seen.member?(x['n'])) | |
seen[x['n']] = nil | |
l2 << x | |
} | |
end | |
return l2 | |
end | |
def coef(l, y0, x0) | |
r = nil | |
begin | |
solve(l, y0, x0) | |
r, e = corr(l, y0, "#{y0}_p") | |
rescue Statsample::Regression::LinearDependency | |
end | |
return {y0 => r} | |
end | |
def test(x0) | |
a = {} | |
l = dist(50, 10e3.to_i) | |
a['c1'] = l.size | |
l = balance2(l) | |
a['c2'] = l.size | |
l.each \ | |
{ | |
|x| | |
x.merge!(data(x['n'])) | |
x.merge!(Hash[data(x['n_2']).to_a.map { |k, v| ["#{k}_2", v] }]) | |
} | |
a.merge!(coef(l, 'x', x0)) | |
f = false | |
x0.each \ | |
{ | |
|x| | |
a.merge!(coef(l, "#{x}_2", x0)) | |
f = true if (a["#{x}_2"].nil?) | |
} | |
if (f) then | |
$stderr.puts(a) | |
return nil | |
end | |
return a | |
end | |
def out(fn, a) | |
return if (a.nil?) | |
f = File.open(fn, 'a') | |
f.puts(a.keys.join("\t")) if (f.size == 0) | |
f.puts(a.values.join("\t")) | |
f.close | |
end | |
def plot(x0, r) | |
f = File.open('plot.cmd', 'w') | |
f.printf("plot #{r} ") | |
x0.each \ | |
{ | |
|x| | |
f.print("'out.txt' using (column('#{x}')) with line,") | |
} | |
f.puts | |
f.close | |
end | |
File.open(fn = 'out.txt', 'w').close | |
x0 = ['d', 'dh', 'dl', 'a0', 'sd0', 'mx0', 'a1', 'sd1', 'mx1'] | |
plot(x0.map { |x| "#{x}_2" } + ['x'], '[][0:1.1]') | |
20.times { out(fn, test(x0)) } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment