Lagrange Programming Neural Networks for Sparse Portfolio Design

Hao Wang; Desmond Hui; Chi-Sing Leung

doi:10.1007/978-3-031-30108-7_4

Lagrange Programming Neural Networks for Sparse Portfolio Design

Hao Wang
, Desmond Hui
, Chi-Sing Leung^*

^*Corresponding author for this work

Department of Electrical Engineering

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 12 - Chapter in an edited book (Author) › peer-review

2 Citations (Scopus)

Abstract

Although there were some works on analog neural networks for sparse portfolio design, the existing works do not allow us to control the number of the selected assets and to adjust the weighting between the risk and return. This paper proposes a Lagrange programming neural network (LPNN) model for sparse portfolio design, in which we can control the number of selected assets. Since the objective function of the sparse portfolio design contains a non-differentiable ℓ₁-norm term, we cannot directly use the LPNN approach. Hence, we propose a new formulation based on an approximation of the ℓ₁-norm. In the theoretical side, we prove that state of the proposed LPNN network globally converges to the nearly optimal solution of the sparse portfolio design. The effectiveness of the proposed LPNN approach is verified by the numerical experiments. Simulation results show that the proposed analog approach is superior to the comparison analog neural network models. c The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

Original language	English
Title of host publication	Neural Information Processing
Subtitle of host publication	29th International Conference, ICONIP 2022 Virtual Event, November 22–26, 2022 Proceedings, Part II
Editors	Mohammad Tanveer, Sonali Agarwal, Seiichi Ozawa, Asif Ekbal, Adam Jatowt
Place of Publication	Cham
Publisher	Springer Nature Switzerland AG
Pages	37-48
ISBN (Electronic)	978-3-031-30108-7
ISBN (Print)	978-3-031-30107-0
DOIs	https://doi.org/10.1007/978-3-031-30108-7_4
Publication status	Published - 2023
Event	29th International Conference on Neural Information Processing (ICONIP 2022) - Virtual, Indore, India Duration: 22 Nov 2022 → 26 Nov 2022 https://iconip2022.apnns.org/index.php

Publication series

Name	Lecture Notes in Computer Science
Volume	13624
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	29th International Conference on Neural Information Processing (ICONIP 2022)
Abbreviated title	29th ICONIP 2022
Place	India
City	Indore
Period	22/11/22 → 26/11/22
Internet address	https://iconip2022.apnns.org/index.php

Bibliographical note

Information for this record is supplemented by the author(s) concerned.

Research Keywords

analog neural network
analog optimization
financial data
Lagrange programming neural network
sparse portfolio optimization

Access Output

10.1007/978-3-031-30108-7_4

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Wang, H., Hui, D., & Leung, C.-S. (2023). Lagrange Programming Neural Networks for Sparse Portfolio Design. In M. Tanveer, S. Agarwal, S. Ozawa, A. Ekbal, & A. Jatowt (Eds.), Neural Information Processing: 29th International Conference, ICONIP 2022 Virtual Event, November 22–26, 2022 Proceedings, Part II (pp. 37-48). (Lecture Notes in Computer Science; Vol. 13624). Springer Nature Switzerland AG. https://doi.org/10.1007/978-3-031-30108-7_4

Wang, Hao ; Hui, Desmond ; Leung, Chi-Sing. / Lagrange Programming Neural Networks for Sparse Portfolio Design. Neural Information Processing: 29th International Conference, ICONIP 2022 Virtual Event, November 22–26, 2022 Proceedings, Part II. editor / Mohammad Tanveer ; Sonali Agarwal ; Seiichi Ozawa ; Asif Ekbal ; Adam Jatowt. Cham : Springer Nature Switzerland AG, 2023. pp. 37-48 (Lecture Notes in Computer Science).

@inbook{9bbaaee878b3444fae5b352b9a989e3c,

title = "Lagrange Programming Neural Networks for Sparse Portfolio Design",

abstract = "Although there were some works on analog neural networks for sparse portfolio design, the existing works do not allow us to control the number of the selected assets and to adjust the weighting between the risk and return. This paper proposes a Lagrange programming neural network (LPNN) model for sparse portfolio design, in which we can control the number of selected assets. Since the objective function of the sparse portfolio design contains a non-differentiable ℓ1-norm term, we cannot directly use the LPNN approach. Hence, we propose a new formulation based on an approximation of the ℓ1-norm. In the theoretical side, we prove that state of the proposed LPNN network globally converges to the nearly optimal solution of the sparse portfolio design. The effectiveness of the proposed LPNN approach is verified by the numerical experiments. Simulation results show that the proposed analog approach is superior to the comparison analog neural network models. c The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.",

keywords = "analog neural network, analog optimization, financial data, Lagrange programming neural network, sparse portfolio optimization",

author = "Hao Wang and Desmond Hui and Chi-Sing Leung",

note = "Information for this record is supplemented by the author(s) concerned.; 29th International Conference on Neural Information Processing (ICONIP 2022), 29th ICONIP 2022 ; Conference date: 22-11-2022 Through 26-11-2022",

