Feature selection - Subset selection

Subset selection evaluates a subset of features as a group for suitability. Subset selection algorithms can be broken up into wrappers, filters, and embedded methods. Wrappers use a search algorithm to search through the space of possible features and evaluate each subset by running a model on the subset. Wrappers can be computationally expensive and have a risk of over fitting to the model. Filters are similar to wrappers in the search approach, but instead of evaluating against a model, a simpler filter is evaluated. Embedded techniques are embedded in, and specific to, a model.

Personnel selection - Selection decisions

Predictors for selection always have less than perfect validity and scatter plots, as well as other forecasting methods such as judgmental bootstrapping, and index models can help us to refine a prediction model as well as identify any mistakes. The criterion cutoff is the point separating successful and unsuccessful performers according to a standard set by the hiring organization. True positives are applied those thought to succeed on the job as a result of having passed the selection test and who have, in fact, performed satisfactorily. True negatives describe those who were correctly rejected based on the measure because they would not be successful employees.

Algorithm selection - Online selection

Online algorithm selection in Hyper-heuristic refers to switching between different algorithms during the solving process. In contrast, (offline) algorithm selection is an one-shot game where we select an algorithm for a given instance only once.

Selection bias - Observer selection

Philosopher Nick Bostrom has argued that data are filtered not only by study design and measurement, but by the necessary precondition that there has to be someone doing a study. In situations where the existence of the observer or the study is correlated with the data, observation selection effects occur, and anthropic reasoning is required.

Cancer selection - Natural selection

Evolution, which is driven by natural selection, is the cornerstone for nearly all branches of biology including cancer biology. In 1859, Charles Darwin's book On the Origin of Species was published, in which Darwin proposed his theory of evolution by means of natural selection. Natural selection is the force that drives changes in the phenotypes observed in populations over time, and is therefore responsible for the diversity amongst all living things. It is through the pressures applied by natural selection upon individuals that leads to evolutionary change over time. Natural selection is simply the selective pressures acting upon individuals within a population due to changes in their environment which picks the traits that are best fit for the selective change.

Personnel selection - Selection decisions

False negatives occur when people are rejected as a result of selection test failure, but would have performed well on the job anyway. Finally, false positives are applied to individuals who are selected for having passed the selection measure, but do not make successful employees. These selection errors can be minimized by increasing the validity of the predictor test.

Selection bias - Observer selection

An example is the past impact event record of Earth: if large impacts cause mass extinctions and ecological disruptions precluding the evolution of intelligent observers for long periods, no one will observe any evidence of large impacts in the recent past (since they would have prevented intelligent observers from evolving). Hence there is a potential bias in the impact record of Earth. Astronomical existential risks might similarly be underestimated due to selection bias, and an anthropic correction has to be introduced.

Selection cutting - Group selection

Another variation of selection silviculture is group selection. Under this system, a number of 'groups', or small openings created by the removal of several adjacent trees, are created in complement to the harvest of scattered individual trees. If the groups created are large enough, and if seed-bed conditions are favorable, this can allow species which are intolerant of shade to regenerate. Group selection is designed to mimic larger, multi-tree mortality events, which in some environments may represent natural disturbance regimes.

Selection (linguistics) - S-selection vs. c-selection

One sometimes encounters the terms s(emantic)-selection and c(ategory)-selection. The concept of c-selection overlaps to an extent with subcategorization. Predicates c-select the syntactic category of their complement arguments - e.g. noun (phrase), verb (phrase), adjective (phrase), etc. - i.e. they determine the syntactic category of their complements. In contrast, predicates s-select the semantic content of their arguments. Thus s-selection is a semantic concept, whereas c-selection is a syntactic one. When the term selection or selectional restrictions appears alone without the c- or s-, s-selection is usually understood.

Selection algorithm - Selection by sorting

By sorting the list or array then selecting the desired element, selection can be reduced to sorting. This method is inefficient for selecting a single element, but is efficient when many selections need to be made from an array, in which case only one initial, expensive sort is needed, followed by many cheap selection operations – O(1) for an array, though selection is O(n) in a linked list, even if sorted, due to lack of random access. In general, sorting requires O(n log n) time, where n is the length of the list, although a lower bound is possible with non-comparative sorting algorithms like radix sort and counting sort.

Ecological selection - Vs. sexual selection

In cases where ecological and sexual selection factors are strongly at odds, simultaneously encouraging and discouraging the same traits, it may also be important to distinguish them as sub-processes within natural selection.

Group selection - Multilevel selection theory

Wilson ties the multilevel selection theory regarding humans to another theory, gene-culture coevolution, by acknowledging that culture seems to characterize a group-level mechanism for human groups to adapt to environmental changes.

