Understanding how utilization of resources evolves over time is an important task with diverse business applications. For example, an analysis of how PC usage evolves over time can help provide the best overall user experience for current customers, can help determine when they need brand new systems vs. upgraded components, and can inform future product design to better anticipate user needs.
    Orion is a serial software package that facilitates evolution analysis of multivariate resource utilization time series data, such as PC usage. Orion aims to characterize how a set of users utilize resources over time by finding prototypical usage patterns (protos) shared by users at different times within their usage history. To do so, it models each user's usage evolution as a sequence of protos and then performs a cross-user usage segmentation, i.e., a segmentation of the sequences of all users such that the error associated with modeling each segment by one of the protos is minimized. The multivariate time series segmentation problem has been previously addressed in the data mining community, yet its goal has been the optimal segmentation of a single time-series. Instead, Orion finds the optimal segmentation of many time-series by a set of previously unknown building blocks, which it learns during the search. The details of the Orion algorithm can be found in the following paper.

paper
tarball
manual
Contact me via email if you need additional information or find any bugs: david [period1] anastasiu [atSign] sjsu [period2] edu.
Please cite our paper if you make use of this program or any of its components in your research.

David C. Anastasiu, Al M. Rashid, Andrea Tagarelli and George Karypis. Understanding Computer Usage Evolution. Proceedings of the 31st IEEE International Conference on Data Engineering (ICDE 2015). 

@inproceedings{anastasiu2014,
	author = {Anastasiu, David C. and Rashid, Al M. and Tagarelli, Andrea and Karypis, George},
	title = {Understanding Computer Usage Evolution},
	booktitle = {31st IEEE International Conference on Data Engineering},
	series = {ICDE '15},
	year = {2015},
	location = {Coex, Seoul, Korea},
	numpages = {12},
}