Dense Center Point Mechanism: A Novel Approach for Multi-expert Decision Integration in Portfolio Management

In portfolio management, acquiring stock market volatility is essential to make informed investment decisions. However, the uncertainty and complexity of the financial market stop individual investors from determining optimal investment strategies. In this paper, we develop a novel method, called Domain Expert Decision Integration System (DEDIS), that allows individual investors to integrate the group decision from various experts. Within this framework, we model experts as nodes in a graph and introduce a dense center point mechanism that rejects noise and integrates expert consensus weights to fit the stock return. To illustrate this intuitively, we introduce five fundamental types of experts and provide a detailed decision integration process. Subsequently, we formulate a dual-objective model based on expert weight vectors to accommodate different investor preferences. Finally, we apply our approach to actual Chinese stock market data, calculating the expert-based portfolio return rate and validating the efficacy of our method in leveraging expert decisions to enhance stock return predictions. Reliability and sensitivity analyses further confirm the robustness and superiority of our proposed model, providing valuable insights for investors seeking to navigate expert advice and extract relevant information for their financial objectives. In practice, our method provides new insights into information integration study and empowers investors to reject noise and access the most useful information for their financial goals.