Cigarette suction resistance measured value correction system

The invention discloses a cigarette suction resistance measured value correction system, which comprises: a position embedding layer used for fusing position embedding data to obtain a second cigarette quality and equipment data set arranged according to a time sequence; the self-attention layer that is used for adopting an attention mechanism to endow weights to the second cigarette quality and the equipment data set and outputting the weights to the feed-forward network layer; the feedforward network layer that is used for endowing a weight value to a feature vector in each piece of input weight-endowing sub-data, fusing the feature vectors to obtain feature fusion sub-data, and stacking the weight-endowing sub-data and the feature fusion sub-data to obtain cigarette quality and equipment sub-data blocks; and the prediction layer that is used for fusing the cigarette quality and the data in the equipment sub-data blocks through the full connection layer, and predicting the suction resistance value at the ne