東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back to Search results for [ null ]

Switch To: Labeled | MARC Mode | ISBD

Predicting Spin-Symmetry Breaking in...

Patidar, Krutika.

Linked to FindBook

Google Book

Amazon

博客來

Predicting Spin-Symmetry Breaking in Organic Photovoltaic Compounds Using a Data Mining Approach.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Predicting Spin-Symmetry Breaking in Organic Photovoltaic Compounds Using a Data Mining Approach./
Author:	Patidar, Krutika.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2020,
Description:	47 p.
Notes:	Source: Masters Abstracts International, Volume: 82-01.
Contained By:	Masters Abstracts International82-01.
Subject:	Chemical engineering. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27996901
ISBN:	9798641331256

Predicting Spin-Symmetry Breaking in Organic Photovoltaic Compounds Using a Data Mining Approach.
Patidar, Krutika.

Predicting Spin-Symmetry Breaking in Organic Photovoltaic Compounds Using a Data Mining Approach. - Ann Arbor : ProQuest Dissertations & Theses, 2020 - 47 p.

Source: Masters Abstracts International, Volume: 82-01.

Thesis (M.S.)--State University of New York at Buffalo, 2020.

This item must not be sold to any third party vendors.

In quantum chemistry, spin-symmetry breaking occurs in electronic structure calculations resulting in significant deviations from the physically valid spin states. However, it is difficult to predict its occurrence in a compound before actually performing these potentially expensive calculations. In this work, we set out to build a data model to prescreen organic compounds with respect to likely instances of spin-symmetry breaking before performing any quantum chemical calculations. We have built a number of machine learning classifier models (utilizing decision tree, random forest, neural network approaches) for this purpose, trained and tested on the molecular systems contained in the Harvard Clean Energy Project database. Our study employs Morgan fingerprints and molecular descriptors of the Dragon library as feature representations. The validation studies use 5-fold cross validation and the test set data provides inferential statistical analysis of accuracy, precision, AUC-ROC scores, and other measures. Based on our evaluation, tree-based algorithms (Decision trees/Random forest) have classified data with 99.9% accuracy and AUC-ROC score of 0.9 when built on Dragon molecular descriptors. We also perform statistical analysis to correlate feature definitions and basic building block structure with spin-symmetry breaking and propose structure property relationship from these analyses. The analysis suggests the presence of thiophene, 1H-pyrrole, naphthalene, or pyrazine as one of the few probable reasons for spin-symmetry breaking in these molecules besides other dominating factors such as conjugated π-bonds, =CH- atom-type bonds. The findings from our work should be useful in aiding quantum chemists to search for a reliable and computationally feasible quantum chemical method for compounds with high probability of spin contamination error.

ISBN: 9798641331256Subjects--Topical Terms:

560457
Chemical engineering.
Subjects--Index Terms:

Data-driven classification

Predicting Spin-Symmetry Breaking in Organic Photovoltaic Compounds Using a Data Mining Approach.
LDR:03116nmm a2200385 4500 001 2273365
005 20201109125227.5
008 220629s2020 ||||||||||||||||| ||eng d
020 $a 9798641331256
035 $a (MiAaPQ)AAI27996901
035 $a AAI27996901
040 $a MiAaPQ $c MiAaPQ
100 1 $a Patidar, Krutika. $3 3550803
245 1 0 $a Predicting Spin-Symmetry Breaking in Organic Photovoltaic Compounds Using a Data Mining Approach.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 47 p.
500 $a Source: Masters Abstracts International, Volume: 82-01.
500 $a Advisor: Hachmann, Johannes.
502 $a Thesis (M.S.)--State University of New York at Buffalo, 2020.
506 $a This item must not be sold to any third party vendors.
520 $a In quantum chemistry, spin-symmetry breaking occurs in electronic structure calculations resulting in significant deviations from the physically valid spin states. However, it is difficult to predict its occurrence in a compound before actually performing these potentially expensive calculations. In this work, we set out to build a data model to prescreen organic compounds with respect to likely instances of spin-symmetry breaking before performing any quantum chemical calculations. We have built a number of machine learning classifier models (utilizing decision tree, random forest, neural network approaches) for this purpose, trained and tested on the molecular systems contained in the Harvard Clean Energy Project database. Our study employs Morgan fingerprints and molecular descriptors of the Dragon library as feature representations. The validation studies use 5-fold cross validation and the test set data provides inferential statistical analysis of accuracy, precision, AUC-ROC scores, and other measures. Based on our evaluation, tree-based algorithms (Decision trees/Random forest) have classified data with 99.9% accuracy and AUC-ROC score of 0.9 when built on Dragon molecular descriptors. We also perform statistical analysis to correlate feature definitions and basic building block structure with spin-symmetry breaking and propose structure property relationship from these analyses. The analysis suggests the presence of thiophene, 1H-pyrrole, naphthalene, or pyrazine as one of the few probable reasons for spin-symmetry breaking in these molecules besides other dominating factors such as conjugated π-bonds, =CH- atom-type bonds. The findings from our work should be useful in aiding quantum chemists to search for a reliable and computationally feasible quantum chemical method for compounds with high probability of spin contamination error.
590 $a School code: 0656.
650 4 $a Chemical engineering. $3 560457
650 4 $a Quantum physics. $3 726746
650 4 $a Artificial intelligence. $3 516317
650 4 $a Organic chemistry. $3 523952
653 $a Data-driven classification
653 $a Machine learning
653 $a Organic photovoltaic compounds
653 $a Spin-symmetry breaking
653 $a Structure-property relationship
690 $a 0542
690 $a 0599
690 $a 0800
690 $a 0490
710 2 $a State University of New York at Buffalo. $b Chemical and Biological Engineering. $3 1023676
773 0 $t Masters Abstracts International $g 82-01.
790 $a 0656
791 $a M.S.
792 $a 2020
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27996901