|
|
|
|
|
|
|
|
|
Mita
| Глоссарий по искусственному интеллекту |
- See: metlife`s intelligent text analyzer. mixture-of-experts (me) models a mixture-of-experts (me) model is a technique that allows a model to include various submodels or ``experts" within the overall model. the experts are combined by a gate function that takes the individual experts` outputs and combines them to provide a final output. these experts are a form of local model that are optimal over restricted sub-domains of the total problem domain. their combination can lead to more accurate models than a single "global" model. an example could be a medical diagnosis program, where each of the experts is a model for a particular disease type. each expert could then predict the likelihood that a given patient has the special disease. the gate function could then combine the individual predictions using a "softmax" function or some other voting function. an extension of the me model is a hierarchical me (hme) model. in this model, the experts are arranged in a tree, so that each expert`s model is a mixture of the models "below" it on the tree. in this sense, classification and regression trees (cart) or a decision tree is a very primitive form of hme. see also: gate function.
- Multilateral interline ticket agreements
|
|
Intelligent, английский
Интеллектуальный
Individual, английский
Физическое лицо
Restricted, английский
Placed on a list that dictates that the trader may not maintain positions, solicit business, or provide indications in a stock, but may serve as broker in agency trades after being properly cleared. traders are so restricted due to investment bank involvement with the company on nonpublic activity (i.e., mergers and acquisitions defense), affiliate ownership, or underwriting activities; signified on the quotron by a flashing "r." a restricted list and the stocks on it should never be conveyed to anyone outside of the trading areas, much less outside the firm. see: grey list.
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
Particular, английский
A частный; конкретный grammar partitioned a разделённый на части sequence 1 одна из базовых операций компьютерной обработки текстов; в каче- стве данных берутся текст либо сегмент текста, грамматика и цель разбора, а на выходе получается удовлетворяющее цели множество значений категориальной отнесённости сегментов в виде, например, дерева разбора. 2 применяется запоминание также и гипотез, выдвигаемых при разбо- ре, и результатов их проверки – м. кэй. 3 используется, в частности, для разбора арифметических выражений. partitioning n членение sentence ~ членение предложений (напр., при переводе )
Likelihood, английский
- Вероятность. правдоподобие.
- Правдоподобие; вероятность
- In probabilistic and statistical methods, the likelihood is a measure of the evidence for the data when a hypothesis is assumed to be true. it is commonly used in bayesian and quasi-bayesian techniques. a related measure, minimum message length, can be derived from information theoretic principles. as an example, suppose you observe a data value x that is assumed to be gaussian. the likelihood for that data when the assumed mean is, say, 5, and the variance is 10, is proportional to e-((x-5)2/2*10), the kernel of a gaussian distribution. a likelihood is typically calculated using the kernel, rather than the complete distribution function, which includes the normalizing constants. see also: minimum message length.
|
Mse criterion, английский
See: mean square error criterion. multi-agent system (mas) see: distributed artificial intelligence. multi-class many classification systems are developed for binary data. multi-class variables are those with more than two categories. this form of variable is also known as multinomial. see also: adaboost.mh, binomial distribution, multi-label. multi-label most class variables belong to a single class. target variables that can belong to more than one class have been called multi-label. an example would be classifications on textual objects that can belong to multiple categories. see also: adaboost.mh, multi-class. multidimensional database (mddb) mddb is a case based reasoning system that is used to diagnosis dysmorphic syndromes, an area with poor medical knowledge by rich information. the system is based on records of over 3000 patients. mddb is commonly used in online analytical processing (olap) and related systems. rather than treat a database of multi-attribute records as a two-dimensional table with rows as records and the attributes (variables) as columns, the data is organized as a k-dimensional rectangle, with one dimension for each attribute. the dimension corresponding to the i -th attribute has ni levels, each corresponding to one of the values that the i -th attribute can assume in this representation. a cell at the intersection of specified values for each of the k attributes contains summary data on all the records that are classified as belonging to that cell. typically, the marginal cells (cells where one or more dimensions have been suppressed) contain information on all of the cells in the missing dimension(s). see also: online analytic processing, http://www.med.unirostock. de/html/english/research/gs52.index.html
Mdlp, английский
See: minimum description length principle.
|
|
|
|
|
|
|
