Tabular explanations example
============================

Learning and explaining German Credit Dataset
---------------------------------------------

.. code:: ipython3

    import pandas as pd
    import numpy as np

    from sklearn import preprocessing
    from sklearn.ensemble import RandomForestClassifier
    from sklearn.model_selection import train_test_split
    from sklearn.linear_model import LogisticRegression

    from xailib.data_loaders.dataframe_loader import prepare_dataframe

    from xailib.explainers.lime_explainer import LimeXAITabularExplainer
    from xailib.explainers.lore_explainer import LoreTabularExplainer
    from xailib.explainers.shap_explainer_tab import ShapXAITabularExplainer

    from xailib.models.sklearn_classifier_wrapper import sklearn_classifier_wrapper


Loading and preparation of data
-------------------------------

We start by reading from a CSV file the dataset to analyze. The table is
loaded by means of the ``DataFrame`` class from the ``pandas`` library.

Among all the attributes of the table, we select the ``class_field``
column that contains the observed class for the corresponding row.

.. code:: ipython3

    source_file = 'datasets/german_credit.csv'
    class_field = 'default'
    # Load and transform dataset 
    df = pd.read_csv(source_file, skipinitialspace=True, na_values='?', keep_default_na=True)

After the data is loaded in memory, we need to extract metadata
information to automatically handle the content withint the table.

The method ``prepare_dataframe`` scans the table and extract the
following information: \* ``df``: is a trasformed version of the
original dataframe, where discrete attributes are transformed into
numerical attributes by using one hot encoding strategy; \*
``feature_names``: is a list containint the names of the features after
the transformation; \* ``class_values``: the list of all the possible
values for the ``class_field`` column; \* ``numeric_columns``: a list of
the original features that contain numeric (i.e. continuous) values; \*
``rdf``: the original dataframe, before the transformation; \*
``real_feature_names``: the list of the features of the dataframe before
the transformation; \* ``features_map``: it is a dictionary pointing
each feature to the original one before the transformation.

.. code:: ipython3

    df, feature_names, class_values, numeric_columns, rdf, real_feature_names, features_map = prepare_dataframe(df, class_field)

Learning a Random Forest classfier
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

We train a RF classifier by using the ``sklearn`` library. We start by
splitting the dataset into a train and test subsets.

.. code:: ipython3

    test_size = 0.3
    random_state = 42
    X_train, X_test, Y_train, Y_test = train_test_split(df[feature_names], df[class_field],
                                                            test_size=test_size,
                                                            random_state=random_state,
                                                            stratify=df[class_field])
    
    

Then we train the model on the training set. Once the model has been
learned, we use a wrapper class to get access to the model for
``XAI lib``

.. code:: ipython3

    bb = RandomForestClassifier(n_estimators=20, random_state=random_state)
    bb.fit(X_train.values, Y_train.values)
    bbox = sklearn_classifier_wrapper(bb)   

Select a new instance to be classfied by the model and print the
predicted class.

.. code:: ipython3

    inst = X_train.iloc[147].values
    print('Instance ',inst)
    print('True class ',Y_train.iloc[8])
    print('Predicted class ',bb.predict(inst.reshape(1, -1)))


.. parsed-literal::

    Instance  [ 15 975   2   3  25   2   1   0   1   0   0   0   1   0   0   0   0   0
       0   0   0   0   0   1   0   0   0   1   0   0   0   0   0   1   0   0
       0   1   0   0   0   0   1   1   0   0   0   0   1   0   0   1   0   0
       1   0   0   1   0   0   1]
    True class  0
    Predicted class  [0]
    

Explaining the prediction
-------------------------

We use the explanators of ``XAI lib`` to provide an explantion for the
classified instance ``inst``. Every explainer of ``XAI lib`` takes in
input the blackbox to be explained with the corresponding feature names,
and a configuration object to initialize the explainer.

SHAP explainer
~~~~~~~~~~~~~~

.. code:: ipython3

    explainer = ShapXAITabularExplainer(bbox, feature_names)
    config = {'explainer' : 'tree', 'X_train' : X_train.iloc[0:100].values}
    explainer.fit(config)

.. code:: ipython3

    exp = explainer.explain(inst)
    # print(exp.exp)

.. code:: ipython3

    exp.plot_features_importance()

LORE explainer
~~~~~~~~~~~~~~

.. code:: ipython3

    explainer = LoreTabularExplainer(bbox)
    config = {'neigh_type':'rndgen', 'size':1000, 'ocr':0.1, 'ngen':10}
    explainer.fit(df, class_field, config)
    exp = explainer.explain(inst)
    print(exp)

.. code:: ipython3

    exp.plotRules()

.. code:: ipython3

    exp.plotCounterfactualRules()

