Extracting features importance

[faisalaleissa]

• MLBlocks version: '0.2.0'
• Python version: Python 3.6.3
• Operating System: MacOS High Sierra

Description

I have been trying to produce the features importance from a classifier (RandomForrest). The feature_importances function doesn't take arguments. So, the MLPipeline doesn't execute (TypeError: 'numpy.ndarray' object is not callable).

What I Did

I left the arguments empty in the primitive definition file in the produce part.

{
    "name": "sklearn.ensemble.RandomForestClassifier1",
    "documentation": "http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html",
    "description": "Scikit-learn RandomForestClassifier.",
    "classifiers": {
        "type": "estimator",
        "subtype": "classifier"
    },
    "modalities": [],
    "primitive": "sklearn.ensemble.RandomForestClassifier",
    "fit": {
        "method": "fit",
        "args": [
            {
                "name": "X",
                "type": "DataFrame"
            },
            {
                "name": "y",
                "type": "Series"
            }
        ]
    },
    "produce": {
        "method": "feature_importances_",
        "args": [],
        "output": [
            {
                "name": "y",
                "type": "Series"
            }
        ]
    },
    "hyperparameters": {
        "fixed": {
            "n_jobs": {
                "type": "int",
                "default": -1
            }
        },
        "tunable": {
            "criterion": {
                "type": "str",
                "default": "entropy",
                "values": ["entropy", "gini"]
            },
            "max_features": {
                "type": "str",
                "default": null,
                "range": [null, "auto", "log2"]
            },
            "max_depth": {
                "type": "int",
                "default": 10,
                "range": [1, 30]
            },
            "min_samples_split": {
                "type": "float",
                "default": 0.1,
                "range": [0.0001, 0.5]
            },
            "min_samples_leaf": {
                "type": "float",
                "default": 0.1,
                "range": [0.0001, 0.5]
            },
            "n_estimators": {
                "type": "int",
                "default": 30,
                "values": [2, 500]
            },
            "class_weight": {
                "type": "str",
                "default": null,
                "range": [null, "balanced"]
            }
        }
    }
}

[csala]

Thanks for reporting this @faisalaleissa

However, I think that the problem is that feature_importances_ is not a method, but rather an attribute, so instead of calling it (which is what MLBlocks will try when you set it as the method), what you would need to do is just "read it", which is not something natively supported by MLBlocks.

So, in order to do this, the only option that you have is to write a python primitive in MLPrimitives that in its fit method calls the RandomForest fit method and in its produce method just takes the attribute from the RandomForest object and returns it.

However, a question arises: what do you intentend to do with the feature_importances_ value?

If what you want to do is a feature selection based on that, I recommend you to have a look at the SelectFromModel class from scikit-learn and the corresponding MLBlocks integration, which you can find here:

• JSON: https://github.com/HDI-Project/MLPrimitives/blob/master/mlblocks_primitives/mlprimitives.feature_selection.ExtraTreesClassifierFeatureSelector.json
• Python adapter: https://github.com/HDI-Project/MLPrimitives/blob/master/mlprimitives/feature_selection.py

[faisalaleissa]

Thanks for the quick reply. I'm working on a classifier reporting library, in which, we compute metrics and evaluate the classifiers. One part is to rank the features that are most affective for the classifier.

I think the feature selector you mentioned will be suitble.

Thanks again.