mirror of
https://github.com/saymrwulf/nvidia-AI.summerschool17-notebooks.git
synced 2026-11-20 09:52:58 +00:00
519 lines
66 KiB
Text
519 lines
66 KiB
Text
|
|
{
|
||
|
|
"metadata": {
|
||
|
|
"name": "105-simple-probability"
|
||
|
|
},
|
||
|
|
"nbformat": 3,
|
||
|
|
"nbformat_minor": 0,
|
||
|
|
"worksheets": [
|
||
|
|
{
|
||
|
|
"cells": [
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Title: Sampling from Probability Distributions\n",
|
||
|
|
"Author: Thomas Breuel\n",
|
||
|
|
"Institution: UniKL"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"\n",
|
||
|
|
"from pylab import *"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [],
|
||
|
|
"prompt_number": 1
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "heading",
|
||
|
|
"level": 1,
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Sampling from Distributions"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"We can easily sample for distributions in Python.\n",
|
||
|
|
"The basic distributions are `randn` (normal density) and `rand` (uniform density)."
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"data = randn(10000)\n",
|
||
|
|
"_=hist(data,bins=100)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "display_data",
|
||
|
|
"png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAD9CAYAAABDaefJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nOnbsmObNmydJ2rdvn959911l
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 4
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "heading",
|
||
|
|
"level": 1,
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Binomial Distributions"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"If we want a binomial variable with $p=0.1$, we can sample as follows:\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"n = 1000\n",
|
||
|
|
"p = 0.1\n",
|
||
|
|
"sum(rand(n)<p)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "pyout",
|
||
|
|
"prompt_number": 69,
|
||
|
|
"text": [
|
||
|
|
"109"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 69
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"p = 0.5\n",
|
||
|
|
"dist = array([sum(rand(n)<p) for i in range(10000)])\n",
|
||
|
|
"_=hist(dist,bins=20)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "display_data",
|
||
|
|
"png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAD9CAYAAABazssqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n8MYbb3DppZdyzz33cPDgQU6d
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 70
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"print mean(dist),var(dist)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "stream",
|
||
|
|
"stream": "stdout",
|
||
|
|
"text": [
|
||
|
|
"500.114 246.746804\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 68
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Theoretical values are:\n",
|
||
|
|
"\n",
|
||
|
|
"$$ \\bar{x} = n ~ p $$\n",
|
||
|
|
"$$ v = n ~ p ~(1-p) $$"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"print n*p\n",
|
||
|
|
"print n*p*(1-p)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "stream",
|
||
|
|
"stream": "stdout",
|
||
|
|
"text": [
|
||
|
|
"500.0\n",
|
||
|
|
"250.0\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 71
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "heading",
|
||
|
|
"level": 1,
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Parameter Estimation"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"We often estimate parameters from measurements. Let's do this for a normal density.\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"data = randn(1000)\n",
|
||
|
|
"print mean(data),var(data)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "stream",
|
||
|
|
"stream": "stdout",
|
||
|
|
"text": [
|
||
|
|
"0.0317909909166 0.877621361124\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 86
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Note that our parameter estimates themselves have a distribution of errors.\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"means = [mean(randn(1000)) for i in range(1000)]\n",
|
||
|
|
"_=hist(means,bins=100)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "display_data",
|
||
|
|
"png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAD9CAYAAACoXlzKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMAzBXkDLFRNUTQBwElUxBbRc
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 87
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "heading",
|
||
|
|
"level": 1,
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Extreme Value Distributions"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Another interesting example is extreme value distributions. They matter a great deal in biology and business, where \"the best\" of a population often \"win\" or are measured. They also matter when people tell you statistics like \"90 percent of population X are women/men, therefore there must be bias\".\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"N = 100000\n",
|
||
|
|
"maxs = array([amax(randn(N)) for i in range(1000)])\n",
|
||
|
|
"_=hist(maxs,bins=100)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "display_data",
|
||
|
|
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAD9CAYAAACY0k3rAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEK4MbEYPVKvsy4o5Xf6NcwaV
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 102
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"This kind of bias (around 20 percent of the standard deviation here) leads to a bias in maxima.\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"max1s = array([amax(randn(N)+0.2) for i in range(1000)])\n",
|
||
|
|
"_=hist(maxs,bins=100,alpha=0.7)\n",
