Experiment_Log.txt

%%%Nov 14th, 2016%%%
Missing Data Experiment: (RMSE)
IDVM:
		k=2, d=10: 2.9539
		k=2, d=5: 2.8353
		k=5, d=10: 2.8056
		k=5, d=8: 2.7966
		k=5, d=4: 2.7591
CCWM:		
		k=1, d=5: 2.7190
		k=1, d=10: 3.0168
		k=2, d=10: 2.7939
		k=1, d=4: 2.6838 (current best)
Notes:
1. RMSE results have little different.
2. d should be around 4, but the correl_matrix/distance_matrix calculation should be larger than d in order to ensure there is no missing value in the d-closest pair when calculating missing value (To guarantee there are d values in missing value filling).
3. Can add graph or table to illustrate the evaluation result (consider later).