多组件学习器实现有机分子沸点的精准预测

doi:10.6023/A22010017

化学学报 ›› 2022, Vol. 80 ›› Issue (6): 714-723.DOI: 10.6023/A22010017 上一篇下一篇

研究论文

多组件学习器实现有机分子沸点的精准预测

刘雨泽^a, 李昆华^a, 黄佳兴^a, 于曦^a^,^b^,^*(), 胡文平^a^,^b^,^*()

a 天津大学化学系天津 300072
b 天津大学天津市分子光电科学重点实验室天津 300072

投稿日期:2022-01-10 发布日期:2022-07-07
通讯作者: 于曦, 胡文平
基金资助:
国家自然科学基金(21973069); 国家自然科学基金(21773169); 天津大学北洋青年学者计划(2018XRX-0007); 天津大学-青海民族大学自主创新合作基金(2021XZC-0064); 青海民族大学青藏高原资源化学与生态环境保护国家民委重点实验室开放基金(2021GKF-1002); 天津大学大学生创新创业训练计划(201910056451)

Accurate Prediction of the Boiling Point of Organic Molecules by Multi-Component Heterogeneous Learning Model

Yuze Liu^a, Kunhua Li^a, Jiaxing Huang^a, Xi Yu^a^,^b(), Wenping Hu^a^,^b()

a Department of Chemistry, Tianjin University, Tianjin 300072
b Tianjin Key Laboratory of Molecular Optoelectronic Sciences, Tianjin University, Tianjin 300072

Received:2022-01-10 Published:2022-07-07
Contact: Xi Yu, Wenping Hu
Supported by:
National Natural Science Foundation of China(21973069); National Natural Science Foundation of China(21773169); PEIYANG Young Scholars Program of Tianjin University(2018XRX-0007); Tianjin University-Qinghai Nationalities University Innovation Cooperation Fund(2021XZC-0064); Qinghai University for Nationalities and Qinghai-Tibet Plateau Resource Chemistry and Ecological Environmental Protection Key Laboratory of National Civil Affairs Commission Open Fund(2021GKF-1002); College Student Innovation and Entrepreneurship Training Program of Tianjin University(201910056451)

文章分享

1. Supporting Information-A22010017.pdf(400KB)

摘要/Abstract

沸点(BP)是有机分子液体的基本物理化学量, 也是化学工业生产中的重要参数. 有机分子的沸点由分子结构决定, 呈现复杂的结构-沸点关系, 函数法(Function Method)、基团贡献法(Group Contribution Method)等传统方法无法应对复杂多样有机分子结构的预测, 应用范围狭窄, 预测精度低. 本研究中, 我们利用基于人工神经网络(ANN)和支持向量机(SVM)的多组件学习器实现有机分子沸点的精准预测. 我们构建了基于可解释性描述符的ANN、基于相关性描述符的ANN及基于复合分子指纹的SVM三个异质模型, 并通过包含4550个各种类别的有机分子沸点的数据集进行训练得到了三个异质性学习器, 最后集成三个学习器对有机分子沸点进行预测. 相比于传统方法和此前的定量结构性质关系(QSPR)模型, 多组件模型结合了三种模型的优点, 展现出很好的预测精度和泛化能力以及低的过拟合, 实现了对多种类型有机分子的沸点的有效预测.

关键词: 化学信息学, 机器学习, 异质学习模型, 人工神经网络, 支持向量机, 集成学习, 有机分子沸点

Boiling point (BP) is the basic physical chemistry quantity of organic molecular liquids, and an important parameter in the chemical industry. The boiling point of organic molecules is determined by the molecular structure, showing a complex structure-property relationship. Traditional methods such as function method and group contribution method cannot cope with the prediction of BP of organic molecules of complex and diverse structures. In this study, we developed an ensemble machine learning method, combining artificial neural networks (ANN) and support vector machines (SVM), to build a multi-component model to realize the boiling point prediction of high accuracy in wide structure space. We developed three heterogeneous models: ANN based on interpretable descriptors, ANN based on correlated descriptors, and SVM based on hybrid molecular fingerprints. The three heterogeneous models were evaluated and optimized using a boiling points data set containing 4550 organic molecules of various categories. The first model (Model A) is an interpretable ANN with selected descriptors possessing physical properties such as molecular composition and molecular chain branches, and with an ANN structure of 19-14-1. The second model (Model B) is an ANN with 35-30-1 structure, of which the descriptors were obtained by using correlation coefficient selecting and bidirectional stepwise regression method. The third model (Model C) is a fingerprint SVM, which had integrated molecular fingerprint information of 1679 bit. Among the three component models, Model A had the best generalization ability and Model B had the best prediction accuracy. The heterogeneous multi-component learner combined with the three models showed a better prediction accuracy and generalization property with low overfitting than that of a single model or homogeneous model. The mean square error on test set of the multi-component model is 12 K by a 5-fold cross validation on 4550 organic molecular data sets, better than all the previous BP prediction methods. The ensemble machine learning strategy we developed here based on heterogeneous models can also be well generalized to the prediction of physical chemical properties beyond BP.