LIME explainer
~~~~~~~~~~~~~~

.. code:: ipython3

    limeExplainer = LimeXAITabularExplainer(bbox)
    config = {'feature_selection': 'lasso_path'}
    limeExplainer.fit(df, class_field, config)
    lime_exp = limeExplainer.explain(inst)
    print(lime_exp.exp.as_list())


.. parsed-literal::

    [('account_check_status=no checking account', -0.03792512128083548), ('duration_in_month', 0.03701527256562679), ('account_check_status=< 0 DM', 0.03144299031649348), ('savings=... < 100 DM', 0.020051934530021572), ('age', -0.019751080001761446), ('credit_history=critical account/ other credits existing (not at this bank)', -0.018970043296280513), ('other_installment_plans=none', -0.018869997928840695), ('other_installment_plans=bank', 0.017658677626390982), ('housing=own', -0.014948467979451343), ('credit_history=delay in paying off in the past', 0.012221985897781883)]
    

.. code:: ipython3

    # limeExplainer.plot_lime_values(lime_exp.as_list(), 5, 10)
    lime_exp.plot_features_importance()



.. raw:: html

    
    <style>
    .vega-bind {
      position: absolute;
      left: 0px;
      top: 0px;
      background-color:#eee;
      padding:10px;
      font-size:10px;
    }
    .chart-wrapper{
      padding-top: 70px;
    }
    
    </style>
    



.. raw:: html

    
    <div id="altair-viz-2ed68f4ea0804764a1e728496079e229"></div>
    <script type="text/javascript">
      (function(spec, embedOpt){
        let outputDiv = document.currentScript.previousElementSibling;
        if (outputDiv.id !== "altair-viz-2ed68f4ea0804764a1e728496079e229") {
          outputDiv = document.getElementById("altair-viz-2ed68f4ea0804764a1e728496079e229");
        }
        const paths = {
          "vega": "https://cdn.jsdelivr.net/npm//vega@5?noext",
          "vega-lib": "https://cdn.jsdelivr.net/npm//vega-lib?noext",
          "vega-lite": "https://cdn.jsdelivr.net/npm//vega-lite@4.8.1?noext",
          "vega-embed": "https://cdn.jsdelivr.net/npm//vega-embed@6?noext",
        };
    
        function loadScript(lib) {
          return new Promise(function(resolve, reject) {
            var s = document.createElement('script');
            s.src = paths[lib];
            s.async = true;
            s.onload = () => resolve(paths[lib]);
            s.onerror = () => reject(`Error loading script: ${paths[lib]}`);
            document.getElementsByTagName("head")[0].appendChild(s);
          });
        }
    
        function showError(err) {
          outputDiv.innerHTML = `<div class="error" style="color:red;">${err}</div>`;
          throw err;
        }
    
        function displayChart(vegaEmbed) {
          vegaEmbed(outputDiv, spec, embedOpt)
            .catch(err => showError(`Javascript Error: ${err.message}<br>This usually means there's a typo in your chart specification. See the javascript console for the full traceback.`));
        }
    
        if(typeof define === "function" && define.amd) {
          requirejs.config({paths});
          require(["vega-embed"], displayChart, err => showError(`Error loading script: ${err.message}`));
        } else if (typeof vegaEmbed === "function") {
          displayChart(vegaEmbed);
        } else {
          loadScript("vega")
            .then(() => loadScript("vega-lite"))
            .then(() => loadScript("vega-embed"))
            .catch(showError)
            .then(() => displayChart(vegaEmbed));
        }
      })({"config": {"view": {"continuousWidth": 400, "continuousHeight": 300}, "axis": {"labelFontSize": 10, "labelLimit": 225.0}}, "vconcat": [{"layer": [{"data": {"name": "data-74da8e1e931767cabbafd8cf64a0bebf"}, "mark": {"type": "rule", "strokeWidth": 3}, "encoding": {"color": {"type": "quantitative", "field": "xValue", "legend": null, "scale": {"domain": [-0.03792512128083548, 0.03792512128083548], "domainMid": 0, "scheme": "redyellowblue"}}, "x": {"type": "quantitative", "axis": {"orient": "top", "titleFontSize": 10}, "field": "xValue", "title": "Select a cutoff range for Feature Importance values "}}}, {"data": {"name": "data-c7bc3ae69e1253e8223f70a11d04b99d"}, "mark": {"type": "rect", "color": "black", "height": 20, "opacity": 0.4, "tooltip": true}, "encoding": {"x": {"type": "quantitative", "field": "x_min"}, "x2": {"field": "x_max"}, "y": {"type": "quantitative", "axis": null, "field": "y"}}, "selection": {"Cutter": {"type": "single", "fields": ["cutoff"], "bind": {"input": "range", "max": 0.03792512128083548, "min": 0, "name": "Importance cutoff value (\u00b1) ", "step": 0.0007585024256167096}, "init": {"cutoff": 0.0}}}, "transform": [{"calculate": "(-Cutter.cutoff)", "as": "x_min"}, {"calculate": "Cutter.cutoff", "as": "x_max"}]}], "height": 20}, {"layer": [{"data": {"name": "data-374d406a2cf248b04a750afc2126e34c"}, "mark": "bar", "encoding": {"color": {"type": "quantitative", "field": "value", "legend": null, "scale": {"domain": [-0.03792512128083548, 0.03792512128083548], "domainMid": 0, "scheme": "blueorange"}}, "tooltip": [{"type": "nominal", "field": "name", "title": "Feature"}, {"type": "quantitative", "field": "value", "title": "Importance"}], "x": {"type": "quantitative", "field": "value", "title": null}, "y": {"type": "nominal", "field": "name", "sort": {"field": "value", "op": "mean", "order": "descending"}, "title": null}}, "selection": {"Cutter": {"type": "single", "fields": ["cutoff"], "bind": {"input": "range", "max": 0.03792512128083548, "min": 0, "name": "Importance cutoff value (\u00b1) ", "step": 0.0007585024256167096}, "init": {"cutoff": 0.0}}}, "transform": [{"filter": "((datum.value > Cutter.cutoff) || (datum.value < (-Cutter.cutoff)))"}]}, {"data": {"name": "data-9c00aa9dc2b1c1a0b4bf9b34ead12925"}, "mark": "rule", "encoding": {"x": {"type": "quantitative", "field": "x"}}}]}], "padding": 10, "$schema": "https://vega.github.io/schema/vega-lite/v4.8.1.json", "datasets": {"data-74da8e1e931767cabbafd8cf64a0bebf": [{"xValue": -0.03792512128083548}, {"xValue": -0.03754587006802713}, {"xValue": -0.037166618855218776}, {"xValue": -0.036787367642410425}, {"xValue": -0.03640811642960207}, {"xValue": -0.03602886521679372}, {"xValue": -0.03564961400398537}, {"xValue": -0.03527036279117702}, {"xValue": -0.03489111157836867}, {"xValue": -0.034511860365560316}, {"xValue": -0.034132609152751965}, {"xValue": -0.033753357939943614}, {"xValue": -0.03337410672713526}, {"xValue": -0.03299485551432691}, {"xValue": -0.03261560430151856}, {"xValue": -0.03223635308871021}, {"xValue": -0.03185710187590186}, {"xValue": -0.031477850663093505}, {"xValue": -0.031098599450285154}, {"xValue": -0.030719348237476803}, {"xValue": -0.03034009702466845}, {"xValue": -0.0299608458118601}, {"xValue": -0.02958159459905175}, {"xValue": -0.029202343386243397}, {"xValue": -0.028823092173435046}, {"xValue": -0.028443840960626694}, {"xValue": -0.028064589747818343}, {"xValue": -0.02768533853500999}, {"xValue": -0.02730608732220164}, {"xValue": -0.02692683610939329}, {"xValue": -0.026547584896584937}, {"xValue": -0.026168333683776586}, {"xValue": -0.025789082470968235}, {"xValue": -0.025409831258159883}, {"xValue": -0.025030580045351532}, {"xValue": -0.02465132883254318}, {"xValue": -0.02427207761973483}, {"xValue": -0.023892826406926478}, {"xValue": -0.023513575194118126}, {"xValue": -0.023134323981309775}, {"xValue": -0.022755072768501423}, {"xValue": -0.022375821555693072}, {"xValue": -0.02199657034288472}, {"xValue": -0.02161731913007637}, {"xValue": -0.021238067917268018}, {"xValue": -0.020858816704459666}, {"xValue": -0.020479565491651315}, {"xValue": -0.020100314278842964}, {"xValue": -0.019721063066034612}, {"xValue": -0.01934181185322626}, {"xValue": -0.01896256064041791}, {"xValue": -0.018583309427609558}, {"xValue": -0.018204058214801207}, {"xValue": -0.017824807001992855}, {"xValue": -0.017445555789184504}, {"xValue": -0.017066304576376153}, {"xValue": -0.0166870533635678}, {"xValue": -0.01630780215075945}, {"xValue": -0.0159285509379511}, {"xValue": -0.015549299725142747}, {"xValue": -0.015170048512334396}, {"xValue": -0.014790797299526044}, {"xValue": -0.014411546086717693}, {"xValue": -0.014032294873909341}, {"xValue": -0.01365304366110099}, {"xValue": -0.013273792448292639}, {"xValue": -0.012894541235484287}, {"xValue": -0.012515290022675936}, {"xValue": -0.012136038809867584}, {"xValue": -0.011756787597059233}, {"xValue": -0.011377536384250882}, {"xValue": -0.01099828517144253}, {"xValue": -0.010619033958634179}, {"xValue": -0.010239782745825828}, {"xValue": -0.009860531533017476}, {"xValue": -0.009481280320209125}, {"xValue": -0.009102029107400773}, {"xValue": -0.008722777894592422}, {"xValue": -0.00834352668178407}, {"xValue": -0.00796427546897572}, {"xValue": -0.007585024256167368}, {"xValue": -0.007205773043359016}, {"xValue": -0.006826521830550665}, {"xValue": -0.006447270617742314}, {"xValue": -0.006068019404933962}, {"xValue": -0.005688768192125611}, {"xValue": -0.0053095169793172595}, {"xValue": -0.004930265766508908}, {"xValue": -0.004551014553700557}, {"xValue": -0.004171763340892205}, {"xValue": -0.003792512128083854}, {"xValue": -0.0034132609152755025}, {"xValue": -0.003034009702467151}, {"xValue": -0.0026547584896587997}, {"xValue": -0.0022755072768504483}, {"xValue": -0.001896256064042097}, {"xValue": -0.0015170048512337456}, {"xValue": -0.0011377536384253942}, {"xValue": -0.0007585024256170428}, {"xValue": -0.0003792512128086914}, {"xValue": -3.400058012914542e-16}, {"xValue": 0.0003792512128080114}, {"xValue": 0.0007585024256163628}, {"xValue": 0.0011377536384247142}, {"xValue": 0.0015170048512330656}, {"xValue": 0.001896256064041417}, {"xValue": 0.0022755072768497683}, {"xValue": 0.0026547584896581197}, {"xValue": 0.003034009702466471}, {"xValue": 0.0034132609152748225}, {"xValue": 0.003792512128083174}, {"xValue": 0.004171763340891525}, {"xValue": 0.004551014553699877}, {"xValue": 0.004930265766508228}, {"xValue": 0.0053095169793165795}, {"xValue": 0.005688768192124931}, {"xValue": 0.006068019404933282}, {"xValue": 0.006447270617741634}, {"xValue": 0.006826521830549985}, {"xValue": 0.007205773043358336}, {"xValue": 0.007585024256166688}, {"xValue": 0.00796427546897504}, {"xValue": 0.00834352668178339}, {"xValue": 0.008722777894591742}, {"xValue": 0.009102029107400093}, {"xValue": 0.009481280320208445}, {"xValue": 0.009860531533016796}, {"xValue": 0.010239782745825148}, {"xValue": 0.010619033958633499}, {"xValue": 0.01099828517144185}, {"xValue": 0.011377536384250202}, {"xValue": 0.011756787597058553}, {"xValue": 0.012136038809866904}, {"xValue": 0.012515290022675256}, {"xValue": 0.012894541235483607}, {"xValue": 0.013273792448291959}, {"xValue": 0.01365304366110031}, {"xValue": 0.014032294873908661}, {"xValue": 0.014411546086717013}, {"xValue": 0.014790797299525364}, {"xValue": 0.015170048512333716}, {"xValue": 0.015549299725142067}, {"xValue": 