Prediction of clearing temperatures of bent-core liquid crystals using decision trees and multivariate adaptive regression splines

Jelena Antanasijević, Viktor Pocajt, Davor Antanasijević, Nemanja Trišović, K. Fodor-Csorba

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Accurate prediction of transition temperature is very helpful for the design of new liquid crystals (LCs) because even small changes in structure can dramatically alter the transition temperature, and therefore the synthesis of LCs should not be governed only by chemical intuition. A quantitative structure–property relationship (QSPR) study was performed on 243 five-ring bent-core LCs in order to predict their clearing temperatures using molecular descriptors. Decision tree and multivariate adaptive regression splines (MARS), techniques well suited for high-dimensional data analysis, were applied to select important descriptors (dimension reduction) and to generate nonlinear models. These techniques were applied both on two-dimensional (2D) descriptors only and on the pool of 2D and 3D descriptors (2&3D). The obtained QSPR models were tested using 15% of available data, and their performance and ability to generalise were analysed using multiple statistical metrics. The best results for the external test set were obtained using the MARS model created with 2&3D descriptors, with a high correlation coefficient of r = 0.95 and a root mean squared error of 7.41 K. All metrics suggest that the proposed QSPR model, generated by MARS, is a robust and satisfactorily accurate approach for the prediction of clearing temperatures of bent-core LCs.

Original languageEnglish
Pages (from-to)1-10
Number of pages10
JournalLiquid Crystals
DOIs
Publication statusAccepted/In press - Mar 15 2016

Fingerprint

Liquid Crystals
clearing
splines
Decision trees
Splines
Liquid crystals
regression analysis
liquid crystals
predictions
Superconducting transition temperature
transition temperature
Temperature
temperature
correlation coefficients
rings
synthesis

Keywords

  • decision tree
  • MARS modelling
  • QSPR
  • transition temperature

ASJC Scopus subject areas

  • Condensed Matter Physics
  • Materials Science(all)
  • Chemistry(all)

Cite this

Prediction of clearing temperatures of bent-core liquid crystals using decision trees and multivariate adaptive regression splines. / Antanasijević, Jelena; Pocajt, Viktor; Antanasijević, Davor; Trišović, Nemanja; Fodor-Csorba, K.

In: Liquid Crystals, 15.03.2016, p. 1-10.

Research output: Contribution to journalArticle

Antanasijević, Jelena ; Pocajt, Viktor ; Antanasijević, Davor ; Trišović, Nemanja ; Fodor-Csorba, K. / Prediction of clearing temperatures of bent-core liquid crystals using decision trees and multivariate adaptive regression splines. In: Liquid Crystals. 2016 ; pp. 1-10.
@article{a0aa053e072d4fea9f8f64899582599b,
title = "Prediction of clearing temperatures of bent-core liquid crystals using decision trees and multivariate adaptive regression splines",
abstract = "Accurate prediction of transition temperature is very helpful for the design of new liquid crystals (LCs) because even small changes in structure can dramatically alter the transition temperature, and therefore the synthesis of LCs should not be governed only by chemical intuition. A quantitative structure–property relationship (QSPR) study was performed on 243 five-ring bent-core LCs in order to predict their clearing temperatures using molecular descriptors. Decision tree and multivariate adaptive regression splines (MARS), techniques well suited for high-dimensional data analysis, were applied to select important descriptors (dimension reduction) and to generate nonlinear models. These techniques were applied both on two-dimensional (2D) descriptors only and on the pool of 2D and 3D descriptors (2&3D). The obtained QSPR models were tested using 15{\%} of available data, and their performance and ability to generalise were analysed using multiple statistical metrics. The best results for the external test set were obtained using the MARS model created with 2&3D descriptors, with a high correlation coefficient of r = 0.95 and a root mean squared error of 7.41 K. All metrics suggest that the proposed QSPR model, generated by MARS, is a robust and satisfactorily accurate approach for the prediction of clearing temperatures of bent-core LCs.",
keywords = "decision tree, MARS modelling, QSPR, transition temperature",
author = "Jelena Antanasijević and Viktor Pocajt and Davor Antanasijević and Nemanja Trišović and K. Fodor-Csorba",
year = "2016",
month = "3",
day = "15",
doi = "10.1080/02678292.2016.1155769",
language = "English",
pages = "1--10",
journal = "Liquid Crystals",
issn = "0267-8292",
publisher = "Taylor and Francis Ltd.",

}

TY - JOUR

T1 - Prediction of clearing temperatures of bent-core liquid crystals using decision trees and multivariate adaptive regression splines

AU - Antanasijević, Jelena

AU - Pocajt, Viktor

AU - Antanasijević, Davor

AU - Trišović, Nemanja

AU - Fodor-Csorba, K.

PY - 2016/3/15

Y1 - 2016/3/15

N2 - Accurate prediction of transition temperature is very helpful for the design of new liquid crystals (LCs) because even small changes in structure can dramatically alter the transition temperature, and therefore the synthesis of LCs should not be governed only by chemical intuition. A quantitative structure–property relationship (QSPR) study was performed on 243 five-ring bent-core LCs in order to predict their clearing temperatures using molecular descriptors. Decision tree and multivariate adaptive regression splines (MARS), techniques well suited for high-dimensional data analysis, were applied to select important descriptors (dimension reduction) and to generate nonlinear models. These techniques were applied both on two-dimensional (2D) descriptors only and on the pool of 2D and 3D descriptors (2&3D). The obtained QSPR models were tested using 15% of available data, and their performance and ability to generalise were analysed using multiple statistical metrics. The best results for the external test set were obtained using the MARS model created with 2&3D descriptors, with a high correlation coefficient of r = 0.95 and a root mean squared error of 7.41 K. All metrics suggest that the proposed QSPR model, generated by MARS, is a robust and satisfactorily accurate approach for the prediction of clearing temperatures of bent-core LCs.

AB - Accurate prediction of transition temperature is very helpful for the design of new liquid crystals (LCs) because even small changes in structure can dramatically alter the transition temperature, and therefore the synthesis of LCs should not be governed only by chemical intuition. A quantitative structure–property relationship (QSPR) study was performed on 243 five-ring bent-core LCs in order to predict their clearing temperatures using molecular descriptors. Decision tree and multivariate adaptive regression splines (MARS), techniques well suited for high-dimensional data analysis, were applied to select important descriptors (dimension reduction) and to generate nonlinear models. These techniques were applied both on two-dimensional (2D) descriptors only and on the pool of 2D and 3D descriptors (2&3D). The obtained QSPR models were tested using 15% of available data, and their performance and ability to generalise were analysed using multiple statistical metrics. The best results for the external test set were obtained using the MARS model created with 2&3D descriptors, with a high correlation coefficient of r = 0.95 and a root mean squared error of 7.41 K. All metrics suggest that the proposed QSPR model, generated by MARS, is a robust and satisfactorily accurate approach for the prediction of clearing temperatures of bent-core LCs.

KW - decision tree

KW - MARS modelling

KW - QSPR

KW - transition temperature

UR - http://www.scopus.com/inward/record.url?scp=84961215303&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961215303&partnerID=8YFLogxK

U2 - 10.1080/02678292.2016.1155769

DO - 10.1080/02678292.2016.1155769

M3 - Article

AN - SCOPUS:84961215303

SP - 1

EP - 10

JO - Liquid Crystals

JF - Liquid Crystals

SN - 0267-8292

ER -