Journal Article ランダムカーネルを用いた条件付確率推定
Conditional Probability Estimation Using Random Kernels

塚本, 太郎  ,  Tsukamoto, Taro

51 ( 6 )  , pp.448 - 450 , 2015-06-13 , 計測自動制御学会 , The Society of Instrument and Control Engineers
This paper proposes a simple approach for probabilistic classification problem. Using kernels with random width as basis for logistic regression, and applying automatic relevance determination (ARD) technique, conditional density is estimated nether with cross-validation, nor with gradient-based adaptive parameter tuning.
形態: 図版あり
Physical characteristics: Original contains illustrations

Number of accesses :  

Other information