Attributes Design descriptors

We will be using the term attributes when dealing with engineering systems. An attribute is a descriptor associated with a specific feature of an engineering system, for example with the diameter of a pipe but also with the type of cross section of this pipe. Therefore, we divide attributes into 

Numerical attributes describe quantitative and measurable features of an engineering system, for example its dimensions. In the case of the pipe mentioned earlier, we could use a numerical attribute diameter of a pipe. It might have, for instance, a value D=12 . Good examples of numerical attributes are length, depth, weight, or speed. 

Symbolic attributes are a little more difficult to define and comprehend. In traditional engineering education, we talk almost exclusively about numerical attributes, and students are quite familiar with them. The concept of symbolic attributes is new. Symbolic attributes describe qualitative, abstract, or conceptual features of an engineering system, and it is for this reason that they are so important in inventive engineering. 

Symbolic attributes can be divided into nominal and structured attributes. Symbolic nominal attributes take values from an unordered set of values. For example, when we consider the symbolic nominal attribute color, it may have the values white, yellow, black, and so on, and we cannot easily order them using a simple criterion. When we consider the symbolic nominal attribute temperature, it may have the values medium, very low, high, low, and so on. These values can be easily ordered or structured and presented in a structured form as a clear sequence of values very low, low, medium, high. We can identify many symbolic attributes that 

We need to know that all numerical attributes could be converted into symbolic attributes. For example, the numerical attribute diameter of a pipe could be converted into a symbolic structured attribute with such values as very small, small, medium, large, and so on. Such conversions are often justified when we do not want to be flooded with technical details and need to develop the big picture of a given designing situation. 

---

For a collection of N compounds, the design matrix is a N X M matrix where the M columns represent a distinct molecular attribute or descriptor. A maximally diverse collection of compounds, therefore, is one that maximizes the value of D. 

Most aquatic oxidation reactions are attributable to well-defined chemical oxidants. As a result, model systems can be designed where second-order rate constants can be determined precisely for families of organic congeners. The comparatively high quality of these data allows mechanistic models of electron transfer to describe aquatic oxidations of environmental interest. Kinetic studies of these processes have produced many QSARs, mostly simple empirical correlations with common convenient descriptors such as the Hammett constant (a), half-wave oxidation potential ( j/2)> energies of the highest occupied molecular orbital ( HOMO), or rate constants for other oxidation reactions as descriptors (Canonica and Tratnyek, 2003). Their predictive power has lead to engineering applications in water treatment and remediation. 

The two main applications of molecular descriptors have been for screens in substructure searching [4] and for attributes in similarity calculations [5]. Historically, many descriptors were developed first for substructure searching (see Chapter 19 of this book) and then used in similarity calculations for applications such as similarity analysis, clustering, combinatorial library design, and data mining (see Chapters 14,... 

---