Related
I have a very large absorbing Markov chain. I want to obtain the fundamental matrix of this chain to calculate the expected number of steps before absortion. From this question I know that this can be calculated by the equation
(I - Q)t=1
which can be obtained by using the following python code:
def expected_steps_fast(Q):
I = numpy.identity(Q.shape[0])
o = numpy.ones(Q.shape[0])
numpy.linalg.solve(I-Q, o)
However, I would like to calculate it using some kind of iterative method similar to the power iteration method used for calculate the PageRank. This method would allow me to calculate an approximation to the expected number of steps before absortion in a mapreduce-like system.
¿Does something similar exist?
If you have a sparse matrix, check if scipy.spare.linalg.spsolve works. No guarantees about numerical robustness, but at least for trivial examples it's significantly faster than solving with dense matrices.
import networkx as nx
import numpy as np
import scipy.sparse as sp
import scipy.sparse.linalg as spla
def example(n):
"""Generate a very simple transition matrix from a directed graph
"""
g = nx.DiGraph()
for i in xrange(n-1):
g.add_edge(i+1, i)
g.add_edge(i, i+1)
g.add_edge(n-1, n)
g.add_edge(n, n)
m = nx.to_numpy_matrix(g)
# normalize rows to ensure m is a valid right stochastic matrix
m = m / np.sum(m, axis=1)
return m
A = sp.csr_matrix(example(2000)[:-1,:-1])
Ad = np.array(A.todense())
def sp_solve(Q):
I = sp.identity(Q.shape[0], format='csr')
o = np.ones(Q.shape[0])
return spla.spsolve(I-Q, o)
def dense_solve(Q):
I = numpy.identity(Q.shape[0])
o = numpy.ones(Q.shape[0])
return numpy.linalg.solve(I-Q, o)
Timings for sparse solution:
%timeit sparse_solve(A)
1000 loops, best of 3: 1.08 ms per loop
Timings for dense solution:
%timeit dense_solve(Ad)
1 loops, best of 3: 216 ms per loop
Like Tobias mentions in the comments, I would have expected other solvers to outperform the generic one, and they may for very large systems. For this toy example, the generic solve seems to work well enough.
I arraived to this answer thanks to #tobias-ribizel's suggestion of using the Neumann series. If we part from the following equation:
Using the Neumann series:
If we multiply each term of the series by the vector 1 we could operate separately over each row of the matrix Q and approximate successively with:
This is the python code I use to calculate this:
def expected_steps_iterative(Q, n=10):
N = Q.shape[0]
acc = np.ones(N)
r_k_1 = np.ones(N)
for k in range(1, n):
r_k = np.zeros(N)
for i in range(N):
for j in range(N):
r_k[i] += r_k_1[j] * Q[i, j]
if np.allclose(acc, acc+r_k, rtol=1e-8):
acc += r_k
break
acc += r_k
r_k_1 = r_k
return acc
And this is the code using Spark. This code expects that Q is a RDD where each row is a tuple (row_id, dict of weights for that row of the matrix).
def expected_steps_spark(sc, Q, n=10):
def dict2np(d, sz):
vec = np.zeros(sz)
for k, v in d.iteritems():
vec[k] = v
return vec
sz = Q.count()
acc = np.ones(sz)
x = {i:1.0 for i in range(sz)}
for k in range(1, n):
bc_x = sc.broadcast(x)
x_old = x
x = Q.map(lambda (u, ol): (u, reduce(lambda s, j: s + bc_x.value[j]*ol[j], ol, 0.0)))
x = x.collectAsMap()
v_old = dict2np(x_old, sz)
v = dict2np(x, sz)
acc += v
if np.allclose(v, v_old, rtol=1e-8):
break
return acc
I have a two-fold homework problem, Implement Karp-Rabin and run it on a test file and the second part:
For the hash values modulo q, explain why it is a bad idea to use q as a power of 2. Can you construct a terrible example e.g. for q=64
and n=15?