0.01592855093795042}, {"xValue": 0.01630780215075877}, {"xValue": 0.01668705336356712}, {"xValue": 0.017066304576375473}, {"xValue": 0.017445555789183824}, {"xValue": 0.017824807001992175}, {"xValue": 0.018204058214800527}, {"xValue": 0.018583309427608878}, {"xValue": 0.01896256064041723}, {"xValue": 0.01934181185322558}, {"xValue": 0.019721063066033932}, {"xValue": 0.020100314278842284}, {"xValue": 0.020479565491650635}, {"xValue": 0.020858816704458986}, {"xValue": 0.021238067917267338}, {"xValue": 0.02161731913007569}, {"xValue": 0.02199657034288404}, {"xValue": 0.022375821555692392}, {"xValue": 0.022755072768500743}, {"xValue": 0.023134323981309095}, {"xValue": 0.023513575194117446}, {"xValue": 0.023892826406925798}, {"xValue": 0.02427207761973415}, {"xValue": 0.0246513288325425}, {"xValue": 0.025030580045350852}, {"xValue": 0.025409831258159203}, {"xValue": 0.025789082470967555}, {"xValue": 0.026168333683775906}, {"xValue": 0.026547584896584257}, {"xValue": 0.02692683610939261}, {"xValue": 0.02730608732220096}, {"xValue": 0.02768533853500931}, {"xValue": 0.028064589747817663}, {"xValue": 0.028443840960626014}, {"xValue": 0.028823092173434366}, {"xValue": 0.029202343386242717}, {"xValue": 0.02958159459905107}, {"xValue": 0.02996084581185942}, {"xValue": 0.03034009702466777}, {"xValue": 0.030719348237476123}, {"xValue": 0.031098599450284474}, {"xValue": 0.031477850663092825}, {"xValue": 0.03185710187590118}, {"xValue": 0.03223635308870953}, {"xValue": 0.03261560430151788}, {"xValue": 0.03299485551432623}, {"xValue": 0.03337410672713458}, {"xValue": 0.033753357939942934}, {"xValue": 0.034132609152751285}, {"xValue": 0.034511860365559636}, {"xValue": 0.03489111157836799}, {"xValue": 0.03527036279117634}, {"xValue": 0.03564961400398469}, {"xValue": 0.03602886521679304}, {"xValue": 0.03640811642960139}, {"xValue": 0.036787367642409745}, {"xValue": 0.037166618855218096}, {"xValue": 0.03754587006802645}], "data-c7bc3ae69e1253e8223f70a11d04b99d": [{"y": 0, "x": -0.5, "x2": 0.5}], "data-374d406a2cf248b04a750afc2126e34c": [{"name": "account_check_status=no checking account", "value": -0.03792512128083548}, {"name": "duration_in_month", "value": 0.03701527256562679}, {"name": "account_check_status=< 0 DM", "value": 0.03144299031649348}, {"name": "savings=... < 100 DM", "value": 0.020051934530021572}, {"name": "age", "value": -0.019751080001761446}, {"name": "credit_history=critical account/ other credits existing (not at this bank)", "value": -0.018970043296280513}, {"name": "other_installment_plans=none", "value": -0.018869997928840695}, {"name": "other_installment_plans=bank", "value": 0.017658677626390982}, {"name": "housing=own", "value": -0.014948467979451343}, {"name": "credit_history=delay in paying off in the past", "value": 0.012221985897781883}], "data-9c00aa9dc2b1c1a0b4bf9b34ead12925": [{"x": 0}]}}, {"mode": "vega-lite"});
    </script>


Learning a different model
--------------------------

Learning a Logistic Regressor
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

We train a Logistic Regression by using the ``sklearn`` library. We
transform the dataset by using a ``Scaler`` to normalize all the
attributes.

.. code:: ipython3

    scaler = preprocessing.StandardScaler().fit(X_train)
    X_scaled = scaler.transform(X_train)
    
    bb = LogisticRegression(C=1, penalty='l2')
    bb.fit(X_scaled, Y_train.values)
    # pass the model to the wrapper to use it in the XAI lib
    bbox = sklearn_classifier_wrapper(bb)

.. code:: ipython3

    # select a record to explain
    inst = X_scaled[182]
    print('Instance ',inst)
    print('Predicted class ',bb.predict(inst.reshape(1, -1)))


.. parsed-literal::

    Instance  [ 2.27797454  3.35504085  0.94540357  1.07634233  0.04854891 -0.72456474
     -0.43411405  1.65027399 -0.61477862 -0.25898489 -0.80681063  4.17385345
     -0.6435382  -0.32533856 -1.03489416 -0.20412415 -0.22941573 -0.33068147
      1.75885396 -0.34899122 -0.60155441 -0.15294382 -0.09298136 -0.46852129
     -0.12038585 -0.08481889 -0.23623492 -1.21387736 -0.36174054 -0.24943031
      2.15526362 -0.59715086 -0.45485883 -0.73610476 -0.43875307  4.23307441
     -0.65242771 -0.23958675 -0.32533856  0.90192655  4.72581563 -0.2259448
     -3.15238005 -0.54212562 -0.70181003 -0.63024248  2.30354212 -0.40586384
      0.49329429 -0.23958675  2.88675135 -1.59227935 -0.46170508  2.46388049
     -1.33747696 -0.13206764 -0.5        -1.21387736  1.21387736 -0.20412415
      0.20412415]
    Predicted class  [1]
    

Explaining the prediction
-------------------------

We use the same explainators as for the previous model. In this case, a
few adjustments are necessary for the initialization of the explanators.
For example, SHAP needs a specific configuration for the linear model we
are using. ### SHAP Explainer

.. code:: ipython3

    explainer = ShapXAITabularExplainer(bbox, feature_names)
    config = {'explainer' : 'linear', 'X_train' : X_scaled[0:100], 'feature_pert' : 'interventional'}
    explainer.fit(config)

.. code:: ipython3

    exp = explainer.explain(inst)
    print(exp)


.. parsed-literal::

    <xailib.explainers.shap_explainer_tab.ShapXAITabularExplanation object at 0x12a72dac8>
    

.. code:: ipython3

    exp.plot_features_importance()



.. raw:: html

    
    <style>
    .vega-bind {
      position: absolute;
      left: 0px;
      top: 0px;
      background-color:#eee;
      padding:10px;
      font-size:10px;
    }
    .chart-wrapper{
      padding-top: 70px;
    }
    
    </style>
    



.. raw:: html

    
    <div id="altair-viz-29414f5752574c7b9e68cddf08aaf2c0"></div>
    <script type="text/javascript">
      (function(spec, embedOpt){
        let outputDiv = document.currentScript.previousElementSibling;
        if (outputDiv.id !== "altair-viz-29414f5752574c7b9e68cddf08aaf2c0") {
          outputDiv = document.getElementById("altair-viz-29414f5752574c7b9e68cddf08aaf2c0");
        }
        const paths = {
          "vega": "https://cdn.jsdelivr.net/npm//vega@5?noext",
          "vega-lib": "https://cdn.jsdelivr.net/npm//vega-lib?noext",
          "vega-lite": "https://cdn.jsdelivr.net/npm//vega-lite@4.8.1?noext",
          "vega-embed": "https://cdn.jsdelivr.net/npm//vega-embed@6?noext",
        };
    
        function loadScript(lib) {
          return new Promise(function(resolve, reject) {
            var s = document.createElement('script');
            s.src = paths[lib];
            s.async = true;
            s.onload = () => resolve(paths[lib]);
            s.onerror = () => reject(`Error loading script: ${paths[lib]}`);
            document.getElementsByTagName("head")[0].appendChild(s);
          });
        }
    
        function showError(err) {
          outputDiv.innerHTML = `<div class="error" style="color:red;">${err}</div>`;
          throw err;
        }
    
        function displayChart(vegaEmbed) {
          vegaEmbed(outputDiv, spec, embedOpt)
            .catch(err => showError(`Javascript Error: ${err.message}<br>This usually means there's a typo in your chart specification. See the javascript console for the full traceback.`));
        }
    
        if(typeof define === "function" && define.amd) {
          requirejs.config({paths});
          require(["vega-embed"], displayChart, err => showError(`Error loading script: ${err.message}`));
        } else if (typeof vegaEmbed === "function") {
          displayChart(vegaEmbed);
        } else {
          loadScript("vega")
            .then(() => loadScript("vega-lite"))
            .then(() => loadScript("vega-embed"))
            .catch(showError)
            .then(() => displayChart(vegaEmbed));
        }
      })({"config": {"view": {"continuousWidth": 400, "continuousHeight": 300}, "axis": {"labelFontSize": 10, "labelLimit": 225.0}}, "vconcat": [{"layer": [{"data": {"name": "data-456173d5f8a5bb144673e82a151da09e"}, "mark": {"type": "rule", "strokeWidth": 3}, "encoding": {"color": {"type": "quantitative", "field": "xValue", "legend": null, "scale": {"domain": [-0.8470079627378541, 0.8470079627378541], "domainMid": 0, "scheme": "redyellowblue"}}, "x": {"type": "quantitative", "axis": {"orient": "top", "titleFontSize": 10}, "field": "xValue", "title": "Select a cutoff range for Feature Importance values "}}}, {"data": {"name": "data-c7bc3ae69e1253e8223f70a11d04b99d"}, "mark": {"type": "rect", "color": "black", "height": 20, "opacity": 0.4, "tooltip": true}, "encoding": {"x": {"type": "quantitative", "field": "x_min"}, "x2": {"field": "x_max"}, "y": {"type": "quantitative", "axis": null, "field": "y"}}, "selection": {"Cutter": {"type": "single", "fields": ["cutoff"], "bind": {"input": "range", "max": 0.8470079627378541, "min": 0, "name": "Importance cutoff value (\u00b1) ", "step": 0.01694015925475708}, "init": {"cutoff": 0.0}}}, "transform": [{"calculate": "(-Cutter.cutoff)", "as": "x_min"}, {"calculate": "Cutter.cutoff", "as": "x_max"}]}], "height": 20}, {"layer": [{"data": {"name": "data-3543b32c4654533cf48a041c66311614"}, "mark": "bar", "encoding": {"color": {"type": "quantitative", "field": "value", "legend": null, "scale": {"domain": [-0.8470079627378541, 0.8470079627378541], "domainMid": 0, "scheme": "blueorange"}}, "tooltip": [{"type": "nominal", "field": "name", "title": "Feature"}, {"type": "quantitative", "field": "value", "title": "Importance"}], "x": {"type": "quantitative", "field": "value", "title": null}, "y": {"type": "nominal", "field": "name", "sort": {"field": "value", "op": "mean", "order": "descending"}, "title": null}}, "selection": {"Cutter": {"type": "single", "fields": ["cutoff"], "bind": {"input": "range", "max": 0.8470079627378541, "min": 0, "name": "Importance cutoff value (\u00b1) ", "step": 0.01694015925475708}, "init": {"cutoff": 0.0}}}, "transform": [{"filter": "((datum.value > Cutter.cutoff) || (datum.value < (-Cutter.cutoff)))"}]}, {"data": {"name": "data-9c00aa9dc2b1c1a0b4bf9b34ead12925"}, "mark": "rule", "encoding": {"x": {"type": "quantitative", "field": "x"}}}]}], "padding": 