Conclusion of these days

Well I think the details about how I welcomed the Spain Professor will not be talked here. He arrived yesterday morning and had a good rest in the hotel for a whole morning then we had lunch together with my teacher. Then he visited our museum and went to the office there he looked at the specimen of Maba and put forward many intuitive impressions on it. Not too much time left so we left the office and separated with my teacher. We spent the last afternoon wandering in the Black Bamboo Garden and he really loves it because he loves to take pictures. We talked a lot about the GM and I find him really enthusiastic at science and he kept criticizing some paleontologist for they usually using mathematical methods in a wrong way. I was for him on that point, many scientists had this problem especially in the biology and paleontology area. The lotus and bamboo is really well there, I myself was also surprised by the scale of the garden cause I never been there before. Then we had dinner together on the Chinese Valentine's Day...Then we came back to rest on our own.

It seems quite ease in the first day but when it comes to the second day which is today, the situation is different. He talked a lot about the specimen in the paper with my tutor in the noon. He is very knowledgeable to know so many specimens and many people in this area but actually he is mainly doing works on GM. And on the afternoon we discussed the statistical methods in GM. Lot of knowledge I learned and it is really helpful to have somebody speak to you face to face especially when many of the concepts is very virtual and hark to practice. So main problem solved were list as follows.

1.Now I am kind of released from the GM theory. Very interesting the story he said about those professors who invented the GM. He said they kind of like "gurus" and many principles in GM is actually might not needed for true analysis. And also it is the analysis part of the GM that makes me puzzled but actually not really settled down because I want to know more methods to utilize the information that we acquired from the shape. Yet still many methods I learned or made clear today.

2.The usage of the PCA on the superimposition data. Including BC and Procrustes methods. Using Past to analysis the data is very simple and efficient, and the figure it produced is also very beautiful and intuitive.

He is really a geek in using PCA. And he explained the use of loading, biplot, minimum span tree and convex hulls very clearly. It is also very powerful in the past to produce beautiful images about the analysis results.

3.Although Ancova is not so difficult to understand but I did not know how to use it well before. And he explain the Ancova with the past with two variables and several characters. The result is pretty good to fit the data. With the regression you can see if there are any difference between the slopes of different groups. Then you deal with the Anova part.But as for the multiple variable part he said that many people tried but ended with bad results in predicting the characters. 

4. EDMA. This is very interesting because the story is also happened between the geeks. They differ with the GM part on analysis the shape. This is a kind of method using an integrated information of distance matrix. And the procedure to calculate a EDMA is not so simple.But for short, I think both are useful in some aspects reviewing the facts.

5. Allometric analysis.  This is a very interesting part that he knows a lot, and I also read some papers about it. But it is also kind if confused by many people in the paleontology area.

6.PLS. This is probably the most difficult part to understand and to use. Because it contains n dependent variables and n independent variables and it is used to explain how the n x affects the n y ...

7.UPGMA. He love to use this method very much because he said it do not change the truth. Well he love to use methods that do not change the data too much. For example the CVA change a lot about the original data to get a maximum separation between groups...But I doubt it weather this kind of method still work efficiently while many new methods with more beautiful calculations and results exist now.

8.Resampling methods. Including Bootstrap and permutation test these kind of resampling methods he like to use them frequently. And so do I. The widely use of computer make it possible to do this and it is widely used bu scientist around the world and are really good methods.

Very simple I described the methods here but I have to play badminton now and I am already late for 30 min. Maybe details in the future and using them in practice.

And very eager to see his presentation tomorrow about the introduction of GM.