{ "metadata": { "name": "", "signature": "sha256:bd1be64a23441d4cf5d632a7270fc48c6c34072fa890c26e4a21fb0d20afc7b6" }, "nbformat": 3, "nbformat_minor": 0, "worksheets": [ { "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "![jpeg](../galleries/kmeans/2.jpg)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Clustering is one example of Unsupervised Learning. Unsupervised Learning is something that we want to learn from the datasets, eventhough we don't know what to label this particular data. This particular group in the plot,named cluster, is something that we want to learn this lesson. Dimensionality Reduction (PCA) also one of the Unsupervised Learning algorithm that I will be discussed in the future blog.\n", "" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In this example, suppose we plot the movies of what Sebastian and Katie like or don't like. As it turns out, they have different taste of movies, hence different type of category. For some company, like Netflix, they want to recommend movies that users like. For that to happen, they want to see which movies the user has been favorited. Then recommend some other movies that fit in to the category. For more information, please check my other [blog post](http://napitupulu-jon.appspot.com/posts/Unsupervised-Learning--Introduction.html)." ] }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "K-Means" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "![jpeg](../galleries/kmeans/3.jpg)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "* Introduced K-Means, one of the most famous algorithm of Unsupervised Learning.\n", "* First we randomly place two (or any other number, definitely less than the number of data point, and in this example, intuitively we pick number two) cluster center.\n", "* Next, we're assigning all the data points to whichever cluster closest to them. \n", "* Then we move into optimize step, where the cluster move to the position closest to all weighted average of distance data points assigned to each cluster.\n", "* Then we're keep iterating assign and optimize until the cluster is no longer move." ] }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "K-Means Limitation" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "![jpeg](../galleries/kmeans/4.jpg)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The initial position for each of the cluster is random (or we can choose). But the initial position can be greatly decided the outcome of the clustering. It can be as we wanted or it doesn't. Some said that that point, in order for not to be at all random, directly assigned of the exact position of the data." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To solve for this, you can make n-times doing clustering for your data. And pick the cluster by vote." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For more intuition and algorithm for doing this, check my other [blog post](http://napitupulu-jon.appspot.com/posts/K-means-algorithm.html).