Page 195 - 데이터과학 무엇을 하는가? 전자책

다음은 다양한 기계학습 알고리즘을 정리한 것이다.

다양한 기계학습 고리
기계학습 분류 고리
Logistic regression, decision tree, nearest-neighbor classifier,
분류 kernel discriminate analysis, neural network, support vector
machine, random forest, boosted tree

Linear regression, regression tree, kernel regression, support
예
vector regression
Principal component analysis, non-negative matrix factorization,
차원(변수) 축소
independent component analysis, manifold learning, SVD
k-means, hierarchical clustering, mean-shift, self-organizing
그룹화
maps(SOMs)
선행학 (Pre- Deep Learning(Stacked Restricted Boltzmann Machine, Stacked
training), Auto-Encoders등을 사용한 Multi layers Neural Nets, Non-linear
2차 분류 Transformation)
Bipartite cross-matching, n-point correlation two-sample testing,
데이터 비교
minimum spanning tree

위 표에 소개된 기계학습 알고리즘들은 꾸준히 발전해 왔는데, 최근
눈에 게 발전한 알고리즘은 다 신경망(Neural Network)을 기초로 발

전한 러 (Deep Learning) 알고리즘이다.

“A fast learning algorithm for deep belief nets, Neural Computation,” Geoffrey
E. Hinton and Simon Osindero, 2006.
“Fast Inference in Sparse Coding Algorithms with Applications to Ob ect Recognition,”
Computational and Biological Learning Lab, Courant Institute, Koray Kavukcuoglu,
Marc’Aurelio Ran ato, and ann LeCun, 2008.
“Pedestrian Detection with Unsupervised Multi-Stage Feature Learning,” in Proc.
International Conference on Computer Vision and Pattern Recognition (CVPR’13), P.
Sermanet, K. Kavukcuoglu, S. Chintala, . LeCun, 2013.

193