Created
June 2, 2017 01:23
-
-
Save vznvzn/960540c87ab900b758b54e52cfe3e406 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) | |
return n.odd? ? (n * 3 + 1) / 2 : n / 2 | |
end | |
def f1(n) | |
n = (n * 3 + 1) / 2 if (n.odd?) | |
n /= 2 if (n.even?) | |
return n | |
end | |
def dense(w, d) | |
w2 = w - 1 | |
a = (0...w2).to_a | |
s = '0' * w2 | |
(1..(d * w - 1)).map { a.delete_at(rand(a.size)) }.each { |x| s[x, 1] = '1' } | |
return ('1' + s).to_i(2) | |
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 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 {'n' => n, 'ns' => ns, 'nl' => nl, 'd' => d(ns), 'dh' => d(nsh), 'dl' => d(nsl)}.merge(asdm1).merge(asdm0) | |
end | |
def dist() | |
w = 100 | |
d = 0.0 | |
c = 250 | |
l = [] | |
(c + 1).times \ | |
{ | |
|x| | |
d = x.to_f / c | |
n = dense(w, d) | |
x = {0 => data(n), 1 => data(f2(n))} | |
x[1]['wr'] = x[1]['nl'].to_f / x[0]['nl'].to_f | |
l << x | |
} | |
$stderr.puts("#{l.size} pts") | |
return l | |
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 = "#{y1}_p" | |
ye = "#{y1}_e" | |
xav = av(l.map { |x| x[y1] }) | |
yav = av(l.map { |x| x[yp] }) | |
tx = ty = txy = e = 0.0 | |
m = nil | |
l.each \ | |
{ | |
|z| | |
x = z[y1] | |
y = z[yp] | |
txy += (x - xav) * (y - yav) | |
tx += (x - xav) ** 2 | |
ty += (y - yav) ** 2 | |
z[ye] = (x - y) | |
e1 = z[ye].abs | |
e += e1 | |
m = [m, e1].compact.max | |
} | |
r = txy / (Math.sqrt(tx) * Math.sqrt(ty)) | |
e /= l.size | |
return {'r' => r, 'e_a' => e, 'e_m' => m} | |
end | |
def dot(x, z) | |
t = z['c'] | |
(z.keys - ['c']).each { |v| t += z[v] * x[v] } | |
return t | |
end | |
def predict(l, vy, z) | |
l.each \ | |
{ | |
|x| | |
x[1]["#{vy}_p"] = dot(x[0], z) | |
} | |
end | |
def fit(l, vx, vy) | |
a = {} | |
vx.each { |v| a[v] = l.map { |x| x[0][v] }.to_vector() } | |
a[vy] = l.map { |x| x[1][vy] }.to_vector() | |
ds = a.to_dataset() | |
r = Statsample::Regression.multiple(ds, vy) | |
# $stderr.puts(r.summary) | |
z = r.coeffs.merge({'c' => r.constant}) | |
predict(l, vy, z) | |
a = corr(l.map { |x| x[1] }, vy) | |
return z.merge!(a) | |
end | |
def fmt(a) | |
a2 = {} | |
a.each \ | |
{ | |
|k, v| | |
a2[k] = v.is_a?(Numeric) ? sprintf("%.3g", v).to_f : v | |
} | |
$stderr.puts(a2.inspect) | |
end | |
def model(v) | |
l = dist() | |
a = {} | |
l2 = (1..l.size).map { {} } | |
(v + ['wr']).each \ | |
{ | |
|x| | |
a[x] = fit(l, v, x) | |
l1 = l.map { |y| y[1][x] } | |
a[x].merge!({'mn' => l1.min, 'mx' => l1.max}) | |
l.size.times \ | |
{ | |
|i| | |
l2[i][x] = (l[i][1]["#{x}_p"] - l[i][1][x]) / a[x]['e_m'] | |
} | |
fmt({'x' => x}.merge(a[x].select { |k, v| ['r', 'e_m', 'e_a', 'mn', 'mx'].member?(k) })) | |
} | |
f = File.open('err.txt', 'w') | |
f.puts(l2[0].keys.join("\t")) | |
l2.each { |x| f.puts(x.values.join("\t")) } | |
f.close | |
a['l'] = l | |
return a | |
end | |
def out(fn, a) | |
return if (fn.nil?) | |
f = File.open(fn, 'a') | |
f.puts(a.keys.join("\t")) if (f.size == 0) | |
if (a.nil?) | |
f.puts | |
else | |
f.puts(a.values.join("\t")) | |
end | |
f.close | |
end | |
def detect(a, v, n, j) | |
x1 = data(n) | |
i = 0 | |
c = 500 | |
r = [3, 2] | |
x1['nl'] = r[0] | |
loop \ | |
{ | |
x2 = {} | |
(v + ['wr']).each \ | |
{ | |
|vy| | |
x2[vy] = dot(x1, a[vy].select { |k, x| (v + ['c']).member?(k) }) | |
x2[vy] += a['l'][j][1]["#{vy}_e"] if (!j.nil?) | |
x2[vy] = a[vy]['mn'] if (x2[vy] < a[vy]['mn']) | |
x2[vy] = a[vy]['mx'] if (x2[vy] > a[vy]['mx']) | |
} | |
x2['nl'] = x1['nl'] * x2['wr'] | |
x1 = x2 | |
out(a['fn'], x1) if (i % 10 == 0) | |
i += 1 | |
break if (x1['nl'] < r[1] || i == c) | |
} | |
out(a['fn'], nil) | |
return i == c ? nil : i | |
end | |
def scan(a, v, j, c) | |
l = [] | |
File.open(a['fn'], 'w').close if (!a['fn'].nil?) | |
(c + 1).times \ | |
{ | |
|i| | |
x = detect(a, v, dense(100, i.to_f / c), j) | |
l << x | |
} | |
return l | |
end | |
def test(a, v) | |
l1 = scan(a, v, nil, 20) | |
a['l'].size.times \ | |
{ | |
|j| | |
l2 = scan(a, v, j, 20) | |
x, y, z = [0, 0, 0] | |
[l1, l2].transpose.each \ | |
{ | |
|l12| | |
if (l12[1].nil?) then | |
z += 1 | |
elsif (l12[1] < l12[0]) | |
x += 1 | |
elsif (l12[1] > l12[0]) | |
y += 1 | |
end | |
} | |
puts([j, 0, 2].join("\t")) | |
puts([j, x, 2].join("\t")) | |
puts | |
puts([j, x, 3].join("\t")) | |
puts([j, x + y, 3].join("\t")) | |
puts | |
puts([j, x + y, 1].join("\t")) | |
puts([j, x + y + z, 1].join("\t")) | |
puts | |
$stdout.flush | |
} | |
end | |
srand(0) | |
v = ['a1', 'a0', 'dh', 'dl', 'sd0', 'sd1', 'mx1'] | |
a = model(v) | |
test(a.dup, v) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment