How can I compare two measurements which are composed of matrices such as [0 1 0 0...]
1 回表示 (過去 30 日間)
古いコメントを表示
I'm currently doing analysis to compare two measurements, so I decided to use cohen's kappa.
The problem is that there are no meaningful values as a result.
For example, I have a two processed data like below.
emg = [0 0 0 0 0 1 1 0 0]; video = [0 0 0 0 0 0 0 0 0];
Do I need to make another row for 0, 1 each?, like
emg = [0 0 0 0 0 0 0 0 0];, emg = [1 1 1 1 1 0 0 1 1];
video = [0 0 0 0 0 0 0 0 0];, video = [1 1 1 1 1 1 1 1 1];
If anyone have an idea about statistical analysis, please give me a hand.
2 件のコメント
dpb
2019 年 3 月 15 日
We don't know enough about your measurements to be able to judge...what is the measurement of and is this a rating system? Cohen's Kappa is to judge consonance between to raters' judgement; doesn't really sound like that's the null hypothesis here.
採用された回答
Jos (10584)
2019 年 3 月 18 日
You are looking for a measure of consistency between two measurements which are represented as binary vectors of the same length? My idea would be to count the discrepancies between them, and perhaps correct for the number of observations:
emg = [0 0 0 0 0 1 1 0 0]; video = [0 0 0 0 0 0 0 0 0];
fraction_of_discrepancy = sum(abs(emg - video)) / numel(emg)
3 件のコメント
Jos (10584)
2019 年 3 月 19 日
I would have expected a single number.
In my formula, a 0 would indeed indicate that the sequences of 0's and 1's are the same in the two vectors.
その他の回答 (0 件)
参考
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!