10, "$schema": "https://vega.github.io/schema/vega-lite/v4.8.1.json", "datasets": {"data-456173d5f8a5bb144673e82a151da09e": [{"xValue": -0.8470079627378541}, {"xValue": -0.8385378831104755}, {"xValue": -0.8300678034830969}, {"xValue": -0.8215977238557184}, {"xValue": -0.8131276442283398}, {"xValue": -0.8046575646009613}, {"xValue": -0.7961874849735827}, {"xValue": -0.7877174053462042}, {"xValue": -0.7792473257188256}, {"xValue": -0.770777246091447}, {"xValue": -0.7623071664640685}, {"xValue": -0.7538370868366899}, {"xValue": -0.7453670072093114}, {"xValue": -0.7368969275819328}, {"xValue": -0.7284268479545543}, {"xValue": -0.7199567683271757}, {"xValue": -0.7114866886997971}, {"xValue": -0.7030166090724186}, {"xValue": -0.69454652944504}, {"xValue": -0.6860764498176615}, {"xValue": -0.6776063701902829}, {"xValue": -0.6691362905629044}, {"xValue": -0.6606662109355258}, {"xValue": -0.6521961313081472}, {"xValue": -0.6437260516807687}, {"xValue": -0.6352559720533901}, {"xValue": -0.6267858924260116}, {"xValue": -0.618315812798633}, {"xValue": -0.6098457331712545}, {"xValue": -0.6013756535438759}, {"xValue": -0.5929055739164973}, {"xValue": -0.5844354942891188}, {"xValue": -0.5759654146617402}, {"xValue": -0.5674953350343617}, {"xValue": -0.5590252554069831}, {"xValue": -0.5505551757796046}, {"xValue": -0.542085096152226}, {"xValue": -0.5336150165248474}, {"xValue": -0.5251449368974689}, {"xValue": -0.5166748572700903}, {"xValue": -0.5082047776427118}, {"xValue": -0.4997346980153332}, {"xValue": -0.49126461838795465}, {"xValue": -0.4827945387605761}, {"xValue": -0.47432445913319754}, {"xValue": -0.465854379505819}, {"xValue": -0.4573842998784404}, {"xValue": -0.44891422025106187}, {"xValue": -0.4404441406236833}, {"xValue": -0.43197406099630475}, {"xValue": -0.4235039813689262}, {"xValue": -0.41503390174154764}, {"xValue": -0.4065638221141691}, {"xValue": -0.3980937424867905}, {"xValue": -0.38962366285941197}, {"xValue": -0.3811535832320334}, {"xValue": -0.37268350360465485}, {"xValue": -0.3642134239772763}, {"xValue": -0.35574334434989774}, {"xValue": -0.3472732647225192}, {"xValue": -0.3388031850951406}, {"xValue": -0.33033310546776207}, {"xValue": -0.3218630258403835}, {"xValue": -0.31339294621300495}, {"xValue": -0.3049228665856264}, {"xValue": -0.29645278695824784}, {"xValue": -0.2879827073308693}, {"xValue": -0.2795126277034907}, {"xValue": -0.27104254807611217}, {"xValue": -0.2625724684487336}, {"xValue": -0.25410238882135505}, {"xValue": -0.2456323091939765}, {"xValue": -0.23716222956659794}, {"xValue": -0.22869214993921938}, {"xValue": -0.22022207031184082}, {"xValue": -0.21175199068446227}, {"xValue": -0.2032819110570837}, {"xValue": -0.19481183142970515}, {"xValue": -0.1863417518023266}, {"xValue": -0.17787167217494804}, {"xValue": -0.16940159254756948}, {"xValue": -0.16093151292019092}, {"xValue": -0.15246143329281237}, {"xValue": -0.1439913536654338}, {"xValue": -0.13552127403805525}, {"xValue": -0.1270511944106767}, {"xValue": -0.11858111478329814}, {"xValue": -0.11011103515591958}, {"xValue": -0.10164095552854102}, {"xValue": -0.09317087590116246}, {"xValue": -0.08470079627378391}, {"xValue": -0.07623071664640535}, {"xValue": -0.06776063701902679}, {"xValue": -0.059290557391648235}, {"xValue": -0.05082047776426968}, {"xValue": -0.04235039813689112}, {"xValue": -0.033880318509512564}, {"xValue": -0.025410238882134006}, {"xValue": -0.01694015925475545}, {"xValue": -0.008470079627376892}, {"xValue": 1.6653345369377348e-15}, {"xValue": 0.008470079627380223}, {"xValue": 0.01694015925475878}, {"xValue": 0.025410238882137337}, {"xValue": 0.033880318509515894}, {"xValue": 0.04235039813689445}, {"xValue": 0.05082047776427301}, {"xValue": 0.059290557391651566}, {"xValue": 0.06776063701903012}, {"xValue": 0.07623071664640868}, {"xValue": 0.08470079627378724}, {"xValue": 0.0931708759011658}, {"xValue": 0.10164095552854435}, {"xValue": 0.11011103515592291}, {"xValue": 0.11858111478330147}, {"xValue": 0.12705119441068002}, {"xValue": 0.13552127403805858}, {"xValue": 0.14399135366543714}, {"xValue": 0.1524614332928157}, {"xValue": 0.16093151292019436}, {"xValue": 0.1694015925475728}, {"xValue": 0.17787167217495126}, {"xValue": 0.18634175180232992}, {"xValue": 0.1948118314297086}, {"xValue": 0.20328191105708704}, {"xValue": 0.21175199068446549}, {"xValue": 0.22022207031184415}, {"xValue": 0.22869214993922282}, {"xValue": 0.23716222956660127}, {"xValue": 0.24563230919397971}, {"xValue": 0.2541023888213584}, {"xValue": 0.26257246844873705}, {"xValue": 0.2710425480761155}, {"xValue": 0.27951262770349394}, {"xValue": 0.2879827073308726}, {"xValue": 0.2964527869582513}, {"xValue": 0.3049228665856297}, {"xValue": 0.31339294621300817}, {"xValue": 0.32186302584038684}, {"xValue": 0.3303331054677655}, {"xValue": 0.33880318509514395}, {"xValue": 0.3472732647225224}, {"xValue": 0.35574334434990107}, {"xValue": 0.36421342397727974}, {"xValue": 0.3726835036046582}, {"xValue": 0.38115358323203663}, {"xValue": 0.3896236628594153}, {"xValue": 0.39809374248679397}, {"xValue": 0.4065638221141724}, {"xValue": 0.41503390174155086}, {"xValue": 0.4235039813689295}, {"xValue": 0.4319740609963082}, {"xValue": 0.44044414062368664}, {"xValue": 0.4489142202510651}, {"xValue": 0.45738429987844376}, {"xValue": 0.4658543795058224}, {"xValue": 0.47432445913320087}, {"xValue": 0.4827945387605793}, {"xValue": 0.491264618387958}, {"xValue": 0.49973469801533665}, {"xValue": 0.5082047776427151}, {"xValue": 0.5166748572700935}, {"xValue": 0.5251449368974722}, {"xValue": 0.5336150165248509}, {"xValue": 0.5420850961522293}, {"xValue": 0.5505551757796078}, {"xValue": 0.5590252554069864}, {"xValue": 0.5674953350343651}, {"xValue": 0.5759654146617436}, {"xValue": 0.584435494289122}, {"xValue": 0.5929055739165007}, {"xValue": 0.6013756535438793}, {"xValue": 0.6098457331712578}, {"xValue": 0.6183158127986362}, {"xValue": 0.6267858924260149}, {"xValue": 0.6352559720533936}, {"xValue": 0.643726051680772}, {"xValue": 0.6521961313081505}, {"xValue": 0.6606662109355291}, {"xValue": 0.6691362905629078}, {"xValue": 0.6776063701902862}, {"xValue": 0.6860764498176647}, {"xValue": 0.6945465294450434}, {"xValue": 0.703016609072422}, {"xValue": 0.7114866886998005}, {"xValue": 0.7199567683271789}, {"xValue": 0.7284268479545576}, {"xValue": 0.7368969275819363}, {"xValue": 0.7453670072093147}, {"xValue": 0.7538370868366931}, {"xValue": 0.7623071664640718}, {"xValue": 0.7707772460914505}, {"xValue": 0.7792473257188289}, {"xValue": 0.7877174053462074}, {"xValue": 0.796187484973586}, {"xValue": 0.8046575646009647}, {"xValue": 0.8131276442283432}, {"xValue": 0.8215977238557216}, {"xValue": 0.8300678034831003}, {"xValue": 0.8385378831104789}, {"xValue": 0.8470079627378574}], "data-c7bc3ae69e1253e8223f70a11d04b99d": [{"y": 0, "x": -0.5, "x2": 0.5}], "data-3543b32c4654533cf48a041c66311614": [{"name": "duration_in_month", "value": 0.8470079627378541}, {"name": "credit_amount", "value": 0.6944189404874552}, {"name": "installment_as_income_perc", "value": 0.3043826317623826}, {"name": "present_res_since", "value": -0.017915878343538215}, {"name": "age", "value": -0.005163583410577356}, {"name": "credits_this_bank", "value": -0.20402960332384254}, {"name": "people_under_maintenance", "value": -0.056755931097354474}, {"name": "account_check_status=0 <= ... < 200 DM", "value": 0.2939399436033418}, {"name": "account_check_status=< 0 DM", "value": -0.21818023774927736}, {"name": "account_check_status=>= 200 DM / salary assignments for at least 1 year", "value": 9.60072863884819e-05}, {"name": "account_check_status=no checking account", "value": 0.3502124003353961}, {"name": "credit_history=all credits at this bank paid back duly", "value": 0.6337640692749543}, {"name": "credit_history=critical account/ other credits existing (not at this bank)", "value": 0.21823305363235895}, {"name": "credit_history=delay in paying off in the past", "value": -0.011253602415574547}, {"name": "credit_history=existing credits paid back duly till now", "value": -0.13292335773836597}, {"name": "credit_history=no credits taken/ all credits paid back duly", "value": -0.029597808453158823}, {"name": "purpose=(vacation - does not exist?)", "value": -0.06856154357807029}, {"name": "purpose=business", "value": 0.003674153425649057}, {"name": "purpose=car (new)", "value": 0.2513672502321305}, {"name": "purpose=car (used)", "value": 0.08468160953031668}, {"name": "purpose=domestic appliances", "value": 0.03854879546694716}, {"name": "purpose=education", "value": 0.0}, {"name": "purpose=furniture/equipment", "value": -2.9446802618582078e-18}, {"name": "purpose=radio/television", "value": 0.006492757319555031}, {"name": "purpose=repairs", "value": 0.005963146482736863}, {"name": "purpose=retraining", "value": 2.3901727085545897e-18}, {"name": "savings=.. >= 1000 DM ", "value": 0.056705276861560266}, {"name": "savings=... < 100 DM", "value": -0.33819273164897024}, {"name": "savings=100 <= ... < 500 DM", "value": -0.012064219528328623}, {"name": "savings=500 <= ... < 1000 DM ", "value": 0.0010553301554828342}, {"name": "savings=unknown/ no savings account", "value": -0.5556654011176968}, {"name": "present_emp_since=.. >= 7 years", "value": 0.0052940900934866746}, {"name": "present_emp_since=... < 1 year ", "value": -0.04654979417893307}, {"name": "present_emp_since=1 <= ... < 4 years", "value": -0.017047143561136142}, {"name": "present_emp_since=4 <= ... < 7 years", "value": 0.09509224367078725}, {"name": "present_emp_since=unemployed", "value": 0.72638253902645}, {"name": "personal_status_sex=female : divorced/separated/married", "value": -0.05151278954058162}, {"name": "personal_status_sex=male : divorced/separated", "value": -0.023431119320486585}, {"name": "personal_status_sex=male : married/widowed", "value": -0.01650280113349013}, {"name": "personal_status_sex=male : single", "value": -0.134994120205601}, {"name": "other_debtors=co-applicant", "value": 0.421695043431072}, {"name": "other_debtors=guarantor", "value": 0.031225103977219948}, {"name": "other_debtors=none", "value": -0.11203996569130724}, {"name": "property=if not A121 : building society savings agreement/ life insurance", "value": 0.04236722282145007}, {"name": "property=if not A121/A122 : car or other, not in attribute 6", "value": 0.041922435650431605}, {"name": "property=real estate", "value": 0.0814789776398865}, {"name": "property=unknown / no property", "value": 0.7764063820618664}, {"name": "other_installment_plans=bank", "value": -0.07246505874542243}, {"name": "other_installment_plans=none", "value": -0.0821908358522365}, {"name": "other_installment_plans=stores", "value": 0.0016952408348179034}, {"name": "housing=for free", "value": -0.7476842519071202}, {"name": "housing=own", "value": 0.023375705983559408}, {"name": "housing=rent", "value": -0.105753305473695}, {"name": "job=management/ self-employed/ highly qualified employee/ officer", "value": -0.1219598759071389}, {"name": "job=skilled employee / official", "value": -0.028367937850261832}, {"name": "job=unemployed/ unskilled - non-resident", "value": 0.002873156187994504}, {"name": "job=unskilled - resident", "value": -0.014458079729267354}, {"name": "telephone=none", "value": -0.0750707003078665}, {"name": "telephone=yes, registered under the customers name ", "value": -0.0750707003078658}, {"name": "foreign_worker=no", "value": 0.04268973216067927}, {"name": "foreign_worker=yes", "value": 0.04268973216067925}], "data-9c00aa9dc2b1c1a0b4bf9b34ead12925": [{"x": 0}]}}, {"mode": "vega-lite"});
    </script>


LORE explainer
~~~~~~~~~~~~~~

