Complete result settings for Retail dataset (table columns are sortable)
The experiments are described in Rooshenas and Lowd, Discriminative Structure Learning of Arithmetic Circuits, AIStats 16
| Split Penalty | Standard Deviation | L1 Penalty | Train LL | Validation LL | Test LL | Node# | Edge# | Feature# | Learning Time |
| 10 | 0.1 | 2 | -10.781008 | -10.905318 | -10.918711 | 2549 | 3706 | 584 | 30.955293 |
| 10 | 0.1 | 1 | -10.779042 | -10.897178 | -10.916361 | 2742 | 4119 | 594 | 30.765322 |
| 10 | 0.5 | 2 | -10.768489 | -10.890084 | -10.908020 | 3402 | 5557 | 620 | 38.239186 |
| 10 | 0.5 | 1 | -10.764746 | -10.884026 | -10.903869 | 4502 | 8129 | 630 | 40.968771 |
| 10 | 0.5 | 0.1 | -10.756101 | -10.880124 | -10.894499 | 8058 | 16320 | 650 | 47.210822 |
| 10 | 0.1 | 0.1 | -10.773762 | -10.897964 | -10.915250 | 2929 | 4503 | 602 | 32.574047 |
| 5 | 0.5 | 2 | -10.735470 | -10.861089 | -10.876853 | 23254 | 51455 | 758 | 113.380763 |
| 5 | 0.5 | 1 | -10.723846 | -10.855062 | -10.863811 | 30433 | 67692 | 818 | 185.929734 |
| 5 | 0.5 | 0.1 | -10.712359 | -10.849894 | -10.861178 | 45015 | 101380 | 876 | 214.131446 |
| 5 | 0.1 | 2 | -10.752122 | -10.879527 | -10.891449 | 7417 | 14688 | 694 | 46.495930 |
| 5 | 0.1 | 0.1 | -10.739395 | -10.869256 | -10.883345 | 13306 | 28410 | 750 | 73.748787 |
| 5 | 0.1 | 1 | -10.746309 | -10.871044 | -10.888187 | 9634 | 19888 | 726 | 52.770976 |
| 2 | 0.5 | 2 | -10.689724 | -10.841953 | -10.851645 | 63382 | 142082 | 1108 | 448.409830 |
| 2 | 0.5 | 1 | -10.671868 | -10.839353 | -10.844606 | 88982 | 202420 | 1268 | 745.258702 |
| 2 | 0.1 | 2 | -10.715591 | -10.850104 | -10.867335 | 33277 | 73808 | 952 | 211.592832 |
| 2 | 0.1 | 1 | -10.700466 | -10.844530 | -10.854459 | 42486 | 95248 | 1090 | 289.817940 |
| 2 | 0.1 | 0.1 | -10.691642 | -10.846914 | -10.853298 | 54969 | 123678 | 1156 | 383.607682 |
| 2 | 0.5 | 0.1 | -10.656439 | -10.828830 | -10.845647 | 132657 | 306813 | 1448 | 1238.593704 |