"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plot_model_performance(rf_orig, data_orig_test.features, y_test)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 3. Bias and Fairness\n",
"\n",
"Today, a problem of the model that can be produce by Machine Learning is bias that data can have. So a question is how to measure those bias and how to avoid them. In python there is a package produced by IBM called [aif360](https://github.com/IBM/AIF360) that can gives us some metrics and algorithms to know if our data / model are bias and to get a fair model.\n",
"\n",
"### 3.1 Metrics\n",
"\n",
"So with aif360 we have some metrics that indicate if our data or model are bias. I will use 5 metrics : \n",
"* Statistical Parity Difference\n",
"* Equal Opportunity Difference\n",
"* Average Absolute Odds Difference\n",
"* Disparate Impact\n",
"* Theil Index\n",
"\n",
"#### 3.1.1 Statistical Parity Difference\n",
"\n",
"This measure is based on the following formula : \n",
"\n",
"$$ Pr(Y=1|D=unprivileged) - Pr(Y=1|D=privileged) $$\n",
"\n",
"Here the bias or *statistical imparity* is the difference between the probability that a random individual drawn from unprivileged is labeled 1 (so here that he has more than 50K for income) and the probability that a random individual from privileged is labeled 1.\n",
"\n",
"So it has to be close to **0** so it will be fair.\n",
"\n",
"Also you can find more details about that here : [One definition of algorithmic fairness: statistical parity](https://jeremykun.com/2015/10/19/one-definition-of-algorithmic-fairness-statistical-parity/)\n",
"\n",
"\n",
"#### 3.1.2 Equal Opportunity Difference\n",
"\n",
"This metric is just a difference between the true positive rate of unprivileged group and the true positive rate of privileged group so it follows this formula :\n",
"\n",
"$$ TPR_{D=unprivileged} - TPR_{D=privileged} $$ \n",
"\n",
"Same as the previous metric we need it to be close to **0**.\n",
"\n",
"#### 3.1.3 Average Absolute Odds Difference\n",
"\n",
"This measure is using both false positive rate and true positive rate to calculate the bias. It's calculating the equality of odds with the next formula :\n",
"\n",
"$$ \\frac{1}{2}[|FPR_{D=unprivileged} - FPR_{D=privileged} | + | TPR_{D=unprivileged} - TPR_{D=privileged}|]$$\n",
"\n",
"It needs to be equal to **0** to be fair.\n",
"\n",
"#### 3.1.4 Disparate Impact\n",
"\n",
"For this metric we use the following formula :\n",
"\n",
"$$ \\frac{Pr(Y=1|D=unprivileged)}{Pr(Y=1|D=privileged)} $$\n",
"\n",
"Like the first metric we use both probabities of a random individual drawn from unprivileged or privileged with a label of 1 but here it's a ratio. \n",
"\n",
"It changes the objective, for the disparate impact it's **1** that we need.\n",
"\n",
"#### 3.1.5 Theil Index\n",
"\n",
"This measure is also known as the generalized entropy index but with $\\alpha$ equals to 1 (more informations on [the Wikipedia page](https://en.wikipedia.org/wiki/Generalized_entropy_index)). So we can calculate it with this formula :\n",
"\n",
"$$ \\frac{1}{n} \\sum_{i=0}^{n} \\frac{b_i}{\\mu} ln \\frac{b_i}{\\mu} $$ \n",
"\n",
"Where $b_i = \\hat{y}_i - y_i + 1 $\n",
"\n",
"So it needs to be close to **0** to be fair.\n",
"\n",
"\n",
"With these differents metrics we can see that for some we need the prediction and for others just the original dataset. This is why we will use 2 classes of the aif360 package : `ClassificationMetric` and `BinaryLabelDatasetMetric`. \n",
"\n",
"For the first one we need the prediction so we can have the following metrics : \n",
"* Equal Opportunity Difference : [documentation link](https://aif360.readthedocs.io/en/latest/modules/metrics.html#aif360.metrics.ClassificationMetric.equal_opportunity_difference)\n",
" \n",
"`equal_opportunity_difference()`\n",
"* Average Absolute Odds Difference : [documentation link](https://aif360.readthedocs.io/en/latest/modules/metrics.html#aif360.metrics.ClassificationMetric.average_abs_odds_difference)\n",
"\n",
"`average_abs_odds_difference()`\n",
"* Theil Index : [documentation link](https://aif360.readthedocs.io/en/latest/modules/metrics.html#aif360.metrics.ClassificationMetric.theil_index)\n",
"\n",
"`theil_index()`\n",
"\n",