.. code:: ipython3

    explainer = LoreTabularExplainer(bbox)
    config = {'neigh_type':'geneticp', 'size':1000, 'ocr':0.1, 'ngen':10}
    explainer.fit(df, class_field, config)
    exp = explainer.explain(inst)
    print(exp)


.. parsed-literal::

    <xailib.explainers.lore_explainer.LoreTabularExplanation object at 0x12bc41a90>
    

.. code:: ipython3

    exp.plotRules()



.. raw:: html

    
    <style>
    .red {
    background-color:firebrick;
    padding:3px 5px 3px 5px;
    border-radius:5px;
    
    color:white;
    }
    .rules{
        margin-top:10px;
        font-weight: 400;
    }
    .rule{
    padding:5px 20px 5px 20px;
    border-radius:5px;
    margin-right:5px;
    font-size:12px;
    line-height:20px;
    display: block;
    margin-bottom: 10px;
    width: fit-content;
    
    color:white;
    background-color:firebrick;
    opacity:0.8;
    }
    </style>
    



.. raw:: html

    
    <h3>Why the predicted value for class <span class='red'>default</span> is <span class='red'>1</span> ?</h3>
    



.. raw:: html

    
    <p>Because all the following conditions happen:</p>
    



.. raw:: html

    <p class='rules'><span class='rule'>age <strong><=</strong> 20.726173400878906</span><span class='rule'>credit amount <strong>></strong> -439.6443485021591</span><span class='rule'>purpose=retraining <strong><=</strong> 0.11524588242173195</span><span class='rule'>duration in month <strong>></strong> -1.9407005310058594</span><span class='rule'>purpose=furniture/equipment <strong><=</strong> 0.18370826542377472</span><span class='rule'>foreign worker=no <strong><=</strong> 0.7168410122394562</span><span class='rule'>purpose=domestic appliances <strong><=</strong> 1.015466570854187</span><span class='rule'>savings=.. >= 1000 DM  <strong><=</strong> 0.7176859378814697</span><span class='rule'>purpose=(vacation - does not exist?) <strong><=</strong> 0.4622504562139511</span><span class='rule'>credit history=critical account/ other credits existing (not at this bank) <strong><=</strong> 0.9085964262485504</span></p>