This is my implementation of the algorithm:
def karp_rabin(text, pattern):
# setup
alphabet = 'ACGT'
d = len(alphabet)
n = len(pattern)
d_n = d**n
q = 2**32-1
m = {char:i for i,char in enumerate(alphabet)}
positions = []
def kr_hash(s):
return sum(d**(n-i-1) * m[s[i]] for i in range(n))
def update_hash():
return d*text_hash + m[text[i+n-1]] - d_n * m[text[i-1]]
pattern_hash = kr_hash(pattern)
for i in range(0, len(text) - n + 1):
text_hash = update_hash() if i else kr_hash(text[i:n])
if pattern_hash % q == text_hash % q and pattern == text[i:i+n]:
positions.append(i)
return ' '.join(map(str, positions))
...The second part of the question is referring to this part of the code/algo:
pattern_hash = kr_hash(pattern)
for i in range(0, len(text) - n + 1):
text_hash = update_hash() if i else kr_hash(text[i:n])
# the modulo q used to check if the hashes are congruent
if pattern_hash % q == text_hash % q and pattern == text[i:i+n]:
positions.append(i)
I don't understand why it would be a bad idea to use q as a power of 2. I've tried running the algorithm on the test file provided(which is the genome of ecoli) and there's no discernible difference.
I tried looking at the formula for how the hash is derived (I'm not good at math) trying to find some common factors that would be really bad for powers of two but found nothing. I feel like if q is a power of 2 it should cause a lot of clashes for the hashes so you'd need to compare strings a lot more but I didn't find anything along those lines either.
I'd really appreciate help on this since I'm stumped. If someone wants to point out what I can do better in the first part (code efficiency, readability, correctness etc.) I'd also be thrilled to hear your input on that.
There is a problem if q divides some power of d, because then only a few characters contribute to the hash. For example in your code d=4, if you take q=64 only the last three characters determine the hash (d**3 = 64).
I don't really see a problem if q is a power of 2 but gcd(d,q) = 1.
Your implementation looks a bit strange because instead of
if pattern_hash % q == text_hash % q and pattern == text[i:i+n]:
you could also use
if pattern_hash == text_hash and pattern == text[i:i+n]:
which would be better because you get fewer collisions.
The Thue–Morse sequence has among its properties that its polynomial hash quickly becomes zero when a power of 2 is the hash module, for whatever polynomial base (d). So if you will try to search a short Thue-Morse sequence in a longer one, you will have a great lot of hash collisions.
For example, your code, slightly adapted:
def karp_rabin(text, pattern):
# setup
alphabet = '01'
d = 15
n = len(pattern)
d_n = d**n
q = 32
m = {char:i for i,char in enumerate(alphabet)}
positions = []
def kr_hash(s):
return sum(d**(n-i-1) * m[s[i]] for i in range(n))
def update_hash():
return d*text_hash + m[text[i+n-1]] - d_n * m[text[i-1]]
pattern_hash = kr_hash(pattern)
for i in range(0, len(text) - n + 1):
text_hash = update_hash() if i else kr_hash(text[i:n])
if pattern_hash % q == text_hash % q : #and pattern == text[i:i+n]:
positions.append(i)
return ' '.join(map(str, positions))
print(karp_rabin('0110100110010110100101100110100110010110011010010110100110010110', '0110100110010110'))
outputs a lot of positions, although only three of then are proper matches.
Note that I have dropped the and pattern == text[i:i+n] check. Obviously if you restore it, the result will be correct, but also it is obvious that the algorithm will do much more work checking this additional condition than for other q. In fact, because there are so many collisions, the whole idea of algorithm becomes not working: you could almost as effectively wrote a simple algorithm that checks every position for a match.
Also note that your implementation is quite strange. The whole idea of polynomial hashing is to take the modulo operation each time you compute the hash. Otherwise your pattern_hash and text_hash are very big numbers. In other languages this might mean arithmetic overflow, but in Python this will invoke big integer arithmetic, which is slow and once again loses the whole idea of the algorithm.
