Coding Question] In terms of K-means clustering algorithm, in this question, you are required to implement this algorithm from scratch (without using external packages or libraries) and use it to cluster the data samples in the iris dataset into 3 clusters based on their petal length and petal width. At each training iteration, it is required to calculate and record the mean distance of data points to their respective cluster centroid. After obtaining the clustering results, it is required to generate figures to vi- sualize the results. In the visualization, you are first required to draw a scatter plot for all the data samples (x-axis corresponds to the petal length while y-axis corresponds to the petal width), and color each sample in red, green, and blue, indicating the cluster it has been assigned to. Then, you are required draw another line plot with x-axis corresponds to the training iteration, and y-axis corresponds to the mean distance to the cluster centroids at that iteration. In the algorithm implementation, no external package or library should be used, while for visualization purpose, you can use any graphics or visualization libraries you prefer (e.g., “ggplot2”). For the initialization, the cluster centroids of the three classes are (1.4, 0.1), (1.3, 0.2), and (1.7, 0.1), respectively. Please run your K-means clustering algorithm for 100 iterations. "Only the R language is allowed for answering the programming questions."
[Coding Question] In terms of K-means clustering
Step by step
Solved in 3 steps with 2 images