Key words: chemical informatics, machine learning, heterogeneous learning model, artificial neural network, supported vector machine, ensemble learning, organic molecular boiling point

引用此文

刘雨泽, 李昆华, 黄佳兴, 于曦, 胡文平. 多组件学习器实现有机分子沸点的精准预测[J]. 化学学报, 2022, 80(6): 714-723.

Yuze Liu, Kunhua Li, Jiaxing Huang, Xi Yu, Wenping Hu. Accurate Prediction of the Boiling Point of Organic Molecules by Multi-Component Heterogeneous Learning Model[J]. Acta Chimica Sinica, 2022, 80(6): 714-723.

导出引用 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

分享此文

图/表 9

参考文献 37

[1]	Walker, J. J. Chem. Soc. 1894, 65, 193. doi: 10.1039/CT8946500193
[2]	Joback, K. G.; Reid, R. C. Chem. Eng. Commun. 1987, 57, 230.
[3]	Wiener, H. J. Am. Chem. Soc. 1947, 69, 17. doi: 10.1021/ja01193a005
[4]	Katritzky, A. R.; Kuanar, M.; Slavov, S.; Hall, C. D.; Karelson, M.; Kahn, I.; Dobchev, D. A. Chem. Rev. 2010, 110, 5714. doi: 10.1021/cr900238d pmid: 20731377
[5]	Zhu, B.; Wu, R.; Yu, X. Acta Chim. Sinica 2020, 78, 1366. (in Chinese) doi: 10.6023/A20070306
	(朱博阳, 吴睿龙, 于曦, 化学学报, 2020, 78, 1366.) doi: 10.6023/A20070306
[6]	Wei, J.; Chu, X.; Sun, X.; Xu, K.; Deng, H.; Chen, J.; Wei, Z.; Lei, M. InfoMat 2019, 1, 338. doi: 10.1002/inf2.12028
[7]	Song, Z.; Chen, X.; Meng, F.; Cheng, G.; Wang, C.; Sun, Z.; Yin, W. J. Chin. Phys. B 2020, 29, 68.
[8]	Wu, W.; Sun, Q. Scientia Sinica: Physica, Mechanica et Astronomica 2018, 48, 58. (in Chinese)
	(吴炜, 孙强, 中国科学: 物理学, 力学, 天文学, 2018, 48, 58.)
[9]	Liu, Y. D.; Yang, Q.; Li, Y.; Zhang, L.; Luo, S. Z. Chin. J. Org. Chem. 2020, 40, 3812. (in Chinese) doi: 10.6023/cjoc202006051
	(刘伊迪, 杨骐, 李遥, 张龙, 罗三中, 有机化学, 2020, 40, 3812.) doi: 10.6023/cjoc202006051
[10]	Fissa, M. R. J. Mol. Graph. Model. 2019, 87, 109. doi: 10.1016/j.jmgm.2018.11.013
[11]	Goll, E. S.; Jurs, P. C. J. Chem. Inf. Model. 1999, 39, 974.
[12]	Beck, B.; Breindl, A.; Clark, T. J. Chem. Infor. Comp. Sci. 2000, 40, 1046.
[13]	Chalk, A. J.; Beck, B.; Clark, T. J. Chem. Infor. Comp. Sci. 2001, 41, 457.
[14]	Gharagheizi, F.; Mirkhani, S. A.; Ilani-Kashkouli, P.; Mohammadi, A. H.; Ramjugernath, D.; Richon, D. Fluid Phase Equilib. 2013, 354, 250. doi: 10.1016/j.fluid.2013.06.034
[15]	Yaws, C. L. Yaws' Critical Property Data for Chemical Engineers and Chemists, 2012, http://app.knovel.com/hotlink/toc/id:kpYCPDCECD/yaws-critical-property/yaws-critical-property
[16]	PubChem-https://pubchem.ncbi.nlm.nih.gov/
[17]	Mauri, A. Ecotoxicological QSARs, Methods in Pharmacology and Toxicology, Roy, K., New York, 2020, pp. 801-820.
[18]	Kubic, W. L.; Jenkins, R. W.; Moore, C. M.; Semelsberger, T. A.; & Sutton, A. D. Ind. Eng. Chem. Res. 2017, 56, 12236. doi: 10.1021/acs.iecr.7b02753
[19]	Kier, L. B.; Hall, L. H. In Molecular ConnectiVity in Chemistry and Drug Research, New York, 1976, pp. 27-39.