I want to try a svm classifier using histogram intersection kernel, for a dataset of 153 images but it takes a long time. This is my code:
a = load('...'); %vectors
b = load('...'); %labels
g = dataset(a,b);
error = crossval(g,libsvc([],proxm([],'ih'),100),10,10);
error1 = crossval(g,libsvc([],proxm([],'ih'),10),10,10);
error2 = crossval(g,libsvc([],proxm([],'ih'),1),10,10);
My implementation of the kernel within the proxm function is:
...
case {'dist_histint','ih'}
[m,d]=size(A);
[n,d1]=size(B);
if (d ~= d1)
error('column length of A (%d) != column length of B (%d)\n',d,d1);
end
% With the MATLAB JIT compiler the trivial implementation turns out
% to be the fastest, especially for large matrices.
D = zeros(m,n);
for i=1:m % m is number of samples of A
if (0==mod(i,1000)) fprintf('.'); end
for j=1:n % n is number of samples of B
D(i,j) = sum(min([A(i,:);B(j,:)]));%./max(A(:,i),B(:,j)));
end
end
I need some matlab optimization for this code!
You can get rid of that kernel loop to calculate D with this bsxfun based vectorized approach -
D = squeeze(sum(bsxfun(#min,A,permute(B,[3 2 1])),2))
Or avoid squeeze with this modification -
D = sum(bsxfun(#min,permute(A,[1 3 2]),permute(B,[3 1 2])),3)
If the calculations of D involve max instead of min, just replace #min with #max there.
Explanation: The way bsxfun works is that it does expansion on singleton dimensions and performs the operation as listed with # inside its call. Now, this expansion is basically how one achieves vectorized solutions that replace for-loops. By singleton dimensions in arrays, we mean dimensions of 1 in them.
In many cases, singleton dimensions aren't already present and for vectorization with bsxfun, we need to create singleton dimensions. One of the tools to do so is with permute. That's basically all about the way vectorized approach stated earlier would work.
Thus, your kernel code -
...
case {'dist_histint','ih'}
[m,d]=size(A);
[n,d1]=size(B);
if (d ~= d1)
error('column length of A (%d) != column length of B (%d)\n',d,d1);
end
% With the MATLAB JIT compiler the trivial implementation turns out
% to be the fastest, especially for large matrices.
D = zeros(m,n);
for i=1:m % m is number of samples of A
if (0==mod(i,1000)) fprintf('.'); end
for j=1:n % n is number of samples of B
D(i,j) = sum(min([A(i,:);B(j,:)]));%./max(A(:,i),B(:,j)));
end
end
reduces to -
...
case {'dist_histint','ih'}
[m,d]=size(A);
[n,d1]=size(B);
if (d ~= d1)
error('column length of A (%d) != column length of B (%d)\n',d,d1);
end
D = squeeze(sum(bsxfun(#min,A,permute(B,[3 2 1])),2))
%// OR D = sum(bsxfun(#min,permute(A,[1 3 2]),permute(B,[3 1 2])),3)
I am assuming the line: if (0==mod(i,1000)) fprintf('.'); end isn't important to the calculations as it does printing of some message.
I'm working on a project for fun and I need an algorithm to do as follows:
Generate a list of numbers of Length n which add up to x
I would settle for list of integers, but ideally, I would like to be left with a set of floating point numbers.
I would be very surprised if this problem wasn't heavily studied, but I'm not sure what to look for.
I've tackled similar problems in the past, but this one is decidedly different in nature. Before I've generated different combinations of a list of numbers that will add up to x. I'm sure that I could simply bruteforce this problem but that hardly seems like the ideal solution.
Anyone have any idea what this may be called, or how to approach it? Thanks all!
Edit: To clarify, I mean that the list should be length N while the numbers themselves can be of any size.
edit2: Sorry for my improper use of 'set', I was using it as a catch all term for a list or an array. I understand that it was causing confusion, my apologies.
This is how to do it in Python
import random
def random_values_with_prescribed_sum(n, total):
x = [random.random() for i in range(n)]
k = total / sum(x)
return [v * k for v in x]
Basically you pick n random numbers, compute their sum and compute a scale factor so that the sum will be what you want it to be.
Note that this approach will not produce "uniform" slices, i.e. the distribution you will get will tend to be more "egalitarian" than it should be if it was picked at random among all distribution with the given sum.
