{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"id": "9bb8d1a0-0e9a-4b0d-a9be-45cdfafcfec2",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" ID | \n",
" x | \n",
" y | \n",
" cluster | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 0 | \n",
" 24.412 | \n",
" 32.932 | \n",
" 2 | \n",
"
\n",
" \n",
" | 1 | \n",
" 1 | \n",
" 35.190 | \n",
" 12.189 | \n",
" 1 | \n",
"
\n",
" \n",
" | 2 | \n",
" 2 | \n",
" 26.288 | \n",
" 41.718 | \n",
" 2 | \n",
"
\n",
" \n",
" | 3 | \n",
" 3 | \n",
" 0.376 | \n",
" 15.506 | \n",
" 0 | \n",
"
\n",
" \n",
" | 4 | \n",
" 4 | \n",
" 26.116 | \n",
" 3.963 | \n",
" 1 | \n",
"
\n",
" \n",
"
\n",
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"customcmap = ListedColormap([\"crimson\", \"mediumblue\", \"darkmagenta\"])\n",
"\n",
"fig, ax = plt.subplots(figsize=(8, 6))\n",
"plt.scatter(x=blobs['x'], y=blobs['y'], s=150,\n",
" c=blobs['cluster'].astype('category'), \n",
" cmap = customcmap)\n",
"ax.set_xlabel(r'x', fontsize=14)\n",
"ax.set_ylabel(r'y', fontsize=14)\n",
"plt.xticks(fontsize=12)\n",
"plt.yticks(fontsize=12)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "d99481fa-88e2-4874-8492-882309a1c498",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" x | \n",
" y | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 24.412 | \n",
" 32.932 | \n",
"
\n",
" \n",
" | 5 | \n",
" 25.893 | \n",
" 31.515 | \n",
"
\n",
" \n",
" | 36 | \n",
" 26.878 | \n",
" 36.609 | \n",
"
\n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" x | \n",
" y | \n",
" centroid | \n",
" error | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 24.412 | \n",
" 32.932 | \n",
" 0 | \n",
" 0.000000 | \n",
"
\n",
" \n",
" | 1 | \n",
" 35.190 | \n",
" 12.189 | \n",
" 1 | \n",
" 211534.211314 | \n",
"
\n",
" \n",
" | 2 | \n",
" 26.288 | \n",
" 41.718 | \n",
" 2 | \n",
" 699.601495 | \n",
"
\n",
" \n",
" | 3 | \n",
" 0.376 | \n",
" 15.506 | \n",
" 0 | \n",
" 776856.744109 | \n",
"
\n",
" \n",
" | 4 | \n",
" 26.116 | \n",
" 3.963 | \n",
" 1 | \n",
" 576327.599678 | \n",
"
\n",
" \n",
"
\n",
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"fig, ax = plt.subplots(figsize=(8, 6))\n",
"plt.scatter(df.iloc[:,0], df.iloc[:,1], marker = 'o', \n",
" c=df['centroid'].astype('category'), \n",
" cmap = customcmap, s=80, alpha=0.5)\n",
"plt.scatter(centroids.iloc[:,0], centroids.iloc[:,1], \n",
" marker = 's', s=200, c=[0, 1, 2], \n",
" cmap = customcmap)\n",
"ax.set_xlabel(r'x', fontsize=14)\n",
"ax.set_ylabel(r'y', fontsize=14)\n",
"plt.xticks(fontsize=12)\n",
"plt.yticks(fontsize=12)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "67c07bc6-d5ad-427b-9cfe-b032ed35feef",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The total error is 11927659.01\n"
]
}
],
"source": [
"print(\"The total error is {0:.2f}\".format(df['error'].sum()))"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "841f165a-89de-4b4e-9fe0-06498088d4f6",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" x | \n",
" y | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 9.889444 | \n",
" 23.242611 | \n",
"
\n",
" \n",
