{"id":"arxiv/2603.21590","url":"https://arxiv.org/abs/2603.21590","tags":["math.ST","cs.LG"],"score":0,"title":"Feature Incremental Clustering with Generalization Bounds","domain":"math.ST","source":"arxiv","authors":["Jing Zhang","Chenping Hou"],"reviews":0,"subject":"math.ST","summary":"In many learning systems, such as activity recognition systems, as new data collection methods continue to emerge in various dynamic environmental applications, the attributes of instances accumulate incrementally, with data being stored in gradually expanding feature spaces. How to design theoretically guaranteed algorithms to effectively cluster this special type of data stream, commonly referred to as activity recognition, remains unexplored. Compared to traditional scenarios, we will face at least two fundamental questions in this feature incremental scenario. (i) How to design…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-23","reactions":{}}