To see the reason you can just picture what the algorithm does in the case of two numbers with a prescribed sum (e.g. 1):
The point P is a generic point obtained by picking two random numbers and it will be uniform inside the square [0,1]x[0,1]. The point Q is the point obtained by scaling P so that the sum is required to be 1. As it's clear from the picture the points close to the center of the have an higher probability; for example the exact center of the squares will be found by projecting any point on the diagonal (0,0)-(1,1), while the point (0, 1) will be found projecting only points from (0,0)-(0,1)... the diagonal length is sqrt(2)=1.4142... while the square side is only 1.0.
Actually, you need to generate a partition of x into n parts. This is usually done the in following way: The partition of x into n non-negative parts can be represented in the following way: reserve n + x free places, put n borders to some arbitrary places, and stones to the rest. The stone groups add up to x, thus the number of possible partitions is the binomial coefficient (n + x \atop n).
So your algorithm could be as follows: choose an arbitrary n-subset of (n + x)-set, it determines uniquely a partition of x into n parts.
In Knuth's TAOCP the chapter 3.4.2 discusses random sampling. See Algortihm S there.
Algorithm S: (choose n arbitrary records from total of N)
t = 0, m = 0;
u = random, uniformly distributed on (0, 1)
if (N - t)*u >= n - m, skip t-th record and increase t by 1; otherwise include t-th record in the sample, increase m and t by 1
if M < n, return to 2, otherwise, algorithm finished
The solution for non-integers is algorithmically trivial: you just select arbitrary n numbers that don't sum up to 0, and norm them by their sum.
If you want to sample uniformly in the region of N-1-dimensional space defined by x1 + x2 + ... + xN = x, then you're looking at a special case of sampling from a Dirichlet distribution. The sampling procedure is a little more involved than generating uniform deviates for the xi. Here's one way to do it, in Python:
xs = [random.gammavariate(1,1) for a in range(N)]
xs = [x*v/sum(xs) for v in xs]
If you don't care too much about the sampling properties of your results, you can just generate uniform deviates and correct their sum afterwards.
Here is a version of the above algorithm in Javascript
function getRandomArbitrary(min, max) {
return Math.random() * (max - min) + min;
};
function getRandomArray(min, max, n) {
var arr = [];
for (var i = 0, l = n; i < l; i++) {
arr.push(getRandomArbitrary(min, max))
};
return arr;
};
function randomValuesPrescribedSum(min, max, n, total) {
var arr = getRandomArray(min, max, n);
var sum = arr.reduce(function(pv, cv) { return pv + cv; }, 0);
var k = total/sum;
var delays = arr.map(function(x) { return k*x; })
return delays;
};
You can call it with
var myarray = randomValuesPrescribedSum(0,1,3,3);
And then check it with
var sum = myarray.reduce(function(pv, cv) { return pv + cv;},0);
This code does a reasonable job. I think it produces a different distribution than 6502's answer, but I am not sure which is better or more natural. Certainly his code is clearer/nicer.
import random
def parts(total_sum, num_parts):
points = [random.random() for i in range(num_parts-1)]
points.append(0)
points.append(1)
points.sort()
ret = []
for i in range(1, len(points)):
ret.append((points[i] - points[i-1]) * total_sum)
return ret
def test(total_sum, num_parts):
ans = parts(total_sum, num_parts)
assert abs(sum(ans) - total_sum) < 1e-7
print ans
test(5.5, 3)
test(10, 1)
test(10, 5)
In python:
a: create a list of (random #'s 0 to 1) times total; append 0 and total to the list
b: sort the list, measure the distance between each element
c: round the list elements
import random
import time
TOTAL = 15
PARTS = 4
PLACES = 3
def random_sum_split(parts, total, places):
a = [0, total] + [random.random()*total for i in range(parts-1)]
a.sort()
b = [(a[i] - a[i-1]) for i in range(1, (parts+1))]
if places == None:
return b
else:
b.pop()
c = [round(x, places) for x in b]
c.append(round(total-sum(c), places))
return c
def tick():
if info.tick == 1:
start = time.time()
alpha = random_sum_split(PARTS, TOTAL, PLACES)
end = time.time()
log('alpha: %s' % alpha)
log('total: %.7f' % sum(alpha))
log('parts: %s' % PARTS)
log('places: %s' % PLACES)
log('elapsed: %.7f' % (end-start))
yields:
[2014-06-13 01:00:00] alpha: [0.154, 3.617, 6.075, 5.154]
[2014-06-13 01:00:00] total: 15.0000000
[2014-06-13 01:00:00] parts: 4
[2014-06-13 01:00:00] places: 3
[2014-06-13 01:00:00] elapsed: 0.0005839
to the best of my knowledge this distribution is uniform
Well, I have this bit of code that is slowing down the program hugely because it is linear complexity but called a lot of times making the program quadratic complexity. If possible I would like to reduce its computational complexity but otherwise I'll just optimize it where I can. So far I have reduced down to:
def table(n):
a = 1
while 2*a <= n:
if (-a*a)%n == 1: return a
a += 1
Anyone see anything I've missed? Thanks!
