Random Forest Model Results

AQI prediction results using Random Forest algorithm

Mean Error (ME)
-1.27
Value to be added
Root Mean Square Error
19.37
RMSE Value
Model Accuracy
71 %
Percentage

Actual vs. Predicted AQI (t+1 hour)

Random Forest Model Scatter Plot

72-Hour AQI Forecast

Time Predicted AQI

LSTM Model Results

AQI prediction results using Long Short-Term Memory networks

Mean Error (ME)
-3.65
Value to be added
Root Mean Square Error
30.17
RMSE Value
Model Accuracy
45 %
Percentage

Actual vs. Predicted AQI (t+1 hour)

LSTM Scatter Plot

72-Hour AQI Forecast

Time Predicted AQI

Transformer Model Results

AQI prediction results using Transformer architecture

Mean Error (ME)
-13.74
Value to be added
Root Mean Square Error
40.26
RMSE Value
Model Accuracy
36 %
Percentage

Actual vs. Predicted AQI (t+1 hour)

Transformer Model Scatter Plot

72-Hour AQI Forecast

Time Predicted AQI

Transformer Model Results

AQI prediction results using Transformer architecture

Mean Error (ME)
3.27
Value to be added
Root Mean Square Error
18.90
RMSE Value
Accuracy
92 %
Percentage

Actual vs. Predicted AQI (t+1 hour)

Transformer Model Scatter Plot

72-Hour AQI Forecast

Time Predicted AQI

Model Comparison

Performance comparison across all three models

Random Forest
Mean Error -1.27
RMSE 19.37
Accuracy 71 %
LSTM
Mean Error -3.65
RMSE 30.17
Accuracy 45 %
Transformer(4 cities)
Mean Error 3.27
RMSE 18.90
Accuracy 92 %

72-Hour Forecast Comparison

Timestamp Random Forest LSTM Transformer (4 Cities)