Group meeting recap [10.28.2005]

• Dr. Inchi Hu was visiting us this week. We met and discussed the information-driven training for multi-class classifier research project.
• On Oct. 28, Yuejing gave a summary presentation on current major methods for classifying multiclasses. The presentation file is posted in the presentation section of the group website.

Programming Skills: Use Batch Command to Change File Names or Extentions

If you want to modify multiple file names or extention names, the batch command below might be helpful:

@echo off echo BEGIN PROCESSING...
for /f "tokens=*" %%a in ('dir /b /s *.txt') do (
echo Current processing file %%a
rename %%a *.dat
)
echo END PROCESSING...

This batch command will change all the *.txt files in current folder and sub-folders to *.dat files.

Planning of Friday's presentation

A temporary plan for Friday's presentation

Summary of multi-classification methods in both machine learning literature and microarray literature:

Two basic ideas:

1. Classification all at once, based on posterior probability of each class.
Including:

• Parametric: discriminant analysis: LDA, QLDA, Logistic Regression
• Nonparmetric: KNN & Prototype Methods, LVQ, FDA, PDA, Neural Network

2. Binary classification methods + methods for combining binary classification result

• 2-class classification methods: SVM, Golub's Weighted Votes
• Methods of combining binary classification: coding system
  pairwise comparison
  One vs all (OVA)

Machine Learning Benchmark Data Summary

Summary of machine learning benchmark datasets
http://statgene.stat.columbia.edu/blog/benchmark.pdf
Summary of Leukemia, Lymphoma, Colon Cancer classification result
http://statgene.stat.columbia.edu/blog/cancer_classification.pdf
Multiclass cancer diagnosis using tumor gene expression signature (compared multiclass classification using SVM, KNN, Golub's Weighted Votes + Pairwise, OVA) [pdf]

Group meeting recap [10.21.2005]

1. Lei presented the latest progress on software developments
2. Yuejing discussed about available machine learning benchmark datasets.
3. Yuejing will give a brief presentation on current machine learning methods next Friday (10.28.2005)

Mathematics and 21st Century Biology (2005)

http://www.nap.edu/books/0309095840/html/R1.html
Published by The National Academies Press

This is report is prepared by the Board on Mathematical Sciences and Their Applications of the National Research Council of The National Academies.

Recommended by: Professor Lo.

This is Cool!

Thanks Lei Cong~
We are now in a blog age!

Group Meeting Recap [10.14.2005]

10-11 AM. In Professor Lo's office.

1. This is the first time Yuejing came to the group meeting.
2. Lei reported current progress on the software development.
3. It is very important to keep up with current scientific research. Professor Lo encouraged everyone to read Science and Nature genetics regularly.
4. Inchi Hu (HKSTU) will be visiting us from 10/23-10/27. He has been working on a classification ideas based on the information score ideas. Professor Lo discussed the preliminary results.
5. Tian proposed that we propose and study the combined association and linkage method for a more general problem--combining different studies. Iulliana said she will look into this problem.

For next meeting:

• Lei will illustrate the current version of the software. [Maybe we should think about a name for the package.]
• Iulliana will bring her textbook on Machine learning.
• Yuejing will conduct a small search for other benchmark data for classification before next meeting.
• Tian will try to run the normalized BHTA screening results on the published simulated example before next meeting.

Brief Guide of How to Use Blogger

Note: You can also find this guide as a word format document at: Link

1. If you have received an email from Lei Cong's Blogger invitation, click the link in your email. Then you will be linked to the Blogger website. If you already had a Blogger account, you can directly logon with it. Otherwise, you need to sign on first. After you logged on, accept the invitation please. After you logged on and accepted the invitation, you will see your Blogger dashboard like this:



2. Click the “Statgene Group Blog” link in the blogs list. You will see the posts that were created by these blog members.



3. You can view those posts by click the “View” link on the right-hand side of the page. Or you can edit or delete your own posts. If you want to create a new post, click “Create” link on the top-left corner. An input interface will show up.



4. You can preview your post by clicking the “Preview” link on the top-right corner of the input window. When you finish, you can click the orange button “Publish Post” to publish your post.

Testing for invitation!

This is a testing post from Lei Cong's Gmail account. If it's working, more invitations will be sent out.

Statgene Group Blog Initially Setup

By Lei, on 10/14/2005.