Other Methods

Many other methods have also been used in functional genomics analysis. Only a short list is included here, with references to published biomedical works incorporating each technique, and using the hierarchy proposed at the start of this chapter:

1. Unsupervised.

a. Feature determination.

i. Principle component analysis and singular value decomposition [189, 95, 68, 149, 5].

b. Cluster determination.

i. Nearest neighbor clustering.

ii. Agglomerative clustering.

A. Fitch dendrogram algorithm [189].

B. Self-organizing tree algorithm [92].

iii. Divisive or partitional clustering.

A. Matrix incision tree [108].

B. Two-way clustering binary tree [4].

C. Coupled two-way clustering [76].

D. Cluster affinity search technique: [20].

c. Network determination.

i. differential equations [44].

ii. Bayesian networks [71].

iii. Hybrid petri networks [128].

iv. Dynamic bayesian networks [134].

v. Boolean regulatory networks [119, 197, 172, 1, 2].

2. Supervised.

a. Single feature or sample determination.

i. Naive bayes classifier [21, 46].

ii. Naive Bayes global relevance [133].

b. Multiple-feature determination.

i. Decision trees: Supervised and multiple-feature determination [59].

ii. Support vector machines [72, 33, 46].

iii. Tree harvesting [86].

Was this article helpful?

0 0

Post a comment