Some issues with bnlearn when using Edge Constrains

[melmasri]

We are running into some issues when running bnlearn.

Here is a screen shot of the issues, and and the config files.

    "benchmark_setup": [
      {
        "data": [
          {
            "graph_id": null,
            "parameters_id": null,
            "data_id": "cleaned_simulated_data.csv",
            "seed_range": null
          }
        ],
        "evaluation": {
          "ggally_ggpairs": false,
          "graph_estimation": {
              "ids": [
                "mmhc-bge-zf",
                "interiamb-zf",
                "gs-zf",
                "tabu-bge",
                "hc-bge"
              ],
              "convert_to": ["cpdag"],
              "graphs": true,
              "adjmats": true,
              "diffplots": false,
              "csvs": true,
              "graphvizcompare": false
            },
          "mcmc_traj_plots": [],
          "mcmc_heatmaps": [],
          "mcmc_autocorr_plots": [],
          "graph_plots": []
        },
        "title": "wbagtest_010"
      }
    ],
    "resources": {
      "data": {},
      "graph": {},
      "parameters": {},
      "structure_learning_algorithms": {
        "bnlearn_mmhc": [
          {
            "id": "mmhc-bge-zf",
            "alpha": [
              0.001,
              0.01,
              0.05,
              0.1
            ],
            "score": "bge",
            "test": "zf",
            "iss": 1,
            "issmu": 1,
            "issw": null,
            "l": 5,
            "k": 1,
            "prior": "uniform",
            "beta": null,
            "timeout": null,
            "edgeConstraints": "geneticConstraints.json"
          }
        ],
        "bnlearn_interiamb": [
          {
            "id": "interiamb-zf",
            "alpha": [
              0.01,
              0.05
            ],
            "test": "zf",
            "B": null,
            "maxsx": null,
            "debug": false,
            "undirected": false,
            "timeout": null,
            "edgeConstraints": "geneticConstraints.json"
          }
        ],
        "bnlearn_gs": [
          {
            "id": "gs-zf",
            "alpha": [
              0.01
            ],
            "test": "zf",
            "B": null,
            "maxsx": null,
            "debug": false,
            "undirected": false,
            "timeout": null,
            "edgeConstraints": "geneticConstraints.json"
          }
        ],
        "bnlearn_tabu": [
          {
            "id": "tabu-bge",
            "score": "bge",
            "iss": 1,
            "issmu": [
              0.0001,
              0.001,
              0.01,
              0.05
            ],
            "issw": null,
            "l": 5,
            "k": 1,
            "prior": "uniform",
            "beta": 1,
            "timeout": null,
            "edgeConstraints": "geneticConstraints.json"
          }
        ],
        "bnlearn_hc": [
          {
            "id": "hc-bge",
            "score": "bge",
            "iss": 1,
            "issmu": [
              0.0001,
              0.001,
              0.01,
              0.05
            ],
            "issw": null,
            "l": 5,
            "k": 1,
            "prior": "uniform",
            "beta": 1,
            "restart": 0,
            "perturb": 1,
            "timeout": null,
            "edgeConstraints": "geneticConstraints.json"
          }
        ]
      }
    }
}

geneticConstraints.json

cleaned_simulated_data.csv

[felixleopoldo]

Hi,

There is not yet a JSON schema for the edge constraints but I don't think you can put null values. Have you tried to remove the fields with null values?

I have created an issue for the JSON schema here.

[melmasri]

@felixleopoldo null values where in the parameters? We don't have null values in the edge constrains

[felixleopoldo]

I saw there are null values in the attached edge constraints JSON file. Maybe that can be the error?