chris-taylor

%% Read in files and extrapolate mortality rate

x = csv.read('life.csv','format', '%f %f');

age  = x{1};
prob = x{2};

model = stats.lm(log(prob(61:end)), log(age(61:100)));
pfun  = @(a) util.if_(a<60, @()prob(a), util.if_(a>121, 1, exp(model.beta(1) + model.beta(2) * log(a))));

chris-taylor

data IOAction a = Return a
                | Put String (IOAction a)
                | Get (String -> IOAction a)

get   = Get Return
put s = Put s (Return ())

seqio :: IOAction a -> (a -> IOAction b) -> IOAction b
seqio (Return a) f = f a
seqio (Put s io) f = Put s (seqio io f)

chris-taylor

classdef Dual
% DUAL Dual numbers class for automatic differentiation.
%
% To do automatic differentiation we introduce a number d such that d^2 == 0.
% Then by a simple application of Taylor's theorem we have
%
%     f(x + d) = f(x) + f'(x) d
%
% with all higher-order terms vanishing. Therefore applying a function to a dual
% number x + d generates both the value f(x) *and* the derivative f'(x).

chris-taylor

module AI.Search.Examples.Graph where

import           Control.Monad
import           Control.Monad.ST
import           Control.Applicative
import           Data.STRef
import           Data.Map (Map, (!))
import qualified Data.Map                                 as Map
import           Data.List (nub)
import           Data.Graph.Inductive (LNode, LEdge, Gr)

chris-taylor

function search(query,path)
% Download full size images from Google image search. Saves image files
% locally. Do not print or republish images without permission.
%
% Based on Python code by Craig Quiter at https://gist.github.com/crizCraig/2816295
%
% Requires the JSONLab package, available from
% http://www.mathworks.com/matlabcentral/fileexchange/33381-jsonlab-a-toolbox-to-encodedecode-json-files-in-matlaboctave
%
% USAGE

chris-taylor

sortBy c=m.map(:[]) -- O(n log n)
 where m[]=[];m[x]=x;m x=m(n x);n[]=[];n[x]=[x];n(x:y:z)=p x y:n z
       p[]y=y;p x[]=x;p (w:x)(y:z)=case c w y of GT->y:p(w:x)z;_->w:p x(y:z)
 
sort x=m(map(:[])x)
 where m[]=[];m[x]=x;m x=m(n x);n[]=[];n[x]=[x];n(x:y:z)=p x y:n z
       p[]y=y;p x[]=x;p (w:x)(y:z)=if w>y then y:p(w:x)z else w:p x(y:z)

chris-taylor

import           Control.Applicative
import qualified Data.Set               as Set
import           AI.Search.Uninformed

step :: Set.Set String -> String -> [] String
step wordlist xs = Set.toList . Set.fromList . filter f $ remove1 xs ++ swap1 xs ++ add1 xs
 where
  f w = Set.member w wordlist
  len = length xs
  letters = ['a'..'z']

chris-taylor

import Control.Probability

expected p = expectation (runProb p)

die = uniform [1..6] :: Bayes Double Integer

game n
  | n == 0    = die
  | otherwise = do
    x <- die

chris-taylor

module FieldExtension where

{-

This module lets you define the exact Fibonacci function using the mathematically
elegant definition in terms of phi = (1 + sqrt 5) / 2, without losing precision
due to calculations being carried out in floating point.

phi :: Q5
phi = 0.5 :+ 0.5

chris-taylor

Heading 1

Some Q code

fib: {[n]
  $[n<2; n; fib[n-1]+fib[n-2]]
}