Abstract: Aiming at the problem of dynamicity of  the wireless local area networks (LAN), which led to the changes of the validity of data labels with time, and required regular  updates and relabeling of data,  increasing the difficulty of data label labeling, we proposed  an adaptive labeling method for multimodal data labels in  wireless local area networks. Firstly, the repetitive wireless LAN multimodal data was cleaned by using dynamic sliding neighbor sorting algorithm, and the multimodal data was fused by using recurrent  neural network to obtain more comprehensive data information. Secondly, the fused wireless LAN  data was divided into deterministic set and fuzzy set, and the deterministic set data was labeled by using support vector machine, and the fuzzy set data was labeled by using K-nearest neighbor (KNN) classifier, thus achieving  the adaptive labeling of wireless LAN multimodal data labels. The experimental results show that the deduplication ratio of the proposed method is always above 12%, the consistency index is 0.992 8, the average absolute percentage error is 0.453 9, the ROC curve is closer to the upper left corner of the coordinate axis, the AUC value is 0.982 4, and the memory occupancy rate is always below 10%. The wireless LAN multimodal data labeling effect is good.

Key words:  , wireless local area network, multimodal data, label annotation, support vector machine, KNN classifier