Dataset for non-infused aroma-based quality identification of Gambung green tea using electronic nose

This experiment focuses on the analysis of green tea aroma using a set of gas sensors. Specifically, the gas sensors selected for this research are the TGS822, TGS2602, TGS2620, MQ138, MQ5, and MQ3. The experiment involved testing a total of 78 different tea samples (chops), with each sample being observed three times. To conduct the experiment, a tea chamber was utilized, capable of accommodating 125 grams of dry green tea. The tea chamber was connected to a sensor chamber through a hose and intake micro air pump. During data acquisition, air from the tea chamber flowed into the sensor chamber for a duration of 60 seconds. Once the airflow from the tea sample was complete, the gas sensors recorded the aroma data for 60 seconds, resulting in 60 data records. These records were then saved into a CSV file for further processing and labeling. The labeling process involved referencing the Indonesian National Standard (SNI) 3945:2016, which defines the quality parameters for green tea according to ISO 11287 Green tea - definition and basic requirements. The SNI 3945:2016 standard specifies both special requirements and general requirements for green tea quality assessment. Special requirements encompass aspects such as water content, soluble ash, ash alkalinity, crude fiber, polyphenols, metal contamination, and microbial contamination. On the other hand, general requirements cover the physical and organoleptic characteristics of the tea, including dryness, steeping water, and steeping dregs. To evaluate the quality of Gambung green tea, an organoleptic test was conducted by a tea tester. The results of this test were used to label the data set obtained from the e-nose. The data set had two labels: quality standard ("good" and "quality defect") for the discrete classification task, and organoleptic score, which combined ratings for dry appearance, brew color, taste, aroma, and dregs of brewing, for the continuous regression task. In summary, this study aimed to predict the quality standard and organoleptic score of green tea samples using gas sensor data. The gas sensors were selected based on their suitability for analyzing the tea's aroma. The labeled data set, obtained through experimentation and organoleptic testing, would serve as the basis for training models for classification and regression tasks.本实验旨在通过一组气体传感器开展绿茶香气分析研究。具体而言，本研究所选用的气体传感器为TGS822、TGS2602、TGS2620、MQ138、MQ5及MQ3。本次实验共测试78份不同的绿茶茶样（chops），每份茶样重复测试三次。 实验采用茶叶舱装载样品，该舱可容纳125克干绿茶。茶叶舱通过软管和进气微型气泵与传感器舱相连。数据采集阶段，茶叶舱内的气体将通入传感器舱，通气流时长为60秒。待茶样释出的气流结束后，气体传感器将采集60秒的香气数据，最终得到60条数据记录。随后将这些数据记录保存至CSV文件中，以便后续处理与标注。 标注环节参考了印度尼西亚国家标准（Indonesian National Standard，SNI）3945:2016，该标准依据ISO 11287《绿茶——定义与基本要求》（Green tea - definition and basic requirements）规定了绿茶的品质参数。SNI 3945:2016标准明确了绿茶品质评估的特殊要求与通用要求：特殊要求涵盖水分含量、可溶性灰分、灰分碱度、粗纤维、多酚类物质、金属污染物及微生物污染物等指标；通用要求则涉及茶叶的物理特性与感官特性，包括干茶状态、冲泡茶汤及冲泡茶渣。 为评估甘邦绿茶的品质，本次实验由专业评茶师开展感官测试，并将测试结果用于标注由电子鼻（e-nose）采集得到的数据集。该数据集包含两类标注：一是用于离散分类任务的品质等级标签（分为"优质"与"品质缺陷"两类），二是用于连续回归任务的感官评分，该评分综合了干茶外观、冲泡茶汤色泽、滋味、香气及冲泡茶渣的评级结果。 综上，本研究旨在通过气体传感器数据预测绿茶茶样的品质等级与感官评分。本次选用的气体传感器均适配茶叶香气分析需求。本研究通过实验与感官测试得到的标注数据集，将作为训练分类与回归模型的基础数据集。