A Novel Combined Investment Recommender System Using Adaptive Neuro-Fuzzy Inference System

Kivonat, rövid leírás

Investment recommendation systems (IRSs) are critical tools used by potential investors to make informed decisions about investment options. However, existing systems have limitations in terms of accuracy and efficiency, leading to a need for more effective and efficient recommendation systems. This dissertation proposes the use of an adaptive neuro-fuzzy inference system (ANFIS) to develop a combined IRS that can provide accurate and efficient investment recommendations for potential investors. The main research question for this study is "How can an ANFIS be utilized to propose an effective and efficient investment recommendation system?" The specific sub-goals of the study are: 1) to categorize and cluster potential investors based on available data to make accurate investment recommendations, 2) to offer customized investment-type services using adaptive neural-fuzzy inference solutions for different categories of potential investors, and 3) to propose a combined recommender system to provide appropriate investment type recommendations for all categorized and clustered potential investors. The dissertation is structured into five chapters. Chapter I provides an overview of the research question and objectives, and Chapter II presents a theoretical framework and literature review, covering existing research on ANFIS in investment recommendation systems. Chapter III explains the methodology used to develop the combined IRS using ANFIS, including data collection, categorization and clustering of potential investors, development of the combined ANFIS model, and evaluation of the proposed system. Chapter IV presents the experimental results and analysis, highlighting the effectiveness of the model in providing appropriate investment-type recommendations for categorized and clustered potential investors. This chapter describes seven experiments that focused on investment recommender systems. Each experiment proposed a unique system that utilized various features of potential investors and their investment type experiences, in addition to employing fuzzy neural inference and the K-Means technique to generate personalized investment recommendations. The first experiment proposed a demographic ANFIS that utilized customer feedback and fuzzy neural inference to generate personalized investment recommendations. The second experiment proposed an automatic recommender system that worked with four key decision factors (KDFs) of potential investors: system value, environmental awareness, high return expectation, and low return expectation. The third experiment used potential investors' financial management traits and investment type for the recommendation. The model was based on an ANFIS, and feedback from knowledge experts and investors was used to improve the system. The fourth experiment used potential investors' experiences data to predict investment outcomes, and the system's performance was evaluated by comparing its recommendations with actual investment outcomes. The fifth experiment proposed an ANFIS-based investment recommendation system based on customers' financial situations, risk tolerance, and investment goals. The sixth experiment investigated the impact of personal characteristics such as age, income, and education level, as well as managerial issues, on investment decisions. The seventh experiment combined and clustered data from the six previous ANFIS systems to provide accurate investment recommendations. The system utilized clustering techniques to group customers with similar financial situations and investment goals, thereby enhancing the personalization of the recommendations. Overall, these experiments propose a novel approach to developing an effective and efficient investment recommendation system using ANFIS. The proposed system has the potential to significantly improve the accuracy and efficiency of investment recommendations, thereby enhancing the decision-making process for potential investors. A comparison of the results with other existing methods and a discussion of the limitations and challenges faced during the development of the system are also included in this chapter. Finally, Chapter V provides a comprehensive discussion of the research findings and their implications, including suggestions for future research. Overall, this dissertation proposes a novel approach to developing an effective and efficient investment recommendation system using ANFIS. The proposed system has the potential to significantly improve the accuracy and efficiency of investment recommendations, thereby enhancing the decision-making process for potential investors.