Bayesian Input Selection for Neural Network Classiiers 1 Bayesian Input Selection for Neural Network Classiiers
نویسندگان
چکیده
In this paper we discuss the use of the Bayesian posterior probability distribution over weight space and Receiver Operating Characteristic curves in a neural network input selection algorithm. The a posteriori distribution is obtained by combining the likelihood function based on training data and a prior distribution based on expert knowledge. To numerically calculate the marginalization, Markov Chain Monte Carlo methods are used. We demonstrate the technique on the problem of ovarian cancer classi-cation. The resulting input selection is then used to train a neural network that signiicantly outperforms the Risk of Malignancy Index, a traditionally used diagnostic aid.
منابع مشابه
Building High Performant Classifiers by Integrating Bayesian Learning, Mutual Information and Committee Techniques - A Case Study in Time Series Prediction
Recently many statistical methods for the design of neural networks were proposed. In this paper we present an integrated approach for building high performant classiiers for time series prediction. The classiier is based on a committee of neural networks, which are constructed to be as independent as possible. Each network is trained using a Bayesian learning rule. The training process is inte...
متن کاملShort term electric load prediction based on deep neural network and wavelet transform and input selection
Electricity demand forecasting is one of the most important factors in the planning, design, and operation of competitive electrical systems. However, most of the load forecasting methods are not accurate. Therefore, in order to increase the accuracy of the short-term electrical load forecast, this paper proposes a hybrid method for predicting electric load based on a deep neural network with a...
متن کاملForecasting of rainfall using different input selection methods on climate signals for neural network inputs
Long-term prediction of precipitation in planning and managing water resources, especially in arid and semi-arid countries such as Iran, has a great importance. In this paper, a method for predicting long-term precipitation using weather signals and artificial neural networks is presented. For this purpose, climatic data (large-scale signals) and meteorological data (local precipitation and tem...
متن کاملUniversity of Cambridge Programme for Industry Neural Networks Summer School Bayesian Methods for Neural Networks: Theory and Applications
Neural networks are parameterized non-linear models used for empirical regression and classi-cation modelling. Their exibility makes them able to discover more general relationships in data than traditional statistical models. Bayesian probability theory provides a unifying framework for data modeling which ooers several beneets. First, the overrtting problem can be solved by using Bayesian met...
متن کاملBiological Data Mining Using Bayesian Neural Networks: A Case Study
Biological data mining is the activity of nding signiicant information in biomolecular data. The signiicant information may refer to motifs, clusters, genes, and protein signatures. This paper presents an example of biological data mining: the recognition of promoters in DNA. We propose a two-level ensemble of classiiers to recognize E. Coli promoter sequences. The rst-level classiiers include ...
متن کامل