.. code:: ipython3

    exp.plotCounterfactualRules()



.. raw:: html

    
    <style>
    .red {
    background-color:firebrick;
    padding:3px 5px 3px 5px;
    border-radius:5px;
    color:white;
    }
    .crules{
        margin-top:10px;
        font-weight: 400;
    }
    .crule{            
    padding:5px 20px 5px 20px;
    border-radius:5px;
    margin-right:5px;
    font-size:12px;
    line-height:20px;
    display: block;
    margin-bottom: 10px;
    width: fit-content;
    
    color:#202020;
    background-color:gold;
    }
    </style>
    



.. raw:: html

    
    <h3>The predicted value for class <span class='red'>default</span> is <span class='red'>1</span>.</h3>
    <h3>It would have been:</h3>
    



.. raw:: html

    
    <div class='crules'>
        <div>
            <h4><span class='red'>0</span> if the following condition holds</h4>
        </br><span class='crule'>age <= 20.726173400878906</span><span class='crule'>credit amount <= -439.6443485021591</span>
        </div>
    </div>
    



.. raw:: html

    
    <div class='crules'>
        <div>
            <h4><span class='red'>0</span> if the following condition holds</h4>
        </br><span class='crule'>age > 20.726173400878906</span><span class='crule'>credit amount <= 26.468921303749084</span><span class='crule'>duration in month <= 5.795059680938721</span><span class='crule'>installment as income perc <= 4.603440999984741</span>
        </div>
    </div>
    