" | 1 | \n",
" 28.435750 | \n",
" 11.546250 | \n",
"
\n",
" \n",
" | 2 | \n",
" 30.759333 | \n",
" 40.311167 | \n",
"
\n",
" \n",
"
\n",
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"fig, ax = plt.subplots(figsize=(8, 6))\n",
"plt.scatter(df.iloc[:,0], df.iloc[:,1], marker = 'o', \n",
" c=df['centroid'].astype('category'), \n",
" cmap = customcmap, s=80, alpha=0.5)\n",
"plt.scatter(centroids.iloc[:,0], centroids.iloc[:,1], \n",
" marker = 's', s=200,\n",
" c=[0, 1, 2], cmap = customcmap)\n",
"ax.set_xlabel(r'x', fontsize=14)\n",
"ax.set_ylabel(r'y', fontsize=14)\n",
"plt.xticks(fontsize=12)\n",
"plt.yticks(fontsize=12)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "de6c03cc-2841-4aa1-b4b2-320fb109d336",
"metadata": {},
"outputs": [],
"source": [
"def kmeans(dset, k=2, tol=1e-4):\n",
" '''\n",
" K-means implementationd for a \n",
" `dset`: DataFrame with observations\n",
" `k`: number of clusters, default k=2\n",
" `tol`: tolerance=1E-4\n",
" '''\n",
" # Let us work in a copy, so we don't mess the original\n",
" working_dset = dset.copy()\n",
" # We define some variables to hold the error, the \n",
" # stopping signal and a counter for the iterations\n",
" err = []\n",
" goahead = True\n",
" j = 0\n",
" \n",
" # Step 2: Initiate clusters by defining centroids \n",
" centroids = initiate_centroids(k, dset)\n",
"\n",
" while(goahead):\n",
" # Step 3 and 4 - Assign centroids and calculate error\n",
" working_dset['centroid'], j_err = centroid_assignation(working_dset, centroids) \n",
" err.append(sum(j_err))\n",
" \n",
" # Step 5 - Update centroid position\n",
" centroids = working_dset.groupby('centroid').agg('mean').reset_index(drop = True)\n",
"\n",
" # Step 6 - Restart the iteration\n",
" if j>0:\n",
" # Is the error less than a tolerance (1E-4)\n",
" if err[j-1]-err[j]<=tol:\n",
" goahead = False\n",
" j+=1\n",
"\n",
" working_dset['centroid'], j_err = centroid_assignation(working_dset, centroids)\n",
" centroids = working_dset.groupby('centroid').agg('mean').reset_index(drop = True)\n",
" return working_dset['centroid'], j_err, centroids"
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "c9d0bd64-f01f-4226-ac89-69509e5b6007",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" x | \n",
" y | \n",
" centroid | \n",
" error | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 24.412 | \n",
" 32.932 | \n",
" 2 | \n",
" 3767.568743 | \n",
"
\n",
" \n",
" | 1 | \n",
" 35.190 | \n",
" 12.189 | \n",
" 1 | \n",
" 1399.889001 | \n",
"
\n",
" \n",
" | 2 | \n",
" 26.288 | \n",
" 41.718 | \n",
" 2 | \n",
" 262.961097 | \n",
"
\n",
" \n",
" | 3 | \n",
" 0.376 | \n",
" 15.506 | \n",
" 0 | \n",
" 2683.086425 | \n",
"
\n",
" \n",
" | 4 | \n",
" 26.116 | \n",
" 3.963 | \n",
" 1 | \n",
" 2723.650198 | \n",
"
\n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" x | \n",
" y | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" 6.322867 | \n",
" 19.559800 | \n",
"
\n",
" \n",
" | 1 | \n",
" 29.330864 | \n",
" 10.432409 | \n",
"
\n",
" \n",
" | 2 | \n",
" 29.304957 | \n",
" 39.050783 | \n",
"
\n",
" \n",
"
\n",
"