Boosted Network Classifiers for Local Feature Selection
Boosted Network Classifiers for Local Feature Selection
复制标题
DOI：
10.1109/tnnls.2012.2214057
复制
发表时间：
2012-09
影响因子：
10.4
通讯作者：
Timothy Hancock;Hiroshi Mamitsuka
中科院分区：
文献类型：
--
作者：
Timothy Hancock;Hiroshi Mamitsuka
与所有模型一样，网络特征选择模型要求对所需特征的大小和结构进行假设。最常见的假设是稀疏性，即整个网络中只有一小部分被认为会产生特定的现象。稀疏性假设是通过正则化模型（如套索）来实现的。然而，假设稀疏性可能不适合许多现实世界的网络，其中具有高度相关的模块。在本文中，我们说明了两种新的优化策略，即，提高期望传播（BEP）和提高消息传递（BMP），直接使用网络结构来估计网络分类器的参数。BEP和BMP是集成方法，通过组合基于局部网络特征构建的各个模型来优化分类性能。BEP和BMP都不假设稀疏解，而是寻求所有网络特征的加权平均值，其中权重用于强调对分类有用的所有特征。在本文中，我们比较了BEP和BMP与网络正则化逻辑回归模型在模拟和真实的生物网络。结果表明，在存在高度相关的网络结构的情况下，假设稀疏性会对网络分类器的准确性和特征选择能力产生不利影响。
Like all models, network feature selection models require that assumptions be made on the size and structure of the desired features. The most common assumption is sparsity, where only a small section of the entire network is thought to produce a specific phenomenon. The sparsity assumption is enforced through regularized models, such as the lasso. However, assuming sparsity may be inappropriate for many real-world networks, which possess highly correlated modules. In this paper, we illustrate two novel optimization strategies, namely, boosted expectation propagation (BEP) and boosted message passing (BMP), which directly use the network structure to estimate the parameters of a network classifier. BEP and BMP are ensemble methods that seek to optimize classification performance by combining individual models built upon local network features. Neither BEP nor BMP assumes a sparse solution, but instead they seek a weighted average of all network features where the weights are used to emphasize all features that are useful for classification. In this paper, we compare BEP and BMP with network-regularized logistic regression models on simulated and real biological networks. The results show that, where highly correlated network structure exists, assuming sparsity adversely effects the accuracy and feature selection power of the network classifier.