A GMDH-type neural network with multi-filter feature selection for the prediction of transition temperatures of bent-core liquid crystals
Apstrakt
A novel strategy for the prediction of the transition temperature of bent-core liquid crystals (LCs) based on the combination of multi filter feature selection and group method of data handling (GMDH) type neural networks is reported. An entire set of 243 compounds was randomly divided into a training set of 207 compounds and a test set of 36 compounds. Descriptors were selected from a pool of 2D, and two pools of 2D and 3D ones, optimized by molecular mechanics (MM) and semi-empirical (SE) method. The reduction of the pool of descriptors was performed using multi filters based on chi square and v-WSH algorithm, while the final subset selection was performed by GMDH algorithm during the learning process. The obtained 2D, MM and SE GMDH models have 11, 13 and 16 descriptors, respectively, and demonstrate good generalization and predictive ability (R-2 = 0.92). The final models were subjected to a randomization test for validation purpose. Those models appear to be not only suitable for ...prediction, but they also allow the identification of key structural features that alter the transition temperature of bent-core LCs.
Izvor:
RSC Advances, 2016, 6, 102, 99676-99684Izdavač:
- Royal Society of Chemistry
Finansiranje / projekti:
- Razvoj i primena metoda i materijala za monitoring novih zagađujućih i toksičnih organskih materija i teških metala (RS-MESTD-Basic Research (BR or ON)-172007)
- Proučavanje sinteze, strukture i aktivnosti organskih jedinjenja prirodnog i sintetskog porekla (RS-MESTD-Basic Research (BR or ON)-172013)
DOI: 10.1039/c6ra15056j
ISSN: 2046-2069
WoS: 000386439800010
Scopus: 2-s2.0-84993945042
Institucija/grupa
Tehnološko-metalurški fakultetTY - JOUR AU - Antanasijević, Davor AU - Antanasijević, Jelena AU - Pocajt, Viktor AU - Ušćumlić, Gordana PY - 2016 UR - http://TechnoRep.tmf.bg.ac.rs/handle/123456789/3305 AB - A novel strategy for the prediction of the transition temperature of bent-core liquid crystals (LCs) based on the combination of multi filter feature selection and group method of data handling (GMDH) type neural networks is reported. An entire set of 243 compounds was randomly divided into a training set of 207 compounds and a test set of 36 compounds. Descriptors were selected from a pool of 2D, and two pools of 2D and 3D ones, optimized by molecular mechanics (MM) and semi-empirical (SE) method. The reduction of the pool of descriptors was performed using multi filters based on chi square and v-WSH algorithm, while the final subset selection was performed by GMDH algorithm during the learning process. The obtained 2D, MM and SE GMDH models have 11, 13 and 16 descriptors, respectively, and demonstrate good generalization and predictive ability (R-2 = 0.92). The final models were subjected to a randomization test for validation purpose. Those models appear to be not only suitable for prediction, but they also allow the identification of key structural features that alter the transition temperature of bent-core LCs. PB - Royal Society of Chemistry T2 - RSC Advances T1 - A GMDH-type neural network with multi-filter feature selection for the prediction of transition temperatures of bent-core liquid crystals EP - 99684 IS - 102 SP - 99676 VL - 6 DO - 10.1039/c6ra15056j ER -
@article{ author = "Antanasijević, Davor and Antanasijević, Jelena and Pocajt, Viktor and Ušćumlić, Gordana", year = "2016", abstract = "A novel strategy for the prediction of the transition temperature of bent-core liquid crystals (LCs) based on the combination of multi filter feature selection and group method of data handling (GMDH) type neural networks is reported. An entire set of 243 compounds was randomly divided into a training set of 207 compounds and a test set of 36 compounds. Descriptors were selected from a pool of 2D, and two pools of 2D and 3D ones, optimized by molecular mechanics (MM) and semi-empirical (SE) method. The reduction of the pool of descriptors was performed using multi filters based on chi square and v-WSH algorithm, while the final subset selection was performed by GMDH algorithm during the learning process. The obtained 2D, MM and SE GMDH models have 11, 13 and 16 descriptors, respectively, and demonstrate good generalization and predictive ability (R-2 = 0.92). The final models were subjected to a randomization test for validation purpose. Those models appear to be not only suitable for prediction, but they also allow the identification of key structural features that alter the transition temperature of bent-core LCs.", publisher = "Royal Society of Chemistry", journal = "RSC Advances", title = "A GMDH-type neural network with multi-filter feature selection for the prediction of transition temperatures of bent-core liquid crystals", pages = "99684-99676", number = "102", volume = "6", doi = "10.1039/c6ra15056j" }
Antanasijević, D., Antanasijević, J., Pocajt, V.,& Ušćumlić, G.. (2016). A GMDH-type neural network with multi-filter feature selection for the prediction of transition temperatures of bent-core liquid crystals. in RSC Advances Royal Society of Chemistry., 6(102), 99676-99684. https://doi.org/10.1039/c6ra15056j
Antanasijević D, Antanasijević J, Pocajt V, Ušćumlić G. A GMDH-type neural network with multi-filter feature selection for the prediction of transition temperatures of bent-core liquid crystals. in RSC Advances. 2016;6(102):99676-99684. doi:10.1039/c6ra15056j .
Antanasijević, Davor, Antanasijević, Jelena, Pocajt, Viktor, Ušćumlić, Gordana, "A GMDH-type neural network with multi-filter feature selection for the prediction of transition temperatures of bent-core liquid crystals" in RSC Advances, 6, no. 102 (2016):99676-99684, https://doi.org/10.1039/c6ra15056j . .