Welcome to tigerliuwu's Blog!

• tensorflow introduction(2) - first demo

  2017-01-04

  Wu Liu

  tensorflow

  tensorflow
  Read All

  ---

• tensorflow introduction(1) - install

  2017-01-03

  Wu Liu

  tensorflow

  tensorflow

  使用virtualenv来安装tensorflow

  Read All

  ---

• tensorflow - resources

  2017-01-02

  Wu Liu

  tensorflow

  tensorflow
  Read All

  ---

• Coursera:Machine Learning by Andrew Ng(6) - SVM

  2016-12-11

  Wu Liu

  machine learning

  coursera  machine learning

  SVM is mainly forcus on the classification for the samples which are very close to the classifier boundary, while LR focus on the global samples to make them far away from the classifier boundary. There are 2 kinds of SVM:

  • SVM without kernel(linear svm)
    • when the \(\Theta * X >= 0\), y = 1; otherwise, y = 0
    • classify the training set with linear line.
  • SVM with kernel(typically gaussian kernel)
    • used to classify the training set with nonlinear line
    • the key is to choose the best C(\(\frac{1}{\lambda}\)) and sigma(\(\delta\))
  Read All

  ---

• Coursera:Machine Learning by Andrew Ng(8) - Anomaly

  2016-12-09

  Wu Liu

  machine learning

  coursera  machine learning

  When the anomaly data is very few, we need to make good use of F1 Score to help us to figure out which model is better.

  Read All

  ---

• Coursera:Machine Learning by Andrew Ng(7) - Compression with PCA

  2016-12-08

  Wu Liu

  machine learning

  coursera  machine learning

  This blog we’re going to talk about principal compononent analysis to compress the original data, PCA is used to reduce the dimension of the samples.

  Read All

  ---