About this Research Topic
Continuous monitoring of large-scale physical, social, and financial phenomena has finally become a reality due to the proliferation of sensor devices, the availability of Internet-of-Things platforms, the abundance of data storage capacity, and the easy provisioning of elastic computational infrastructure that can support large-scale statistical analysis on numerous and long-running sequences of data measurements. This confluence of factors has given rise to the emerging area of Big Data Time Series research, which is concerned with the analysis, design, implementation and operationalization of algorithms, statistical techniques, and software systems that monitor, model, analyze, and predict the evolution of a variety of systems over time and at a higher frequency and accuracy than what has been feasible only a few short years ago.
Techniques for time series analysis have similarly evolved from a manual and model, or assumption, based paradigm to being data-driven and fully automated, enabling decision making and business operations to continuously incorporate the output of time series models and analytics, for managing the performance of the business or physical system of interest. For example, in supply chain management, large-scale demand forecasting for the products of a business across regions helps determine the levels and locations of inventory where it will be stored. Similarly, in cloud computing, estimated future utilization of resources across hundreds of thousands of computing nodes helps guide capacity planning and upgrades, whereas in large manufacturing facilities such as semiconductor plants, the various stages of the production process are continuously monitored through online measurements and time series models to adjust the various process parameters towards achieving target levels of yield and quality.
This Research Topic solicits novel and original research work on techniques, algorithms, systems, and applications of large-scale analysis of time series data pertaining to any domain, whether that is physical, social, or financial. The solicited papers:
· may describe algorithmic innovations pertaining to the analysis of large and diverse time series data.
· may address system design issues and their implemented solutions towards handling voluminous amounts of time series data that is continuously collected, processed, and stored.
· may shed insights on large time-series data sets collected from real physical, social, or financial systems.
· may introduce streaming pattern discovery techniques in univariate or multivariate time-series datasets.
Emphasis is given on the volume (size), velocity, variety, and veracity of time series data that should go beyond the “classical” techniques and applications of the past when measurements were manual, infrequent, and sparse.
Keywords: Big Data, Time Series, Machine Learning, Statistical Models, Data Management
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
