A multi-attribute decision-making fusion model for stock trading with customizable investor personality traits in a picture fuzzy environment

In this paper, a fuzzy logic-based machine learning (ML) algorithm is introduced. This proposed ML algorithm accepts picture fuzzy sets (PFS) as the fuzzified input and incorporates genetic algorithm (GA) during the training process. The proposed ML algorithm is then incorporated into two well-known decision-making methods, namely the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) and Evaluation Based on Distance from Average Solution (EDAS). These two decision-making methods and the proposed ML algorithm are then applied to solve a multi-attribute decision-making (MADM) problem related to the evaluation and ranking of public listed companies based on their stock performance, in accordance with investors' personalities. The actual daily closing stock price of five public listed companies from the big market capitalization (Big Cap) category traded in the Kuala Lumpur Stock Exchange (KLSE) for a period of 10 years is used as the datasets for this study. Monte Carlo simulation is used to verify the accuracy of the results. In addition, a comprehensive comparative study of some recent PFS-based decision-making methods in the existing literature and the proposed methods is conducted, and all the typical instances of the investors' personalities are observed. The results obtained through this comparative study corroborates the results obtained via the proposed methods, and this proves the effectiveness of the proposed methods. The differences in the results obtained via the different methods are analyzed and discussed, and this again proves that the results obtained via the proposed methods are effective and consistent with the judgments of human experts. (c) 2023 Elsevier B.V. All rights reserved.

Automated summary

This paper introduces a fuzzy logic-based machine learning algorithm and applies it to solve the evaluation and ranking problem of public listed companies. The algorithm incorporates a genetic algorithm during the training process and is integrated into two popular decision-making methods. Empirical results demonstrate the effectiveness and consistency of the proposed methods in multi-attribute decision-making.