Data underlying the publication: Microstructural features governing fracture of a two-dimensional amorphous solid identified by machine learning

Datasets for paper "Microstructural features governing fracture of a two-dimensional amorphous solid identified by machine learning".<br>We applied isotropic dilational strain to a densely packed monolayer of attractive colloidal microspheres, resulting in fracture. Using brightfield microscopy and particle tracking, it is possible to examine the microstructural evolution of the monolayer during fracturing. Furthermore, using a quantified representation of the microstructure in combination with a machine learning algorithm, we calculate the likelihood of regions of the monolayer to be on a crack line.<br>The raw data from 20 independent experiments are provided here. The original datasets were sliced for ease of handling the file size, resulting in 79 separate files included here. Each of the 79 datasets consists of two files:(1) particle trajectories of all particles in the monolayer(2) classification labels for Machine Learning: particles receive a label '0' if the do not participate in the fracture, or '1' if they do.<br>The names of both files (1) and (2) begin with the same string, containing information on the experiment, and end with 'positions' and 'labels' respectively. A README file containing all details accompanies the datasets.

本数据集为论文《由机器学习识别的二维非晶固体断裂调控微观结构特征》的配套数据。我们对紧密堆积的吸引力型胶体微球单层施加各向同性膨胀应变，诱导其发生断裂。借助明场显微镜与粒子追踪技术，可观测该单层在断裂过程中的微观结构演化过程。此外，结合微观结构的量化表征与机器学习（Machine Learning）算法，我们可计算该单层中各区域处于裂纹线上的概率。
本数据集包含20组独立实验的原始数据。为便于控制文件体积，原始数据集已被拆分为79个独立文件，即本数据集所包含的全部文件。这79个数据集均包含两类文件：(1) 该单层内所有粒子的运动轨迹文件；(2) 机器学习（Machine Learning）分类标签文件：未参与断裂的粒子标注为0，参与断裂的粒子标注为1。两类文件的文件名均以包含实验信息的相同字符串作为前缀，分别以'positions'和'labels'作为后缀。本数据集附带一份README文件，其中包含全部相关细节说明。