System and Method for Better Visualization and Analytics of Multi-variate Heterogeneous Time-Series Data

Tech ID: 20B186

Competitive Advantages

  • Making it easier to train subsequent machine learning/AI algorithms in tasks involving prediction or classification
  • Providing a reliable and robust representation for the time-series data as noise and other outliers are removed during the weighted averaging operation.
  • Providing a scalable solution where one can analyze time-series data with any length in duration including any number of variables.
  • Improving memory allocation  and  computational  speeds  to  due  to compressing  of potentially very high-volume data into a rich yet compact representative version.

Summary

Whenever a sensor network is implemented, for instance in a transportation setting, especially for a multitude of different scenarios including shipment locations, durations and modes, the data being collected by these sensors invariably becomes heterogeneous with different profiles and durations.  This creates multi-variate time-series datasets spanning time periods which can be different by orders of magnitude and subsequently a challenging scenario for anyone who wants to gain actionable insight from vast amounts of data. In this work we propose a brand-new approach to how we can better summarize and visualize multi-variate time series data coming from numerous sources in a sensor network distributed across a  wide  range  of  application  scenarios to  gather  actionable  and  robust  insight  while ensuring accurate analytics down the information chain including prediction and classification of data. This representative signal  can  be  either  for  a  particular  location  in  the  sensor network distributed across different applications or, vice versa, it could be for the application itself  distributed  across  different  locations  in  the  sensor  network  which  provides greater flexibility when different points of view are needed to analyze a situation. This method has ability to reduce operational costs by unlocking the true advantage of using sensor networks in the production and distribution of goods

Cold-chain time-series Based Machine Learning Pipeline. The images show the real environment for data collection in this work.

Desired Partnerships

  • License 
  • Sponsored Research 
  • Co-Development 

­

Technology Transfer
TTOinfo@usf.edu
(813) 974-0994

Researcher(s)

Patents