Traditionally we validate an analytical model by simulation and validate a simulation model by measurements/testbed.
What about validation of measurement results? Are they 100% accurate?
Measurement results can be validated in several ways.
Compare with similar results published by other researchers. Repeat the measurements/experiments/field trials to validate the results.
Unfortunately, Analytic models can not provide 100% accurate results because there are so many variables in indoor environment such as occupancy, moisture, temperature, building materials, furniture materials and so on. Each space has its own unique characteristic. So, what can we do? Empirical approach will provide the most accurate results. You can try two different RF meter to validate your data.
We usually validate measurement results using residuals. Residuals are the difference between the values obtained from the measurements and other measurements carried out in different environments. Sometimes it may be necessary to gather repeated data in the same environment or conditions and use them as a basis for the comparison. Plots of the residuals usually give a graphic illustration of the consistency of the data. If the residuals alternate about the zero position (between positive and negative values about zero) then the measurements represents the initial measurement significantly. However a quantitative way to validate the model is to use the RMS errors observed by comparing the two measurement values. If a model was developed from the field measurements, the model can also be validated using the residuals obtained by subtracting the model generated values from the validation data values. RMS errors can also be used. F-test and t test can also provide additional means of checking the model performance.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire