Prediction of skiing time by structured regression algorithm
Апстракт
In this paper, the application of Gaussian conditional random fields (GCRF) in the case of prediction skiing time between ski gates in ski center Kopaonik, is presented. Gaussian conditional random fields is well-known structured regression method that exploits advantages of unstructured predictors and combines them with the information concerning correlation between outputs. Four different unstructured predictors were used: ridge regression, LASSO regression, Random forest regression and support vector machine regression. Even thought, only 18 features are used for prediction of skiing time, GCRF achieved better results, concerning R2 and mean absolute error, compared to unstructured predictors.
Кључне речи:
structured regression / Gaussian conditional random fields / GCRF / skiingИзвор:
7th International Symposium on Industrial Engineering – SIE 2018, 27th-28th September, 2018, 180-183Издавач:
- University of Belgrade - Faculty of Mechanical Engineering, Department of Industrial Engineering
Колекције
Институција/група
Mašinski fakultetTY - CONF AU - Petrović, Andrija AU - Bugarić, Uglješa AU - Delibašić, Boris AU - Ivetić, Igor PY - 2018 UR - https://machinery.mas.bg.ac.rs/handle/123456789/5878 AB - In this paper, the application of Gaussian conditional random fields (GCRF) in the case of prediction skiing time between ski gates in ski center Kopaonik, is presented. Gaussian conditional random fields is well-known structured regression method that exploits advantages of unstructured predictors and combines them with the information concerning correlation between outputs. Four different unstructured predictors were used: ridge regression, LASSO regression, Random forest regression and support vector machine regression. Even thought, only 18 features are used for prediction of skiing time, GCRF achieved better results, concerning R2 and mean absolute error, compared to unstructured predictors. PB - University of Belgrade - Faculty of Mechanical Engineering, Department of Industrial Engineering C3 - 7th International Symposium on Industrial Engineering – SIE 2018, 27th-28th September T1 - Prediction of skiing time by structured regression algorithm EP - 183 SP - 180 UR - https://hdl.handle.net/21.15107/rcub_machinery_5878 ER -
@conference{ author = "Petrović, Andrija and Bugarić, Uglješa and Delibašić, Boris and Ivetić, Igor", year = "2018", abstract = "In this paper, the application of Gaussian conditional random fields (GCRF) in the case of prediction skiing time between ski gates in ski center Kopaonik, is presented. Gaussian conditional random fields is well-known structured regression method that exploits advantages of unstructured predictors and combines them with the information concerning correlation between outputs. Four different unstructured predictors were used: ridge regression, LASSO regression, Random forest regression and support vector machine regression. Even thought, only 18 features are used for prediction of skiing time, GCRF achieved better results, concerning R2 and mean absolute error, compared to unstructured predictors.", publisher = "University of Belgrade - Faculty of Mechanical Engineering, Department of Industrial Engineering", journal = "7th International Symposium on Industrial Engineering – SIE 2018, 27th-28th September", title = "Prediction of skiing time by structured regression algorithm", pages = "183-180", url = "https://hdl.handle.net/21.15107/rcub_machinery_5878" }
Petrović, A., Bugarić, U., Delibašić, B.,& Ivetić, I.. (2018). Prediction of skiing time by structured regression algorithm. in 7th International Symposium on Industrial Engineering – SIE 2018, 27th-28th September University of Belgrade - Faculty of Mechanical Engineering, Department of Industrial Engineering., 180-183. https://hdl.handle.net/21.15107/rcub_machinery_5878
Petrović A, Bugarić U, Delibašić B, Ivetić I. Prediction of skiing time by structured regression algorithm. in 7th International Symposium on Industrial Engineering – SIE 2018, 27th-28th September. 2018;:180-183. https://hdl.handle.net/21.15107/rcub_machinery_5878 .
Petrović, Andrija, Bugarić, Uglješa, Delibašić, Boris, Ivetić, Igor, "Prediction of skiing time by structured regression algorithm" in 7th International Symposium on Industrial Engineering – SIE 2018, 27th-28th September (2018):180-183, https://hdl.handle.net/21.15107/rcub_machinery_5878 .