China
Africa
Explore chapters and articles related to this topic
Motoo Kimura (1924–1994)
Published in Krishna Dronamraju, A Century of Geneticists, 2018
Tomoko Ohta emphasized the importance of nearly neutral mutations, in particular, slightly deleterious mutations. The population dynamics of nearly neutral mutations is essentially the same as that of neutral mutations unless the absolute magnitude of the selection coefficient is greater than 1/N, where N is the effective population size with respect to selection. The value of N may therefore affect how many mutations can be treated as neutral and how many as deleterious.
A novel bagging approach for variable ranking and selection via a mixed importance measure
Published in Journal of Applied Statistics, 2018
Chun-Xia Zhang, Jiang-She Zhang, Guan-Wei Wang, Nan-Nan Ji
Variable selection is always an important topic in statistical modeling since it can significantly enhance estimation and prediction accuracy, and improve interpretability. Currently, large amounts of high-dimensional data have emerged in many research and application areas because it is much more convenient to collect data than ever before. However, more variables do not necessarily imply that higher prediction accuracy can be obtained. At the same time, a large number of variables often prevent us from explaining how the exploratory variables influence our interested outcome. When facing with high-dimensional data, it is commonly believed that there are only a few variables playing an important role, namely, the model is usually sparse. As a result, it is particularly important to detect truly important variables to achieve high accuracy for prediction purpose or to identity relevant variables for interpretation purpose. In order to achieve this objective, variable selection has become an indispensable tool. At present, a large variety of variable selection techniques have been developed, such as subset selection [2,18], coefficient shrinkage [6,8,10,22,24,37,38], variable screening [9,15] and so on. Similar to the common practice in variable selection literature [2,6,8–10,15,18,22,24,37,38], we focus on variable selection in linear regression models in this paper. Generally speaking, a linear regression model can be expressed as n observations of the covariates 1), σ is unknown.