development of different machine learning models with vector quantization technique and Cuckoo search optimization algorithm, Fall 2021
Greenhouse gases are air pollutants which cause warming of the global climate. These gases like methane and carbon dioxide can be stored in the form of compact gas hydrates to reduce their harmful effects. Prediction of gas hydrate formation condition is very important in gas hydrate production and storage in industries. This study aims at developing machine learning methods based on support vector regression and adaptive boosting to predict the gas hydrates formation condition. In this regard, SVR, AdaBoost.R2, VQ-SVR, VQ-AdaBoost.R2, CS-VQ-SVR and CS-VQ-AdaBoost.R2 models have been compared to obtain one model with the best performance. A cuckoo search algorithm has been used to determine the optimal values of the models’ hyper-parameters. For reducing the computation time and improving the accuracy and robustness of the models, vector quantization technique has been used. As a result, since the values of the coefficient of determination and root mean square error for CS-VQ-SVR model are 0.0215 and 0.9995, respectively, and the best agreement between predicted and actual values in this model’s graphs has obtained, the CS-VQ-SVR model has the best accuracy and robustness among other developed models in predicting hydrate formation pressure with time. These results show that machine learning is viable to predict the conditions of gas hydrate formation and prevent greenhouse gasses emission.