課程資訊

Computer Intensive Statistics in Ecology

99-2

Ocean5052

241EU1920

Ceiba 課程網頁
http://ceiba.ntu.edu.tw/992ecol_stat

THIS IS AN ADVANCED COURSE INTENDED FOR SENIOR UNDERGRADUATE AND GRADUATE STUDENTS WITH KNOWLEDGE OF BASIC STATISTICS INCLUDING RANDOM VARIABLES, ANALYSIS OF VARIANCE, REGRESSION ANALYSIS, AND RANK-BASED NON-PARAMETRIC STATISTICS. WE WILL DISCUSS SEVERAL COMPUTER-INTENSIVE STATISTICAL METHODS. WE WILL DISCUSS THE THEORY, ASSUMPTION, AND APPLICATION OF THESE METHODS IN ECOLOGICAL PROBLEMS. THE COURSE IS DESIGNED FOR HAND-ON WORK. STUDENTS NEED TO GET FAMILIAR WITH AT LEAST ONE COMPUTER LANGUAGE TO DO THE STATISTICS. MOST OF WORK CAN BE DONE WITH MATLAB, BUT ANY OTHER PROGRAMMING LANGUAGE WILL DO EQUALLY WELL. SOMETIMES, WE WILL USE WELL-DEVELOPED SOFTWARE WHEN THE COMPUTATION IS TOO COMPLICATED AND BEYOND THE BASIC LEVEL. THERE WILL BE DEDICATED TIME EVERY WEEK FOR STUDENTS TO PRESENT THEIR WORKS AND TO DISCUSS THE APPLICATION OF THESE METHODS ON REAL WORLD PROBLEMS. THE TOPICS MAY INCLUDE:
1. INTRODUCTION TO RANDOM VARIABLES
2. DISTRIBUTION AND RANDOM NUMBER GENERATOR
3. DIMENSION REDUCTION METHODS
4. MONTE CARLO METHOD
5. PERMUTATION, BOOTSTRAP, JACKKNIFE, SUB-SAMPLING AND RE-SAMPLING
6. INTERPOLATION, OPTIMIZATION, MINIMIZATION,
7. MAXIMUM LIKELIHOOD
8. CATEGORICAL AND REGRESSION TREE
9. KERNEL SMOOTHING
10. SIMPLE NEURAL NETWORK
11. MISSING DATA
12. STOCHASTIC TIME SERIES ANALYSIS
13. SPECTRAL ANALYSIS
14. FRACTAL
15. NONLINEAR TIME SERIES ANALYSIS

The objectives are to provide students computational skills for sophisticated statistical methods that are often required for biological questions.

Students need to get familiar with at least one computer language to do the statistics. There will be dedicated time every week for students to present their works and to discuss the application of these methods on real world problems.

Office Hours

(僅供參考)

 No. 項目 百分比 說明 1. homework 70% 2. presentation 30%

 課程進度
 週次 日期 單元主題 Week 1 02/24 Introduction to Matlab programming and plotting Week 2 03/03 Random variables, distribution, random number generator, statistical identity Week 3 03/10 Bootstrap Week 4 03/17 Jackknife Week 5 03/24 Bootstrapped confidence limits Week 6 03/31 Permutation and randomization Week 7 04/07 Minimization Week 8 04/14 Dimension reduction methods Week 9 04/21 TBD Week 10 04/28 Classification Week 11 5/5 Adaptive linear learning Week 12 5/12 Multi-layer perceptron Week 13 5/19 Simulation, Monte Carol methods, and surrogate test, ODE solver Week 14 5/26 Maximal likelihood Week 15 6/2 Introduction to Bayesian analysis Week 16 Model selection and AIC Week 17 Stochastic time series analysis and Spectral analysis Week 18 final discussion