Real-time data text mining based on Gravitational Search Algorithm

Short messages are one of the milestones on the web especially on social media (SM). Due to the widespread circulation of SM, it already turns into excessively painful capturing outmost relevant and significant information for certain users. One of the main motivations of this work is that many users may need an inclusive brief of all comments without reading the entire list of short messages for decision making. In this work, mining in big social media data is formulated for the first time into a multi objective optimization (MOO) task to extract the essence of a text. Since some users may demand the brief at any moment, several groups of dissimilar short messages are established based on graph coloring mechanism. Six interesting feature are formalized to exhibit more interactive messages. A Gravitational Search Algorithm (GSA) is employed to satisfy several important objectives for generating a concise summary. The problem was picked by using the Normal Boundary Intersection (NBI) mechanism to trade-off among different features. Additionally, to satisfy real-time needs, an inventive incremental grouping task is modelled to update the existing colors. From exhaustive experimental results, the proposed approach outperformed other strong comparative methods. (C) 2019 Elsevier Ltd. All rights reserved.