Summary of problems in discussion

Summary, 2021-01-29

The aims or problmes are as follows:

  1. Write a R package to implement the algorithm about the estimation of the current model;

  2. Compare the performance of three methods: 1)the Proposed, 2) conducting GMM after PCA dimension reduction, and 3) conducting GMM on the latent variable, where GMM denotes Gaussian mixture model.

  3. Write the model settings and simulation results into a document, and viualization is preferred. Model settings include that \(p\) is set as \(1000, 2000\), \(n\) is set as \(2000, 4000\), \(K\) is set as \(2, 4, 6\).

  4. Plot the estimates of latent variables \(z\).

  5. Learn the method of data preprocessing and Yang Yi Shixiong can provide somde demo codes and real data.

  6. Read some papers about the area of Single-cell.

  7. Think about adding the Hidden Gaussian Markov fields into model, then use ICM algorithm to implement it.s

  8. Read the book about Gaussian Process.

Summary, 2021-02-05

The aims or problmes are as follows:

  1. Check the computation complexity, i.e. the relationship between computing time and \(n,p,K\) and \(q\).

  2. Increase \(n\) and \(p\), such as \(n=4000,5000\) and \(p=2000\).

  3. Look up to the settting of signal-noise ratio.

Summary, 2021-03-11

The aims are as follows:

  1. Think about adding the Hidden Gaussian Markov fields into model, then use ICM algorithm to implement it.s