LIME explainer
~~~~~~~~~~~~~~

.. code:: ipython3

    limeExplainer = LimeXAITabularExplainer(bbox)
    config = {'feature_selection': 'lasso_path'}
    limeExplainer.fit(df, class_field, config)
    lime_exp = limeExplainer.explain(inst)
    print(lime_exp.exp.as_list())


.. parsed-literal::

    [('other_debtors=co-applicant', -1.3046177878918616e-09), ('credit_history=all credits at this bank paid back duly', -1.0114574629252053e-09), ('present_emp_since=unemployed', -8.87554096296626e-10), ('other_debtors=none', 7.43754044231906e-10), ('housing=for free', -4.4157786564097103e-10), ('property=unknown / no property', -3.275710719845092e-10), ('credit_amount', 3.271233788564153e-10), ('job=management/ self-employed/ highly qualified employee/ officer', -3.164190703926506e-10), ('housing=own', 2.8902027822084106e-10), ('savings=unknown/ no savings account', -2.604277452741881e-10), ('job=skilled employee / official', 2.3808188198617575e-10), ('foreign_worker=yes', 2.365347360238489e-10), ('telephone=none', 2.2048259721367863e-10), ('age', 2.171945479826713e-10), ('savings=... < 100 DM', 2.1116662177987812e-10), ('credits_this_bank', 1.9999632029038067e-10), ('credit_history=existing credits paid back duly till now', 1.9243622007776865e-10), ('people_under_maintenance', 1.902008911572941e-10), ('purpose=car (new)', -1.7104663723358493e-10), ('account_check_status=no checking account', 1.6584313433238958e-10), ('account_check_status=0 <= ... < 200 DM', -1.639544710042764e-10), ('credit_history=critical account/ other credits existing (not at this bank)', 1.317487567892989e-10), ('job=unskilled - resident', 1.307761159896724e-10), ('other_installment_plans=stores', 1.2347569776391545e-10), ('foreign_worker=no', 1.1825353902253505e-10), ('present_emp_since=1 <= ... < 4 years', 1.1478921168922655e-10), ('property=if not A121/A122 : car or other, not in attribute 6', 1.1222769011436428e-10), ('personal_status_sex=female : divorced/separated/married', 1.1002871894681165e-10), ('savings=100 <= ... < 500 DM', 1.0982251402773794e-10), ('purpose=domestic appliances', 1.0567984890752028e-10), ('present_res_since', 9.869484730455045e-11), ('account_check_status=>= 200 DM / salary assignments for at least 1 year', 9.721716212812873e-11), ('present_emp_since=.. >= 7 years', 9.327030468700815e-11), ('installment_as_income_perc', 9.192261925231111e-11), ('property=real estate', 9.180043418264463e-11), ('purpose=(vacation - does not exist?)', 8.974505020571898e-11), ('account_check_status=< 0 DM', 8.848004118893571e-11), ('purpose=retraining', 8.80910843922895e-11), ('purpose=education', 8.803520453193465e-11), ('purpose=business', 8.330599059469541e-11), ('housing=rent', 7.975475868460632e-11), ('property=if not A121 : building society savings agreement/ life insurance', 7.826524390749874e-11), ('other_debtors=guarantor', 7.385760952840171e-11), ('present_emp_since=... < 1 year ', 7.338094381227495e-11), ('duration_in_month', 6.689756440260244e-11), ('purpose=car (used)', 6.582965568284186e-11), ('job=unemployed/ unskilled - non-resident', 6.473736018584135e-11), ('present_emp_since=4 <= ... < 7 years', 6.230002403518189e-11), ('purpose=furniture/equipment', 5.974714318917145e-11), ('purpose=radio/television', 5.909852887925919e-11), ('credit_history=delay in paying off in the past', 5.620862803354922e-11), ('savings=500 <= ... < 1000 DM ', 5.582941358078461e-11), ('other_installment_plans=none', 5.501318386790144e-11), ('personal_status_sex=male : married/widowed', 5.500125372750834e-11), ('telephone=yes, registered under the customers name ', -5.495252929908006e-11), ('purpose=repairs', 5.2177896575440796e-11), ('savings=.. >= 1000 DM ', 4.0557757647139625e-11), ('personal_status_sex=male : divorced/separated', 3.627184253632623e-11), ('personal_status_sex=male : single', -2.9862189862658355e-11), ('credit_history=no credits taken/ all credits paid back duly', 2.8131802175589855e-11), ('other_installment_plans=bank', 1.9548368945624186e-11)]
    

.. code:: ipython3

    lime_exp.plot_features_importance()



.. raw:: html

    
    <style>
    .vega-bind {
      position: absolute;
      left: 0px;
      top: 0px;
      background-color:#eee;
      padding:10px;
      font-size:10px;
    }
    .chart-wrapper{
      padding-top: 70px;
    }
    
    </style>
    



.. raw:: html

    
    <div id="altair-viz-888610c43438438b9ef686360a3ff500"></div>
    <script type="text/javascript">
      (function(spec, embedOpt){
        let outputDiv = document.currentScript.previousElementSibling;
        if (outputDiv.id !== "altair-viz-888610c43438438b9ef686360a3ff500") {
          outputDiv = document.getElementById("altair-viz-888610c43438438b9ef686360a3ff500");
        }
        const paths = {
          "vega": "https://cdn.jsdelivr.net/npm//vega@5?noext",
          "vega-lib": "https://cdn.jsdelivr.net/npm//vega-lib?noext",
          "vega-lite": "https://cdn.jsdelivr.net/npm//vega-lite@4.8.1?noext",
          "vega-embed": "https://cdn.jsdelivr.net/npm//vega-embed@6?noext",
        };
    
        function loadScript(lib) {
          return new Promise(function(resolve, reject) {
            var s = document.createElement('script');
            s.src = paths[lib];
            s.async = true;
            s.onload = () => resolve(paths[lib]);
            s.onerror = () => reject(`Error loading script: ${paths[lib]}`);
            document.getElementsByTagName("head")[0].appendChild(s);
          });
        }
    
        function showError(err) {
          outputDiv.innerHTML = `<div class="error" style="color:red;">${err}</div>`;
          throw err;
        }
    
        function displayChart(vegaEmbed) {
          vegaEmbed(outputDiv, spec, embedOpt)
            .catch(err => showError(`Javascript Error: ${err.message}<br>This usually means there's a typo in your chart specification. See the javascript console for the full traceback.`));
        }
    
        if(typeof define === "function" && define.amd) {
          requirejs.config({paths});
          require(["vega-embed"], displayChart, err => showError(`Error loading script: ${err.message}`));
        } else if (typeof vegaEmbed === "function") {
          displayChart(vegaEmbed);
        } else {
          loadScript("vega")
            .then(() => loadScript("vega-lite"))
            .then(() => loadScript("vega-embed"))
            .catch(showError)
            .then(() => displayChart(vegaEmbed));
        }
      })({"config": {"view": {"continuousWidth": 400, "continuousHeight": 300}, "axis": {"labelFontSize": 10, "labelLimit": 225.0}}, "vconcat": [{"layer": [{"data": {"name": "data-ce582a2c277beead1759820d7cd3f420"}, "mark": {"type": "rule", "strokeWidth": 3}, "encoding": {"color": {"type": "quantitative", "field": "xValue", "legend": null, "scale": {"domain": [-1.3046177878918616e-09, 1.3046177878918616e-09], "domainMid": 0, "scheme": "redyellowblue"}}, "x": {"type": "quantitative", "axis": {"orient": "top", "titleFontSize": 10}, "field": "xValue", "title": "Select a cutoff range for Feature Importance values "}}}, {"data": {"name": "data-c7bc3ae69e1253e8223f70a11d04b99d"}, "mark": {"type": "rect", "color": "black", "height": 20, "opacity": 0.4, "tooltip": true}, "encoding": {"x": {"type": "quantitative", "field": "x_min"}, "x2": {"field": "x_max"}, "y": {"type": "quantitative", "axis": null, "field": "y"}}, "selection": {"Cutter": {"type": "single", "fields": ["cutoff"], "bind": {"input": "range", "max": 1.3046177878918616e-09, "min": 0, "name": "Importance cutoff value (\u00b1) ", "step": 2.609235575783723e-11}, "init": {"cutoff": 0.0}}}, "transform": [{"calculate": "(-Cutter.cutoff)", "as": "x_min"}, {"calculate": "Cutter.cutoff", "as": "x_max"}]}], "height": 20}, {"layer": [{"data": {"name": "data-88db20dec88e237822ea09b14e3327a7"}, "mark": "bar", "encoding": {"color": {"type": "quantitative", "field": "value", "legend": null, "scale": {"domain": [-1.3046177878918616e-09, 1.3046177878918616e-09], "domainMid": 0, "scheme": "blueorange"}}, "tooltip": [{"type": "nominal", "field": "name", "title": "Feature"}, {"type": "quantitative", "field": "value", "title": "Importance"}], "x": {"type": "quantitative", "field": "value", "title": null}, "y": {"type": "nominal", "field": "name", "sort": {"field": "value", "op": "mean", "order": "descending"}, "title": null}}, "selection": {"Cutter": {"type": "single", "fields": ["cutoff"], "bind": {"input": "range", "max": 1.3046177878918616e-09, "min": 0, "name": "Importance cutoff value (\u00b1) ", "step": 2.609235575783723e-11}, "init": {"cutoff": 0.0}}}, "transform": [{"filter": "((datum.value > Cutter.cutoff) || (datum.value < (-Cutter.cutoff)))"}]}, {"data": {"name": "data-9c00aa9dc2b1c1a0b4bf9b34ead12925"}, "mark": "rule", "encoding": {"x": {"type": "quantitative", "field": "x"}}}]}], "padding": 10, "$schema": "https://vega.github.io/schema/vega-lite/v4.8.1.json", "datasets": {"data-ce582a2c277beead1759820d7cd3f420": [{"xValue": -1.3046177878918616e-09}, {"xValue": -1.291571610012943e-09}, {"xValue": -1.2785254321340243e-09}, {"xValue": -1.2654792542551056e-09}, {"xValue": -1.252433076376187e-09}, {"xValue": -1.2393868984972683e-09}, {"xValue": -1.2263407206183496e-09}, {"xValue": -1.213294542739431e-09}, {"xValue": -1.2002483648605122e-09}, {"xValue": -1.1872021869815936e-09}, {"xValue": -1.174156009102675e-09}, {"xValue": -1.1611098312237562e-09}, {"xValue": -1.1480636533448376e-09}, {"xValue": -1.1350174754659189e-09}, {"xValue": -1.1219712975870002e-09}, {"xValue": -1.1089251197080815e-09}, {"xValue": -1.0958789418291629e-09}, {"xValue": -1.0828327639502442e-09}, {"xValue": -1.0697865860713255e-09}, {"xValue": -1.0567404081924069e-09}, {"xValue": -1.0436942303134882e-09}, {"xValue": -1.0306480524345695e-09}, {"xValue": -1.0176018745556508e-09}, {"xValue": -1.0045556966767322e-09}, {"xValue": -9.915095187978135e-10}, {"xValue": -9.784633409188948e-10}, {"xValue": -9.654171630399761e-10}, {"xValue": -9.523709851610575e-10}, {"xValue": -9.393248072821388e-10}, {"xValue": -9.262786294032201e-10}, {"xValue": -9.132324515243015e-10}, {"xValue": -9.001862736453828e-10}, {"xValue": -8.871400957664641e-10}, {"xValue": -8.740939178875454e-10}, {"xValue": -8.610477400086268e-10}, {"xValue": -8.480015621297081e-10}, {"xValue": -8.349553842507894e-10}, {"xValue": -8.219092063718708e-10}, {"xValue": -8.088630284929521e-10}, {"xValue": -7.958168506140334e-10}, {"xValue": -7.827706727351147e-10}, {"xValue": -7.697244948561961e-10}, {"xValue": -7.566783169772774e