Development of an Investment Recommender System Using Factor Analysis, ANFIS, and MMNN

Author:

Asemi Asefeh1,Asemi Adeleh2,Ko Andrea1

Affiliation:

1. Corvinus University of Budapest

2. Univeriti Malaya

Abstract

Abstract

The main goal is to present two investment recommender systems (IRS), by combining clustering, factor analysis, Adaptive Neuro Fuzzy Inference System (ANFIS), and Multimodal Neural Network (MNN). The aim is to merge each method with advanced techniques to improve the precision and efficiency of investment recommendations. To develop and implement the IRS, clustering and factor analysis are initially used to detect patterns and connections among variables aiding in grouping individuals into several categories. Then ANFIS is developed in MATLAB using data derived from factor analysis to prove rules for recommending clusters of investment types. Furthermore, MNN was created using Python making use of TensorFlow and Keras libraries using same data for ANFIS. This network is pre-trained with data to predict investment types. The performance of both models is assessed by metrics RMSE and MSE on test data to gauge their accuracy of recommendations. An assessment of the IRSs illustrates its effectiveness in offering investment recommendations. Both models highlight promising performance as shown by the error rates on the test data. By combining clustering, factor analysis, ANFIS and MNN a holistic strategy appears for tailoring investment advice. This approach effectively merged methods with innovative machine learning (ML) and deep learning (DL) techniques. This paper proposes the personalized IRSs that are useful for investment advice. By integrating clustering, factor analysis, ANFIS, and MNN, IRS provides a unique approach with using Explainable artificial intelligence (XAI) to increase the accuracy of investment recommendations. These systems use the strengths of each method in combining them.

Publisher

Springer Science and Business Media LLC

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3