Write and test a code on MATLAB to form 'k' clusters of given 'N' data points,...

Transcribed Image Text:

Write and test a code on MATLAB to form 'k' clusters of given 'N' data points, where each data point is of 'd' dimensions. The code should follow the given steps: a) Open a text file "input.txt" and read data from it. Each data point in one line with comma separated entries. The number of entries in each line tells you the dimension of the data. b) Take the input of number of clusters 'k' from the user c) Initialize centroids by first shuffling the dataset and randomly selecting 'k' data points for the centroids without replacement. d) Keep iterating until the change in centroids is less than a threshold 'e'. i. Compute the sum of the squared distance between data points and all centroids ii. Assign each data point to the closest centroid i. Compute the centroids for the clusters by taking the average of all the data points that belong to one cluster. e) Write a text file "output.txt" containing the location of 'k' centroids Grad-+g Scheme: Problem 1: Opening and reading text file: "input.txt" Displaying an error message when the file does not exist [1 point] Loading the data points Taking the input of number of clusters 'k' from the user Random initialization of the centroids Estimation of the squared distance of each data point Assignment of data point to cluster centroid Updated estimation of centroids from new clusters Writing the text file "output.txt" [50 points] [2 points] [5 points] [2 points] [5 points] [15 points] [5 points] [10 points] [5 points] Write and test a code on MATLAB to form 'k' clusters of given 'N' data points, where each data point is of 'd' dimensions. The code should follow the given steps: a) Open a text file "input.txt" and read data from it. Each data point in one line with comma separated entries. The number of entries in each line tells you the dimension of the data. b) Take the input of number of clusters 'k' from the user c) Initialize centroids by first shuffling the dataset and randomly selecting 'k' data points for the centroids without replacement. d) Keep iterating until the change in centroids is less than a threshold 'e'. i. Compute the sum of the squared distance between data points and all centroids ii. Assign each data point to the closest centroid i. Compute the centroids for the clusters by taking the average of all the data points that belong to one cluster. e) Write a text file "output.txt" containing the location of 'k' centroids Grad-+g Scheme: Problem 1: Opening and reading text file: "input.txt" Displaying an error message when the file does not exist [1 point] Loading the data points Taking the input of number of clusters 'k' from the user Random initialization of the centroids Estimation of the squared distance of each data point Assignment of data point to cluster centroid Updated estimation of centroids from new clusters Writing the text file "output.txt" [50 points] [2 points] [5 points] [2 points] [5 points] [15 points] [5 points] [10 points] [5 points]

Answer rating: 100% (QA)

function clusters clusterCenters kMeansClusteringdataSetnumClustersnumIterations Simple implementati... View the full answer