"Then for the metrics that don't require prediction (the second class) we can use them with the following functions : \n",
"* Statistical Parity Difference : [documentation link](https://aif360.readthedocs.io/en/latest/modules/metrics.html#aif360.metrics.BinaryLabelDatasetMetric.statistical_parity_difference)\n",
"\n",
"`statistical_parity_difference()`\n",
"* Disparate Impact : [documentation link](https://aif360.readthedocs.io/en/latest/modules/metrics.html#aif360.metrics.ClassificationMetric.disparate_impact)\n",
"\n",
"`disparate_impact()`\n",
"\n",
"\n",
"Now I construct a `DataFrame` that will keep the values of the differents metrics I talked just above with a function. "
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"def get_fair_metrics_and_plot(data, model, plot=True, model_aif=False):\n",
" pred = model.predict(data).labels if model_aif else model.predict(data.features)\n",
" # fair_metrics function available in the metrics.py file\n",
" fair = fair_metrics(data, pred)\n",
"\n",
" if plot:\n",
" # plot_fair_metrics function available in the visualisations.py file\n",
" # The visualisation of this function is inspired by the dashboard on the demo of IBM aif360 \n",
" plot_fair_metrics(fair)\n",
" display(fair)\n",
" \n",
" return fair"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
},
{
"data": {
"text/markdown": [
"### Check bias metrics :"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"A model can be considered bias if just one of these five metrics show that this model is biased."
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"#### For the Sex attribute :"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"With default thresholds, bias against unprivileged group detected in **3** out of 5 metrics"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"#### For the Race attribute :"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"With default thresholds, bias against unprivileged group detected in **4** out of 5 metrics"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"fair = get_fair_metrics_and_plot(data_orig_test, rf_orig)"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [],
"source": [
"# This DataFrame is created to stock differents models and fair metrics that we produce in this notebook\n",
"algo_metrics = pd.DataFrame(columns=['model', 'fair_metrics', 'prediction', 'probs'])\n",
"\n",
"def add_to_df_algo_metrics(algo_metrics, model, fair_metrics, preds, probs, name):\n",
" return algo_metrics.append(pd.DataFrame(data=[[model, fair_metrics, preds, probs]], columns=['model', 'fair_metrics', 'prediction', 'probs'], index=[name]))"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [],
"source": [
"probs = rf.predict_proba(data_orig_test.features)\n",
"preds = rf.predict(data_orig_test.features)\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, rf, fair, preds, probs, 'Origin')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"As we can see, this first model is biased. The next step is to answer the question : How to fix it ?\n",
"\n",
"### 3.2 How to fix it ?\n",
"\n",
"AIF360 use 3 types of algorithms :\n",
"* Pre-processing algorithms : they are used before training the model\n",
"* In-processing algorithms : they are fair classifiers so it's during the training\n",
"* Post-processing algorithms : they are used after training the model\n",
"\n",
"\n",
"\n",
"This is the **fairness pipeline**. An example instantiation of this generic pipeline consists of loading data into a dataset object, transforming\n",
"it into a fairer dataset using a fair pre-processing algorithm, learning a classifier from this transformed dataset, and obtaining\n",
"predictions from this classifier. Metrics can be calculated on the original, transformed, and predicted datasets as well as between the\n",
"transformed and predicted datasets. Many other instantiations are also possible (more information on [the aif360 paper](https://arxiv.org/pdf/1810.01943.pdf)).\n",
"\n",
"#### 3.2.1 Pre-processing algorithms \n",
"\n",
"There are 4 pre-processing algorithms but for 3 of them **there is a problem** : it will work only if the dataset have one protected attribute but here we have 2 : *Sex* and *Race*. But let's have a quick intro for all of them and at the end I will use the one that work for my case.\n",
"\n",
