T/F Questions. Justify your answer. (a) Spectral clustering is the solution of a

Question

T/F Questions. Justify your answer.

(a) Spectral clustering is the solution of a normalized min cut problem.

(b) Spectral clustering doesn't require to specify the number of clusters beforehand.

(c) Graph Laplacian is positive semi-denite, i.e., all the eigenvalues of Graph Laplacian are

positive.

(d) EM algorithm is used when we have an incomplete" dataset.

(e) M step in EM algorithm corresponds to maximizing the expected log likelihood.

(f) EM algorithm is an iterative algorithm.

(g) (2 ) The E step in EM algorithm corresponds to nding the log likelihood.

(h) K-means clustering performs soft clustering and EM algorithm performs hard clustering.

(i) For a multivariate Gaussian distribution, any eigenvalue of its covariance matrix provide

variance information along the corresponding eigenvector.

(j) Each iteration of EM algorithm decreases the incomplete-data log likelihood.

Explanation / Answer

(a). True

(b). True

(c). True

(d). False

(e). True

(f). True

(g). False

(h). False

(i)

(j). False

Related Questions

Navigate

• Browse (All)
• Browse T
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.