Extreme Learning Machine for Bank Clients Classification

Abstract

In this paper, a classification mode for commercial bank clients' classification using the extreme learning machine (ELM) algorithm is proposed to study the commercial banks VIP loss. Firstly, we adopt the existing data sets of banks to train the ELM model; then, customer classification algorithm and its parameters are selected for classification purpose. Lastly, comparative analysis with existed methods are also compared, which showed that its advantages with the traditional gradient algorithm and other classification algorithm, which further indicate that ELM algorithm can not only overcome their drawbacks but also has faster learning rate, higher rate of accuracy, and better generalization.

EWS-GCN: Edge Weight-Shared Graph Convolutional Network for Transactional Banking Data

“…(1) The authors of [17] limit the choice to the same matrix W k ≡ W for all the channels and discuss different variants of constructing functions α k (x i , x j , e ij ). In this work, while staying almost in the generality of equation (1) we propose the particular instantiation well-adapted to the peculiarities of our data.…”

“…Risk management plays a core role in financial institutions, and banks actively invest in practices related to risk assessment. Currently, Machine Learning (ML) techniques became a standard instrument in banking for such tasks as client classification [1] and clustering [2]. Classical interpretable clients' assessment ML models usually require significant domain knowledge and laborious manual work to design features and the resulting model.…”

Diagnosis of diabetes mellitus using extreme learning machine

“…Even though ELM was classified as new algorithm, its application to some predictions have been made, such as predicting material properties [10], demand forecasting [4], predict in retail industry [11], and bank client classification [12].…”

Dynamic configuration and approximation capabilities of extreme learning machines

“…Almost twenty years later, Huang et al [62] ELM has recently attracted the attention from more and more researchers and variants of ELM have been proposed [11,12,21,25,26,33,80,87,92,115,120,123,124,131,138]. ELM and its variants have been applied to various types of applications including bank clients classification [22], sales forecasting in fashion retailing [121,129], sales forecasting in industry retailing [15], bacterial protein subcellular localization prediction [71], protein secondary structure prediction [126], microarray gene expression cancer diagnosis [133], prediction of melting point of organic compounds [6], fault diagnosis on hydraulic tube tester [46], power utility non-technical loss analysis [99], land cover classification [101], terrain modelling [130], plant species identification [132], water treatment process [136], face recognition [96,139], human action recognition [92], image deblurring with filters [127], transmission line protection [86], no-reference image quality assessment [122], reversible watermarking [24], testing correct model specification [17], XML document classification [137] and classificati...…”