Knowledge discovery of consensus and conflict interval-based temporal patterns: A novel group decision approach

Temporal pattern mining problems, developed from sequential pattern mining problems, have recently been discussed frequently regarding the gathering of temporal sequences and aggregating them in order to gain insight into consensus decision-making. Existing temporal pattern mining problems reveal only point-based relations; however, in reality, several interval-based circumstances exist, which enable precisely describing temporal relationships. Practical applications include the order and duration of investors purchasing stocks and portfolio management. This study proposes a novel model and its associated algorithm for identifying consensus and conflict patterns from user-provided subjective interval-based temporal sequences. We conducted an experiment on stock investments in the semiconductor industry by drawing on collected authentic datasets and user ratings to demonstrate the model's effectiveness. The experimental results reveal six consensus patterns and one pair of conflict patterns from the participants' subjective investment intuitions, which is consistent with common sense concerning the semiconductor stock market.