Combining similarity measures

The second category of matchers use machine learning techniques to combine similarity measures. However, they share almost the same parameters than the first category. 

MatchPlanner is a schema matcher based on decision trees to combine similarity measures. Although it has been extended by machine learning techniques to generate these decision trees [Duchateau 2009], users can provide their own decision trees to the system [Duchateau et al. 2008a], There is no weight on the measures, but their order and position in the decision tree are crucial. Manually designing such decision trees is interesting when one wants to promote time performance or use a specific similarity measure. 

To sum up, many tools have designed their own strategies to combine similarity measures. However, most of them are based on weighted functions that the users may have to tune. 

The previous section gathers tools that support users to manually find the best strategy, i.e., the best method for combining similarity measures and its optional parameters such as weights. This last part is dedicated to the tools that automatically discover the best strategy eTuner and YAM. 

In general, different similarity measures yield different rankings, except when they are monotonic. Improved results are obtained by using data fusion methods to combine the rankings resulting from different coefficients. 

For specimens where gradients in the ms etic moment are of interest, similar arguments apply. Here, however, two separate reflectivity experiments are performed in which the incident neutrons are polarized parallel and perpendicular to the surfiice of the specimen. Combining reflectivity measurements under these two polarization conditions in a manner similar to that for the unpolarized case permits the determination of the variation in the magnetic moments of components parallel and perpendicular to the film surface. This is discussed in detail by Felcher et al. and the interested reader is referred to the literature. 

Similar to supply inlets, no measurements exist for evaluating the inlets specific influence on contaminant concentration. The available measurements for the combinations are the same as for exhaust hoods, i.e., capture efficiencies and similar measures. Sometimes the performance of a combined system can be approximated from the performance of the incoming supply inlet and exhaust hood. 

Pan, D., Iyer, M., Liu, J., Li, Y., Hopfinger, A. J., Constructing optimum blood brain barrier QSAR models using a combination of 4D-molecular similarity measures and cluster analysis descriptors. J. Chem. Inf. Model. 2004, 44, 2083-2098. 

The result of the clustering procedure depends on which procedure is applied and on the similarity measures used. Each gives a different view of the complex reality in the data set. It is therefore highly recommended that a clustering method is combined with a PC A or PLS display (see Chapters 17, 31 and 35) and, if possible, that several clustering methods and several types of similarity are used. 

Vigorous hydrolysis of neomycin results in the formation of furfural which may be quantitated in a number of different ways- 5 156. Dutcher et all57 use(j a UV spectrophotometric procedure, measuring the absorbance of the solution at 280nm. A similar procedure combined with measurement of the optical rotation has been used by Brooks et all58 to determine the neomycin B and C content... 

Ginn CMR, Willett P, Bradshaw J. (2000) Combination of Molecular Similarity Measures Using Data Fusion. Perspect. Drug Discov. Des. 20 1-16. 

Ginn, C. M. R., Willett, R, and Bradshaw, J. (2000) Combination of molecular similarity measures using data fusion. Perspec. Drug Disc. Design 20, 1-16. 

Because there are phases but only (7 - 1) interfaces, this experiment alone cannot provide sufficient information to calculate all the Dej given the phase diagram and measurements of interfacial positions as a function of time. However, when its results are combined with measured times for complete dissolution in water of individual drops of surfactant of known initial radius injected through glass micropipettes into similar cells placed on a conven-... 

An interesting question is how H3 is formed on the emitter surface and whether H3 molecules can exist on the surface. This question can be investigated with a measurement of the appearance energy of Hj ions. Jason etal.264 find Hj in field ionization of condensed layers of hydrogen, and measure the appearance energy to be 12.7 eV. This value is 2.9 eV smaller than that of H2. Ernst Block conclude265 from a similar measurement in field ionization mass spectrometry of hydrogen that an H 3 ion is formed at the moment when a chemisorbed H atom combines... 

Abstract. A multi-wavelength, multi-sensor Look-Up-Table (LUT) algorithm has been developed to retrieve information about stratospheric aerosols from satellite-based observations of particulate extinction. Specifically, the LUT algorithm combines extinction measurements from SAGE n with similar measurements from the CLAES instrument, and uses the composite spectra in month-latitude-altitude bins to retrieve values and uncertainties of particle effective radius, surface area, volume and size distribution width. 

The extension of equilibrium measurements to normally reactive carbocations in solution followed two experimental developments. One was the stoichiometric generation of cations by flash photolysis or radiolysis under conditions that their subsequent reactions could be monitored by rapid recording spectroscopic techniques.3,4,18 20 The second was the identification of nucleophiles reacting with carbocations under diffusion control, which could be used as clocks for competing reactions in analogy with similar measurements of the lifetimes of radicals.21,22 The combination of rate constants for reactions of carbocations determined by these methods with rate constants for their formation in the reverse solvolytic (or other) reactions furnished the desired equilibrium constants. 

---