Estimating Redundancy Information of Selected Features in Multi-dimensional Pattern Classification
This paper proposes a novel criterion for estimating the redundancy information of selected feature sets in multi-dimensional pattern classification. An appropriate feature selection process typically maximizes the relevancy of features to each class and minimizes the redundancy of features between selected features. Unlike to the relevancy information that can be measured by mutual information, however, it is difficult to estimate the redundancy information because its dynamic range is varied by the characteristics of features and classes.
By utilizing the conceptual diagram of the relationship between candidate features, selected features, and class variables, this paper proposes a new criterion to accurately compute the amount of redundancy. Specifically, the redundancy term is estimated by conditional mutual information between selected and candidate features to each class variable, which does not need a cumbersome normalization process as the conventional algorithm does. The proposed algorithm is implemented into a speech/music discrimination system to evaluate classification performance. Experimental results by varying the number of selected features verify that the proposed method shows higher classification accuracy than conventional algorithms.
|30||International Journal||Yoomi Hur, Jonathan S. Abel, Young-Cheol Park, Dae Hee Youn "Techniques for Synthetic Reconfiguration of Microphone Arrays" in Journal of the AES, vol.59, issue 6, pp.404-418, 2011|
|29||International Journal||Chi-Sang Jung, Hyunson Seo, Hong-Goo Kang "Estimating Redundancy Information of Selected Features in Multi-dimensional Pattern Classification" in Pattern Recognition Letters, vol.32, issue 4, pp.590-596, 2011|
|28||International Journal||Dong-il Hyun, Donggeum Lee, Youngcheol Park, Dae Hee Youn, Jeongil Seo "Joint Channel Coding Based on Principal Component Analysis" in ETRI Journal, vol.32, issue 5, pp.831-834, 2010|
|27||International Journal||Min-Seok Choi, Hong-Goo Kang "A Two-channel Noise Estimator for Speech Enhancement in Highly Non-stationary Environment" in IEEE Transactions on Audio, Speech, and Language Processing, vol.19, issue 4, pp.905-915, 2011|
|26||International Journal||Chi-Sang Jung, Moo Young Kim, Hong-Goo Kang "Selecting Feature Frames for Automatic Speaker Recognition Using Mutual Information" in IEEE Transactions on Audio, Speech, and Language Processing, vol.18, issue 6, pp.1332-1340, 2010|
|25||International Journal||Bong-Jin Lee, Chi-Sang Jung, Jeung-Yoon Choi, Hong-Goo Kang "On the Importance of Transition Regions for Automatic Speaker Recognition" in IEICE Transactions on Information and Systems, vol.E93-D, No.1, pp.197-200, 2010|
|24||International Journal||Jae-Seong Lee, Chang-Joon Lee, Young-Cheol Park, Dae Hee Youn "Efficient FFT Algorithm for Psychoacoustic Model of the MPEG-4 AAC" in IEICE Transactions on Information and Systems, vol.E92-D, No.12, pp.2535-2539, 2009|
|23||International Journal||Chang-Heon Lee, Hyen-O Oh, Hong-Goo Kang "On the Study of Noise Allocation for Speech Signal in Low Bit-Rate Audio Coding" in IEEE Signal Processing Letters, vol.16, issue 10, pp.849-852, 2009|
|22||International Journal||Bong-Jin Lee, Jeung-Yoon Choi, Hong-Goo Kang "Phonetically optimized speaker modeling for robust speaker recognition" in The Journal of the Acoustical Society of America, vol.126, issue 3, 2009|
|21||International Journal||Tacksung Choi, Sunkuk Moon, Young-Cheol Park, Dea Hee Youn, Seokpil Lee "A GMM-Based Feature Selection Algorithm for Multi-Class Classification" in IEICE Transactions on Information and Systems, vol.E92-D. No.8, pp.1584-1587, 2009|