xboard

# This is a very simple Python 2.7 implementation of the Information Set Monte Carlo Tree Search algorithm.
# The function ISMCTS(rootstate, itermax, verbose = False) is towards the bottom of the code.
# It aims to have the clearest and simplest possible code, and for the sake of clarity, the code
# is orders of magnitude less efficient than it could be made, particularly by using a 
# state.GetRandomMove() or state.DoRandomRollout() function.
# 
# An example GameState classes for Knockout Whist is included to give some idea of how you
# can write your own GameState to use ISMCTS in your hidden information game.
# 
# Written by Peter Cowling, Edward Powley, Daniel Whitehouse (University of York, UK) September 2012 - August 2013.

xboard

"""
Scanner: match text to generate tokens.
Adam Blinkinsop <blinks@acm.org>

First, construct a scanner with the tokens you'd like to match described as
keyword arguments, using Python-syntax regular expressions.
WARNING: Group syntax in these expressions has an undefined effect.

>>> simple = Scan(ID=r'\w+')

xboard

"""
NMF by coordinate descent, designed for sparse data (without missing values)
"""

# Author: Mathieu Blondel <mathieu@mblondel.org>
# License: BSD 3 clause

import numpy as np
import scipy.sparse as sp
import numba

xboard

"""Kernel K-means"""

# Author: Mathieu Blondel <mathieu@mblondel.org>
# License: BSD 3 clause

import numpy as np

from sklearn.base import BaseEstimator, ClusterMixin
from sklearn.metrics.pairwise import pairwise_kernels
from sklearn.utils import check_random_state

xboard

import scipy as sp
from scipy import optimize as opt
def nnlr(X, y, C):
    """
    Non-negative Logistic Regression with L2 regularizer
    """
    def lr_cost(X, y, theta, C):
        m = len(y)
        return (1./m) * (sp.dot(-y, sp.log(sigmoid(sp.dot(X, theta)))) \

xboard

## Corey Chivers, 2012 ##
sim_bayes<-function(p=0.5,N=100,y_lim=20,a_a=2,a_b=10,b_a=8,b_b=3)
{
   ## Simulate outcomes in advance   
   outcomes<-sample(1:0,N,prob=c(p,1-p),replace=TRUE)
   success<-cumsum(outcomes)

   for(frame in 1:N)
   {
      png(paste("plots/",1000+frame,".png",sep=""))