|
|
|
|
|
|
|
|
|
Bound and collapse
Глоссарий по искусственному интеллекту |
Bound and collapse is a two-step algorithm for learning a bayesian network (bn) in databases with incomplete data. the two (repeated) steps are bounding of the estimates with values that are consistent with the current state, followed by a collapse of the estimate bounds using a convex combination of the bounds. implemented in the experimental program bayesian knowledge discoverer. see also: bayesian knowledge discover, http://kmi.open.ac.uk/projects/bkd/
|
|
Incomplete, английский
A неполный application incongruence n псхл. неконгруэнтность
Consistent, английский
Combination, английский
- Комбинация, сочетание
- Комбинация, сочетание сомсм communications counter- measures меры противодействия радиосвязи противника 163
- Сочетание; комбинация (сочетание взаимосвязанных признаков изобретения).
- Сочетание; комбинация (сочетание взаимосвязанных признаков изобретения)
- Отношение
- N сочетание word ~ словосочетание
- Комбинация; соединение; сочетание; объединение; набор; система
- Сочетание; соединение; комбинация; смешение; состав ~ of errors сумма ошибок load ~ сочетание нафузок
- Series of two or more fences within 39 feet 4 inches of each other that must be taken as a pair, an in-and-out.
- Applies to derivative products. arrangement of options involving two long or two short positions with different expiration dates or strike (exercise) prices. see: straddle.
- Мотоцикл с прицепной коляской
- Vehicle consisting or two or more separable units, of which each part need not be independently mobile, for example a motorcycle and side
Experimental, английский
- Экспериментальный
- A эксперимен- тальный phonetics
Discoverer, английский
|
Boblo, английский
Boblo is an expert system based on bayesian networks used to detect errors in parental identification of cattle in denmark. the model includes both representations of genetic information (rules for comparing phenotypes) as well as rules for laboratory errors. see also: graphical model.
An example boxplot, английский
Branch-and-bound search branch-and-bound searches are used to improve searches through a tree representation of a solution space. as the algorithm progresses through a tree, it maintains a list of all partial paths that have been previously evaluated. at each iteration, it chooses the best (lowest cost) path that is currently known and expands that to its next level, scoring each of the new possible paths. these new paths are added to the list of possible paths replacing their common ancestor, and the process is reevaluated at the current best path. when a solution has been found, it may be improved by reevaluating the stored paths to eliminate more expensive solutions. the remaining paths can then be evaluated until they provide either a better solution or become more expensive than the best known solution.
|
|
|
|
|
|
|