Algorithm 2 LSTM Training Methods

Require: ILI, Respiratory, and AQI Dataset integration Ensure: Sum up the number of disease cases per week and order them   1:$df=df[:100]\leftarrow$Call the first 100 observation data   2:$MinMaxScaler\leftarrow$Normalize dataset with MinMaxScaler   3:$trainsize=int\left(len\right(df)\ast 0.8)$   4:$valsize=len\left(df\right)-trainsize$   5:$train,val=df[0:trainsize,:],df[trainsize:len(df),:]\leftarrow$Training and validation data partition   6:$previous=5\leftarrow$Number of previous parameter   7:$X\_train,Y\_train=create\_dataset(train,previous)$   8:$X\_val,Y\_val=create\_dataset(val,previous)$   9:$model=tf.keras.Sequential\left(\right)\leftarrow$Create LSTM model 10:$model.add\left(LSTM\right(4,input\_shape=(1,previous)\left)\right)\leftarrow$Add LSTM layers with 4 hidden layers 11:$model.add\left(Dense\right(1\left)\right)\leftarrow$Add Dense Layer 12:$model.compile($ loss=’mean_squared_error’, optimizer=’adam’$)\leftarrow$Compile the model using Adam optimizer 13:$model.fit($ X_train, Y_train, epochs=150, batch_size=1, verbose=2$)\leftarrow$Train the model in 150 epochs 14:$RMSE\leftarrow$Calculate the RMSE