求翻译：对相关数据进行预处理，所有指标进行一致化与无量纲化处理，使用三倍标准差检验法进行异常数据的剔出，最终获得xx个样本数据，将这xx个样本划分为训练样本和测试样本，随机抽取35%(33个)作为训练样本集，用于构造SVMs集成模型，其余的65%(60个)作为测试样本集，用于模型泛化能力的检验。对于可持续发展能力的评价集中，本文采用模糊集来定义输出级集合，充分结合该领域专家的经验，将其分成三个等级：可持续发展性强(Ⅰ)、可持续发展性一般(Ⅱ)、可持续发展性弱(Ⅲ)。是什么意思？

Pre-processing of data, all indicators of the same treatment with the non-dimensional, three times the standard deviation using the test method for abnormal data bound, ultimately, xx sample data, this xx samples into training and test samples samples, randomly selected 35% (33) as the training set,

The relevant data for pre-processing, all of the indicators for the homogenization and dimensionless quantities of 3 times, use standard deviation of the test for data anomalies, which was ultimately removed from the sample data, xx xx samples which will be divided into training and test samples, ra

正在翻译，请等待...

On related data for pretreatment, all indicators for consistent of and non outline of of processing, using three times times standard difference inspection method for exception data of ruled out, eventually get XX a samples data, will this XX a samples Division for training samples and test samples,

Pre-processing of data, all indicators of the same treatment with the non-dimensional, three times the standard deviation using the test method for abnormal data bound, ultimately, xx sample data, this xx samples into training and test samples samples, randomly selected 35% (33) as the training set,