EDIT: I forgot to mention: n is always a prime number.
EDIT 2: Here is my new improved program (thank's for all the contributions!):
def table(n):
if n == 2: return 1
if n%4 != 1: return
a1 = n-1
for a in range(1, n//2+1):
if (a*a)%n == a1: return a
EDIT 3: And testing it out in its real context it is much faster! Well this question appears solved but there are many useful answers. I should also say that as well as those above optimizations, I have memoized the function using Python dictionaries...
Ignoring the algorithm for a moment (yes, I know, bad idea), the running time of this can be decreased hugely just by switching from while to for.
for a in range(1, n / 2 + 1)
(Hope this doesn't have an off-by-one error. I'm prone to make these.)
Another thing that I would try is to look if the step width can be incremented.
Take a look at http://modular.fas.harvard.edu/ent/ent_py .
The function sqrtmod does the job if you set a = -1 and p = n.
You missed a small point because the running time of your improved algorithm is still in the order of the square root of n. As long you have only small primes n (let's say less than 2^64), that's ok, and you should probably prefer your implementation to a more complex one.
If the prime n becomes bigger, you might have to switch to an algorithm using a little bit of number theory. To my knowledge, your problem can be solved only with a probabilistic algorithm in time log(n)^3. If I remember correctly, assuming the Riemann hypothesis holds (which most people do), one can show that the running time of the following algorithm (in ruby - sorry, I don't know python) is log(log(n))*log(n)^3:
class Integer
# calculate b to the power of e modulo self
def power(b, e)
raise 'power only defined for integer base' unless b.is_a? Integer
raise 'power only defined for integer exponent' unless e.is_a? Integer
raise 'power is implemented only for positive exponent' if e < 0
return 1 if e.zero?
x = power(b, e>>1)
x *= x
(e & 1).zero? ? x % self : (x*b) % self
end
# Fermat test (probabilistic prime number test)
def prime?(b = 2)
raise "base must be at least 2 in prime?" if b < 2
raise "base must be an integer in prime?" unless b.is_a? Integer
power(b, self >> 1) == 1
end
# find square root of -1 modulo prime
def sqrt_of_minus_one
return 1 if self == 2
return false if (self & 3) != 1
raise 'sqrt_of_minus_one works only for primes' unless prime?
# now just try all numbers (each succeeds with probability 1/2)
2.upto(self) do |b|
e = self >> 1
e >>= 1 while (e & 1).zero?
x = power(b, e)
next if [1, self-1].include? x
loop do
y = (x*x) % self
return x if y == self-1
raise 'sqrt_of_minus_one works only for primes' if y == 1
x = y
end
end
end
end
# find a prime
p = loop do
x = rand(1<<512)
next if (x & 3) != 1
break x if x.prime?
end
puts "%x" % p
puts "%x" % p.sqrt_of_minus_one
The slow part is now finding the prime (which takes approx. log(n)^4 integer operation); finding the square root of -1 takes for 512-bit primes still less than a second.
Consider pre-computing the results and storing them in a file. Nowadays many platforms have a huge disk capacity. Then, obtaining the result will be an O(1) operation.
(Building on Adam's answer.)
