【摘要】 基于GEOROC数据库全球闪长岩中角闪石的矿物化学成分，利用机器学习中支持向量机分类器，本文对汇聚边缘（CM）和板内裂谷（IV）两种环境中的闪长质岩浆岩进行了角闪石成分对比研究。结果显示：角闪石的化学成分能够区分上述两种构造环境，并依此建立区分不同构造背景的模型。该分类模型建立在超过1 700条来自两种构造背景下的闪长岩中的角闪石数据的训练与测试优化基础上，且分类精确度达到87%。总体而言，通过角闪石的矿物化学成分再计算获得的离子数(如MgC、MnC、Al^IV、Fe³⁺等)以及相应的温度、压力与氧逸度是区分CM与IV两类构造背景的主要指标。将相应的分类模型应用于扬子西北缘旺苍米仓山地区形成于～760 Ma的闪长岩与辉长闪长岩中的角闪石，分类结果显示其来自板内裂谷环境。结合前人研究及角闪石主、微量元素地球化学特征，本文认为该区的闪长岩与辉长闪长岩可能形成于板内伸展裂谷环境。更多还原

【Abstract】 Based on the mineral chemical composition of amphibole in diorites in the GEOROC database, and using the support vector machine classifier in machine learning, this paper has carried out a comparative study on the composition of amphibole in the diorite magmatic rocks in the convergent margin（CM）and intraplate rift（IV）environments. The results show that the composition of amphibole can well distinguish the two tectonic environments and establish corresponding identification models. The model is based on the training and testing optimization of more than 1 700 amphibole data from diorite under two tectonic settings, and the classification accuracy reaches 87%. In general, the recalculated compositional parameters(such as Mg_C, Mn_C, Al^IV, Fe³⁺, etc.) and the calculated temperature, pressure and oxygen fugacity obtained by recalculating the mineral chemical composition of amphibole are the main influencing factors for the differentiation of CM and IV tectonic settings. The corresponding classification model is applied to hornblende in diorite and gabbrodiorite formed at～760 Ma in the Micangshan area, Wangcang, on the northwestern margin of the Yangtze block. The classification results show that it comes from the intraplate rift environment. Based on previous studies and geochemical characteristics of major and trace elements of amphibole, this paper believes that diorite and gabbrodiorite in this area may have formed in the intracontinental extensional rift environment after arc-continent convergence.更多还原