\n", "\n", "To play around in the visualization of K-Means, check this [link](http://www.naftaliharris.com/blog/visualizing-k-means-clustering/)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For setting parameter in sklearn, parameters that we have discussed so far, check [sklearn documentation](http://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html)." ] }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "Mini Project" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As usual, because this blog post are the note that I have taken from Udacity course, you can see the link of the course for this note at the bottom of the page. Here I attack some of the problem they have at their mini project." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "n this project, we\u2019ll apply k-means clustering to our Enron financial data. Our final goal, of course, is to identify persons of interest; since we have labeled data, this is not a question that particularly calls for an unsupervised approach like k-means clustering.\n", "\n", "Nonetheless, you\u2019ll get some hands-on practice with k-means in this project, and play around with feature scaling, which will give you a sneak preview of the next lesson\u2019s material." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The starter code can be found in k_means/k_means_cluster.py, which reads in the email + financial (E+F) dataset and gets us ready for clustering. You\u2019ll start with performing k-means based on just two financial features--take a look at the code, and determine which features the code uses for clustering.\n", "\n", "Run the code, which will create a scatterplot of the data. Think a little bit about what clusters you would expect to arise if 2 clusters are created." ] }, { "cell_type": "code", "collapsed": false, "input": [ "%load k_means_cluster.py" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 2 }, { "cell_type": "code", "collapsed": false, "input": [ "%pylab inline" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "Populating the interactive namespace from numpy and matplotlib\n" ] }, { "output_type": "stream", "stream": "stderr", "text": [ "WARNING: pylab import has clobbered these variables: ['clf']\n", "`%matplotlib` prevents importing * from pylab and numpy\n" ] } ], "prompt_number": 4 }, { "cell_type": "code", "collapsed": false, "input": [ "#!/usr/bin/python \n", "\n", "\"\"\" \n", " skeleton code for k-means clustering mini-project\n", "\n", "\"\"\"\n", "\n", "\n", "\n", "\n", "import pickle\n", "import numpy\n", "import matplotlib.pyplot as plt\n", "import sys\n", "sys.path.append(\"../tools/\")\n", "from feature_format import featureFormat, targetFeatureSplit\n", "\n", "\n", "\n", "\n", "def Draw(pred, features, poi, mark_poi=False, name=\"image.png\", f1_name=\"feature 1\", f2_name=\"feature 2\"):\n", " \"\"\" some plotting code designed to help you visualize your clusters \"\"\"\n", "\n", " ### plot each cluster with a different color--add more colors for\n", " ### drawing more than 4 clusters\n", " colors = [\"b\", \"c\", \"k\", \"m\", \"g\"]\n", " for ii, pp in enumerate(pred):\n", " plt.scatter(features[ii][0], features[ii][1], color = colors[pred[ii]])\n", "\n", " ### if you like, place red stars over points that are POIs (just for funsies)\n", " if mark_poi:\n", " for ii, pp in enumerate(pred):\n", " if