Brain is a marvelous organ that controls all functions of the body and processes signal from the outside world. It is intuitive to hypothesize there is a relationship between body weight and brain weight for animals. In general, the larger the body weight the larger the brain weight. This relationship can be observed during the growth of any animal. While the body weight increases as an animal grows from baby to adult, its brain weight also increases. However, the relationship is non-linear. Small mammals have relatively larger brains than big ones. This fact suggests that brain weight should be modeled using the power law with power index between 0 and 1 when building transfer function involving animals with significantly different body weight. In this paper, we use the STEAMS approach to investigate the existence of body and brain relationship not for an individual species or a single animal but among many species. Statistical modeling techniques including simple linear regression, regression with transformation, non-linear regression, neutral network, sub-setting sample data into training data set and validation set are studied in this paper. The accuracy and effectiveness of each method is discussed in detail.
0 Comments
|
AuthorCategories |