RL Model Improvement Dashboard
Observe the high-level findings the team has gathered and monitor project progress
Created on August 2|Last edited on August 25
Comment
Table of Contents (click to expand)
IntroductionExperiment 1: Hyperparameter OptimizationCurrent Best ResultsCurrent Best Results Against BenchmarkExperiment 2: Addition of Twitter DataRolling Sharpe Ratio: With vs. Without Twitter DataWithout Twitter Data: Cumulative Returns vs. BaselineWith Twitter Data: Cumulative Returns vs. BaselineMean Reward ComparisonHardware Utilization Comparison
Introduction
This report is intended for a team lead or technical manager. The purpose is to offer high-level findings with key metrics and KPIs so a lead can determine if the methodology behind the new findings is feasible for production.
Experiment 1: Hyperparameter Optimization
Current Best Results
Run set
22
Current Best Results Against Benchmark
Run set
22
Experiment 2: Addition of Twitter Data
Rolling Sharpe Ratio: With vs. Without Twitter Data
Run set
2
Without Twitter Data: Cumulative Returns vs. Baseline
Run: absurd-elevator-34
1
With Twitter Data: Cumulative Returns vs. Baseline
Run: glad-wood-33
1
Mean Reward Comparison
Run set
2
Hardware Utilization Comparison
Run set
2
Add a comment