poi[ii]:\n", " plt.scatter(features[ii][0], features[ii][1], color=\"r\", marker=\"*\")\n", " plt.xlabel(f1_name)\n", " plt.ylabel(f2_name)\n", " plt.savefig(name)\n", " plt.show()\n", "\n", "\n", "\n", "### load in the dict of dicts containing all the data on each person in the dataset\n", "data_dict = pickle.load( open(\"../final_project/final_project_dataset.pkl\", \"r\") )\n", "### there's an outlier--remove it! \n", "data_dict.pop(\"TOTAL\", 0)\n", "\n", "\n", "### the input features we want to use \n", "### can be any key in the person-level dictionary (salary, director_fees, etc.) \n", "feature_1 = \"salary\"\n", "feature_2 = \"exercised_stock_options\"\n", "poi = \"poi\"\n", "features_list = [poi, feature_1, feature_2]\n", "data = featureFormat(data_dict, features_list )\n", "poi, finance_features = targetFeatureSplit( data )\n", "\n", "\n", "### in the \"clustering with 3 features\" part of the mini-project,\n", "### you'll want to change this line to \n", "### for f1, f2, _ in finance_features:\n", "### (as it's currently written, line below assumes 2 features)\n", "for f1, f2 in finance_features:\n", " plt.scatter( f1, f2 )\n", "plt.show()\n", "\n", "\n", "\n", "from sklearn.cluster import KMeans\n", "features_list = [\"poi\", feature_1, feature_2]\n", "data2 = featureFormat(data_dict, features_list )\n", "poi, finance_features = targetFeatureSplit( data2 )\n", "clf = KMeans(n_clusters=2)\n", "pred = clf.fit_predict( finance_features )\n", "Draw(pred, finance_features, poi, name=\"clusters_before_scaling.pdf\", f1_name=feature_1, f2_name=feature_2)\n", "\n", "\n", "### cluster here; create predictions of the cluster labels\n", "### for the data and store them to a list called pred\n", "\n", "try:\n", " Draw(pred, finance_features, poi, mark_poi=False, name=\"clusters.pdf\", f1_name=feature_1, f2_name=feature_2)\n", "except NameError:\n", " print \"no predictions object named pred found, no clusters to plot\"\n", "\n", "\n", "\n", "\n", "\n" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEGCAYAAAB2EqL0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIiIiIiIiIiIiIiIiIiIiIiIi8v8BtuLCbDxg5EQAAAAASUVORK5CYII=\n", "text": [ "" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAAEVCAYAAADU/lMpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text": [ "" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAAEVCAYAAADU/lMpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text": [ "" ] } ], "prompt_number": 5 }, { "cell_type": "markdown", "metadata": {}, "source": [ "Deploy k-means clustering on the financial_features data, with 2 clusters specified as a parameter. Store your cluster predictions to a list called pred, so that the Draw() command at the bottom of the script works properly. In the scatterplot that pops up, are the clusters what you expected?" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Add a third feature to features_list, \u201ctotal_payments\". Now rerun clustering, using 3 input features instead of 2 (obviously we can still only visualize the original 2 dimensions). Do any points switch clusters? How many? This new clustering, using 3 features, couldn\u2019t have been guessed by eye--it was the k-means algorithm that identified it.