causality

Report for Florin Popescu

Download Fact sheet

 Profile of features selected

Figure 1: Profile of features selected. (Click to enlarge). Legend: dcause=direct cause, deffect=direct effects, ocauses=other causes (indirect), oeffects=other effects (indirect), spouses=parent of a direct effect, orelatives=other relatives, unrelated=completely irrelevant

 Tscore performance of all entries

Figure 2: Tscore of all entries made as a function of entry order. (Click to enlarge). The dashed lines indicate the 25 and 50 percentiles.

 Fscore performance of all entries

Figure 3: Fscore of all entries made as a function of entry order. (Click to enlarge). The dashed lines indicate the 25 and 50 percentiles.

Performances:

Fnum=number of features (*=sorted list, **=table of results. Fnum given for best Tscore if table of results submitted), Fscore=MB AUC, Tscore=AUC of test predictions; Top Ts=best score among all valid last entries made by participants. Max Ts=best score reachable, including reference entries using the knowledge of true causal relationships not available to participants.
* Entry Method Fnum Fscore Tscore (Ts) Top Ts Max Ts <Tscore> Rank
CINA0 1495 optimally compressive regression 25/132 * 0.6087 0.7769±0.0051 0.9788 0.9788
CINA1 1495 optimally compressive regression 25/132 * 0.7639 0.7322±0.0052 0.8977 0.8977 0.7488 10
CINA2 1495 optimally compressive regression 25/132 * 0.7639 0.7372±0.0052 0.8157 0.8910

Return to Causation and Predition Challenge Analysis