WEB SEARCH ANALYSIS ON SEVERAL LEVELS OF EVIDENCE BY BURSTINESS AWARE

The issue of effectively indexing and finding such material becomes increasingly crucial as the quantity and scale of huge time stamped collections (such as sequences of digitalized newspapers, magazines, and blogs) grows. Event detection in the context of such collections has been widely studied in relation to term burstiness as a mechanism. In this study, we investigate how the search process might be improved by further using the burstiness information. We describe a brand-new method that makes use of principles from discrepancy theory to represent the burstiness of a phrase. The time periods of maximal burstiness for a given term may then be identified using a parameter-free, linear-time technique that we can construct. Finally, we provide a detailed analysis of our methodology in light of various circumstances and explain the first burstiness-driven search framework.