Look at the Wikipedia page on quadratic reciprocity:
x^2 ≡ −1 (mod p) is solvable if and only if p ≡ 1 (mod 4).
Then you can avoid the search of a root precisely for those odd prime n's that are not congruent with 1 modulo 4:
def table(n):
if n == 2: return 1
if n%4 != 1: return None # or raise exception
...
Based off OP's second edit:
def table(n):
if n == 2: return 1
if n%4 != 1: return
mod = 0
a1 = n - 1
for a in xrange(1, a1, 2):
mod += a
while mod >= n: mod -= n
if mod == a1: return a//2 + 1
It looks like you're trying to find the square root of -1 modulo n. Unfortunately, this is not an easy problem, depending on what values of n are input into your function. Depending on n, there might not even be a solution. See Wikipedia for more information on this problem.
Edit 2: Surprisingly, strength-reducing the squaring reduces the time a lot, at least on my Python2.5 installation. (I'm surprised because I thought interpreter overhead was taking most of the time, and this doesn't reduce the count of operations in the inner loop.) Reduces the time from 0.572s to 0.146s for table(1234577).
def table(n):
n1 = n - 1
square = 0
for delta in xrange(1, n, 2):
square += delta
if n <= square: square -= n
if square == n1: return delta // 2 + 1
strager posted the same idea but I think less tightly coded. Again, jug's answer is best.
Original answer: Another trivial coding tweak on top of Konrad Rudolph's:
def table(n):
n1 = n - 1
for a in xrange(1, n // 2 + 1):
if (a*a) % n == n1: return a
Speeds it up measurably on my laptop. (About 25% for table(1234577).)
Edit: I didn't notice the python3.0 tag; but the main change was hoisting part of the calculation out of the loop, not the use of xrange. (Academic since there's a better algorithm.)
Is it possible for you to cache the results?
When you calculate a large n you are given the results for the lower n's almost for free.
One thing that you are doing is repeating the calculation -a*a over and over again.
Create a table of the values once and then do look up in the main loop.
Also although this probably doesn't apply to you because your function name is table but if you call a function that takes time to calculate you should cache the result in a table and just do a table look up if you call it again with the same value. This save you the time of calculating all of the values when you first run but you don't waste time repeating the calculation more than once.
I went through and fixed the Harvard version to make it work with python 3.
http://modular.fas.harvard.edu/ent/ent_py
I made some slight changes to make the results exactly the same as the OP's function. There are two possible answers and I forced it to return the smaller answer.
import timeit
def table(n):
if n == 2: return 1
if n%4 != 1: return
a1=n-1
def inversemod(a, p):
x, y = xgcd(a, p)
return x%p
def xgcd(a, b):
x_sign = 1
if a < 0: a = -a; x_sign = -1
x = 1; y = 0; r = 0; s = 1
while b != 0:
(c, q) = (a%b, a//b)
(a, b, r, s, x, y) = (b, c, x-q*r, y-q*s, r, s)
return (x*x_sign, y)
def mul(x, y):
return ((x[0]*y[0]+a1*y[1]*x[1])%n,(x[0]*y[1]+x[1]*y[0])%n)
def pow(x, nn):
ans = (1,0)
xpow = x
while nn != 0:
if nn%2 != 0:
ans = mul(ans, xpow)
xpow = mul(xpow, xpow)
nn >>= 1
return ans
for z in range(2,n) :
u, v = pow((1,z), a1//2)
if v != 0:
vinv = inversemod(v, n)
if (vinv*vinv)%n == a1:
vinv %= n
if vinv <= n//2:
return vinv
else:
return n-vinv
tt=0
pri = [ 5,13,17,29,37,41,53,61,73,89,97,1234577,5915587277,3267000013,3628273133,2860486313,5463458053,3367900313 ]
for x in pri:
t=timeit.Timer('q=table('+str(x)+')','from __main__ import table')
tt +=t.timeit(number=100)
print("table(",x,")=",table(x))
print('total time=',tt/100)
This version takes about 3ms to run through the test cases above.
For comparison using the prime number 1234577
OP Edit2 745ms
The accepted answer 522ms
The above function 0.2ms