"Because this notebook is an exploration of fairness into models I will only use 1 protected attribute : Sex, it will allow me to use almost all the algorithms (but for your information the *Reweighing* algo can works with all protected so that's why I will create a function that gives me the protected and unprotected attributes.\n",
"\n",
"You can find all the code information on [the documentation](https://aif360.readthedocs.io/en/latest/modules/preprocessing.html#)\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [],
"source": [
"def get_attributes(data, only_one=True):\n",
" unprivileged_groups = []\n",
" privileged_groups = []\n",
" \n",
" for attr in data.protected_attribute_names:\n",
" idx = data.protected_attribute_names.index(attr)\n",
" privileged_groups.append({attr:data.privileged_protected_attributes[idx]}) \n",
" unprivileged_groups.append({attr:data.unprivileged_protected_attributes[idx]}) \n",
" if only_one:\n",
" break\n",
"\n",
" return privileged_groups, unprivileged_groups"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*****\n",
"##### 3.2.1.1 Disparate impact remover\n",
"Source : [Feldman et al., 2015](https://dl.acm.org/citation.cfm?doid=2783258.2783311)\n",
"\n",
"Disparate impact remover is a preprocessing technique that edits feature values increase group fairness while preserving rank-ordering within groups.\n",
"If you want to see how it works you can take a look on [an example Notebook from the GitHub of AIF360](https://github.com/IBM/AIF360/blob/master/examples/demo_disparate_impact_remover.ipynb).\n",
"\n",
"Some code example :\n",
"```\n",
"from aif360.algorithms.preprocessing import DisparateImpactRemover\n",
"DIR = DisparateImpactRemover()\n",
"data_transf_train = DIR.fit_transform(data_orig_train)\n",
"```\n",
"*****\n",
"##### 3.2.1.2 Learning fair representations\n",
"Source : [Zemel et al., 2013](http://proceedings.mlr.press/v28/zemel13.html)\n",
"\n",
"Learning fair representations is a pre-processing technique that finds a latent representation which encodes the data well but obfuscates information about protected attributes.\n",
"You can find more information on [the demo notebook](https://github.com/IBM/AIF360/blob/master/examples/demo_lfr.ipynb)."
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"250 41993.58928923252\n",
"500 33883.92930598489\n",
"750 33625.13761517554\n",
"1000 33625.135392161435\n",
"1250 33611.92929107428\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"privileged_groups, unprivileged_groups = get_attributes(data_orig_train, only_one=True)\n",
"\n",
"LFR = LFR(unprivileged_groups=unprivileged_groups, privileged_groups=privileged_groups, k=1, verbose=0)\n",
"# LFR.fit(data_orig_train)\n",
"data_transf_train = LFR.fit_transform(data_orig_train)\n",
"\n",
"# Train and save the model\n",
"rf_transf = train_model(rf, data_transf_train)\n",
"\n",
"data_transf_test = LFR.transform(data_orig_test)\n",
"fair = get_fair_metrics_and_plot(data_transf_test, rf_transf, plot=False)\n",
"\n",
"# algo_metrics = add_to_df_algo_metrics(algo_metrics, rf_transf, fair, 'LFR')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*****\n",
"##### 3.2.1.3 Optimized preprocessing\n",
"Source : [Calmon et al., 2017](http://papers.nips.cc/paper/6988-optimized-pre-processing-for-discrimination-prevention)\n",
"\n",
"Optimized preprocessing is a preprocessing technique that learns a probabilistic transformation that edits the features and labels in the data with group fairness, individual distortion, and data fidelity constraints and objectives.\n",
"There is also [a demo notebook on the aif360 GitHub](https://github.com/IBM/AIF360/blob/master/examples/demo_optim_data_preproc.ipynb).\n",
"\n",
"*To be honest I tried to work with this one but it's more complicated : it uses options that you have to configure yourself and I don't really find how to choose it. Also it use an Optimizer and I didn't find how to build this class. (I didn't read the paper about this algorithm)*\n",
"*****\n",
"##### 3.2.1.4 Reweighing\n",
"Source : [Kamiran and Calders, 2012](https://link.springer.com/article/10.1007%2Fs10115-011-0463-8)\n",
"\n",
"Reweighing is a preprocessing technique that Weights the examples in each (group, label) combination differently to ensure fairness before classification."
