ML_Path

This is to record my path heading to ML

To Do

• [] RNN and LSTM

• [] Introduce LSTM by Li Hongyi here

• [] Residual Network

• Linear regression VS Logistic Regress Activation Layer在Keras中的表示

• [] Logistic Regression VS no-hidden NN without/with activation

• [] L1 regularization vs L2 Regularization

• [] Keras中的linear actication与None activation是否一致，如果一致，为啥不同线性层的叠加效果会好很多，多个线性层的叠加不应该等效成一个线性层呢

  2.linear algebra online course https://www.bilibili.com/video/av15463995/?p=1

Resouces

Kaggle https://www.kaggle.com/

Linear Algebra

MIT https://www.bilibili.com/video/av15463995
李宏毅 https://www.bilibili.com/video/av22727915?from=search&seid=8427486326617068898

Conclustion

1. Linear Regression vs Logistical Regression
   1. Linear Regression是真的回归问题，输出y是一个连续值
   2. 而Logistical Regression其实是一个分类问题，参考Andrew关于logistical regression的介绍，而之所以名字中带有regression是由于历史原因，所以逻辑回归虽然名字中有回归二字，但其实质是分类问题