Statistical significance of gridded anomaly matrix

2 次查看（过去 30 天）

Claire 2015-1-14

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/169872-statistical-significance-of-gridded-anomaly-matrix

评论： Sanatan Binisia 2021-4-2

I have two 2D matrices which contain latitude/longitude temperature data, with dimensions 22x12. I.e matrix A(22x12) and matrix B(22x12).

I would like to determine if my matrix A is statistically significant from matrix B.

Matrix B was created by changing a parameter in matrix A, so I was looking at using a student's paired t-test, which does not assume that your two datasets are independent.

The issue I am having is in trying to apply the t-test to my data.

I have tried to use the 'ttest' function but the analysis seems to be done along columns

[hlat,plat]=ttest(A,B);

ans is a vector with 12 points.

What I actually want is to test for significance at each grid point, so that I can determine where (in terms of latitude and longitude coordinates), my model is giving statistically significant results. I.e. I'd like a 22x12 output of the significance test. I then want to use this to mark the statistically significant regions in an anomaly plot.

I'm not sure if/how to use 'ttest' to do this, or if I am using the wrong function all together.

Any help would be greatly appreciated! Thanks!