基于复杂网络的机器学习方法(machine learning methods based on complex networks)
(1)收集基于向量的数据集并对其进行预处理(Gather the vector-based data set and preprocess it accordingly)
数据的预处理一般包括:
- 属性转换(attribute transformation):
缩放(scaling),
标准化(normalization),
标准化( standardization),
中心化(demeaning),
组合(combination),
分解( decomposition),
聚合(aggregation) - 删除(deletion)
- 数据清洗(cleaning)
移除异常值(removal of outliers),
添补缺失值(imputation of missing attributes ) - 采样(sampling)
- 以及其它许多预处理操作
(2)将基于向量的数据转化为基于网络的数据(Transform the vector-based into network-based data).
- 网络构建技术(Network Construction Techniques)
- 生成的网络(generated network)
(3)将基于网络的机器学习任务应用到基于矢量数据构建的网络中(Apply the network-based machine learning task in the network constructed from
the vector-based data)
以上是《Machine Learning in Complex Networks》阅读笔记!