-10}, {"xValue": -7.436321390983587e-10}, {"xValue": -7.305859612194401e-10}, {"xValue": -7.175397833405214e-10}, {"xValue": -7.044936054616027e-10}, {"xValue": -6.91447427582684e-10}, {"xValue": -6.784012497037654e-10}, {"xValue": -6.653550718248467e-10}, {"xValue": -6.52308893945928e-10}, {"xValue": -6.392627160670093e-10}, {"xValue": -6.262165381880907e-10}, {"xValue": -6.13170360309172e-10}, {"xValue": -6.001241824302533e-10}, {"xValue": -5.870780045513347e-10}, {"xValue": -5.74031826672416e-10}, {"xValue": -5.609856487934973e-10}, {"xValue": -5.479394709145786e-10}, {"xValue": -5.3489329303566e-10}, {"xValue": -5.218471151567413e-10}, {"xValue": -5.088009372778226e-10}, {"xValue": -4.95754759398904e-10}, {"xValue": -4.827085815199853e-10}, {"xValue": -4.696624036410666e-10}, {"xValue": -4.5661622576214794e-10}, {"xValue": -4.4357004788322927e-10}, {"xValue": -4.305238700043106e-10}, {"xValue": -4.174776921253919e-10}, {"xValue": -4.0443151424647325e-10}, {"xValue": -3.913853363675546e-10}, {"xValue": -3.783391584886359e-10}, {"xValue": -3.6529298060971723e-10}, {"xValue": -3.5224680273079856e-10}, {"xValue": -3.392006248518799e-10}, {"xValue": -3.261544469729612e-10}, {"xValue": -3.1310826909404255e-10}, {"xValue": -3.0006209121512387e-10}, {"xValue": -2.870159133362052e-10}, {"xValue": -2.7396973545728653e-10}, {"xValue": -2.6092355757836786e-10}, {"xValue": -2.478773796994492e-10}, {"xValue": -2.348312018205305e-10}, {"xValue": -2.2178502394161184e-10}, {"xValue": -2.0873884606269317e-10}, {"xValue": -1.956926681837745e-10}, {"xValue": -1.8264649030485583e-10}, {"xValue": -1.6960031242593715e-10}, {"xValue": -1.5655413454701848e-10}, {"xValue": -1.435079566680998e-10}, {"xValue": -1.3046177878918114e-10}, {"xValue": -1.1741560091026247e-10}, {"xValue": -1.0436942303134379e-10}, {"xValue": -9.132324515242512e-11}, {"xValue": -7.827706727350645e-11}, {"xValue": -6.523088939458778e-11}, {"xValue": -5.2184711515669105e-11}, {"xValue": -3.913853363675043e-11}, {"xValue": -2.609235575783176e-11}, {"xValue": -1.3046177878913089e-11}, {"xValue": 5.583469134732937e-24}, {"xValue": 1.3046177878924256e-11}, {"xValue": 2.6092355757842928e-11}, {"xValue": 3.91385336367616e-11}, {"xValue": 5.218471151568027e-11}, {"xValue": 6.523088939459894e-11}, {"xValue": 7.827706727351762e-11}, {"xValue": 9.132324515243629e-11}, {"xValue": 1.0436942303135496e-10}, {"xValue": 1.1741560091027363e-10}, {"xValue": 1.304617787891923e-10}, {"xValue": 1.4350795666811098e-10}, {"xValue": 1.5655413454702965e-10}, {"xValue": 1.6960031242594832e-10}, {"xValue": 1.82646490304867e-10}, {"xValue": 1.9569266818378566e-10}, {"xValue": 2.0873884606270434e-10}, {"xValue": 2.21785023941623e-10}, {"xValue": 2.348312018205417e-10}, {"xValue": 2.4787737969946035e-10}, {"xValue": 2.60923557578379e-10}, {"xValue": 2.739697354572977e-10}, {"xValue": 2.8701591333621637e-10}, {"xValue": 3.0006209121513504e-10}, {"xValue": 3.131082690940537e-10}, {"xValue": 3.261544469729724e-10}, {"xValue": 3.3920062485189106e-10}, {"xValue": 3.5224680273080973e-10}, {"xValue": 3.652929806097284e-10}, {"xValue": 3.7833915848864707e-10}, {"xValue": 3.9138533636756575e-10}, {"xValue": 4.044315142464844e-10}, {"xValue": 4.174776921254031e-10}, {"xValue": 4.3052387000432176e-10}, {"xValue": 4.4357004788324043e-10}, {"xValue": 4.566162257621591e-10}, {"xValue": 4.696624036410778e-10}, {"xValue": 4.827085815199964e-10}, {"xValue": 4.957547593989151e-10}, {"xValue": 5.088009372778338e-10}, {"xValue": 5.218471151567525e-10}, {"xValue": 5.348932930356711e-10}, {"xValue": 5.479394709145898e-10}, {"xValue": 5.609856487935085e-10}, {"xValue": 5.740318266724272e-10}, {"xValue": 5.870780045513458e-10}, {"xValue": 6.001241824302645e-10}, {"xValue": 6.131703603091832e-10}, {"xValue": 6.262165381881018e-10}, {"xValue": 6.392627160670205e-10}, {"xValue": 6.523088939459392e-10}, {"xValue": 6.653550718248579e-10}, {"xValue": 6.784012497037765e-10}, {"xValue": 6.914474275826952e-10}, {"xValue": 7.044936054616139e-10}, {"xValue": 7.175397833405325e-10}, {"xValue": 7.305859612194512e-10}, {"xValue": 7.436321390983699e-10}, {"xValue": 7.566783169772886e-10}, {"xValue": 7.697244948562072e-10}, {"xValue": 7.827706727351259e-10}, {"xValue": 7.958168506140446e-10}, {"xValue": 8.088630284929633e-10}, {"xValue": 8.219092063718819e-10}, {"xValue": 8.349553842508006e-10}, {"xValue": 8.480015621297193e-10}, {"xValue": 8.610477400086379e-10}, {"xValue": 8.740939178875566e-10}, {"xValue": 8.871400957664753e-10}, {"xValue": 9.00186273645394e-10}, {"xValue": 9.132324515243126e-10}, {"xValue": 9.262786294032313e-10}, {"xValue": 9.3932480728215e-10}, {"xValue": 9.523709851610686e-10}, {"xValue": 9.654171630399873e-10}, {"xValue": 9.78463340918906e-10}, {"xValue": 9.915095187978247e-10}, {"xValue": 1.0045556966767433e-09}, {"xValue": 1.017601874555662e-09}, {"xValue": 1.0306480524345807e-09}, {"xValue": 1.0436942303134993e-09}, {"xValue": 1.056740408192418e-09}, {"xValue": 1.0697865860713367e-09}, {"xValue": 1.0828327639502554e-09}, {"xValue": 1.095878941829174e-09}, {"xValue": 1.1089251197080927e-09}, {"xValue": 1.1219712975870114e-09}, {"xValue": 1.13501747546593e-09}, {"xValue": 1.1480636533448487e-09}, {"xValue": 1.1611098312237674e-09}, {"xValue": 1.174156009102686e-09}, {"xValue": 1.1872021869816047e-09}, {"xValue": 1.2002483648605234e-09}, {"xValue": 1.213294542739442e-09}, {"xValue": 1.2263407206183608e-09}, {"xValue": 1.2393868984972794e-09}, {"xValue": 1.252433076376198e-09}, {"xValue": 1.2654792542551168e-09}, {"xValue": 1.2785254321340354e-09}, {"xValue": 1.2915716100129541e-09}], "data-c7bc3ae69e1253e8223f70a11d04b99d": [{"y": 0, "x": -0.5, "x2": 0.5}], "data-88db20dec88e237822ea09b14e3327a7": [{"name": "other_debtors=co-applicant", "value": -1.3046177878918616e-09}, {"name": "credit_history=all credits at this bank paid back duly", "value": -1.0114574629252053e-09}, {"name": "present_emp_since=unemployed", "value": -8.87554096296626e-10}, {"name": "other_debtors=none", "value": 7.43754044231906e-10}, {"name": "housing=for free", "value": -4.4157786564097103e-10}, {"name": "property=unknown / no property", "value": -3.275710719845092e-10}, {"name": "credit_amount", "value": 3.271233788564153e-10}, {"name": "job=management/ self-employed/ highly qualified employee/ officer", "value": -3.164190703926506e-10}, {"name": "housing=own", "value": 2.8902027822084106e-10}, {"name": "savings=unknown/ no savings account", "value": -2.604277452741881e-10}, {"name": "job=skilled employee / official", "value": 2.3808188198617575e-10}, {"name": "foreign_worker=yes", "value": 2.365347360238489e-10}, {"name": "telephone=none", "value": 2.2048259721367863e-10}, {"name": "age", "value": 2.171945479826713e-10}, {"name": "savings=... < 100 DM", "value": 2.1116662177987812e-10}, {"name": "credits_this_bank", "value": 1.9999632029038067e-10}, {"name": "credit_history=existing credits paid back duly till now", "value": 1.9243622007776865e-10}, {"name": "people_under_maintenance", "value": 1.902008911572941e-10}, {"name": "purpose=car (new)", "value": -1.7104663723358493e-10}, {"name": "account_check_status=no checking account", "value": 1.6584313433238958e-10}, {"name": "account_check_status=0 <= ... < 200 DM", "value": -1.639544710042764e-10}, {"name": "credit_history=critical account/ other credits existing (not at this bank)", "value": 1.317487567892989e-10}, {"name": "job=unskilled - resident", "value": 1.307761159896724e-10}, {"name": "other_installment_plans=stores", "value": 1.2347569776391545e-10}, {"name": "foreign_worker=no", "value": 1.1825353902253505e-10}, {"name": "present_emp_since=1 <= ... < 4 years", "value": 1.1478921168922655e-10}, {"name": "property=if not A121/A122 : car or other, not in attribute 6", "value": 1.1222769011436428e-10}, {"name": "personal_status_sex=female : divorced/separated/married", "value": 1.1002871894681165e-10}, {"name": "savings=100 <= ... < 500 DM", "value": 1.0982251402773794e-10}, {"name": "purpose=domestic appliances", "value": 1.0567984890752028e-10}, {"name": "present_res_since", "value": 9.869484730455045e-11}, {"name": "account_check_status=>= 200 DM / salary assignments for at least 1 year", "value": 9.721716212812873e-11}, {"name": "present_emp_since=.. >= 7 years", "value": 9.327030468700815e-11}, {"name": "installment_as_income_perc", "value": 9.192261925231111e-11}, {"name": "property=real estate", "value": 9.180043418264463e-11}, {"name": "purpose=(vacation - does not exist?)", "value": 8.974505020571898e-11}, {"name": "account_check_status=< 0 DM", "value": 8.848004118893571e-11}, {"name": "purpose=retraining", "value": 8.80910843922895e-11}, {"name": "purpose=education", "value": 8.803520453193465e-11}, {"name": "purpose=business", "value": 8.330599059469541e-11}, {"name": "housing=rent", "value": 7.975475868460632e-11}, {"name": "property=if not A121 : building society savings agreement/ life insurance", "value": 7.826524390749874e-11}, {"name": "other_debtors=guarantor", "value": 7.385760952840171e-11}, {"name": "present_emp_since=... < 1 year ", "value": 7.338094381227495e-11}, {"name": "duration_in_month", "value": 6.689756440260244e-11}, {"name": "purpose=car (used)", "value": 6.582965568284186e-11}, {"name": "job=unemployed/ unskilled - non-resident", "value": 6.473736018584135e-11}, {"name": "present_emp_since=4 <= ... < 7 years", "value": 6.230002403518189e-11}, {"name": "purpose=furniture/equipment", "value": 5.974714318917145e-11}, {"name": "purpose=radio/television", "value": 5.909852887925919e-11}, {"name": "credit_history=delay in paying off in the past", "value": 5.620862803354922e-11}, {"name": "savings=500 <= ... < 1000 DM ", "value": 5.582941358078461e-11}, {"name": "other_installment_plans=none", "value": 5.501318386790144e-11}, {"name": "personal_status_sex=male : married/widowed", "value": 5.500125372750834e-11}, {"name": "telephone=yes, registered under the customers name ", "value": -5.495252929908006e-11}, {"name": "purpose=repairs", "value": 5.2177896575440796e-11}, {"name": "savings=.. >= 1000 DM ", "value": 4.0557757647139625e-11}, {"name": "personal_status_sex=male : divorced/separated", "value": 3.627184253632623e-11}, {"name": "personal_status_sex=male : single", "value": -2.9862189862658355e-11}, {"name": "credit_history=no credits taken/ all credits paid back duly", "value": 2.8131802175589855e-11}, {"name": "other_installment_plans=bank", "value": 1.9548368945624186e-11}], "data-9c00aa9dc2b1c1a0b4bf9b34ead12925": [{"x": 0}]}}, {"mode": "vega-lite"});
    </script>