An Integrated Smart Pond Water Quality Monitoring and Fish Farming Recommendation Aquabot System

. 2024 Jun 6;24(11):3682. doi: 10.3390/s24113682

Algorithm 1 Proposed Custom Ensemble Method

1:
Input: $D_{r}$ = Dataset; $n_m o d e l s$ = Number of base models;
2:
Output: Trained ensemble model merging bagging, boosting, and stacking.
3:
function Ensemble Pseudocode( $D_{r}$ , $n_m o d e l s$ )
4:
Initialization:
5:
Set $n_m o d e l s$ to the desired number of base models.
6:
Initialize an empty list $m o d e l_l i s t$ to store the trained base models.
7:
for $i = 1$ to $n_m o d e l s$ do
8:
Training the Base Model:
9:
Train a base model to predict class labels on the entire training dataset $D_{r}$ .
10:
Append the trained model to $m o d e l_l i s t$ .
11:
end for
12:
Making Predictions:
13:
Initialize an empty list $p r e d i c t i o n s$ to store predictions from each base model.
14:
for each base model in $m o d e l_l i s t$ do
15:
Use the base model to predict the class labels for the testing dataset.
16:
Append the predictions to $p r e d i c t i o n s$ .
17:
end for
18:
Combining Predictions:
19:
For each instance in the testing dataset:
20:
Count the class labels predicted by each base model.
21:
Select the most frequent class label as the final prediction.
22:
Evaluating Ensemble Performance:
23:
Evaluate the ensemble’s performance on the test dataset.
24:
end function