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"privileged_groups, unprivileged_groups = get_attributes(data_orig_train, only_one=False)\n",
"\n",
"RW = Reweighing(unprivileged_groups=unprivileged_groups, privileged_groups=privileged_groups)\n",
"# RW.fit(data_orig_train)\n",
"data_transf_train = RW.fit_transform(data_orig_train)\n",
"\n",
"# Train and save the model\n",
"rf_transf = train_model(rf, data_transf_train)\n",
"\n",
"data_transf_test = RW.transform(data_orig_test)\n",
"fair = get_fair_metrics_and_plot(data_orig_test, rf_transf, plot=False)\n",
"probs = rf_transf.predict_proba(data_orig_test.features)\n",
"preds = rf_transf.predict(data_orig_test.features)\n",
"\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, rf_transf, fair, preds, probs, 'Reweighing')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 3.2.2 In-processing algorithms\n",
"\n",
"It exists 3 in-processing algorithms with aif360. But **I will only use 2** : *Adversarial Debiasing* wich is using TensorFlow and *Prejudice Remover Regularizer*. I'm not using the last one : *ART Classifier*, because I didn't found any documentation that show how it works (maybe I didn't search enough 🙄).\n",
"\n",
"You can find all the code information on [the documentation](https://aif360.readthedocs.io/en/latest/modules/inprocessing.html)\n",
"*****\n",
"##### 3.2.2.1 Adversarial Debiasing\n",
"Source : [Zhang et al., 2018](http://www.aies-conference.com/wp-content/papers/main/AIES_2018_paper_162.pdf)\n",
"\n",
"Adversarial debiasing is an in-processing technique that learns a classifier to maximize prediction accuracy and simultaneously reduce an adversary’s ability to determine the protected attribute from the predictions. This approach leads to a fair classifier as the predictions cannot carry any group discrimination information that the adversary can exploit."
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"epoch 0; iter: 0; batch classifier loss: 2.027651; batch adversarial loss: 0.696884\n",
"epoch 1; iter: 0; batch classifier loss: 0.535001; batch adversarial loss: 0.708483\n",
"epoch 2; iter: 0; batch classifier loss: 0.486652; batch adversarial loss: 0.671215\n",
"epoch 3; iter: 0; batch classifier loss: 0.452246; batch adversarial loss: 0.645171\n",
"epoch 4; iter: 0; batch classifier loss: 0.502617; batch adversarial loss: 0.609100\n",
"epoch 5; iter: 0; batch classifier loss: 0.423729; batch adversarial loss: 0.640453\n",
"epoch 6; iter: 0; batch classifier loss: 0.418263; batch adversarial loss: 0.621099\n",
"epoch 7; iter: 0; batch classifier loss: 0.382579; batch adversarial loss: 0.585946\n",
"epoch 8; iter: 0; batch classifier loss: 0.372982; batch adversarial loss: 0.592329\n",
"epoch 9; iter: 0; batch classifier loss: 0.341219; batch adversarial loss: 0.644888\n",
"epoch 10; iter: 0; batch classifier loss: 0.435262; batch adversarial loss: 0.581868\n",
"epoch 11; iter: 0; batch classifier loss: 0.435852; batch adversarial loss: 0.649803\n",
"epoch 12; iter: 0; batch classifier loss: 0.461696; batch adversarial loss: 0.640365\n",
"epoch 13; iter: 0; batch classifier loss: 0.493184; batch adversarial loss: 0.623321\n",
"epoch 14; iter: 0; batch classifier loss: 0.428212; batch adversarial loss: 0.655655\n",
"epoch 15; iter: 0; batch classifier loss: 0.349279; batch adversarial loss: 0.662475\n",
"epoch 16; iter: 0; batch classifier loss: 0.401383; batch adversarial loss: 0.642877\n",
