智能系统学报
智能繫統學報
지능계통학보
CAAI TRANSACTIONS ON INTELLIGENT SYSTEMS
2015年
1期
27-36
,共10页
仝伯兵%王士同%梅向东
仝伯兵%王士同%梅嚮東
동백병%왕사동%매향동
稀疏贝叶斯%两层分类%距离学习%视频烟雾检测%最近邻算法%有限样本%泛化性%时间效率
稀疏貝葉斯%兩層分類%距離學習%視頻煙霧檢測%最近鄰算法%有限樣本%汎化性%時間效率
희소패협사%량층분류%거리학습%시빈연무검측%최근린산법%유한양본%범화성%시간효솔
parse Bayesian%twol-evel classification%distance learning%video smoke detection%KNN%finite sam-ples%generalization%time efficiency
在有限样本下距离量的选择对最近邻算法( K-nearest neighbor ,KNN)算法有重要影响。针对以前距离量学习泛化性不强以及时间效率不高的问题,提出了一种稀疏条件下的两层分类算法( sparsity-inspired two-level classifi-cation algorithm ,STLCA)。该算法分为高低2层,在低层使用欧氏距离确定一个未标记的样本局部子空间;在高层,用稀疏贝叶斯在子空间进行信息提取。由于其稀疏性,在噪声情况下有很好的稳定性,可泛化性强,且时间效率高。通过在噪声数据以及在视频烟雾检测中的应用表明,STLCA算法能取得更好的效果。
在有限樣本下距離量的選擇對最近鄰算法( K-nearest neighbor ,KNN)算法有重要影響。針對以前距離量學習汎化性不彊以及時間效率不高的問題,提齣瞭一種稀疏條件下的兩層分類算法( sparsity-inspired two-level classifi-cation algorithm ,STLCA)。該算法分為高低2層,在低層使用歐氏距離確定一箇未標記的樣本跼部子空間;在高層,用稀疏貝葉斯在子空間進行信息提取。由于其稀疏性,在譟聲情況下有很好的穩定性,可汎化性彊,且時間效率高。通過在譟聲數據以及在視頻煙霧檢測中的應用錶明,STLCA算法能取得更好的效果。
재유한양본하거리량적선택대최근린산법( K-nearest neighbor ,KNN)산법유중요영향。침대이전거리량학습범화성불강이급시간효솔불고적문제,제출료일충희소조건하적량층분류산법( sparsity-inspired two-level classifi-cation algorithm ,STLCA)。해산법분위고저2층,재저층사용구씨거리학정일개미표기적양본국부자공간;재고층,용희소패협사재자공간진행신식제취。유우기희소성,재조성정황하유흔호적은정성,가범화성강,차시간효솔고。통과재조성수거이급재시빈연무검측중적응용표명,STLCA산법능취득경호적효과。
The selection of distance greatly affects KNN algorithm as it relates to finite samples due to weak generali-zation and low time efficiency in the previous learning of distance .In this paper , a new sparsity-inspired two-level classification algorithm (STLCA) is proposed.This proposed algorithm is divided into two levels:high and low.It uses Euclidean distance at the low-level to determine an unlabeled sample local subspace and at the high level it u-ses sparse Bayesian to extract information from subspace .Due to the sparsity in noise conditions , STLCA can have good stability , strong generalization and high time efficiency .The results showed that the STLCA algorithm can a-chieve better results through the application in noise data and video smoke detection .