Electronic nose dataset for beef quality monitoring under an uncontrolled environment
Description
Table 1 is a list of sensors used in this experiment. The sensory classes of beef are described in Table 2 and the example of ground-truth data is shown in Fig.2. Then, Fig. 1 denotes the schematic experimental design. Fig.3 implies the example of ambient conditions in this experiment. The sample of e-nose signals is depicted in Fig.4. This dataset consists of five series correspond to five beef cuts where one series contains 2160 minutes of measurement points. Every series is distributed in comma-separated value (csv). The first row contains the column header as follows: • Minute: time of measurement point (minute); • Class: discrete label of beef quality [“excellent”, “good”, “acceptable”, “spoiled”]; • TVC: continuous label of microbial population (log10 cfu/g); • MQ_: sensor resistance value of a particular gas sensor (Ω); • Humidity: relative humidity (%) in the sample chamber; • Temperature: temperature (C) in the sample chamber. The dataset is sorted ascendingly by column “minute” to represent the time-series data in beef quality monitoring. The experimental dates for each series are as follows: • Time-series 1 (TS1.csv): May 12, 2016 to May 13, 2016 • Time-series 2 (TS2.csv): October 8, 2018 to October 9, 2018 • Time-series 3 (TS3.csv): October 11, 2018 to October 12, 2018 • Time-series 4 (TS4.csv): October 13, 2018 to October 14, 2018 • Time-series 5 (TS5.csv): October 15, 2018 to October 16, 2018.