Clustering of features

Clustering of objects is related to the Q technique. Usually, distance measures are used for this. Clustering of features is called R technique. The basis for that is the computation of the correlation matrix. 

Figure Cl.1.4. Photoelectron spectra of V, ,(A= 17, 27, 43, and 65) at 6.42 eV photon energy, compared to tire bulk photoelectron spectmm of V(100) surface at 21.21 eV photon energy. The cluster spectra reveal tire appearance of bulk features at and how tire cluster spectral features evolve toward tire bulk. The bulk spectmm is referenced to tire Fenni level. Wu H, Desai S R and Wang L S 1996 Phys. Rev. Lett. 77 2436, figure 2.

The largest protonated cluster of water molecules yet definitively characterized is the discrete unit lHi306l formed serendipitously when the cage compound [(CyHin)3(NH)2Cll Cl was crystallized from a 10% aqueous hydrochloric acid solution. The structure of the cage cation is shown in Fig. 14.14 and the unit cell contains 4 [C9H,8)3(NH)2aiCUHnOfiiai- The hydrated proton features a short. symmetrical O-H-0 bond at the centre of symmetry und 4 longer unsymmetrical O-H - 0 bonds to 4... 

To address the shortcoming of the simple MPI model, Medvedev and coworkers [40] developed a hybrid OpenMP/MPI method that takes advantage of both distributed and shared memory features of these clusters of multiprocessor nodes. The features of this model are ... 

The applicability of a clustering algorithm to pattern recognition is entirely dependent upon the clustering characteristics of the patterns in the representation space. This structural dependence emphasizes the importance of representation. An optimal representation uses pattern features that result in easily identified clustering of the different pattern classes in the representation space. At the other extreme, a poor choice of representation can result in patterns from all classes being uniformly distributed with no discernible class structure. 

Most of the G-protein-coupled receptors are homologous with rhodopsin however, other quantitatively minor families as well as some individual receptors do not share any of the structural features common to the rhodopsin family (Figure 2.3). The most dominant of these are the glucagon/VIP/caldtonin receptor family, or family B (which has approximately 65 members), and the metabotropic glutamate receptor family, or family C (which has approximately 15 members), as well as the frizzled/smoothened family of receptors. Thus, the only structural feature that all G-protein-coupled receptors have in common is the seven-transmembrane helical bundle. Nevertheless, most non-rhodopsin-like receptors do have certain minor structural features in common with the rhodopsin-like receptors — for example, a disulfide bridge between the top of TM-III and the middle of extracellular loop-3, and a cluster of basic residues located just below TM-VI. 

The Bayesian classifier works by building approximate probability distributions for a set of n features using examples of each class. To illustrate, if there are three classes, each described by 10 features (for the purposes of this discussion, a feature is just a real number) then the classifier will try to model three probability distributions in 10-dimensional space. These distributions can be thought of as spheres or clusters in feature space. The process... 

---