"epoch 17; iter: 0; batch classifier loss: 0.375208; batch adversarial loss: 0.645407\n",
"epoch 18; iter: 0; batch classifier loss: 0.352355; batch adversarial loss: 0.653866\n",
"epoch 19; iter: 0; batch classifier loss: 0.381276; batch adversarial loss: 0.634348\n",
"epoch 20; iter: 0; batch classifier loss: 0.368791; batch adversarial loss: 0.568455\n",
"epoch 21; iter: 0; batch classifier loss: 0.362154; batch adversarial loss: 0.618681\n",
"epoch 22; iter: 0; batch classifier loss: 0.327374; batch adversarial loss: 0.627193\n",
"epoch 23; iter: 0; batch classifier loss: 0.417629; batch adversarial loss: 0.574505\n",
"epoch 24; iter: 0; batch classifier loss: 0.344165; batch adversarial loss: 0.621186\n",
"epoch 25; iter: 0; batch classifier loss: 0.335994; batch adversarial loss: 0.655801\n",
"epoch 26; iter: 0; batch classifier loss: 0.373382; batch adversarial loss: 0.586958\n",
"epoch 27; iter: 0; batch classifier loss: 0.376127; batch adversarial loss: 0.709685\n",
"epoch 28; iter: 0; batch classifier loss: 0.410763; batch adversarial loss: 0.635980\n",
"epoch 29; iter: 0; batch classifier loss: 0.472456; batch adversarial loss: 0.580156\n",
"epoch 30; iter: 0; batch classifier loss: 0.445524; batch adversarial loss: 0.585510\n",
"epoch 31; iter: 0; batch classifier loss: 0.466579; batch adversarial loss: 0.604531\n",
"epoch 32; iter: 0; batch classifier loss: 0.365391; batch adversarial loss: 0.645607\n",
"epoch 33; iter: 0; batch classifier loss: 0.358953; batch adversarial loss: 0.573522\n",
"epoch 34; iter: 0; batch classifier loss: 0.467547; batch adversarial loss: 0.612585\n",
"epoch 35; iter: 0; batch classifier loss: 0.467910; batch adversarial loss: 0.556397\n",
"epoch 36; iter: 0; batch classifier loss: 0.329569; batch adversarial loss: 0.585394\n",
"epoch 37; iter: 0; batch classifier loss: 0.429180; batch adversarial loss: 0.615465\n",
"epoch 38; iter: 0; batch classifier loss: 0.363523; batch adversarial loss: 0.616181\n",
"epoch 39; iter: 0; batch classifier loss: 0.309973; batch adversarial loss: 0.591368\n",
"epoch 40; iter: 0; batch classifier loss: 0.395164; batch adversarial loss: 0.624585\n",
"epoch 41; iter: 0; batch classifier loss: 0.405081; batch adversarial loss: 0.610189\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"privileged_groups, unprivileged_groups = get_attributes(data_orig_train, only_one=True)\n",
"\n",
"# sess.close()\n",
"# tf.reset_default_graph()\n",
"sess = tf.Session()\n",
"\n",
"debiased_model = AdversarialDebiasing(privileged_groups = privileged_groups,\n",
" unprivileged_groups = unprivileged_groups,\n",
" scope_name='debiased_classifier',\n",
" num_epochs=42,\n",
" debias=True,\n",
" sess=sess)\n",
"\n",
"debiased_model.fit(data_orig_train)\n",
"\n",
"fair = get_fair_metrics_and_plot(data_orig_test, debiased_model, plot=False, model_aif=True)\n",
"data_pred = debiased_model.predict(data_orig_test)\n",
"\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, debiased_model, fair, data_pred.labels, data_pred.scores, 'AdvDebiasing')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*****\n",
"##### 3.2.2.2 Prejudice Remover Regularizer\n",
"Source : [Kamishima et al., 2012](https://rd.springer.com/chapter/10.1007/978-3-642-33486-3_3)\n",
"\n",
"Prejudice remover is an in-processing technique that adds a discrimination-aware regularization term to the learning objective."
