|
Article · December 23 doi: 10. 30871/jaic v7 6809 citations read authorsBog'liq 01#Повторная калибровка модельного емкостного датчика для измерения влажности почвы IoTModel
Adjusted
R Squared
Linear Regression
0.821
Polynomial Regression Order 2
0.927
Polynomial Regression Order 3
0.945
Table 4 shows that polynomial regression order 3 has the
highest Adjusted R Squared value. Adjusted R Squared of
0.945 means that 94.5% of the variation in volumetric soil
moisture content can be explained by the model. The
polynomial regression order 3 equation is then used to predict
soil moisture.
D. Re-calibration Application
Based on the results of the analysis that has been carried
out, an interactive web application is assemble using the shiny
package in R. An interactive web application uses an event
observation system, namely specific actions that must be
taken from the user before calculating an expression, on the
analyze button. The web application can be accessed and used
on pages that have links:
https://statisticsontraining1.shinyapps.io/soilmoisture/
.
The re-calibration application algorithm to produce the best
model is as follows.
Figure 2. Algorithm for recalibration application
Figure 2 shows the stages of the gravimetric test results
modelling process. Recalibration will use gravimetric test
input to produce the best model coefficients that will be used
for capacitive sensor recalibration. The input panel display is
as follows:
Figure 3. Input Panel of Re-calibration Application
Figure 3 shows that the input panel on the web dashboard
application divided into three columns. The first column is to
calculate the gravimetric water content in dry soil sample. The
second column is the mass of the soil (g) to the increase in
water volume. The third column is the raw data of the soil
moisture sensor (v). At the bottom, there is an analyze button
which will trigger the modelling process. The output panel
display is as follows:
Start
End
Data
Linear Model
Regression
Quadratic Model
Regression
Gravimetric
Equations
Model Selection
using Adjusted
R Squared
|
| |