Group selection - Multilevel selection theory

MLS theory can be used to evaluate the balance between group selection and individual selection in specific cases. An experiment by William Muir compared egg productivity in hens, showing that a hyper-aggressive strain had been produced through individual selection, leading to many fatal attacks after only six generations; by implication, it could be argued that group selection must have been acting to prevent this in real life. Group selection has most often been postulated in humans and, notably, eusocial Hymenoptera that make cooperation a driving force of their adaptations over time and have a unique system of inheritance involving haplodiploidy that allows the colony to function as an individual while only the queen reproduces.

Ecological selection - Vs. sexual selection

For another example, in a region devastated by nuclear radiation, such as the Bikini Atoll, capacity to survive gamma rays to sexual maturity and (for the female) to term is a key ecological selection factor, although it is neither "natural" nor sexual. Some would call this too artificial selection, not natural or ecological, as the radiation does not enter the ecology as a factor save due to man's effort. Ambiguous artificial-plus-ecological factors may reasonably be called "environmental", and the term environmental selection may be preferable in these cases.

Balancing selection - Frequency-dependent selection

Frequency-dependent selection occurs when the fitness of a phenotype is dependent on its frequency relative to other phenotypes in a given population. In positive frequency-dependent selection the fitness of a phenotype increases as it becomes more common. In negative frequency-dependent selection the fitness of a phenotype decreases as it becomes more common. For example, in prey switching, rare morphs of prey are actually fitter due to predators concentrating on the more frequent morphs. As predation drives the demographic frequencies of the common morph of prey down, the once rare morph of prey becomes the more common morph. Thus, the morph of advantage now is the morph of disadvantage. This may lead to boom and bust cycles of prey morphs. Host-parasite interactions may also drive negative frequency-dependent selection, in alignment with the Red Queen hypothesis. For example, parasitism of freshwater New Zealand snail (Potamopyrgus antipodarum) by the trematode Microphallus sp. results in decreasing frequencies of the most commonly hosted genotypes across several generations. The more common a genotype became in a generation, the more vulnerable to parasitism by Microphallus sp. it became. Note that in these examples that no one phenotypic morph, nor one genotype is entirely extinguished from a population, nor is one phenotypic morph nor genotype selected for fixation. Thus, polymorphism is maintained by negative frequency-dependent selection.

Feature selection - Correlation feature selection

The Correlation Feature Selection (CFS) measure evaluates subsets of features on the basis of the following hypothesis: "Good feature subsets contain features highly correlated with the classification, yet uncorrelated to each other". The following equation gives the merit of a feature subset S consisting of k features:

Selection algorithm - Online selection algorithm

Online selection may refer narrowly to computing the kth smallest element of a stream, in which case partial sorting algorithms (with k + O(1)) space for the k smallest elements so far) can be used, but partition-based algorithms cannot be.

Selection algorithm - Partition-based selection

Linear performance can be achieved by a partition-based selection algorithm, most basically quickselect. Quickselect is a variant of quicksort – in both one chooses a pivot and then partitions the data by it, but while Quicksort recurses on both sides of the partition, Quickselect only recurses on one side, namely the side on which the desired kth element is. As with Quicksort, this has optimal average performance, in this case linear, but poor worst-case performance, in this case quadratic. This occurs for instance by taking the first element as the pivot and searching for the maximum element, if the data is already sorted. In practice this can be avoided by choosing a random element as pivot, which yields almost certain linear performance. Alternatively, a more careful deterministic pivot strategy can be used, such as median of medians. These are combined in the hybrid introselect algorithm (analogous to introsort), which starts with Quickselect but falls back to median of medians if progress is slow, resulting in both fast average performance and optimal worst-case performance of O(n).

Ecological selection - Vs. sexual selection

Differentiating ecological selection from sexual is useful especially in such extreme cases; Above examples demonstrate exceptions rather than a typical selection in the wild. In general, ecological selection is assumed to be the dominant process in natural selection, except in highly cognitive species that do not, or do not always, pair bond, e.g. walrus, gorilla, human. But even in these species, one would distinguish cases where isolated populations had no real choice of mates, or where the vast majority of individuals died before sexual maturity, leaving only the ecologically selected survivor to mate—regardless of its sexual fitness under normal sexual selection processes for that species.

Group selection - Multilevel selection theory

in which b k is the benefit to kin (b in the original equation) and b e is the benefit accruing to the group as a whole. He then argues that, in the present state of the evidence in relation to social insects, it appears that b e >rb k, so that altruism needs to be explained in terms of selection at the colony level rather than at the kin level. However, kin selection and group selection are not distinct processes, and the effects of multi-level selection are already accounted for in Hamilton's rule, rb>c, provided that an expanded definition of r, not requiring Hamilton's original assumption of direct genealogical relatedness, is used, as proposed by E. O. Wilson himself.