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"debiased_model = PrejudiceRemover(sensitive_attr=\"Sex\", eta = 25.0)\n",
"debiased_model.fit(data_orig_train)\n",
"\n",
"fair = get_fair_metrics_and_plot(data_orig_test, debiased_model, plot=False, model_aif=True)\n",
"data_pred = debiased_model.predict(data_orig_test)\n",
"\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, debiased_model, fair, data_pred.labels, data_pred.scores, 'PrejudiceRemover')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 3.2.3 Post-processing algorithms\n",
"\n",
"It exists 3 post-processing algorithms with aif360. \n",
"\n",
"You can find all the code information on [the documentation](https://aif360.readthedocs.io/en/latest/modules/postprocessing.html)\n",
"\n",
"I prepare the predict dataset for all the algorithms that require a dataset with predict scores and labels."
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [],
"source": [
"data_orig_test_pred = data_orig_test.copy(deepcopy=True)\n",
"\n",
"# Prediction with the original RandomForest model\n",
"scores = np.zeros_like(data_orig_test.labels)\n",
"scores = rf.predict_proba(data_orig_test.features)[:,1].reshape(-1,1)\n",
"data_orig_test_pred.scores = scores\n",
"\n",
"preds = np.zeros_like(data_orig_test.labels)\n",
"preds = rf.predict(data_orig_test.features).reshape(-1,1)\n",
"data_orig_test_pred.labels = preds\n",
"\n",
"def format_probs(probs1):\n",
" probs1 = np.array(probs1)\n",
" probs0 = np.array(1-probs1)\n",
" return np.concatenate((probs0, probs1), axis=1)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*****\n",
"##### 3.2.3.1 Calibrated Equality of Odds\n",
"Source : [Pleiss et al., 2017](https://papers.nips.cc/paper/7151-on-fairness-and-calibration)\n",
"\n",
"Calibrated equalized odds postprocessing is a post-processing technique that optimizes over calibrated classifier score outputs to find probabilities with which to change output labels with an equalized odds objective."
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"privileged_groups, unprivileged_groups = get_attributes(data_orig_train, only_one=True)\n",
"\n",
"cost_constraint = \"fnr\" # \"fnr\", \"fpr\", \"weighted\"\n",
"\n",
"CPP = CalibratedEqOddsPostprocessing(privileged_groups = privileged_groups,\n",
" unprivileged_groups = unprivileged_groups,\n",
" cost_constraint=cost_constraint,\n",
" seed=42)\n",
"\n",
"CPP = CPP.fit(data_orig_test, data_orig_test_pred)\n",
"data_transf_test_pred = CPP.predict(data_orig_test_pred)\n",
"\n",
"fair = fair_metrics(data_orig_test, data_orig_test_pred, pred_is_dataset=True)\n",
"\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, \n",
" CPP, \n",
" fair, \n",
" data_transf_test_pred.labels, \n",
" format_probs(data_transf_test_pred.scores), \n",
" 'CalibratedEqOdds')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*****\n",
"##### 3.2.3.2 Equality of Odds\n",
"Source : [Hardt et al., 2016](https://papers.nips.cc/paper/6374-equality-of-opportunity-in-supervised-learning)\n",
"\n",
"Equalized odds postprocessing is a post-processing technique that solves a linear program to find probabilities with which to change output labels to optimize equalized odds."
