..il AT&T 6:15 PM jlo7p88famt75w Questions 1 (80 points). Implement &-Mea algoritlhan for clustering (Algoeitn 13.1 in the textbook). Your peogram should take a comma-sepazated data tale, and the auaer of dustes(&-value) commsand line parameters. There is one optional parameter, which is another filennme which lists the id of initial centroids. For example, without the optional parametes, we may ran yoar program /eycMeana datafi2e x 3 In that case, it should cluster the data points in datafile.tat into 3 dlusters The data fie datale.ext lists one data point in each line. The values of a data points along diffecrent dimen- sions are separated by a comma. No id is given for a data point, beat yon me that the i of a data point is the same as its line number in the input file assuming that the line umbering starts from 1. For initial cetroid you should use k random data points. Now, with the optional parameter, we can also run yoar program as below nykMeans datafile.txt5 centroidfile.tat Say, the centroidfile.txt contains the Sollowing integers for a clustering tak with kS 19 371 390 In that case you must use the points with id 5, 19,201, 371 aad 390 as your initial ceutrolds. For all cases, use e-0.0001 as your stopping condition Yar program output should consists of the following information L. The number of data points in the input file, the dimension, and the value of k .The umber of iterations the 3. The final mean of each cluster and the SSE score (sum of square error) . The final cluster nssignment of all thhe points. 3. The final sze of each cluster. program took for couvergence 2 (20 points) Execute your program on the iris.txt dataset (will be available in piazza resouroe section), Rm it for at least 10 times with different randons initializations of centroid (make sure to randomine your random seeds so that different rus of the program starts with a difSeeent random set of eentroids). Now, evasate the elasteeing uing purity-hased evalnation (e 17.1.1 in textbook) and report the best purity soore. Note that for parity based evaluation you neod to kuow the true label of a ata points Here is the tre label assignments on the iris.txt dataset: 1-30 class 1, 51-100 class 2, and 101-150: class 3 ..il AT&T 6:15 PM jlo7p88famt75w Questions 1 (80 points). Implement &-Mea algoritlhan for clustering (Algoeitn 13.1 in the textbook). Your peogram should take a comma-sepazated data tale, and the auaer of dustes(&-value) commsand line parameters. There is one optional parameter, which is another filennme which lists the id of initial centroids. For example, without the optional parametes, we may ran yoar program /eycMeana datafi2e x 3 In that case, it should cluster the data points in datafile.tat into 3 dlusters The data fie datale.ext lists one data point in each line. The values of a data points along diffecrent dimen- sions are separated by a comma. No id is given for a data point, beat yon me that the i of a data point is the same as its line number in the input file assuming that the line umbering starts from 1. For initial cetroid you should use k random data points. Now, with the optional parameter, we can also run yoar program as below nykMeans datafile.txt5 centroidfile.tat Say, the centroidfile.txt contains the Sollowing integers for a clustering tak with kS 19 371 390 In that case you must use the points with id 5, 19,201, 371 aad 390 as your initial ceutrolds. For all cases, use e-0.0001 as your stopping condition Yar program output should consists of the following information L. The number of data points in the input file, the dimension, and the value of k .The umber of iterations the 3. The final mean of each cluster and the SSE score (sum of square error) . The final cluster nssignment of all thhe points. 3. The final sze of each cluster. program took for couvergence 2 (20 points) Execute your program on the iris.txt dataset (will be available in piazza resouroe section), Rm it for at least 10 times with different randons initializations of centroid (make sure to randomine your random seeds so that different rus of the program starts with a difSeeent random set of eentroids). Now, evasate the elasteeing uing purity-hased evalnation (e 17.1.1 in textbook) and report the best purity soore. Note that for parity based evaluation you neod to kuow the true label of a ata points Here is the tre label assignments on the iris.txt dataset: 1-30 class 1, 51-100 class 2, and 101-150: class 3
