MachineLearner

To collect different machine learning algorithms here for future revision

Let's talk about Math

Maxium likelihood: just to get a best Θ which can MAX P(X|Θ).
- To do this we can the derivation of the function P(X|Θ), usually log is a useful way to get rid of the exponential component for things like gaussian distribution.
  Probabilistic Graphical Model
  
  I've learnt that from Carl's machine learning course in 2018-2019. Just want to have a deep understanding now in 2020. It is easy to forget things if you don't use that. But I am still need to find a key that how to recap it quickly. (I reckon I should use emacs to edit those now, but I feel that emacs editor is not really good for doing this job?)
paper to read - from Zhihu
- An Introduction to Probabilistic Graphical Models
  Clustering
  
  cluster quality
issue #6

keywords: core, border, outlier, eps, min_sample
Youtube videos: 1, 2
Jupyter notebooks: 1, 2
explained core label in DBSCAN algorithm
issue #2: dbscan result
KMeans & KNN

I was really familiar with those two algorithms. But I need to do some revision now.

MME
- bias on (uniform distribution)
MLE

Test statistic

I learned test statistic from Partick's anomaly detection course[notes]. At the beginning, I have no idea what it is. Thanks to a course from Taiwai（劉惠美－數理統計學）which could be found here. 老师在第一节课中说统计中有两个概念：点估计和假设检验。
rejection region
- issue #3: how to transfer the region to a probability
examples: different variance
- 7-2: 假设检验-page 12
- 第 3 章假设检验-例 3.8

This section will cover things I don't really understand and hopefully I can add as much as possible and fit them into normal sections

When I did my interview at Oxford University, they asked me some questions that I cannot answer. (which means that I still have too much to learn!)

TODO: learn & update VC dimention part (due by: 25th Aug)

TODO: add things