[20]	Dash, M.; Liu, H. Intell. Data Anal., 1997, 1, 131. doi: 10.3233/IDA-1997-1302
[21]	Roy, P. P.; Leonard, J. T.; Roy, K. Chemometr. Intell. Lab. Syst. 2008, 90, 31. doi: 10.1016/j.chemolab.2007.07.004
[22]	MacKay, J. C. Neural Comput. 1992, 4, 415 doi: 10.1162/neco.1992.4.3.415
[23]	Zhou, Z. H. Ensemble Methods: Foundations and Algorithms, Chapman and Hall/CRC, 2012, p. 129
[24]	Zhou, L.; Wang, B.; Jiang, J.; Pan, Y.; & Wang, Q. Chemometr. Intell. Lab. Syst. 2017, 167, 190. doi: 10.1016/j.chemolab.2017.06.009
[25]	Robert, T. J. R. Stat. Soc. Ser. A 1996, 58, 267.
[26]	Sheela, K.; Deepa, S. N. Math. Probl. Eng. 2013, 2013, 11.
[27]	Eriksson, L.; Jaworska, J.; Worth, A. P.; Cronin, M. T. D.; McDowell, R. M.; Gramatica, P. Environ. Health Perspect. 2003, 111, 1361. doi: 10.1289/ehp.5758
[28]	Needham, D. E.; Wei, I. C.; Seybold, P. G. J. Am. Chem. Soc. 1988, 110, 4186. doi: 10.1021/ja00221a015
[29]	Balaban, A. T.; Ciubotariu, D.; Medeleanu, M. J. Chem. Infor. Comp. Sci. 1991, 313, 517.
[30]	Stanton, D. T. J. Chem. Infor. Comp. Sci. 2000, 40, 81.
[31]	Hall, L. H.; Kier, L. B. J. Chem. Infor. Comp. Sci. 1995, 35, 1039.
[32]	Randić, M.; Balaban, A. T.; Basak, S. J. Chem. Infor. Comp. Sci. 2001, 41, 593.
[33]	Katritzky, A. R.; Mu, L.; Lobanov, V. S.; Karelson, M. J. Phys. Chem. 1996, 100, 10400. doi: 10.1021/jp953224q
[34]	Zhou, C. Y.; Nie, C. M.; Li, S.; Li, Z. H. J. Comput. Chem. 2007, 28, 2413. doi: 10.1002/jcc.20540
[35]	Katritzky, A. R.; Lobanov, V. S.; Karelson, M. J. Chem. Infor. Comp. Sci. 1998, 38, 28.
[36]	Varamesh, A.; Hemmati-Sarapardeh, A.; Dabir, B.; Mohammadi, A. H. J. Mol. Liq. 2017, 242, 59. doi: 10.1016/j.molliq.2017.06.039
[37]	Sola, D.; Ferri, A.; Banchero, M.; Manna, L.; Sicardi, S. Fluid Phase Equilib. 2008, 263, 33. doi: 10.1016/j.fluid.2007.09.022

No.	Name	Type	Detail
1	Log P	Molecular properties	Octanol-water partition coeff.
2	nROH	Functional group	Number of hydroxyl groups
3	nROR	Functional group	Number of ethers
4	nR=O	Functional group	Number of carbonyl
5	nRNH₂	Functional group	Number of 1st amine
6	nRNHR	Functional group	Number of 2st amine
7	nRNR₂	Functional Group	Number of 3st amine
8	nRCN	Functional Group	Number of nitriles
9	nSatom	Constitutional indices	Number of Surfur atom
10	nFatom	Constitutional indices	Number of Fluorine atom
11	nClatom	Constitutional indices	Number of Chlorine atom
12	nBratom	Constitutional indices	Number of Bromine atom
13	nIatom	Constitutional indices	Number of Iodine atom
14	nCatom	Constitutional indices	Number of carbon atom
15	HA	Structural Parameter	Number of acceptor atom for H-bond
16	HD	Structural Parameter	Number of donor atom for H-bond
17	nCharge	Constitutional indices	Total charge
18	Chi2n	Topological indices	2-Path kier shape index
19	AromR	Ring descriptors	Aromatic Ratio

