The most important protein attributes (features) in structure of different HA subtypes selected by different attribute weighting algorithms.

Total number of attribute weighting algorithms which have announced the certain attribute important (weight higher than 0.5, Table S4). This table presents the number of algorithms that selected the attribute. Weighting algorithms were PCA, SVM, Relief, Uncertainty, Gini index, Chi Squared, Deviation, Rule, Information Gain, and Information Gain Ratio.

已为某属性赋予高于0.5的权重（即判定该属性为重要属性，详见附表S4）的属性加权算法总数量。本表格统计了选中该属性的算法数量。本次涉及的属性加权算法包括主成分分析（PCA）、支持向量机（SVM）、Relief算法、不确定性度量、基尼指数（Gini index）、卡方检验（Chi Squared）、偏差度量、规则算法、信息增益（Information Gain）以及信息增益比（Information Gain Ratio）。