1. The k-means algorithm always converges to the same result irrespective of the starting points (cluster centroids). True or False? 2. For a given dataset, the k-means algorithm finds the globally...

Please disregard APA citation request. I have attached the assignment. Please read and answer ALL questions carefully and precisely. Questions consist of multiple choice, true/false, and short answers. Thanks! I need this complete in the next couple of hours as it is due before the end of the day.


1. The k-means algorithm always converges to the same result irrespective of the starting points (cluster centroids). True or False? 2. For a given dataset, the k-means algorithm finds the globally best clustering. True or False? 3. When we run the PCA dimensionality reduction algorithm, its effect is similar to what we will get if we were to load the dataset in Excel and delete some columns (dimensions). For instance, if we load a 10 column/dimensional dataset and use PCA to reduce it to 2 dimensions, the result is similar to loading it in Excel and deleting 8 columns. True or False 4. In what way is a decision tree worse than a neural network for learning the XOR function? 5. If a set JK does not satisfy the support threshold in an association mining context, which of the following will definitely also NOT satisfy the support threshold? a. JL b. FJK c. KL d. J 6. It is possible for a dataset to have the conditional independence assumption (of classes given features) be violated and yet have the Naive Bayes classifier provide the right classification. True or False 7. An association mining/frequent itemset algorithm is searching through an exponential space of itemsets. For instance, if there 1000 items sold in a store, it will search through 2^1000 itemsets. What makes such algorithms feasible in practice? 8. A nearest neighbor algorithm is called lazy because why? 9. If we run k-means on a dataset with 100 observations (instances) with a setting of k=5 clusters, then each of the clusters will have 20 observations. True or False? 10. If we cookup 100 data points according to the formula y = 3x +2, and feed these data points into a linear regressor (where y is the dependent variable, and x is the independent variable i.e. predictor ) that optimizes the Least Squares error, then the relationship learnt by the regressor need not necessarily be y=3x+2. True or False?