|Missing Observations in Time Series Data with Applications to Building Systems Monitors|
Ryan Elmore, National Renewable Energy Lab
Monday, October 15, 2012
4:00pm, room 223 Weber Building
The problem of "Missing Data" is common when dealing with data streams from building
energy/environmental monitors. The missingness may be due to one (or more) of several
factors -- power outage, system failure, recording (database) error, etc. -- and the result is
inevitably an incomplete data set. In this talk, we will discuss several methods for handling
missing observations in several data streams that we are collecting on the NREL campus.
We will present the results of a simulation study on method effectiveness and discuss their
implications on post-imputation clustering. The "statistical" part of this talk will be preceded
by a brief discussion of NREL's Computational Sciences Center and some possible future work.