|
||
|
|
"_=hist(max1s,bins=100,alpha=0.7)\n",
|
||
|
|
"sum(max1s>maxs)*1.0/len(maxs)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "pyout",
|
||
|
|
"prompt_number": 103,
|
||
|
|
"text": [
|
||
|
|
"0.70599999999999996"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"output_type": "display_data",
|
||
|
|
"png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAD9CAYAAABOd5eOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nC6BZ+LRqnZ9ElPyYzImIBMBk
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 103
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Interestingly, you can get much greater extreme value differences without any bias at all, simply by having more variance in one population than in another.\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"maxs = array([amax(randn(10000)) for i in range(1000)])\n",
|
||
|
|
"max2s = array([amax(1.1*randn(10000)) for i in range(1000)])\n",
|
||
|
|
"_=hist(maxs,bins=100,alpha=0.7)\n",
|
||
|
|
"_=hist(max2s,bins=100,alpha=0.7)\n",
|
||
|
|
"sum(max2s>maxs)*1.0/len(maxs)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "pyout",
|
||
|
|
"prompt_number": 104,
|
||
|
|
"text": [
|
||
|
|
"0.82399999999999995"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"output_type": "display_data",
|
||
|
|
"png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD9CAYAAACsq4z3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nuroabW1tCxEvERFFETOhFxYW
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 104
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "heading",
|
||
|
|
"level": 1,
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Conditional Distributions"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Let's look at another important example from statistics. Consider a population of size $N = 1000000$:\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"N = 1000000"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [],
|
||
|
|
"prompt_number": 105
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Now assume that about 1:1000 of the population are infected with some disease.\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"infected = (rand(N)<1e-3)\n",
|
||
|
|
"sum(infected)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "pyout",
|
||
|
|
"prompt_number": 106,
|
||
|
|
"text": [
|
||
|
|
"1045"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 106
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Also, assume we have a test that is 99 percent accurate. We give the test to the entire population and measure the results.\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"testresult = where(rand(N)<1e-2,1-infected,infected)\n",
|
||
|
|
"sum(testresult)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "pyout",
|
||
|
|
"prompt_number": 107,
|
||
|
|
"text": [
|
||
|
|
"11120"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 107
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Now, let's look at the population of people whose test result is positive.\n",
|
||
|
|
"What fraction of those people are infected?\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [
|
||
|
|
"sum(infected[testresult==True])*1.0/sum(testresult==True)"
|
||
|
|
],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"output_type": "pyout",
|
||
|
|
"prompt_number": 108,
|
||
|
|
"text": [
|
||
|
|
"0.092985611510791363"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"prompt_number": 108
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"(Implications)\n",
|
||
|
|
"\n",
|
||
|
|
"That's not good: less than 10 percent of the population that tests positive is actually infected. Ninety percent of those who test positive are _false positives_ even though the test is 99 percent accurate.\n",
|
||
|
|
"\n",
|
||
|
|
"This is a common issue in medicine, anti-terrorism, and many other problems:\n",
|
||
|
|
"\n",
|
||
|
|
"- testing everybody for HIV infection\n",
|
||
|
|
"- testing everybody for prostate cancer / breast cancer\n",
|
||
|
|
"- widespread Internet surveillance for \"terrorism\"\n",
|
||
|
|
"\n",
|
||
|
|
"All of these end up generating enormous numbers of false positives."
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"(Bayesian Calculation)\n",
|
||
|
|
"\n",
|
||
|
|
"Mathematically, we write this in terms of _conditional probabilities_:\n",
|
||
|
|
"\n",
|
||
|
|
"prior probability: \n",
|
||
|
|
"\n",
|
||
|
|
"$$P(i)$$\n",
|
||
|
|
"\n",
|
||
|
|
"probability of test: \n",
|
||
|
|
"\n",
|
||
|
|
"$$P(t=0|i=0) = 0.99$$\n",
|
||
|
|
"$$P(t=1|i=1) = 0.99$$\n",
|
||
|
|
"\n",
|
||
|
|
"Bayes formula:\n",
|
||
|
|
"\n",
|
||
|
|
"$$ P(i=1|t=1) = \\frac{P(t=1|i=1) P(i=1)}{P(t=1)} = \\frac{P(t=1|i=1) P(i=1)}{P(t=1|i=0) P(i=0) + P(t=1|i=1) P(i=1) } $$\n",
|
||
|
|
"\n",
|
||
|
|
"$$ ~~~~~~ = \\frac{0.99 \\cdot 0.001}{ 0.01 \\cdot 0.999 + 0.99 \\cdot 0.001 } \\approx \\frac{0.001}{0.01} = 0.1 $$\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"(Derivation of Bayes Formula)\n",
|
||
|
|
"\n",
|
||
|
|
"Joint density ($c$ and $x$ could be any variables):\n",
|
||
|
|
"\n",
|
||
|
|
"$$ p(c,x) $$\n",
|
||
|
|
"\n",
|
||
|
|
"Conditional density:\n",
|
||
|
|
"\n",
|
||
|
|
"$$ p(c,x) = P(c|x) ~ p(x) $$\n",
|
||
|
|
"\n",
|
||
|
|
"Bayes formula:\n",
|
||
|
|
"\n",
|
||
|
|
"$$ P(c|x) ~ p(x) ~ = ~ p(x|c) ~ P(c) $$\n",
|
||
|
|
"\n",
|
||
|
|
"$$ P(c|x) = \\frac{p(x|c) P(c)}{p(x)} $$"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"collapsed": false,
|
||
|
|
"input": [],
|
||
|
|
"language": "python",
|
||
|
|
"metadata": {},
|
||
|
|
"outputs": []
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"metadata": {}
|
||
|
|
}
|
||
|
|
]
|
||
|
|
}
|