Prediction_Of_Total_Prodiced_Biogas_Flow

The code in here is voluteering work;
Recursive Support Vector Regressor model is built aimig to measure and show the perfomance of data collected to predict Biogaz flow.
The codes are as follows:

Recursive_SVM_Feature_selective_for_Biogas_Flow_Prediction.ipynb

The code presnet Support Vector Regression (SVR) model used on a dataset that aims at predicting Total preduction of Biogaz Flow. The SVR model recursively eliminate features from the dataset based on their SVR linear coeffiecient values, more spesifically the model eliminate features with the smallest resulting coeffecient weights.
Results of the model are still found to be high and thus normalizing functions are implemented in the later work / codes.

Recursive_SVM_scaled_Feature_selective_for_Biogas_Flow_Prediction.ipynb

The code is similar to the one above. This time, however, the data was scaled by using a normalized function, 'MinMax_Scalaer' function in sklearn library.

Recursive_SVM_Standard_scaled_Feature_selective_for_Biogas_Flow_Prediction_Results.ipynb

The code includes the results of the model with regards to Mean Square Error, Root Mean Square Error, Mean Absolute Percentage Error, Nash–Sutcliffe Model Efficiency, and Accuracy Score.

Recursive_SVM_Standard_scaled_Feature_selective_for_Biogas_Flow_Prediction_Results2.ipynb

The code includes the results of Percent bias (PBIAS), Average Bias Deviation, and Index of Agreement. The C parameter for the model was edited to aquire better results, mainly in terms of 'Index of Agreement' measure.

Cite

If you use the dataset, please consider citing the folowing paper: