摘自 ai 100 天周俊川/陳明佑簡報
tSNE 作法
example : tSNE on MNIST, cluster 非常清楚
截圖自 Coursera
example : MNIST in different perplexities
截圖自 Coursera
通常先降維再做 tSNE
是絕佳的視覺化工具
可用來當特徵
小心解讀跑出來的結果
補充資料的連結 :
Matrix Factorization:
Overview of Matrix Decomposition methods (sklearn)
http://scikit-learn.org/stable/modules/decomposition.html
t-SNE:
Multicore t-SNE implementation <--米哥的課硬要放上狄哥的 github 連結
https://github.com/DmitryUlyanov/Multicore-TSNE
Comparison of Manifold Learning methods (sklearn)
http://scikit-learn.org/stable/auto_examples/manifold/plot_compare_methods.html
How to Use t-SNE Effectively (distill.pub blog) <-- 強烈建議到此一遊
https://distill.pub/2016/misread-tsne/
tSNE homepage (Laurens van der Maaten)
https://lvdmaaten.github.io/tsne/
Example: tSNE with different perplexities (sklearn)
http://scikit-learn.org/stable/auto_examples/manifold/plot_t_sne_perplexity.html#sphx-glr-auto-examples-manifold-plot-t-sne-perplexity-py
Interactions:
Facebook Research's paper about extracting categorical features from trees
https://research.fb.com/publications/practical-lessons-from-predicting-clicks-on-ads-at-facebook/
Example: Feature transformations with ensembles of trees (sklearn)
http://scikit-learn.org/stable/auto_examples/ensemble/plot_feature_transformation.html