An Improved Classifier Chain Algorithm for Multi-label Classification of Big Data Analysis

The widely known classifier chains method for multi-label classification, which is based on the binary relevance (BR) method, overcomes the disadvantages of BR and achieves higher predictive performance, but still retains important advantages of BR, most importantly low time complexity. Nevertheless, despite its advantages, it is clear that a randomly arranged chain can be poorly ordered. We overcome this issue with a different strategy: Several times K-means algorithms are employed to get the correlations between labels and to confirm the order of binary classifiers. The algorithm ensure the right correlations be transmitted persistently as great as possible by improve the earlier predictions accuracy. The experimental results on the Reuters-21578 text chat data set and image data set show that the approach is efficient and appealing in most cases.

[1]  Charles Elkan,et al.  Learning and Inference in Probabilistic Classifier Chains with Beam Search , 2012, ECML/PKDD.

[2]  Cyrus Shahabi,et al.  Authentication of k Nearest Neighbor Query on Road Networks , 2014, IEEE Transactions on Knowledge and Data Engineering.

[3]  S. Angel Latha Mary,et al.  AUTHENTICATION OF K NEAREST NEIGHBOR QUERY ON ROAD NETWORKS , 2015 .

[4]  Kun Zhang,et al.  Multi-label learning by exploiting label dependency , 2010, KDD.

[5]  Min-Ling Zhang,et al.  A Review on Multi-Label Learning Algorithms , 2014, IEEE Transactions on Knowledge and Data Engineering.

[6]  Stephen Tyree,et al.  Stochastic Neighbor Compression , 2014, ICML.

[7]  Geoff Holmes,et al.  Classifier chains for multi-label classification , 2009, Machine Learning.

[8]  Xuelong Li,et al.  Negative Samples Analysis in Relevance Feedback , 2007, IEEE Transactions on Knowledge and Data Engineering.

[9]  J. Bezdek,et al.  Generalized k -nearest neighbor rules , 1986 .

[10]  Zhi-Hua Zhou,et al.  ML-KNN: A lazy learning approach to multi-label learning , 2007, Pattern Recognit..

[11]  Jonathan Qiang Jiang,et al.  Predicting Protein Function by Multi-Label Correlated Semi-Supervised Learning , 2012, IEEE/ACM Transactions on Computational Biology and Bioinformatics.