Correlation test for model outputs

In summary, the conversation is about building a daily model in Excel to predict water demand, and comparing it to a more accurate model that runs on smaller time intervals. The question is about which statistical approach to use for comparing the results of the two models, with suggestions of using either CORRELATION or PEARSON functions. The person also brings up the possibility of using a linear regression model for analyzing the data.
  • #1
Richard_R
14
0
Hello all,

I am currently building a model in Excel for predicting domestic water demand on a daily basis. The daily model will be compared to a more accurate model which runs on timesteps of 5 minutes (the latter is data intensive however which is why we are building a daily model as it requires less data).

Does anyone know what the correct correlation/statistical approach is to compare the daily model results with the 5 minute model? I have aggregated the results from the 5 minute model to a daily timestep so results can be directly compared, i.e.

http://sudsolutions.co.uk/misc/model_results.PNG

Excel has CORRELATION and PEARSON statistical functions so was wondering if I need to use one of these. The results aren't in the form of y=mx+c so I don't think I want an r^2 "goodness of fit" test (or do I?).

Thanks in advance for any help.

Regards
Rob
 
Physics news on Phys.org
  • #2
PEARSON or CORRELATION seem to be the way to go for your data (from excel help I think they do the same thing?)

Correlation can broadly refer to either linear or nonlinear correlation. I suppose you are interested in the linear case, and thus can use a linear regression model in the form of y = mx + c to analyze the data.
 

FAQ: Correlation test for model outputs

What is a correlation test for model outputs?

A correlation test for model outputs is a statistical method used to determine the strength of the relationship between two or more variables in a model. It measures how closely the variables are related to each other and whether the relationship is positive or negative.

When should a correlation test for model outputs be used?

A correlation test for model outputs should be used when you want to understand the relationship between two or more variables in a model. It is commonly used in scientific research, data analysis, and predictive modeling to identify patterns and make predictions.

How is a correlation test for model outputs performed?

A correlation test for model outputs is typically performed using a statistical software or a calculator. The most commonly used method is Pearson's correlation coefficient, which measures the linear relationship between two continuous variables. Other methods include Spearman's rank correlation coefficient and Kendall's tau coefficient, which are used for non-parametric data.

What do the results of a correlation test for model outputs mean?

The results of a correlation test for model outputs provide information about the strength and direction of the relationship between the variables. The correlation coefficient ranges from -1 to 1, with 0 indicating no correlation, 1 indicating a perfect positive correlation, and -1 indicating a perfect negative correlation. The closer the value is to 1 or -1, the stronger the relationship between the variables.

Can a correlation test for model outputs determine causation?

No, a correlation test for model outputs cannot determine causation. Correlation only measures the degree of association between variables, not causation. Other factors and variables may be involved in the relationship, and further research is needed to establish causality.

Similar threads

I Coefficient sign flip in linear regression with correlated predictors
Replies
13
Views
3K
I Akaike Information Criterion Vs Likelihood Ratio Test
Replies
5
Views
3K
A How to correctly analyze the results of a preference study?
Replies
13
Views
2K
Issues with SLS Model for Rubber and Foam Compression Tests
Replies
5
Views
452
I Significant correlation, not significant coefficient
Replies
4
Views
2K
Comparing Near-Infrared Spectra: What Stat Method?
Replies
5
Views
1K
Hypothesis testing of a complex deterministic model
Replies
19
Views
3K
F test for R^2 when comparing two models
Replies
2
Views
2K
How to automate tests on a physical model without real data
Replies
4
Views
970
Comparing model fits for many populations - F statistic
Replies
1
Views
1K

Hot Threads

Recent Insights

Back
Top