No.	Name	Type	Detail
1	Log P	Molecular properties	Octanol-water partition coeff.
2	nROH	Functional group	Number of hydroxyl groups
3	nROR	Functional group	Number of ethers
4	nR=O	Functional group	Number of carbonyl
5	nRNH₂	Functional group	Number of 1st amine
6	nRNHR	Functional group	Number of 2st amine
7	nRNR₂	Functional Group	Number of 3st amine
8	nRCN	Functional Group	Number of nitriles
9	nSatom	Constitutional indices	Number of Surfur atom
10	nFatom	Constitutional indices	Number of Fluorine atom
11	nClatom	Constitutional indices	Number of Chlorine atom
12	nBratom	Constitutional indices	Number of Bromine atom
13	nIatom	Constitutional indices	Number of Iodine atom
14	nCatom	Constitutional indices	Number of carbon atom
15	HA	Structural Parameter	Number of acceptor atom for H-bond
16	HD	Structural Parameter	Number of donor atom for H-bond
17	nCharge	Constitutional indices	Total charge
18	Chi2n	Topological indices	2-Path kier shape index
19	AromR	Ring descriptors	Aromatic Ratio

No.	Descriptor^a	Type^b	Detail
1	Sv	Constitutional indices	Sum of VDW volumes
2	Sp	Constitutional indices	Sum of polarizabilities
3	nBO	Constitutional indices	Number of non-H bonds
4	SCBO	Constitutional indices	Sum of conventional bond orders
5	nH	Constitutional indices	Numbers of Hydrogen atoms
6	pilD	Walk and path counts	Conventional bond order ID number
7	IDET	Information indices	Total information content on the distance equality
8	HVcpx	Information indices	graph vertex complexity index
9	TlC1	Information indices	Total Information Content index
10	ClC0	Information indices	Complementary Information Content index
11	ATS4m	2D autocorrelations	B-M autocorrelation of lag 4 (log function) weighted by mass
12	ATS2p	2D autocorrelations	B-M autocorrelation of lag 2 (log function) weighted by polarizability
13	ATS1li	2D autocorrelations	B-M autocorrelation of lag 1 (log function) weighted by ionization potential
14	ATS3i	2D autocorrelations	B-M autocorrelation of lag 3 (log function) weighted by ionization potential
15	ATSC1i	2D autocorrelations	Centred B-M autocorrelation of lag 1 weighted by ionization potential
16	MATS1i	2D autocorrelations	Moran autocorrelation of lag 1 weighted by ionization potential
17	GGl10	2D autocorrelations	topological charge index of order 10
18	SpMin3_Bh(m)	Burden eigenvalues	smallest eigenvalue n. 3 of Burden matrix weighted by mass
19	SpMin4_Bh(m)	Burden eigenvalues	smallest eigenvalue n. 4 of Burden matrix weighted by mass
20	P_VSA_m_2	P_VSA-like descriptors	P_VSA-like on mass, bin 2
21	P_VSA_e_2	P_VSA-like descriptors	P_VSA-like on Sanderson electronegativity, bin 2
22	P_VSA_p_1	P_VSA-like descriptors	P_VSA-like on polarizability, bin 1
23	P_VSA_ppp_L	P_VSA-like descriptors	P_VSA-like on potential pharmacophore points, L-lipophilic
24	P_VSA_ppp_ar	P_VSA-like descriptors	P_VSA-like on potential pharmacophore points, ar-aromatic atoms
25	P_VSA_charge_6	P_VSA-like descriptors	P_VSA-like on partial charges, bin 6
26	nCbH	Functional group counts	number of unsubstituted benzene C(sp2)
27	CATS2D_03_LL	Pharmacophore descriptors	CATS2D Lipophilic-Lipophilic at lag 03
28	Qpos	Charge descriptors	total positive charge
29	RPCG	Charge descriptors	relative positive charge
30	MLOGP2	Molecular properties	squared logP
31	LOGPcons	Molecular properties	Octanol-water partition coeff. (consensus)
32	PDI	Molecular properties	packing density index
33	BLTD48	Molecular properties	Verhaar Daphnia base-line toxicity from MLOGP (mmol/L)
34	DLS_cons	Drug-like indices	DRAGON consensus drug-like score
35	MDEC-23	MDE descriptors	molecular distance edge between all secondary and tertiary carbons

