Immunity-Based Computational Models - Immunological Computation: Theory and Applications - page 46

Table 3.1 Immunity-Based Computational Models and Specifi c

Immunological Concepts

Immunological Concepts
and Entities

Computational
Problem Immunity-Based Models

Self or nonself
recognition T cell

Negative selection algorithms
(Forrest et al., 1994)

Anomaly, fault, and
change detection

Idiotypic network,
immune memory,
and B cell

Immune network theory
(Hunt and Cooke, 1995)

Learning (supervised
and unsupervised)

Clonal expansion, affi nity
maturation, and B cell

Clonal selection algorithm
(De Castro and Von Zuben,
2000)

Search and
optimization

DT (Aickelin and Cayzer, 2002) Innate immunity Defense strategy

Table 3.2 Machine Learning versus Immunology Terminology

Machine Learning Immune Models

Detectors, clusters, classifi ers,
and strings

T cells, B cells, and antibodies

Positive samples, training data,
and patterns

Self-cells, self-molecules, and immune
cells

Incoming data, verifying data
samples, and test data

Antigens, pathogens, and epitopes

Distance and similarity measures Affi nity measure in the shape-space

String-matching rule Complementary rule and other rules

Shape − space

Immunity-based

processes

Affinity

Immunological models

Figure 3.1 Major components of immunity-based models.

3.1 Shape-Space and Affi nity
Perelson and Oster (1979) introduced a shape-space (or representation space) con-
cept to represent antibody or antigen binding (see Figure 3.2). Accordingly, anti-
gens and antibodies are characterized by their physicochemical binding properties,
which are represented as coordinate points in such space, typically, a Euclidean