An efficient query platform for streaming and dynamic natural graphs

摘要: Massive scale data streaming is now prevalent and can be used to dynamically build large graphs which are then efficiently analyzable for insightful information. In situations where real-time analytics required approximate outcomes within time bounds may desirable. We have identified graph summarization TCM sketching in particular as a good technique data. provides set of metrics such Average Relative Error, Number Effective Queries, Effectiveness Queries Confusion Matrix queries on graphs. propose extensions the model automatic sketch creation while being constructed evaluate approach with different policies query combinations. The proposed framework works well 80% 90% efficiency ±3 deviations from exact results.