\n", "\n", "(You'll need to change the code that makes the scatterplot to accommodate 3 features instead of 2, see the comments in the starter code for instructions on how to do this.)" ] }, { "cell_type": "code", "collapsed": false, "input": [ "%load k_means_cluster.py" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 6 }, { "cell_type": "code", "collapsed": false, "input": [ "#!/usr/bin/python \n", "\n", "\"\"\" \n", " skeleton code for k-means clustering mini-project\n", "\n", "\"\"\"\n", "\n", "\n", "\n", "\n", "import pickle\n", "import numpy\n", "import matplotlib.pyplot as plt\n", "import sys\n", "sys.path.append(\"../tools/\")\n", "from feature_format import featureFormat, targetFeatureSplit\n", "\n", "\n", "\n", "\n", "def Draw(pred, features, poi, mark_poi=False, name=\"image.png\", f1_name=\"feature 1\", f2_name=\"feature 2\"):\n", " \"\"\" some plotting code designed to help you visualize your clusters \"\"\"\n", "\n", " ### plot each cluster with a different color--add more colors for\n", " ### drawing more than 4 clusters\n", " colors = [\"b\", \"c\", \"k\", \"m\", \"g\"]\n", " for ii, pp in enumerate(pred):\n", " plt.scatter(features[ii][0], features[ii][1], color = colors[pred[ii]])\n", "\n", " ### if you like, place red stars over points that are POIs (just for funsies)\n", " if mark_poi:\n", " for ii, pp in enumerate(pred):\n", " if poi[ii]:\n", " plt.scatter(features[ii][0], features[ii][1], color=\"r\", marker=\"*\")\n", " plt.xlabel(f1_name)\n", " plt.ylabel(f2_name)\n", " plt.savefig(name)\n", " plt.show()\n", "\n", "\n", "\n", "### load in the dict of dicts containing all the data on each person in the dataset\n", "data_dict = pickle.load( open(\"../final_project/final_project_dataset.pkl\", \"r\") )\n", "### there's an outlier--remove it! \n", "data_dict.pop(\"TOTAL\", 0)\n", "\n", "\n", "### the input features we want to use \n", "### can be any key in the person-level dictionary (salary, director_fees, etc.) \n", "feature_1 = \"salary\"\n", "feature_2 = \"exercised_stock_options\"\n", "feature_3 = \"total_payments\"\n", "poi = \"poi\"\n", "features_list = [poi, feature_1, feature_2,feature_3]\n", "data = featureFormat(data_dict, features_list )\n", "poi, finance_features = targetFeatureSplit( data )\n", "\n", "\n", "### in the \"clustering with 3 features\" part of the mini-project,\n", "### you'll want to change this line to \n", "### for f1, f2, _ in finance_features:\n", "### (as it's currently written, line below assumes 2 features)\n", "for f1, f2,f3 in finance_features:\n", " plt.scatter( f1, f2,f3 )\n", "plt.show()\n", "\n", "\n", "\n", "from sklearn.cluster import KMeans\n", "features_list = [\"poi\", feature_1, feature_2,feature_3]\n", "data2 = featureFormat(data_dict, features_list )\n", "poi, finance_features = targetFeatureSplit( data2 )\n", "clf = KMeans(n_clusters=2)\n", "pred = clf.fit_predict( finance_features )\n", "Draw(pred, finance_features, poi, name=\"clusters_before_scaling.pdf\", f1_name=feature_1, f2_name=feature_2)\n", "\n", "\n", "### cluster here; create predictions of the cluster labels\n", "### for the data and store them to a list called pred\n", "\n", "try:\n", " Draw(pred, finance_features, poi, mark_poi=False, name=\"clusters.pdf\", f1_name=feature_1, f2_name=feature_2)\n", "except