Lahar Markovian Stream Reversal Application
What is Lahar?
Lahar is a Markovian Streams project at the University of Washington. Markovian Streams consist of correlated, uncertain, ordered data. They can be constructed with probabilistic models from a variety of sources, including RFID and GPS. Storing and querying such data has many applications, including theft detection and healthcare.
Project Summary
At this time, Lahar has the capability of processing Markovian Streams in a chronologically forward direction. For some queries on location data, it may be more efficient to process these streams from end to beginning. The purpose of my project to design an application for reversing Markovian streams and queries, and to determine how backward processing impacts the efficiency of Lahar.
Project Goals
- To develop a stand-alone application for reversing and processing Markovian Streams.
- To determine if backward processing improves the efficiency of querying Markovian Streams, and to what extent.
- To determine what properties of queries make them more suitable for backward or forward stream processing.
Documents
Initial Project Proposal
Abstract
Final Paper
Poster for Grace Hopper