I have got a minute-by-minute indoor air quality (pollutant concentration) monitoring data set for 3 hours. For the same period, I have minute-by-minute modeling dataset. Would somebody please suggest the most appropriate statistical test to evaluate the accuracy of modelled dataset. Is simple pearson correlation coefficient good enough? The data sets are not normal as per Kolmogorov-Smirnov and shapiro-wilk tests (although, n=180).
according to you data you should use non parametric test because data lost very important assumption , i mean normality distribution so you can use two sample non parametric like sign test to see if they is difference or not. also if you want correlation use spearman's rank .
thanks
Thank you Abdullahi and Mercedec for your valuable responses :)
Regards.
- Badges
- Science topic