NameError:\n", " print \"no predictions object named pred found, no clusters to plot\"\n", "\n", "\n", "\n", "\n", "\n" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEGCAYAAAB2EqL0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAEbxJREFUeJzt3X2QXWVhx/HvkhehTUtI6QTILq4loGbGSqQTloiTpZU2\nZDS2M0yJoxMNnUJpGaidUV7EYf1La221KWPItMAErKDFTgbaUBoZLmAdUzEvBkg0QWiTWELHAEqi\nIym3fzzPzZ7c3Lt5bp6759zLfj8zZ/e55zz3Ps/Nbs7vnuc5Zw9IkiRJkiRJkiRJkiRJkiRJ6iF3\nAvuB7Ql1/wbYEpfvAy9NYr8kST3mPcBC0gKj6FrgH7rfHUlSLxvm6MA4B3gIeBJ4HHhri+d8C/id\nSe+ZJKmnDHN0YDwCzI/lC+PjojcDPwIGJr1nkjTFTa+6AxOYBVwE/FNh3cymOivi9npZnZIk9YZh\nxo8wfpVw9DCRzcDIZHZIkhSc1KXXmUY4Y+nBNttXA7uAbYSJ7RQ/AZ4DLo+PB4DfLGx/G3Aa8O1O\nOytJ6ly3AuN64BlaDw0tI8xDnAtcBaxp8xr3Eiaw3wrsAVYBHwL+CNgKPAUsL9S/Ij5HktQnBoFv\nAJfQ+gjjdsLOvWEnMLeEfkmSuqgbRxhfAD4OvN5m+zzCEUPDXkLISJL6SG5gvA94kTB/MdGprc3b\nPKtJkvpM7mm1iwnzCsuAkwlnNt0NrCzU2QcMFR4PxnVNzqnDs5ndkaQp5VnGr1XrK0toPYexDNgQ\nyyO0P6upDvU+XW7tgT7Y//5d7L/9P9Gl3NGabl+41+j81fH7WkJYLAN2AwcJZz9JkvpMNwPjsbhA\nCIqia7vYjiSpAt26DmOKG626A5lGq+5AptGqO5BptOoOZBqtugOZRqvuQN/opT/aVy95OE6S+tzA\nkS9l8AhDkpTEwJAkJTEwJElJDAxJUhIDQ5KUxMCQJCUxMCRJSQwMSVISA0OSlMTAkCQlMTAkSUkM\nDElSEgNDkpTEwJAkJckNjJOBTcBW4BngMy3qjAKvAFvicktmm5KkCuTece/nwCXAofha3wQujt+L\nHgOWZ7YlSapQN4akDsXvM4FpwIEWdXrpRk2SpBPQjcA4iTAktR94lDA0VVQHFgPbgA3Agi60KUkq\nWe6QFMDrwPnAqcDDhDmLWmH7ZmCIcCRyGbAeOK/1S40VyqN4r11JKqpx9O61XN0eKvoU8DPg8xPU\neQ64gGOHrryntyR1pL/u6X06MDuWTwEuJZwJVTSX8Te0KJZbzXNIknpY7pDUmcA6QvCcBNwDPAJc\nHbevBS4HrgEOE4alVmS2KUmqQC+dveSQlCR1pL+GpCRJU4SBIUlKYmBIkpIYGJKkJAaGJCmJgSFJ\nSmJgSJKSGBiSpCQGhiQpiYEhSUpiYEiSkhgYkqQkBoYkKYmBIUlKYmBIkpIYGJKkJLmBcTKwCdgK\nPAN8pk291cAuYBuwMLNNSVIFcm/R+nPgEsKtV6cD3wQujt8blgHzgXOBC4E1wEhmu5KkknVjSOpQ\n/D4TmAYcaNq+nHDfbwhHI7OBuV1oV5JUom4ExkmEIan9wKOEoamiecCewuO9wGAX2pUklSh3SArg\ndeB84FTgYWAUqDXVab5Jeb31S40VyqNxkSQFNY7dvZaneUee61PAz4DPF9bdTniH98XHO4ElhCOS\nonrbHJEktTBw5EsZcoekTifMSQCcAlwKbGmq8wCwMpZHgJc5NiwkST0ud0jqTMKE9klxuQd4BLg6\nbl8LbCCcKbUbOAisymxTklSB0g5lEjgkJUkd6a8hKUnSFGFgSJKSGBiSpCQGhiQpiYEhSUpiYEiS\nkhgYkqQkBoYkKYmBIUlKYmBIkpIYGJKkJAaGJCmJgSFJSmJgSJKSGBiSpCQGhiQpSW5gDAGPAk8D\nTwHXtagzCrxCuHXrFuCWzDYlSRXIvUXra8DHgK3ALOC7wEZgR1O9x4DlmW1JkiqUe4TxAiEsAF4l\nBMVZLer10q1gJUknoJtzGMPAQmBT0/o6sBjYBmwAFnSxTUlSSXKHpBpmAfcD1xOONIo2E+Y6DgGX\nAeuB81q/zFihPBoXSVJQi0s1ujFUNAP4F+Ah4IsJ9Z8DLgAONK2vh4MRSVKagSNfypA7JDUA3AE8\nQ/uwmMv4G1oUy81hIUnqcblDUu8GPgx8j3DKLMDNwNmxvBa4HLgGOEwYllqR2aYkqQK9dPaSQ1KS\n1JH+GpKSJE0RBoYkKYmBIUlKYmBIkpIYGJKkJAaGJCmJgSFJSmJgSJKSGBiSpCQGhiQpiYEhSUpi\nYEiSkhgYkqQkBoYkKYmBIUlKkhsYQ8CjwNPAU8B1beqtBnYB24CFmW1KkiqQe8e914CPAVuBWcB3\ngY3AjkKdZcB84FzgQmANMJLZriSpZLlHGC8QwgLgVUJQnNVUZzmwLpY3AbMJ9/mWJPWRbs5hDBOG\nmzY1rZ8H7Ck83gsMdrFdSVIJuhUYs4D7gesJRxrNmu856827JanP5M5hAMwAvg58GVjfYvs+wuR4\nw2Bc18JYoTwaF0lSUItLNZo/+Z/I89cBPyZMfreyDLg2fh8BvkjrSe+6Bx6S1ImBI19Kay3DxcDj\nwPcY39vfDJwdy2vj99uApcBBYBWwucVrGRiS1JH+CoxuMjAkqSPlBoZXekuSkhgYkqQkBoYkKYmB\nIUlKYmBIkpIYGJKkJAaGJCmJgSFJSmJgSJKSGBiSpCQGhiQpiYEhSUpiYEiSkhgYkqQkBoYkKYmB\nIUlK0o3AuBPYD2xvs30UeAXYEpdbutCmJKlk07vwGncBfwfcPUGdx4DlXWhLklSRbhxhPAG8dJw6\nvXQrWEnSCShjDqMOLAa2ARuABSW0KUnqsm4MSR3PZmAIOARcBqwHziuhXUlSF5URGD8tlB8CvgTM\nAQ4cW3WsUB6NiyQpqMWlGt2aWxgGHgTe0WLbXOBFwtDUIuBrsX6zeqgiSUozcORLGbpxhHEvsAQ4\nHdgD3ArMiNvWApcD1wCHCcNSK7rQpiSpZL109pJHGJLUkXKPMLzSW5KUxMCQJCUxMCRJSQwMSVIS\nA0OSlMTAkCQlMTAkSUkMDElSEgNDkpTEwJAkJTEwJElJDAxJUhIDQ5KUxMCQJCUxMCRJSQwMSVKS\n3MC4E9gPbJ+gzmpgF7ANWJjZniSpIrmBcRewdILty4D5wLnAVcCazPYkSRXJDYwngJcm2L4cWBfL\nm4DZwNzMNiVJFZjsOYx5wJ7C473A4CS3KUmaBNNLaKP5BuX19lXHCuXRuEiSglpcqjHZgbEPGCo8\nHozr2hib3N5IUl8b5egP0p8utfXJHpJ6AFgZyyPAy4SzqiRJfSb3CONeYAlwOmGu4lZgRty2FthA\nOFNqN3AQWJXZniSpIs3zC1WqTzi9IUlqMnDkSxm80luSlMTAkCQlMTAkSUkMDElSEgNDkpTEwJAk\nJTEwJElJDAxJUhIDQ5KUxMCQJCUxMCRJSQwMSVISA0OSlMTAkCQlMTAkSUkMDElSkm4ExlJgJ7AL\nuKHF9lHgFWBLXG7pQpuSpJLl3qJ1GnAb8F5gH/Adwn28dzTVewxYntmWJKlCuUcYiwj3634eeA24\nD/hAi3q9dCtYSdIJyA2MecCewuO9cV1RHVgMbAM2AAsy25QkVSB3SKqeUGczMAQcAi4D1gPnta46\nViiPxkWSFNTiUo3coaIRwl5+aXx8E/A68JcTPOc54ALgQNP6elr+SJKCgSNfypA7JPUkcC4wDMwE\nriBMehfNZfwNLYrl5rCQJPW43CGpw8C1wMOEM6buIJwhdXXcvha4HLgm1j0ErMhsU5JUgV46e8kh\nKUnqSH8NSUmSpggDQ5KUxMCQJCUxMCRJSQwMSVISA0OSlMTAkCQlMTAkSUkMDElSEgNDkpTEwJAk\nJTEwJElJDAxJUhIDQ5KUxMCQJCUxMCRJSboRGEuBncAu4IY2dVbH7duAhV1oU5JUstzAmAbcRgiN\nBcAHgbc31VkGzCfc+/sqYE1mm5KkCuQGxiJgN/A88BpwH/CBpjrLgXWxvAmYDczNbFeSVLLcwJgH\n7Ck83hvXHa/OYGa7kqSSTc98fj2xXvNNyts8b6xQHo2LJCmoxaUauYGxDxgqPB4iHEFMVGcwrmth\nLLM7kvRGNsrRH6Q/XWrruUNSTxIms4eBmcAVwANNdR4AVsbyCPAysD+zXUlSyXKPMA4D1wIPE86Y\nugPYAVwdt68FNhDOlNoNHARWZbYpSapA89xClerpUyKSpLgLL20/7pXekqQkBoYkKYmBIUlKYmBI\nkpIYGJKkJAaGJCmJgSFJSmJgSJKSGBiSpCQGhiQpiYEhSUpiYEiSkhgYkqQkBoYkKYmBIUlKknMD\npTnAV4E3A88Df0i4m16z54GfAP8HvAYsymhTklSRnCOMG4GNwHnAI/FxK3XCTWgX8oYNi1rVHchU\nq7oDmWpVdyBTreoOZKpV3YFMtao70DdyAmM5sC6W1wG/P0HdXrqz3ySoVd2BTLWqO5CpVnUHMtWq\n7kCmWtUdyFSrugN9Iycw5gL7Y3l/fNxKHfgG8CTwxxntSZIqdLw5jI3AGS3Wf7LpcZ32N+R+N/A/\nwK/H19sJPNFBHyVJPSBnqGgnYW7iBeBM4FHgbcd5zq3Aq8Bft9i2Gzgnoz+SNNU8C8yvuhMpPgfc\nEMs3Ap9tUeeXgF+J5V8G/gP43cnvmiSpl8whzE38APh3YHZcfxbwr7H8G8DWuDwF3FRyHyVJkiT1\nq78CdgDbgH8GTi1suwnYRZj3KA5JXQBsj9v+trD+TYSLAncB3yZcHNjwEcJRzQ+AlYX1bwE2xefc\nB8zIfUOJlhLe1y7Gh+jKMkSYO3qacAR3XVw/h3CCQfPRH5Tzs+jUNGAL8GAf9n82cD/hd/8Z4MI+\n6v9NhN+d7cBXYlu93Pc7CWdkbi+sq7q/nex3WvV/qu43uZTx03Q/y/icxgLCkNQMYJgwud2YbP9P\nxi/m20DY+QL8KfClWL6C8EYg/HI8S/ilmB3LjX/grxGuNgdYA/xJ/ls6rmmE9zNMeH9bgbeX0G7D\nGcD5sTwL+H5s/3PAJ+L6GyjvZ1H8z9qJvwD+EXggPu6n/q8Drozl6YTfx37o/zDwQ8JOBsKO5iM9\n3vf3EC4ALu5wq+rviex3WvV/Ku43j/EHwJdj+SaO/uT9b8AI4cyqHYX1K4DbC3UujOXpwP/G8gcJ\nb6rh9vi8gVin8Q8/El9jsl3U1M6NtL/ivQzrgfcSPpE0ros5Iz6Gcn4WnRokzIVdwvgRRr/0/1TC\nTrdZP/R/DuEDxmnxdR8k7Lx6ve/DHL3DrbK/J7Lfae5/Uc/vNyfrjw9eSUg+CJPgewvb9gLzWqzf\nF9cTv++J5cPAK8CvTfBacwh/x+r1Fq81mYr9LPanCsOETy+baH9RZRk/i059Afg44z87+qj/byH8\nh7sL2Az8PeFswH7o/wHC6e3/DfyI8P9nY5/0vajK/nZ7v9Pz+81OA2MjIR2bl/cX6nwS+AVhTLQM\n7S4YfKO3XTQL+DpwPfDTpm0TXVRZtfcBLxLmL9pdE9TL/Z8OvIswDPAu4CDHHmH2av/PAf6c8EHj\nLMLv0Ieb6vRq39sps7/dbqcv9pudBsalwDtaLI2hhI8Cy4APFZ6zjzA52zBISLh9sdy8vvGcs2O5\nMS784xavNRTXHSCMzTXez2BcP9la9Wdvm7qTZQYhLO4hDElB+KTVuEL/TMJOGSb/Z9Hpe19M+Jtk\nzwH3Ar8d30e/9H9vXL4TH99PCI4X+qD/vwV8K77WYcKE60V90veiqn5Xurnf+ShTa78JhImXp4HT\nm9Y3Jm9mEg7hn2X80+QmwpjbAMdO3jTG3FZw9OTNDwlv8rRCGcLkzRWxfDvlTN5MJ7yfYcL7K3vS\newC4mzCsU9TuosqyfhYnYgnjHzz6qf+PE/5iM8BY7Hs/9P+dhDPrToltrgP+rA/6Psyxk95V9rfT\n/U5z/6fifhMIp2X9F2F4YQvjs/UANxNm+XcCv1dY3zg9bDewurD+TYQ30jg9bLiwbVVcv4twVkdD\n8fSwr1Le6WGXESYPd1P+hYkXE8YftzL+776U9hdVQjk/ixOxhPGzpPqp/+8kHGEUT4vsl/5/gvHT\natcR/s/0ct/vJcy3/IIwVr+qB/rbyX6nuf9XMnX3m5IkSZIkSZIkSZIkSZIkSZIkSZIk6f8BIHUd\nD6cmg4wAAAAASUVORK5CYII=\n", "text": [ "" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAAEVCAYAAADU/lMpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRERERERERERERERS+j8mkLkjW5hcKQAAAABJRU5ErkJggg==\n", "text": [ "" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAAEVCAYAAADU/lMpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRERERERERERERERS+j8mkLkjW5hcKQAAAABJRU5ErkJggg==\n", "text": [ "" ] } ], "prompt_number": 11 }, { "cell_type": "heading", "level": 4, "metadata": {}, "source": [ "4 points switch clusters" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In the next lesson, we\u2019ll talk about feature scaling. It\u2019s a type of feature preprocessing that you should perform before some classification and regression tasks. Here\u2019s a sneak preview that should call your attention to the general outline of what feature scaling does.\n", "\n", "What are the maximum and minimum values taken by the \u201cexercised_stock_options\u201d feature used in this example?\n", "\n", "(NB: if you look at finance_features, there are some \"NaN\" values that have been cleaned away and replaced with zeroes--so while those might look like the minima, it's a bit deceptive because they're more like points for which we don't have information, and just have to put in a number. So for this question, go back to data_dict and look for the maximum and minimum numbers that show up there, ignoring all the \"NaN\" entries.)" ] }, { "cell_type": "code", "collapsed": false, "input": [ "ex_stok = []\n", "for users in data_dict:\n", " val = data_dict[users][\"exercised_stock_options\"]\n", " if val == 'NaN':\n", " continue\n", " ex_stok.append(val)\n", "print max(ex_stok)\n", "print min(ex_stok)\n", " " ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "34348384\n", "3285\n" ] } ], "prompt_number": 13 }, { "cell_type": "markdown", "metadata": {}, "source": [ "What are the maximum and minimum values taken by \u201csalary\u201d?" ] }, { "cell_type": "code", "collapsed": false, "input": [ "salary = []\n", "for users in data_dict:\n", " val = data_dict[users][\"salary\"]\n", " if val == 'NaN':\n", " continue\n", " salary.append(val)\n", " \n", "print max(salary)\n", "print min(salary)" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "1111258\n", "477\n" ] } ], "prompt_number": 15 }, { "cell_type": "markdown", "metadata": {}, "source": [ "The plot on the next slide shows the exact same clustering code that you just wrote, but in this example we applied feature scaling before performing the clustering.\n", "\n", "Notice that now the range of the features has changed to [0.0, 1.0]. That's the only change we've made.\n", "\n", "In the next lesson you\u2019ll learn a lot more about what feature scaling means, but for now, just look at the effect on the clusters--which point(s) switch their associated cluster?" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "![jpeg](../galleries/kmeans/5.jpg)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "> **REFERENCE**:\n", "\n", "> * https://www.udacity.com/course/viewer#!/c-ud120/l-2293728536/e-3009398697/m-3024388568\n" ] } ], "metadata": {} } ] }