{ "cells": [ { "cell_type": "markdown", "id": "20fce00f", "metadata": {}, "source": [ "### Data Analysis\n", "#### Bayes' rule\n", "
\n", "https://github.com/ostad-ai/Machine-Learning\n", "
Explanation: https://www.pinterest.com/HamedShahHosseini/Machine-Learning/background-knowledge" ] }, { "cell_type": "markdown", "id": "9b279ce8", "metadata": {}, "source": [ "The **Bayes' rule** for two events $A$ and $B$:

\n", "$P(A|B)=\\frac{P(B|A) \\cdot P(A)}{P(B)}, P(B) \\neq 0$\n", "

For our example, we measure the probability of a person to be infected if the test is positive
\n", "
$P(Infected|PositiveTest)=\\frac{P(PositiveTest|Infected) \\cdot P(Infected)}{P(PositiveTest)}, P(PositiveTest) \\neq 0$
\n", "where $P(PositiveTest)=P(PositiveTest|Infected) \\cdot P(Infected)+P(PositiveTest|not Infected) \\cdot P(not Infected)$" ] }, { "cell_type": "code", "execution_count": 1, "id": "20acd6ae", "metadata": {}, "outputs": [], "source": [ "def prob_infected_if_test_positive(p_true_positive,p_infected,p_false_positive):\n", " numerator=p_true_positive*p_infected\n", " denominator=numerator+p_false_positive*(1-p_infected)\n", " if denominator!=0:\n", " return numerator/denominator\n", " else:\n", " return None" ] }, { "cell_type": "code", "execution_count": 2, "id": "74ae0061", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Probability of person being infected if the test is positive: 0.08411214953271028\n" ] } ], "source": [ "p_true_positive=0.9 # if the person is infected, the test is positive for 90% of cases\n", "p_infected=0.02 # the fraction of people in the population to be infected\n", "p_false_positive=0.2 # if the person is not infected, the test is positive for 20% of cases\n", "# compute the probability of being infected if the test is positive\n", "result=prob_infected_if_test_positive(p_true_positive,p_infected,p_false_positive)\n", "print(f'Probability of person being infected if the test is positive: {result}')" ] }, { "cell_type": "markdown", "id": "4276ec61", "metadata": {}, "source": [ "**Law of total probability:**

\n", "$P(B)=\\sum_i P(B|A_i)P(A_i)$\n", "

**Example:** We have three vases.\n", "- vase one has 3 red balls and 6 blue balls.\n", "- vase two has 2 red balls and 8 blue balls\n", "- vase three has 6 red balls and 3 blue balls\n", "\n", "We may choose each vase with the equal probability. Then, what is the probability of drawing a red ball from vases?" ] }, { "cell_type": "code", "execution_count": 3, "id": "3c8266f5", "metadata": {}, "outputs": [], "source": [ "def law_of_total_prob(p_vases=[1/3,1/3,1/3],reds=[3,2,6],blues=[6,8,3]):\n", " p_reds=[]; p_drawn_red=0; p_drawn_blue=0\n", " for red,blue in zip(reds,blues):\n", " p_reds.append(red/(red+blue))\n", " for p_red,p_vase in zip(p_reds,p_vases):\n", " p_drawn_red+=p_red*p_vase\n", " p_drawn_blue+=(1-p_red)*p_vase\n", " return p_drawn_red,p_drawn_blue" ] }, { "cell_type": "code", "execution_count": 4, "id": "06446a2d", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Probability of drawing a red ball: 0.39999999999999997\n" ] } ], "source": [ "print(f'Probability of drawing a red ball: {law_of_total_prob()[0]}')" ] }, { "cell_type": "code", "execution_count": null, "id": "4ae0ff9e", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.1" } }, "nbformat": 4, "nbformat_minor": 5 }