LightGBM: An effective decision tree gradient boosting method to predict customer loyalty in the finance industry

Marcos Roberto Machado; Salma Karray; Ivaldo Tributino De Sousa

doi:10.1109/ICCSE.2019.8845529

LightGBM: An effective decision tree gradient boosting method to predict customer loyalty in the finance industry

Marcos Roberto Machado, Salma Karray, Ivaldo Tributino De Sousa

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

111 Citations (Scopus)

Abstract

This study presents an implementation of a Machine Learning model to predict customer loyalty for a financial company. We compare the accuracy of two Gradient Boosting Decision Tree Models: XGBoosting and the LightGBM algorithm, which has not yet been used for customer loyalty prediction. We apply these methods to predict credit card customers' loyalty scores for a financial company. The dataset has been made available through a Kaggle's competition. We assess customer loyalty prediction accuracy through RMSE and find that LightGBM performs better than XGBoosting.

Original language	English
Title of host publication	14th International Conference on Computer Science and Education, ICCSE 2019
Publisher	IEEE
Pages	1111-1116
Number of pages	6
ISBN (Electronic)	978-1-7281-1846-8
DOIs	https://doi.org/10.1109/ICCSE.2019.8845529
Publication status	Published - 23 Sept 2019
Externally published	Yes
Event	14th International Conference on Computer Science and Education, ICCSE 2019 - Toronto, Canada Duration: 19 Aug 2019 → 21 Aug 2019 Conference number: 14

Conference

Conference	14th International Conference on Computer Science and Education, ICCSE 2019
Abbreviated title	ICCSE 2019
Country/Territory	Canada
City	Toronto
Period	19/08/19 → 21/08/19

Keywords

Customer loyalty
GBDT models
LightGBM
Machine learning
n/a OA procedure

Access to Document

10.1109/ICCSE.2019.8845529

Cite this

@inproceedings{5e83de556481498cae59f0f82e6c22e1,

title = "LightGBM: An effective decision tree gradient boosting method to predict customer loyalty in the finance industry",

abstract = "This study presents an implementation of a Machine Learning model to predict customer loyalty for a financial company. We compare the accuracy of two Gradient Boosting Decision Tree Models: XGBoosting and the LightGBM algorithm, which has not yet been used for customer loyalty prediction. We apply these methods to predict credit card customers' loyalty scores for a financial company. The dataset has been made available through a Kaggle's competition. We assess customer loyalty prediction accuracy through RMSE and find that LightGBM performs better than XGBoosting.",

keywords = "Customer loyalty, GBDT models, LightGBM, Machine learning, n/a OA procedure",

author = "Machado, {Marcos Roberto} and Salma Karray and {De Sousa}, {Ivaldo Tributino}",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 14th International Conference on Computer Science and Education, ICCSE 2019, ICCSE 2019 ; Conference date: 19-08-2019 Through 21-08-2019",

year = "2019",

month = sep,

day = "23",

doi = "10.1109/ICCSE.2019.8845529",

language = "English",

pages = "1111--1116",

booktitle = "14th International Conference on Computer Science and Education, ICCSE 2019",

publisher = "IEEE",

address = "United States",

}

Machado, MR, Karray, S & De Sousa, IT 2019, LightGBM: An effective decision tree gradient boosting method to predict customer loyalty in the finance industry. in 14th International Conference on Computer Science and Education, ICCSE 2019., 8845529, IEEE, pp. 1111-1116, 14th International Conference on Computer Science and Education, ICCSE 2019, Toronto, Canada, 19/08/19. https://doi.org/10.1109/ICCSE.2019.8845529

LightGBM: An effective decision tree gradient boosting method to predict customer loyalty in the finance industry. / Machado, Marcos Roberto; Karray, Salma; De Sousa, Ivaldo Tributino.
14th International Conference on Computer Science and Education, ICCSE 2019. IEEE, 2019. p. 1111-1116 8845529.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - LightGBM

T2 - 14th International Conference on Computer Science and Education, ICCSE 2019

AU - Machado, Marcos Roberto

AU - Karray, Salma

AU - De Sousa, Ivaldo Tributino

N1 - Conference code: 14

PY - 2019/9/23

Y1 - 2019/9/23

N2 - This study presents an implementation of a Machine Learning model to predict customer loyalty for a financial company. We compare the accuracy of two Gradient Boosting Decision Tree Models: XGBoosting and the LightGBM algorithm, which has not yet been used for customer loyalty prediction. We apply these methods to predict credit card customers' loyalty scores for a financial company. The dataset has been made available through a Kaggle's competition. We assess customer loyalty prediction accuracy through RMSE and find that LightGBM performs better than XGBoosting.

AB - This study presents an implementation of a Machine Learning model to predict customer loyalty for a financial company. We compare the accuracy of two Gradient Boosting Decision Tree Models: XGBoosting and the LightGBM algorithm, which has not yet been used for customer loyalty prediction. We apply these methods to predict credit card customers' loyalty scores for a financial company. The dataset has been made available through a Kaggle's competition. We assess customer loyalty prediction accuracy through RMSE and find that LightGBM performs better than XGBoosting.

KW - Customer loyalty

KW - GBDT models

KW - LightGBM

KW - Machine learning

KW - n/a OA procedure

UR - http://www.scopus.com/inward/record.url?scp=85073245949&partnerID=8YFLogxK

U2 - 10.1109/ICCSE.2019.8845529

DO - 10.1109/ICCSE.2019.8845529

M3 - Conference contribution

AN - SCOPUS:85073245949

SP - 1111

EP - 1116

BT - 14th International Conference on Computer Science and Education, ICCSE 2019

PB - IEEE

Y2 - 19 August 2019 through 21 August 2019

ER -

LightGBM: An effective decision tree gradient boosting method to predict customer loyalty in the finance industry

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this