Optical Constants of Ginseng Tablets

This dataset consists of all the optical constant data and labels, supporting the findings in manuscript " A Label-Free and Data-Efficient Botanical Characterization Framework using Terahertz Spectroscopy and Soft-Voting Ensembles." Upon unzipping the compressed file, there would be 4 folders, "Scheme_1" to 4, denoting the 4 data selection schemes for input feature ablation study in this manuscript. Within each folder, there are 24 csv files, with 6 train data, 6 test data, 6 train labels and 6 test lebels. 1. Train data are denoted as "data_train_n.csv", n = 1~6. These are the train data for cross-validation fold "n". 2. Similarly, train data are denoted as "data_test_n.csv" 3. Train labels are named as "label_train_n.csv", in the form of a single row form. Each row denotes the ginseng species in corresponding "data_train_n.csv" 4. Similarly, test lebels are named as "label_test_n.csv"本数据集涵盖全部光学常数（optical constant）数据与标签，用以支撑研究论文《基于太赫兹光谱（Terahertz Spectroscopy）与软投票集成（Soft-Voting Ensembles）的无标签数据高效植物表征框架》（A Label-Free and Data-Efficient Botanical Characterization Framework using Terahertz Spectroscopy and Soft-Voting Ensembles）的研究结论。解压该压缩包后，将得到4个文件夹，依次为Scheme_1至Scheme_4，分别对应本文中开展输入特征消融实验（ablation study）所使用的4种数据选取方案。每个文件夹内包含24个CSV文件，分别为6组训练数据、6组测试数据、6组训练标签与6组测试标签。 1. 训练数据以"data_train_n.csv"命名，其中n取值为1~6，分别对应交叉验证折（cross-validation fold）第n折的训练数据集。 2. 同理，测试数据以"data_test_n.csv"命名。 3. 训练标签命名为"label_train_n.csv"，采用单行格式存储，每行内容对应"data_train_n.csv"中样本的人参品种标签。 4. 同理，测试标签命名为"label_test_n.csv"。