Classification
-> regression, ranking, reinforcement learning, detection
Logistic Classifier
WX + b = y
-> a, b, c
p = 0.7, 0.2, 0.1
p = probability which closed to 1.0 – 0
S(yi) = e^yi / Σi e^yj
Any scores turned to probability
ソフトウェアエンジニアの技術ブログ:Software engineer tech blog
随机应变 ABCD: Always Be Coding and … : хороший