Algorithmic Analysis on Optimization Oriented Hot Event Detection and Product Recommendation

Abstract

“A hot topic is said to be a topic that people widely talk about during a specific time”. However, there was no precise description of the hot topic and hotness evaluation standards. This paper presents a novel system that concerns on Hot Topic Detection (HTD) integrated with the recommendation o products as well.  The proposed hot event detection model is performed via (i) Pre-processing (ii) Feature Selection and weighting (iii) Text Model Construction (iv) Cluster-based topic Identification. Initially, the keywords are extracted from every tweet. Subsequently, the feature vector space is evaluated that is then subjected for portraying text model construction. At last, clustering is carried out via optimization logic that meant for detecting the hot topic. Particularly, two diverse clustering processes are carried out: micro-clustering and macro-clustering, where the optimal centroid is selected by Monarch Butterfly Optimization (MBO). Following the hot event detection, the adopted model concerns the recommendation of the product associated with the discussed topic. Finally, algorithmic analysis is done by varying the weighting element of the MBO model from 0.2, 0.4, 0.6 and 0.8.