Prediction of Turkish mutual funds’ net asset value using the fund portfolio distribution

Accurate prediction of mutual funds’ net asset value (NAV) has become increasingly important for investors. Mutual fund investors will be significantly supported by the development of models that accurately predict the future performances of mutual funds. Using these models will facilitate the selec...

Full description

Saved in:
Bibliographic Details
Published in:Neural computing & applications 2023-09, Vol.35 (26), p.18873-18890
Main Authors: Yılmaz, Ümit, Orbak, Âli Yurdun
Format: Article
Language:English
Subjects:
Algorithms
Artificial Intelligence
Artificial neural networks
Asset management
Computational Biology/Bioinformatics
Computational Science and Engineering
Computer Science
Data Mining and Knowledge Discovery
Error analysis
Image Processing and Computer Vision
Investments
Original Article
Performance evaluation
Prediction models
Probability and Statistics in Computer Science
Regularization
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Accurate prediction of mutual funds’ net asset value (NAV) has become increasingly important for investors. Mutual fund investors will be significantly supported by the development of models that accurately predict the future performances of mutual funds. Using these models will facilitate the selection of suitable mutual funds for investors who want to invest in the medium and long term. The aim of this study, using artificial neural networks and nonlinear autoregressive networks with exogenous inputs (NARX) methods and Levenberg–Marquardt (LM), Bayesian regularization (BR), and scaled conjugate gradient training algorithms, is to predict the NAV of two Turkish mutual funds, which are Deniz Asset Management First Variable Fund (DBP) and İstanbul Asset Management Short-Term Bonds and Bills Fund, with the funds’ their portfolio distributions. For this purpose, prediction models were developed with these methods, training algorithms, and some specific hyperparameters and applied to the datasets of the funds examined in the study. The performances of the developed models were compared according to the method and training algorithm pairs for each fund. For performance evaluation, mean squared error, mean absolute percent error, and coefficient of correlation statistical measures are used. From the result, it can be clearly suggested that the NARX-BR pair outperforms other models for DBP, and the NARX-LM pair outperforms other models for IST.
ISSN:0941-0643
1433-3058
DOI:10.1007/s00521-023-08716-5