Data Set of Extracted Summary Statistics from Equipment Sensor Data

This data set was generated in accordance with the semiconductor industry and contains values of summary statistics from sensor recordings of the high-precision and high-tech production equipment. Basically, the semiconductor production consists of hundreds of process steps performing physical and chemical operations on so-called wafers, i.e. slices based on semiconductor material. In the production chain, each process equipment is equipped with several sensors recording physical parameters like gas flow, temperature, voltage, etc., resulting in so-called sensor data. Out of the sensor data, values of summary statistics are extracted. These are values like mean, standard deviation and gradients. To keep the entire production as stable as possible, these values are used to monitor the whole production in order to intervene in case of deviations. After the production, each device on the wafer is tested in the most careful way resulting in so-called wafer test data. In some cases, suspicious patterns occur in the wafer test data potentially leading to failure. In this case the root cause must be found in the production chain. For this purpose, the given data is provided. The aim is to find correlations between the wafer test data and the values of summary statistics in order to identify the root cause. The given data is divided into four data sets: "XTrain.csv", "YTrain.csv", "XTest.csv" and "YTest.csv". "XTrain.csv" and "XTest.csv" represent the values of summary statistics originating in the production chain separated for the purpose of training and validating a statistical model. Included are 114 observations of 77 parameters (values of summary statistics). The "YTrain.csv" and "YTest.csv" contain the corresponding wafer test data (144 observations of one parameter).本数据集贴合半导体行业场景生成，涵盖高精度高科技生产设备的传感器记录汇总统计量（summary statistics）数值。半导体生产包含数百道工序，对晶圆（wafer）——以半导体材料制成的薄片——实施物理与化学加工。在生产链路中，每台制程设备均配备多组传感器，用于记录气体流量、温度、电压等物理参数，由此生成传感器数据（sensor data）。从传感器数据中可提取汇总统计量数值，包括均值、标准差、梯度等指标。为维持整体生产稳定性，此类指标被用于监控全生产流程，以便在出现偏差时及时介入干预。生产工序完成后，需对晶圆上的每颗器件开展严苛测试，由此生成晶圆测试数据（wafer test data）。部分场景下，晶圆测试数据中会出现异常模式，可能引发器件失效，此时需在生产链路中定位根本原因（root cause）。本数据集正是为该场景需求而提供，其目标为挖掘晶圆测试数据与汇总统计量数值间的关联关系，以实现问题根源的精准识别。本数据集共分为四个数据文件，分别为"XTrain.csv"、"YTrain.csv"、"XTest.csv"与"YTest.csv"。其中"XTrain.csv"与"XTest.csv"为源自生产链路的汇总统计量数值，按统计模型训练与验证的需求进行拆分，包含77项参数（汇总统计量数值）的114条观测样本。"YTrain.csv"与"YTest.csv"则包含对应的晶圆测试数据，为1项参数的144条观测样本。