Popular Post

Search This Blog

Monday, August 28, 2017

#imPACt 2017 conference program is published. 2. "The Curse of P90..." (#CMGnews):

See the full program HERE

I love Percentiles! My anomaly detection tool (SonR) now has option to use percentiles to calculate UCL and LCL (Control Limits).

Let's go a listen about  percentiles at the following CMG international conference session:

The Curse of P90: An Elegant Way to Overcome it Without Magic 
CAP
Over the decades of development of methodologies and metrics for IT capacity planning and performance analysis, percentile terminology has become the lingua franca of the field. It makes sense: percentiles are easy to interpret, not sensitive to outliers, and directly usable for approximating the distribution of the variable being measured for stochastic simulations. However, depending on which percentile is used, we can miss important information, like multimodality of the metric's distribution. Another, less obvious, downside of relying on percentiles comes into play when we size infrastructure for a high percentile of demand (e.g., p90). Given that it takes time to order, manufacture, receive, and install infrastructure, this means that we need to answer the statistically nontrivial question, "what will this percentile of demand be in one to three years?" This paper discusses the issues that arise in answering it and proposes an elegant way of resolving them.
Presenter bio: Alexander Gilgur is a Data Scientist and Systems Analyst with over 20 years of experience in a wide variety of domains - Control Systems, Chemical Industry, Aviation, Semiconductor manufacturing, Information Technologies, and Networking - and a solid track record of implementing his innovations in production. He has authored and co-authored a number of know-hows, publications, and patents. Alex enjoys applying the beauty of Math and Statistics to solving capacity and performance problems and is interested in non-stationary processes, which make the core of IT problems today. Presently, he is a Network Data Scientist at Facebook and an occasional faculty member at UC Berkeley's MIDS program. He is also a father, a husband, a skier, a soccer player, a sport psychologist, a licensed soccer coach, a licensed professional engineer (PE), and a music aficionado. Alex's technical blog is at http://alexonsimanddata.blogspot.com.