Lawrence Berkeley National Lab - Modeling Access Pattern of Large Scientific Data Repository with Machine Learning

The main research goal is to develop methods and tools for data access pattern analysis and prediction. It includes following tasks:
* Pre-processing of the dCache HPSS logs is completed (5/2020).
* Data exploration to study data for access patterns (e.g. clustering)
* Development of a machine learning method for to predict the access pattern (with error analysis)
* Development of model update and synchronization method for more data

View project submission here.