(no subject)

[watertank]

...apply knowledge representation and Semantic Web technologies to problems in discovering and integrating ecological data and data analysis techniques. These technologies rely on ontologies that appropriately capture and encode scientific knowledge from the domains of interest.

the two major uses of ontologies are accessing and analyzing ecologically important information. The first activity uses the ontologies to describe ecological and environmental data sets in sufficient detail to permit automation of the discovery of data sets relevant to addressing a particular scientific question. The second activity uses the ontologies to describe data analysis tools so that the semantic mediation system can assist in the selection of tools and creation of scientific workflows given semantic descriptions of the incoming data and/or the desired results.

Richard J. Williams, Neo Martinez and Jennifer Golbeck. Ontologies for ecoinformatics. Web Semantics: Science, Services and Agents on the World Wide Web. Volume 4, Issue 4, December 2006, Pages 237-242.

Comments

[watertank.livejournal.com]

The traits of entities and interactions that a scientist chooses to observe are typically influenced by the scientist's theories and hypotheses. The same is true of the way in which the scientist subdivides the observed world into individual entities and interactions between those entities. However, scientifically interesting traits that capture scientists’ interest typically change more frequently than the entities possessing the traits. The ontology is designed around this understanding, defining properties independently of entities and processes. This flexible architecture allows new properties to be defined and attached to entities and interactions without having to refine the underlying entity and interaction classes.

=== separating data collection from classification.
as the volume of data increases and interpretations diverge there's no way to re-do classification manually.

[watertank.livejournal.com]

Another useful example is the distinction between the terms predator and parasite. “Predator” refers to an animal that kills and eats other animals. “Parasite” also refers to an animal that consumes some species, termed the host, but this interaction occurs for an extended period of time and usually does not kill the host. Thus, predators and parasites can be distinguished by both the relative duration of the interaction between the two organisms and whether that interaction results in the immediate death of the organism being consumed.

These and many other idiosyncratic descriptions of feeding behaviors are systematically broken into several independent components concerning what is being eaten, how it is eaten, and the effect of the eating. These components include taxonomic categorization, trophic level, the part of the prey consumed, the relative duration of the feeding interaction, and whether the feeding interaction leads to the death of the prey. The meaning of feeding terms can be captured by one or more of these descriptors and terms may have more meaning than is currently captured in our feeding ontology. Still, the ontology provides a useful and relatively rich way of defining many of the terms used to describe feeding behaviors.

=== stretching the size of a predator and the time of "eating" its prey.

[watertank.livejournal.com]

Ecology is modeled as a science of entities and interactions occurring within an environment. This structure greatly facilitates the description of the many models in ecology. The model ontology allows the rules of the interactions of entities to be specified. Models are categorized using the common ecological divisions of individual, population, community and ecosystem, with further divisions within each of these broad categories.

Models are considered to be composed of model entities, interactions and parameters. Each of these is further subdivided like the underlying model, into individual, population, community and ecosystem model concepts. Parameters can be associated with entities or interactions.

== see von Bertalanffy's system definition ( objects, attributes, interactions).

[watertank.livejournal.com]

A food web is a network with directed links (DirectedNetwork), where each node represents a set of organisms and each link represents a feeding interaction, in which biomass and therefore energy is transferred from some prey or otherwise classified organisms to the consumer organisms.

== transfer of mass/energy. where's information?

[watertank.livejournal.com]

Our model of the empirical scientific process of measurement is based on two fundamental concepts: the observation and what is being observed. The observation can be of an item or its trait; in other words, a scientist either observes the existence of something or measures some trait of the entity being observed.

A trait links two instances, the Trait that was observed and the entity that the Trait was measured on. By decoupling the entities and traits, it is possible to extend the ontology by introducing a new trait without having to change the definition of the entity that was observed. This accommodates the typical mode of innovation in scientific research, in which novel traits are commonly developed, whereas the entities that the scientist studies, while also evolving as scientific understanding develops, change much more slowly. The value of a trait frequently has units associated with it, and so there is a separate ontology describing units and dimensions.

[watertank.livejournal.com]

Ecology, as a field where extracting useful data from others’ research is a challenge, is a prime example of were Semantic Web-based data sharing is potentially valuable. Currently, this work is largely independent of the web, other than as a distribution mechanism. The food web instance data used in the WoW project is collected into a single centralized knowledge base rather than being drawn from a distributed set of data. This is mainly a result of the fact that the semantic web and associated tools are in their infancy and the ecological community is not currently producing data with semantic markup and publishing it on the semantic web.


=== an information packaging issue