Decision trees defined

Data, computer versus human view, 56-58 Data sets, see Example data sets DCLS, seeDirea da ical least squares (DCLS) Decision trees defining the problem, 9-11 muiiivaiiate calibration, 186-188 pattern recognition. 62-64 Definitions. 5-7 Degrees of freedom PLS F-test. 304 SLMCA F-test. 152-153 Dendrogram construaion of. 65-71 definition of 65... 

Classification describes the process of assigning an instance or property to one of several given classes. The classes are defined beforehand and this class assignment is used in the learning process, which is therefore supervised. Statistical methods and decision trees (cf. Section 9.3) are also widely used for classification tasks. 

This task consists of defining target levels for indoor and outdoor conditions based on requirements for laws and regulations, human health, production processes and equipment, and type of premises and construction. Target levels should also be defined for the ventilation system. For the decision tree, see Fig. 3.6. 

Although decision trees contain a number of attractive features, including competitive accuracy, when considered strictly as classification devices (Saraiva and Stephanopoulos, 1992a), the most important point for our purposes is that each of the tree s terminal nodes identifies a particular hyperrectangle in the decision space, X, associated with a given y value. For example, node M defines a y = excellent rectangle that corresponds to the following rule ... 

In Fig. 4 we present the final induced decision tree, as well as the partition of the (xi,X4) plane defined by its leaves, together with a projection of all the available (x, y) pairs on the same plane. These two decision variables are clearly influencing the current performance of the refinery unit, and the decision tree leaves perform a reasonable partition of the plane. To achieve better performance, we must look for operating zones that will result in obtaining mostly y = 3 values. Terminal nodes 2... 

Fig. 4. (a) Induced decision tree (b) partition of the plane defined by its leaves. 

Two groups of objects can be separated by a decision surface (defined by a discriminant variable). Methods using a decision plane and thus a linear discriminant variable (corresponding to a linear latent variable as described in Section 2.6) are LDA, PLS, and LR (Section 5.2.3). Only if linear classification methods have an insufficient prediction performance, nonlinear methods should be applied, such as classification trees (CART, Section 5.4), SVMs (Section 5.6), or ANNs (Section 5.5). 

The decision tree method of Cramer et al. (1978) was based on the toxicological data then available and used a series of 33 questions, each leading either to another question or to classification into one of three classes of presumptive toxicity. The questions were primarily based on chemical stmcmre, but namral occurrence in body tissues or fluids as well as natural occurrence in traditional foods was also considered. The three classes of substances were defined as follows ... 

A generalized controller operational flowchart is included in Figure 2. This flowchart shows relative battery state-of-charge (SOC) to be the starting criteria for the decision tree. SOC is defined as the ratio of ampere-hours (Ah) stored in the battery array at any given time to the maximum Ah capacity of the array. During this process the actual controller subroutine also includes the voltage and current that is available and/or required by each component. However, for the purpose of explanation, the power available or required by various system components can be related to battery SOC. 

Figure 23.7 shows a simple example of a decision tree where the majority of active compounds can be defined by a single set of rules. In more complex trees, one active group of compounds may be defined by a set of rules which state... 

At the start of such a search a detailed market analysis is required. The assessment of the technology position is needed, and the company s own strategy must be defined. To help in this decision making it is useful to collect the knowledge of the whole organization into a decision tree on which decisions are clearly framed and the criteria for making any particular decision are made clear [11]. The next section pro-... 

The assumption of a constant contribution to activity of a particular substituent is embodied in modern computational off-shoots of QSAR. For example, one report described a decision tree that separates active from inactive molecules by using the presence or count of pre-defined substructures as descriptors of the molecules. 

---