| 題目：eXclusive Component Analysis -- Theories and Applications 主講人：Dr. Chao-Hui HUANG ( Bioinformatics Institute (BII), Agency for Science, Technology and Research (A*STAR), Singapore) 時間：103年12月3日(星期三13：30-15：00) 地點：三峽校區法學院大樓2F01教室 The abstract: an enhancement of the Independent Component Analysis (ICA), named eXclusive Component Analysis (XCA) and its applications on image statistics and machine learning is presented. XCA is especially effective when comparing the characteristics of several datasets. This effectiveness arise from XCA's ability to find exclusive features of each dataset as well as features common to all datasets. This paper also present a new set of classification methods, XCABoost, that are developed to utilize the exclusive components of datasets elucidated by XCA. XCABoost methods are formulated as an ensemble of linear and non-linear classifiers and as a classifier for Multiple Instance Learning. The XCABoost methods is benchmarked using well known datasets and compared to 28 other classifiers with good performance. |
| 相關連結: http://web.bii.a-star.edu.sg/~huangch/research_topics/index.html |
