todo.txt

1. Simulation.py needs some refactoring
	- Alot of complexity with it handling both simulation setup and dataset setup
	- See 6

2. Add functionality to save algorithms in YahooRewardManager, DatasetRewardManager 
	- Similar implemenations should be in CF_Bandit repo (YahooMain.py)
	- MLP implementation will be different

3. Potentially add MSE recording and plotting in Dataset/Yahoo RewardManagers 
	- I think we need to change createReccommendation in lib/BaseAlg.py to get access to prediction
	- All algorithms decide functions will need to return the picked article and the predicted value 

5. There is some inconsitency between algorithms that return a list of picked articles and a single instance of an article
	- Similarly only some algorithms have a page rank implemenation (paramater k in decide functions)
	- It is also confusing that k can only be one for algorithms on dataset

6. Issues with hidden_dimension set to > 0 when some algorithms are run. 
	- Ex: when running all algorithms with simulation and hidden_dim = 4 errors throw
	- I believe it is an issued in Simulation.py