0.25
0.5
0.75
1.25
1.5
1.75
2
Arbitrage of Forecasting Experts
Published on Feb 14, 20181148 Views
Forecasting is an important task across several domains. Its generalised interest is related to the uncertainty and evolving structure of time series. Forecasting methods are typically designed to
Related categories
Chapter list
Arbitrated Ensemble for Time Series Forecasting00:00
Outline - 100:22
Outline - 200:38
Time Series Forecasting00:39
State-of-the-art01:59
Combining experts03:28
Forecasters’ varying relative performance05:35
Main goal06:30
Our proposal06:43
Outline - 307:52
Arbitration - 108:02
Arbitration - 208:58
Arbitration - 309:26
Arbitration - 410:10
Related Work11:37
At each time-step - 112:52
At each time-step - 213:36
At each time-step - 315:38
Expert’s inter-dependence - 117:02
Expert’s inter-dependence - 218:09
Controlling expert’s redundancy - 119:02
Controlling expert’s redundancy - 220:15
Outline - 423:05
Time series23:13
Time series modeling25:08
Experimental setup - 126:05
Experimental setup - 226:54
Experimental setup - 329:17
Results: comparing ADE to state of the art approaches - 131:20
Results: comparing ADE to state of the art approaches - 232:28
Results: combination versus selection33:14
Results: out-of-bag observations34:24
Results: dynamic ensemble composition35:15
Results: deployment strategies - 135:58
Results: deployment strategies - 236:41
Results: sensitivity analysis37:44
Results: sensitivity analysis on Ω and λ38:22
Results: sensitivity analysis on ensemble size39:28
Results: sensitivity analysis40:12
Outline - 540:59
Conclusions41:01
Reproducibility42:26