year = "2023",

doi = "10.1007/978-3-031-30108-7\_4",

language = "English",

isbn = "978-3-031-30107-0",

series = "Lecture Notes in Computer Science",

publisher = "Springer Nature Switzerland AG",

pages = "37--48",

editor = "Mohammad Tanveer and Sonali Agarwal and Seiichi Ozawa and Asif Ekbal and Adam Jatowt",

booktitle = "Neural Information Processing",

address = "Switzerland",

url = "https://iconip2022.apnns.org/index.php",

}

Wang, H, Hui, D & Leung, C-S 2023, Lagrange Programming Neural Networks for Sparse Portfolio Design. in M Tanveer, S Agarwal, S Ozawa, A Ekbal & A Jatowt (eds), Neural Information Processing: 29th International Conference, ICONIP 2022 Virtual Event, November 22–26, 2022 Proceedings, Part II. Lecture Notes in Computer Science, vol. 13624, Springer Nature Switzerland AG, Cham, pp. 37-48, 29th International Conference on Neural Information Processing (ICONIP 2022), Indore, India, 22/11/22. https://doi.org/10.1007/978-3-031-30108-7_4

Lagrange Programming Neural Networks for Sparse Portfolio Design. / Wang, Hao; Hui, Desmond; Leung, Chi-Sing.
Neural Information Processing: 29th International Conference, ICONIP 2022 Virtual Event, November 22–26, 2022 Proceedings, Part II. ed. / Mohammad Tanveer; Sonali Agarwal; Seiichi Ozawa; Asif Ekbal; Adam Jatowt. Cham: Springer Nature Switzerland AG, 2023. p. 37-48 (Lecture Notes in Computer Science; Vol. 13624).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 12 - Chapter in an edited book (Author) › peer-review

TY - CHAP

T1 - Lagrange Programming Neural Networks for Sparse Portfolio Design

AU - Wang, Hao

AU - Hui, Desmond

AU - Leung, Chi-Sing

N1 - Information for this record is supplemented by the author(s) concerned.

PY - 2023

Y1 - 2023

N2 - Although there were some works on analog neural networks for sparse portfolio design, the existing works do not allow us to control the number of the selected assets and to adjust the weighting between the risk and return. This paper proposes a Lagrange programming neural network (LPNN) model for sparse portfolio design, in which we can control the number of selected assets. Since the objective function of the sparse portfolio design contains a non-differentiable ℓ1-norm term, we cannot directly use the LPNN approach. Hence, we propose a new formulation based on an approximation of the ℓ1-norm. In the theoretical side, we prove that state of the proposed LPNN network globally converges to the nearly optimal solution of the sparse portfolio design. The effectiveness of the proposed LPNN approach is verified by the numerical experiments. Simulation results show that the proposed analog approach is superior to the comparison analog neural network models. c The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

AB - Although there were some works on analog neural networks for sparse portfolio design, the existing works do not allow us to control the number of the selected assets and to adjust the weighting between the risk and return. This paper proposes a Lagrange programming neural network (LPNN) model for sparse portfolio design, in which we can control the number of selected assets. Since the objective function of the sparse portfolio design contains a non-differentiable ℓ1-norm term, we cannot directly use the LPNN approach. Hence, we propose a new formulation based on an approximation of the ℓ1-norm. In the theoretical side, we prove that state of the proposed LPNN network globally converges to the nearly optimal solution of the sparse portfolio design. The effectiveness of the proposed LPNN approach is verified by the numerical experiments. Simulation results show that the proposed analog approach is superior to the comparison analog neural network models. c The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

KW - analog neural network

KW - analog optimization

KW - financial data

KW - Lagrange programming neural network

KW - sparse portfolio optimization

UR - https://www.scopus.com/pages/publications/85161676911

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85161676911&origin=recordpage

U2 - 10.1007/978-3-031-30108-7_4

DO - 10.1007/978-3-031-30108-7_4

M3 - RGC 12 - Chapter in an edited book (Author)

AN - SCOPUS:85161676911

SN - 978-3-031-30107-0

T3 - Lecture Notes in Computer Science

SP - 37

EP - 48

BT - Neural Information Processing

A2 - Tanveer, Mohammad

A2 - Agarwal, Sonali

A2 - Ozawa, Seiichi

A2 - Ekbal, Asif

A2 - Jatowt, Adam

PB - Springer Nature Switzerland AG

CY - Cham

T2 - 29th International Conference on Neural Information Processing (ICONIP 2022)

Y2 - 22 November 2022 through 26 November 2022

ER -

Wang H, Hui D, Leung CS. Lagrange Programming Neural Networks for Sparse Portfolio Design. In Tanveer M, Agarwal S, Ozawa S, Ekbal A, Jatowt A, editors, Neural Information Processing: 29th International Conference, ICONIP 2022 Virtual Event, November 22–26, 2022 Proceedings, Part II. Cham: Springer Nature Switzerland AG. 2023. p. 37-48. (Lecture Notes in Computer Science). Epub 2023 Apr 13. doi: 10.1007/978-3-031-30108-7_4