No.	Descriptor^a	Type^b	Detail
1	Sv	Constitutional indices	Sum of VDW volumes
2	Sp	Constitutional indices	Sum of polarizabilities
3	nBO	Constitutional indices	Number of non-H bonds
4	SCBO	Constitutional indices	Sum of conventional bond orders
5	nH	Constitutional indices	Numbers of Hydrogen atoms
6	pilD	Walk and path counts	Conventional bond order ID number
7	IDET	Information indices	Total information content on the distance equality
8	HVcpx	Information indices	graph vertex complexity index
9	TlC1	Information indices	Total Information Content index
10	ClC0	Information indices	Complementary Information Content index
11	ATS4m	2D autocorrelations	B-M autocorrelation of lag 4 (log function) weighted by mass
12	ATS2p	2D autocorrelations	B-M autocorrelation of lag 2 (log function) weighted by polarizability
13	ATS1li	2D autocorrelations	B-M autocorrelation of lag 1 (log function) weighted by ionization potential
14	ATS3i	2D autocorrelations	B-M autocorrelation of lag 3 (log function) weighted by ionization potential
15	ATSC1i	2D autocorrelations	Centred B-M autocorrelation of lag 1 weighted by ionization potential
16	MATS1i	2D autocorrelations	Moran autocorrelation of lag 1 weighted by ionization potential
17	GGl10	2D autocorrelations	topological charge index of order 10
18	SpMin3_Bh(m)	Burden eigenvalues	smallest eigenvalue n. 3 of Burden matrix weighted by mass
19	SpMin4_Bh(m)	Burden eigenvalues	smallest eigenvalue n. 4 of Burden matrix weighted by mass
20	P_VSA_m_2	P_VSA-like descriptors	P_VSA-like on mass, bin 2
21	P_VSA_e_2	P_VSA-like descriptors	P_VSA-like on Sanderson electronegativity, bin 2
22	P_VSA_p_1	P_VSA-like descriptors	P_VSA-like on polarizability, bin 1
23	P_VSA_ppp_L	P_VSA-like descriptors	P_VSA-like on potential pharmacophore points, L-lipophilic
24	P_VSA_ppp_ar	P_VSA-like descriptors	P_VSA-like on potential pharmacophore points, ar-aromatic atoms
25	P_VSA_charge_6	P_VSA-like descriptors	P_VSA-like on partial charges, bin 6
26	nCbH	Functional group counts	number of unsubstituted benzene C(sp2)
27	CATS2D_03_LL	Pharmacophore descriptors	CATS2D Lipophilic-Lipophilic at lag 03
28	Qpos	Charge descriptors	total positive charge
29	RPCG	Charge descriptors	relative positive charge
30	MLOGP2	Molecular properties	squared logP
31	LOGPcons	Molecular properties	Octanol-water partition coeff. (consensus)
32	PDI	Molecular properties	packing density index
33	BLTD48	Molecular properties	Verhaar Daphnia base-line toxicity from MLOGP (mmol/L)
34	DLS_cons	Drug-like indices	DRAGON consensus drug-like score
35	MDEC-23	MDE descriptors	molecular distance edge between all secondary and tertiary carbons

Fingerprint	Length	Radical	MSE
ECFP^a	1024	2-2	378
ECFP	2048	2-2	398
ECFP	2048	2-4	420
PFP^b	1024	2-6	685

多组件学习器实现有机分子沸点的精准预测

Accurate Prediction of the Boiling Point of Organic Molecules by Multi-Component Heterogeneous Learning Model

RichHTML

PDF

补充材料

可视化

摘要/Abstract

引用此文

分享此文

图/表 9

参考文献 37

相关文章 15

编辑推荐

相关统计

本文评价

Model composition	R²_train	R²_test	MAE_train	MAE_test	MSE_train	MSE_test
Model A	0.9885	0.9793	9.23	10.84	186.7	333.9
Model B	0.995	0.9833	5.61	8.53	81.4	270.1
Model C	0.9951	0.9789	6.01	10.45	83.2	350.1
Heterogenous^a	0.9953	0.9872	5.52	7.95	77.8	209.8
Heterovoted^b	0.9953	0.9864	5.55	8.1	73.8	214.6
Homogenous^c	0.9955	0.9848	5.16	7.96	79.9	245.6

