{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "### Exercise 5-1: PCA" ] }, { "cell_type": "code", "execution_count": 60, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[-3, -2, -1, 0, 1, 2, -2, -1, 0, 1, 2, -2, -1, 0, 1, 2, 3],\n", " [-2, -1, 0, 1, 2, 3, -2, -1, 0, 1, 2, -3, -2, -1, 0, 1, 2]])" ] }, "execution_count": 60, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import numpy as np\n", "x = np.array( [ (-3,-2), (-2,-1), (-1,0), (0,1), (1,2), (2,3), (-2,-2), (-1,-1), (0,0), (1,1), (2,2), (-2,-3), (-1,-2), (0,-1), (1,0), (2,1), (3,2)]).T\n", "x" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "*(a)* Compute the covariance matrix M." ] }, { "cell_type": "code", "execution_count": 61, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[ 3. , 2.625],\n", " [ 2.625, 3. ]])" ] }, "execution_count": 61, "metadata": {}, "output_type": "execute_result" } ], "source": [ "cov_m = np.cov(x, rowvar=True)\n", "cov_m" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "*(b)* Compute the eigenvalues and eigenvectors of M ." ] }, { "cell_type": "code", "execution_count": 62, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[ 0.375 5.625]\n", "[[-0.70710678 0.70710678]\n", " [ 0.70710678 0.70710678]]\n" ] } ], "source": [ "eigenvalues, normalized_eigenvectors = np.linalg.eigh(cov_m)\n", "print(eigenvalues)\n", "print(normalized_eigenvectors)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "*(c)* Find the smallest eigenvalue and find the related eigenvector as well. The resulted eigenvector builds the\n", "basis for the new subspace.\n", "Note: Shouldn't we pick the biggest eigenvalue(s) for PCA?" ] }, { "cell_type": "code", "execution_count": 63, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0" ] }, "execution_count": 63, "metadata": {}, "output_type": "execute_result" } ], "source": [ "min_eigenvalue_index = np.argmin(eigenvalues)\n", "min_eigenvalue_index" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "*(d)* Transform vectors of X in this new subspace. $y = W^T \\times x$" ] }, { "cell_type": "code", "execution_count": 64, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([-0.70710678, 0.70710678])" ] }, "execution_count": 64, "metadata": {}, "output_type": "execute_result" } ], "source": [ "W = normalized_eigenvectors[min_eigenvalue_index]\n", "W" ] }, { "cell_type": "code", "execution_count": 65, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([ 0.70710678, 0.70710678, 0.70710678, 0.70710678, 0.70710678,\n", " 0.70710678, 0. , 0. , 0. , 0. ,\n", " 0. , -0.70710678, -0.70710678, -0.70710678, -0.70710678,\n", " -0.70710678, -0.70710678])" ] }, "execution_count": 65, "metadata": {}, "output_type": "execute_result" } ], "source": [ "W.T.dot(x)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 @ /development/datamining", "language": "python", "name": "datamining" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.5.2" } }, "nbformat": 4, "nbformat_minor": 2 }