Greenhouse Temp-Humidity Prediction with Multi-scale Convolution and Sparse

Dataset Abstract: Greenhouse Indoor Temp-Humidity Time-Series Data for 30-Minute Interval Prediction**  \r\nThis dataset comprises continuous time-series data collected from a greenhouse environment, with samples recorded at 30-minute intervals. It includes multiple sensor parameters: time, temperature (*temp*), humidity (*shidu*), light intensity (*gq*), carbon dioxide (*co2*), and other environmental indicators. Engineered to support research on intelligent greenhouse temp-humidity prediction via multi-scale convolution and sparse attention-extended LSTM (xLSTM) models, this dataset offers a high-temporal-resolution, multi-dimensional resource for validating advanced deep learning algorithms in agricultural environmental sensing and time-series forecasting tasks.