1010912024, 08:16 4 tnport necessary Libraries Anport natplotlib.pyplot as pit frm skiearn.cluster inport Kans Innport Necessary Libraries Inthe fist code cell import the necessary libraries like numpy, pandas, matplatb, and sciktlearn forthe kMeans algorithm. keMeansipynb ~ Colab ‘ep 2: Create a Dataset For demonstration, lets create a simple 20 dataset using numpy. 4 creating 2 sample 20 cataset with 2 clusters np. randon.seec(@) Xe npovstack((np.randon.cornal (loe=[2, 2], scale np.randon.nornal oe+(8, 8], 531 af = pacoatarrane(Xy cotumase['X°, °Y"]) 1 visualize the asta pit.scatter(af{'x"], dV], c='blue', piestitle('Sanple 20 bate!) pit xlabelx") pitiylabelcy") Bitsshont) a Sample 2D Data ‘Step 4: Apply the keMeans Algorithm Now, apply the k-Means clustering algorithm using scik-earn We wil set k=2 since we want two clusters Y canerte 01am qumbers using nl # poply keteans clustering keans’ = oheans(n_elusterse2) leans. F480) 1 Get the cluster centers (centroies) prine(Centrotds:", centroids) 1 Get the cluster Labels (to which cluster each point belongs) Aabels = areons labels, Drine("Lavels:", labels) centroids: [{7.99963924 6.68137373] (2-bowao3oe 2951408831] ntpsifolab research google. comidrve/1WaStGINcOULatig6X_ZZIXK20QwcKarR-#sorolTo=JFP40EOxOpAnkpriniMode=true 5, sizen(50, 209. 5) sizee(58, 2009) a Chose 181010912024, 08:16 keMeansipynt - Colao 20eccecccseececseses0ecss) 1/4er/oeai/ Lib /python}.10/cist-packages/skleatn/clusver/_kreans.py:2416: Futureaarning: The defSult value of “r_inst” will change érom 3 Stper(). creck parano_vs_input(h, derault a tnita20) ‘Stop 5: Visualize the Clustering Results Next, visualize the clusters along with the centroids een(ef"X"], dF['Y'], e=labels, $258, cxape"vieidis") feentrotds[, 8), centrotés(, 1, pisvtitle(k-neans Clustering with ke2") oclabel(X") peeylebelW) pit. shont) ea, 9200, alghan8.75, narker='x') k-Means Clustering with k=2 ‘Step 6: Evaluate the Model (Optional) You can evaluate how wel the clusters are formed using metrics like Inertia (witin-luster sum of ‘squares), Lower inertia means better clustering python ‘calculate inertia (aun of squared distances to the nearest cluster center) print("Thertio:', keeans- inertia.) ‘Step 7: Experiment with Different k Values You can also experiment with different values of kto find the optimal numberof clusters using the Elbow Method, anertias = (] ° keyalues = range(a, 19) srcons » Weans(o_clusters-k) ireans- #5800) Aner #1as.append(kneans. sere) he snerta vs number of clusters (Eabow Method) (values, inertias, "bo-') ‘ithe(Eltow Method for optinalk") label('Wamber oF clusters (k)') spabel{"inertia") ntpsifolab research google. comidrve/1WaStGINcOULatig6X_ZZIXX20QwcKarR-#sorolTo=JFP40EOxOpAnkpriniMode=true 281010912024, 08:16 pie show) ase/loeat//pythont.30/stst-packages/sklearn/cluster/_kewans.py 2816: Stper(). check parang_9s_tnput(x, deraulta_tnites8) asr/loeaifto/pyenon}-i0/dist-packages/sklearn/cluster/_keeans.py:2816 Stper(). check 2arans_vs_input (x, default intt-20) asr/locai7i/pyenors. /dist-packages/sklean/cluster/_kewans.oy:2416 ‘Stper(), check 2arans_vs_input (x, geeaultn_intt=20) sasr/locai?i/pyenon3. 7 ist-packages/skleaen/cluster/_keeans.py:2416 Super), check parans_vs_input (x, geeaultn_intt=i8) Jase/loeai7it/pyenon)-io/dtst-packages/sklearn/cluseer/_keeans.py:2816: Stper(). check parans_vs_input(x, derault_n_init-28) ase laeai?to/python}i0/dist-packages/skleann/cluster/_keeans.py:2816 Stper(). check 2arans_vs_tnput(x, default intt-s0) sasr/locai7iby/pytnon3. a7 dist-packages/skleaen/cluster/_keeans.py:2416 Super), check parans_vs_input Gk, geeault_n_intt=20) Jase /locai7b/pythons. i/dist-packages/skl ean cluster/_kneans.py:2416 Stper(). creck parano_9s_ input (x, derault_tnit-28) asr/loeai?to/pyenon}-10/3st-packages/sklearn/cluster/_keeans.py:2416 Stper(). creck parang_vsinput(x, gerault a initai8) Elbow Method for Optimal k keMeansipynb - Colab Fatureaening: Faturetoening: Fatureaoning: 750 1500 2000 inertia Number of clusters) Smport pandas 28 ps tram skiearn.cluster import Weans 1 step 1: coente dataset np. randon.seea(@) np.vstack((np.randon.rornal(loe~(2, 2], scaler rnp.randon.nornal Qoe+(8, 8], sealent 5, sizen(50, 209. 1 sizes(58, 2003) GF = pd.oataFrane(X, cotuanselX°, °YJ) pe pie pie pie pie seatten(afi'®'), a['Y"], blue’, {ithe( ‘Sonple 29 Date") abel XD ylavel('¥") shew) 4 step 3: Apply keeans leneans = xheana(nelusters-2) keane F008) 4 step 4: visualize the clusters pis pe ne pe pe ntpssioolab es Scatter(af('X'], df['Y"], crlabels, 5°58, crage'viriais") scatter(centroids{:, @]y centroids{:, 1], ce'red", 59208, alpharB.75, markers ste leeane Clustering with k=2") saabel(*X") sabel(-¥") Show) the default value of inst The default value of “init The defsutt value of “inst The default value of “init the dofoult value of “inst rch google. comdrive!I WaSGINCOULatig6X_ZZIX2QQwcKark-scrollTo=JFP4oEOxOpAngpriniMode=ttue wan wan wan change charge charge 351010912024, 08:16 keMeansipynt - Colao 4 step 5: bow Rethod (Options) inertias = (1 values = ronge(a, 32) sveans = Kieans(_clusters-k) rears. #3200) sner2ias,append(eneans. inertia) plot(e.valves, inertias, “bo-") ‘ste eibow Hethod for Optimal") rasbelCMamer of clusters (K)") sjlabel(“irer how) ntpsifolab research google. comidrve/1WaStGINcOULatig6X_ZZIXX20QwcKarR-#sorolTo=JFP40EOxOpAnkprintMode=true1010912024, 08:16 Sample 2D Data uper()-—check:parane_we Sopot, cea) k-Means Clustering with k: tnse-30) keMeansipynb - Colab 2 asn/loeai/4o/python3.10/ctst-packages/sklearn/ cl ‘super() «check parama_vs_input(hy defautten, init eo/_kewans.y:3416: Futureoontn: ) ee/_keans.9y 4416: Futureantng: he default value of“ intt” will change fo) ntpsifolab research google. comidrve/1WaStGINcOULatig6X_ZZIXX20QwcKarR-#sorolTo=JFP40EOxOpAnkprintMode=true ange #0 55