Classification results

The three introduced network structures were trained with the training data set and tested with the test dataset. The backpropagation network reaches its best classification result after 70000 training iterations ... 

Table 2 Classification results of the Backpropagation Network in percent...

Often the goal of a data analysis problem requites more than simple classification of samples into known categories. It is very often desirable to have a means to detect oudiers and to derive an estimate of the level of confidence in a classification result. These ate things that go beyond sttictiy nonparametric pattern recognition procedures. Also of interest is the abiUty to empirically model each category so that it is possible to make quantitative correlations and predictions with external continuous properties. As a result, a modeling and classification method called SIMCA has been developed to provide these capabihties (29—31). 

The method has several advantages, the first being its mathematical simplicity, which does not prevent it from yielding classification results as good and often better than the much more complex methlods discussed in other sections of this chapter. Moreover, it is free from statistical assumptions, such as normality of the distribution of the variables. 

In general, there has been good agreement between PyMS-derived classification results and results from standard taxonomic methods. This agreement was seen from comparative studies on Carnobacterium,57 Fusobacterium,5s Pep-tostreptococcus,59 Photobacterium,60 Prevotella,61 Rothia,62 Streptococcus,63 and Streptomyces.m PyMS was also used, together with molecular and numerical... 

Classification of the measured variables included in NA1 and NA2 as redundant. The other measurements are categorised as nonredundant. Measured variable classification results for this example are in Table 7. 

Dry bean flour fractions produced by dry roasting, milling and air classification resulted in versatile food ingredients. Fractions possessed good functional and nutritional properties which were found to be acceptable in a variety of food systems. These processes and products appear to have potential for improving nutritive status through improved dry bean utilization. 

Classification Results of the Glass Vessels Data Using LDA... 

The classification is based on the different conversion concepts cited in the literature, such as fuel-bed mode (batch and continuous), fuel-bed configuration (cocurrent, countercurrent, and crosscurrent), and some new concepts presented by the author, such as fuel-bed movement (fixed, moving, and mixed) and fuel-bed composition (homogeneous and heterogeneuos). The classification resulted in 18 types of updraft conversion systems, according to Figure 32. Some of them are more or less hypothetical, while others are found in practice, see section B 3.4 below. 

Cluster analysis Is used to determine the particle types that occur in an aerosol. These types are used to classify the particles in samples collected from various locations and sampling periods. The results of the sample classifications, together with meteorological data and bulk analytical data from methods such as instrunental neutron activation analysis (INAA). are used to study emission patterns and to screen samples for further study. The classification results are used in factor analysis to characterize spatial and temporal structure and to aid in source attribution. The classification results are also used in mass balance comparisons between ASEM and bulk chemical analyses. Such comparisons allow the combined use of the detailed characterizations of the individual-particle analyses and the trace-element capability of bulk analytical methods. 

A particilarly powerful use of the classification results is in factor analysis. This will help to uncover interrelationships anong the particle types and will provide additional information for source attribution. The results of the factor analysis are also helpful for judging the significance of the cluster analysis, in that if the occupations of two similar particle types are uncorrelated over several samples then this indicates that the particle types and the clusters from which they are derived are significantly different. 

Table II. Classification Results for Chandler. Arizona, as percent of total particles classified.

Since SIMCA is a class modeling method, class assignment is based on fit of the unknowns to the class models. This assignment allows the classification result that the unknown is none of the described classes, and has the advantage of providing the relative geometric portion of the newly classified object. This makes it possible to assess or quantitate the test sample in terms of external variables that are available for the training sets. 

Several paediction diagnostic tools are discussed below and a summaiy is found at fije end of the section in Table These tools are used to assess the reliabfflty of the classification results. 

Raw Measurement Plot If an on-line analyzer were soning the containers, unknown 3 would have been rejected and the spectrum stored for further evaluation. The spectrum of this unknown is plotted in Figure 4.56 along with the average of the training set spectra for PVC and PET. Unknown 3 has features from both of these classes and, therefore, the classification results are not surprising. 

We list here some classification results for maps on the sphere or on the plane. 

We now list some existence and classification results for isohedral (r, q)-polycycles obtained by using the previous formalism. 

The following theorem, which is a slight generalization of Theorem 5.2.1, is very helpful in deriving classification results. 

The major skeleton, elementary polycycles, and classification results... 

As evidenced by these data, the impact of the carotenoid features on the classification results is remarkable. Photodestruction of the carotenoid components is an efficient alternative to omitting all spectral regions where carotenoid bands may superimpose other features the latter being proposed, e.g. in studies on microorganisms [61]. The width of some of the carotenoid bands would force us to discard numerous important other spectral features hidden beneath the intense carotenoid signals. 

---