Type of compound	N	Descriptor	Method	R²	S	Reference
Alkanes	94	Tls(W, P)	MLR		0.97	[3]
Alkanes	74	Tls(5)	MLR	0.999	1.86	[28]
Alkanes(C2~C7)	72	Tls(LOVI’s)	RA	0.994	3.9	[29]
Furans, thiophenestrahydrofuran	209	Tls, electronic, geometrical	RA	0.969	11.2	[30]
Alkanes, alcohols	245	Atom-type E	MLR		8	[31]
Alcohols	58	Weighted path related Tls	MRA	0.978	3.64	[32]
Diverse organic compounds	298	Constitutional, topological, geometrical and CPSA(8)	MLR	0.954	16.15	[33]
Hydrocarbons	143	Tls	MLR	0.9821	7.2549	[34]
Diverse organic compounds (contain N, O, F, Cl, Br and I)	612	Constitutional, topological, geometrical and CPSA(8)	MLR	0.965	15.5	[35]
Diverse organic compounds	450	Molecular weight and specific gravity	RBF		18.78	[36]
Single ring, fused rings, halogens OH, COOH, COOR, CON, CN, NH₂ and NO₂	155	Tls	BMLR	0.9864	9.1	[37]
Diverse organic compounds	14216	Mathematical Selection	ANN	0.943	22	[14]
Diverse organic compounds	4550	Tls, Constitutional, Fingerprints	Composite ML	0.9953	10.0873	Present work

[1]	戚兴怡, 胡耀峰, 王若愚, 杨雅清, 赵宇飞. 机器学习在新材料筛选方面的应用进展[J]. 化学学报, 2023, 81(2): 158-174.
[2]	韩逸之, 蓝建慧, 刘学, 石伟群. 基于机器学习势函数的熔盐体系分子动力学研究进展[J]. 化学学报, 2023, 81(11): 1663-1672.
[3]	程敏, 王诗慧, 罗磊, 周利, 毕可鑫, 戴一阳, 吉旭. 面向乙烷/乙烯分离的金属有机框架膜的大规模计算筛选[J]. 化学学报, 2022, 80(9): 1277-1288.
[4]	王诗慧, 薛小雨, 程敏, 陈少臣, 刘冲, 周利, 毕可鑫, 吉旭. 机器学习与分子模拟协同的CH₄/H₂分离金属有机框架高通量计算筛选[J]. 化学学报, 2022, 80(5): 614-624.
[5]	蔡铖智, 李丽凤, 邓小梅, 李树华, 梁红, 乔智威. 基于机器学习和高通量计算筛选金属有机框架的甲烷/乙烷/丙烷分离性能[J]. 化学学报, 2020, 78(5): 427-436.
[6]	朱博阳, 吴睿龙, 于曦. 人工智能助力当代化学研究[J]. 化学学报, 2020, 78(12): 1366-1382.
[7]	刘治鲁, 李炜, 刘昊, 庄旭东, 李松. 金属有机骨架的高通量计算筛选研究进展[J]. 化学学报, 2019, 77(4): 323-339.
[8]	谢江安, 梅虎, 吕娟, 潘显超, 王青, 张亚兰. 基于VHSE 结构表征的蛋白酶体酶切位点预测及酶切特异性研究[J]. 化学学报, 2012, 70(03): 318-324.
[9]	禹新良. 自由基共聚合中单体Q-e活性参数预测[J]. 化学学报, 2010, 68(22): 2264-2272.
[10]	刘东,章文军,许禄. 手性羟酸和氨基酸类化合物的构效关系研究[J]. 化学学报, 2009, 67(2): 145-150.
[11]	包鑫,戴连奎. 加权最小二乘支持向量机稳健化迭代算法及其在光谱分析中的应用[J]. 化学学报, 2009, 67(10): 1081-1086.
[12]	乔澍,谢昆,付川,祁俊生. 芳香胺类化合物N—H键离解能的定量构效关系研究[J]. 化学学报, 2009, 67(10): 1109-1115.
[13]	刘涛,宋哲,焦春波,刘伟,朱鸣华,王晓钢. 基于SVM方法分析蛋白酶体裂解位点的特异性[J]. 化学学报, 2008, 66(21): 2341-2347.
[14]	饶含兵, 李泽荣, 陈晓梅, 李象远^{. 基于支持向量学习机的HIV-1蛋白酶抑制剂的活性预测[J]. 化学学报, 2007, 65(3): 197-202.}
[15]	程存归,田玉梅,金文英. 应用小波特征提取肺癌组织FTIR的支持向量机分类方法研究[J]. 化学学报, 2007, 65(22): 2539-2543.