2014-bucket-of-model-for-workload-prediction-2014.rtf

{\rtf1\ansi\ansicpg1252\cocoartf1138\cocoasubrtf470
{\fonttbl\f0\fswiss\fcharset0 Helvetica;}
{\colortbl;\red255\green255\blue255;}
\paperw11900\paperh16840\margl1440\margr1440\vieww10800\viewh8400\viewkind0
\pard\tx566\tx1133\tx1700\tx2267\tx2834\tx3401\tx3968\tx4535\tx5102\tx5669\tx6236\tx6803\pardirnatural

\f0\fs24 \cf0 The paper propose a bucket of models method for selecting the best predictor for workload prediction. It then used to estimate the demand.\
\
The core is a decision tree, which classify the learning algorithm via different metrics (or what they called prediction objectives), e.g., prediction period, overhead etc. The user should input these parameter / or collected online and then the tree would reach one or more learning algorithms\
\
In case it end up with more learning algorithms, it then choose the one produces the smallest MASE as the global error.\
\
\
bad\
\
no qos modelling actually, and of course, no conflict objective}