Wrapper method

SVM-BFS — Application of Wrapper Method and Floating Search Method... 

Wrapper methods utilize the prediction ability of some learning machine (SVM in this book) to evaluate the feature subset. Compared with other methods, wrapper method can assure to get a feature subset with higher accuracy by using the specified learning machine (SVM here). The principle of wrapper methods is to minimize... 

Floating search method proposed by Pudil et al. [112] is used as feature subset generation procedure, and has been proved to be one of the best subset generation algorithms for moderately large or small data sets. Backward floating search (BFS) method needs more computer time than forward floating search method, but it can treat interactive features well, so we combine it with SVM based on wrapper method and name it SVM-BFS method. 

Nowadays, as the rapid development of data collection techniques, more and more data are collected, and wrapper methods cannot meet the need of rapid data processing. Then, a number of input pruning methods using SVM have been developed. Guyon and his co-workers proposed to use optimal brain damage as the feature evaluation method and combined it with recursive feature elimination to perform gene selection [64],... 

Note The outline presented by Pierson is, based on the method of W. Huff, M. Halik l F. Pristera described in PATR 2472 (Dec 1957) of the Feltman Research Laboratories, Picatinny Arsenal, Dover, NJ and also on the paper of F. Pristera (of PicArsn) in Applied Spectro-j scopy 7, 115 (1953) (See also Ref 18, pp7-9) Analysis of Wrapper for paraffin, NG, paper] ash, and moisture is described in detail, by Pierson (Ref 11, pi364) [... 

B. Physical Tests, Which include a)Apparent specific gravity and wrapper-to -explosive ratio. These values are detd for several cartridges taken at random from each shipment of expls. Methods of testing are... 

A method has been reported for the determination of calcium, copper, iron, magnesium, potassium, sodium and zinc in cellulose [169]. The sample (10 g) was air-dried and then ashed at 575°C until all the carbon was removed. Hydrochloric acid (5 ml of 6M) was added to the residue and evaporated to dryness twice before taking up the sample in a third aliquot, diluting to 100 ml and aspiration into an air /acetylene flame. It is likely that volatile elements such as cadmium may be lost at such an elevated ashing temperature and temperatures below 500°C may be preferable. Alternatively wet ashing with nitric acid has been proposed for the determination of aluminium, cadmium, potassium and zinc in pressed boards [170] or sodium in gypsum glass board [171]. For the determination of lead in confection wrappers, the sample may be treated with concentrated nitric acid at 70—80°C and diluted for flame analysis [172]. In the full method, the wrapper was wiped clean with a damp tissue, cut up to 0.5 X 0.5 mm pieces and dried at 110°C (for paper, for plastic 80°C) for 1 h. The sample (0.5 g) was heated with concentrated nitric acid (1ml) at... 

For this, the internal behavior of public and protected methods" " is modeled by collaboration diagrams (cf. top-right quadrant of Fig. 5.62). Such collaboration diagrams describe on the one hand, how the homogenizing wrapper s data structure is modified. Therefore, the private methods (see phase 3), e.g. for instantiating an object, are used. On the other hand, the methods of the technical wrapper s export interface (see phase 2) are used to manipulate the wrapped tool accordingly. 

Based upon this trace, three methods, constituting the technical wrapper s export interface, are specified by selecting specific subsequences out of the trace ... 

The intended homogenized data model, realized by the wrapper to be designed, is addressed in Fig. 5.65. The class Document represents a Comos PT flowsheet, indicated by the return value of method id(), which returns a reference to an instance of class IComosDDocument. 

Figure 5.66 demonstrates the final step of wrapper development, namely the specification of the mapping between the given data model (cf. Fig. 5.63) and that of the homogenizing wrapper (cf. Fig. 5.65) . Therefore, the methods of the technical wrapper s export interface (cf. Fig. 5.64) and the private methods of the homogenizing wrapper s classes (cf. Fig. 5.65) are applied.

Firstly, within method speciflcation addDevicesO (cf. box 1 of Fig. 5.66) the reference to the instance of class IComosDDocument, representing the flowsheet, is queried (1.1) and the technical wrapper is asked for the number of devices within this flowsheet (1.2). Next, a loop is defined (1.3), bounded by the return value of (1.2), calling method addDeviceO of class IComosDDocument in each iteration. The loop variable, indicating the index of the currently handled device, serves as input parameter for method addDeviceO. 

Finally, method checkDeviceO of class Type has to be modeled (cf. box 3 of Fig. 5.66). Again the reference to the instance of class IComosDDocument is queried (1.1) and the technical wrapper is asked for a reference to the... 

Until now, our approach to model-driven wrapper development deals with tools and applications offering COM interfaces. Conceptually, COM is a component technology following the object-oriented paradigm. COM components represent subsystems, i.e. a set of classes including attribute and method definitions, relationships between classes, and a set of interfaces to access the subsystem. COM offers an external description of application interfaces by means of a type library. Such interface descriptions of components included in a type library serve as input for the analyzer to comprehend a component and its various parts syntactically. 

---