Temporal Aspects in Big Data: Keynote Address
Jixin Ma · 2022
Scenario recognition is an important technology in Data Science and Artificial Intelligence. It reveals historical patterns, trends and associations related to human behavior and time by computing and analyzing large-scale data sets. Scenario recognition can be used to solve the problems of prediction, decision-making, diagnosis, process management and supervision, historical reconstruction and so on. However, accurate representing and matching scene patterns in big data is a complex problem. This talk will introduce some basic concepts with respects to time, fluent, state and scenario, and related theories and application examples, as well as a framework for scene pattern representation and matching. The framework can effectively represent relative and absolute time knowledge, and solve the calculation problem of temporal similarity by transforming time similarity into graph similarity. As a special case, we will demonstrate how to apply the framework to represent the traffic light control problem of human intersection in time series and state series.