Equivalent Number of Degrees of Freedom for Neural Networks

نویسندگان

  • Salvatore Ingrassia
  • Isabella Morlini
چکیده

In Ingrassia and Morlini (2005) we have suggested the notion of equivalent number of degrees of freedom (e.d.f.) to be used in neural network modeling from small datasets. This concept is mainly based on some theoretical results given in Bartlett (1998). It is much smaller than the total number of parameters and it does not depend on the number of input variables. We generalize our previous results and discuss the use of e.d.f. in the general framework of multivariate nonparametric model selection. Through numerical simulations, we also investigate the behavior of model selection criteria like AIC, GCV, BIC/SBC and UEV, when the e.d.f. is used instead of the total number of the adaptive parameters in the model.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Kinematic Synthesis of Parallel Manipulator via Neural Network Approach

In this research, Artificial Neural Networks (ANNs) have been used as a powerful tool to solve the inverse kinematic equations of a parallel robot. For this purpose, we have developed the kinematic equations of a Tricept parallel kinematic mechanism with two rotational and one translational degrees of freedom (DoF). Using the analytical method, the inverse kinematic equations are solved for spe...

متن کامل

The Performance of an Hexahedron C* Element in Finite Element Analysis

The performance of an 8-noded hexahedron C1* element in elasticity is investigated. Three translational displacements and their derivatives as strain in each direction are considered as degrees of freedom (d.o.f.’s) at each node. The geometric mapping is enforced using a C0 element with no derivative as nodal d.o.f.’s . The stiffness matrix of the element is also computed using a transformation...

متن کامل

Mechanical System Modelling of Robot Dynamics Using a Mass/Pulley Model

The well-known electro-mechanical analogy that equates current, voltage, resistance, inductance and capacitance to force, velocity, damping, spring constant and mass has a shortcoming in that mass can only be used to simulate a capacitor which has one terminal connected to ground. A new model that was previously proposed by the authors that combines a mass with a pulley (MP) is shown to simulat...

متن کامل

Estimating Equivalent Kernels for Neural Networks: A Data Perturbation Approach

We describe the notion of "equivalent kernels" and suggest that this provides a framework for comparing different classes of regression models, including neural networks and both parametric and non-parametric statistical techniques. Unfortunately, standard techniques break down when faced with models, such as neural networks, in which there is more than one "layer" of adjustable parameters. We ...

متن کامل

Degrees of Freedom in Deep Neural Networks

In this paper, we explore degrees of freedom in deep sigmoidal neural networks. We show that the degrees of freedom in these models are related to the expected optimism, which is the expected difference between test error and training error. We provide an efficient Monte-Carlo method to estimate the degrees of freedom for multi-class classification methods. We show that the degrees of freedom i...

متن کامل

A Super - Element Based on Finite Element Method for Latticed Columns Computational Aspect and Numerical Results

This paper presents a new super-element with twelve degrees of freedom for latticed columns. This elements is developed such that it behaves, with an acceptable approximation, in the same manner as a reference model does. The reference model is constructed by using many Solid elements. The cross section area, moments of inertia, shear coefficient and torsoinal rigidity of the developed new elem...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2006