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"privileged_groups, unprivileged_groups = get_attributes(data_orig_train, only_one=True)\n",
"\n",
"EOPP = EqOddsPostprocessing(privileged_groups = privileged_groups,\n",
" unprivileged_groups = unprivileged_groups,\n",
" seed=42)\n",
"\n",
"EOPP = EOPP.fit(data_orig_test, data_orig_test_pred)\n",
"data_transf_test_pred = EOPP.predict(data_orig_test_pred)\n",
"\n",
"fair = fair_metrics(data_orig_test, data_transf_test_pred, pred_is_dataset=True)\n",
"\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, \n",
" EOPP, \n",
" fair, \n",
" data_transf_test_pred.labels, \n",
" format_probs(data_transf_test_pred.scores), \n",
" 'EqOdds')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*****\n",
"##### 3.2.3.3 Reject Option Classification\n",
"Source : [Kamishima et al., 2012](https://ieeexplore.ieee.org/document/6413831/)\n",
"\n",
"Reject option classification is a postprocessing technique that gives favorable outcomes to unpriviliged groups and unfavorable outcomes to priviliged groups in a confidence band around the decision boundary with the highest uncertainty."
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n",
"WARNING:root:There are some instances in the dataset which are not designated as either privileged or unprivileged. Are you sure this is right?\n"
]
}
],
"source": [
"privileged_groups, unprivileged_groups = get_attributes(data_orig_train, only_one=True)\n",
"\n",
"ROC = RejectOptionClassification(privileged_groups = privileged_groups,\n",
" unprivileged_groups = unprivileged_groups)\n",
"\n",
"ROC = ROC.fit(data_orig_test, data_orig_test_pred)\n",
"data_transf_test_pred = ROC.predict(data_orig_test_pred)\n",
"\n",
"fair = fair_metrics(data_orig_test, data_transf_test_pred, pred_is_dataset=True)\n",
"\n",
"algo_metrics = add_to_df_algo_metrics(algo_metrics, \n",
" ROC, \n",
" fair, \n",
" data_transf_test_pred.labels, \n",
" format_probs(data_transf_test_pred.scores), \n",
" 'RejectOption')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 3.3 Compare all the algorithms for one attribute"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [],
"source": [
"def compare_fair_metrics(algo_metrics, attr='Sex'):\n",
" \n",
" df_metrics = pd.DataFrame(columns=algo_metrics.loc['Origin','fair_metrics'].columns.values)\n",
" for fair in algo_metrics.loc[:,'fair_metrics']:\n",
" df_metrics = df_metrics.append(fair.loc[attr], ignore_index=True)\n",
"\n",
" df_metrics.index = algo_metrics.index.values\n",
" \n",
" display(df_metrics)\n",
" plot_fair_metrics_plotly(df_metrics)\n",
" score = score_fair_metrics(df_metrics)\n",
" plot_score_fair_metrics(score)"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"plot_compare_model_performance(algo_metrics, data_orig_test)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can see that the original model still has a great accuracy and F1 Score. I remind you that here I didn't optimize the models to be excellent I used default configurations.\n",
"\n",
"Now you have to ask yourself the question, **what is more important : performance or fair ?**\n",
"\n",
"## 4. Interpretable model\n",
"### 4.1 Feature importance\n",
"\n",
"The question of **model interpretation is becoming really important** and the most common way to explain a model is to use **feature importance**. It's usefull to get an insight of what the model is using in feature to predict something but it has limits. First of all let's see how it works.\n",
"\n",
"Feature importance changes according to the model that we're using. It's important to know that **it exists differents ways to measure feature importance** according to the model so I will introduce some measure for 3 types of models : Linear Regression, Decision Tree and Tree models.\n",
"\n",
"#### 4.1.1 Linear Regression\n",
"\n",
"For the Linear Regression if we have standardized the features the coefficient represents the influence of the feature on the model. So for a Linear model with a formula : \n",
"\n",
"$$h_\\theta(X) = \\theta_0x_0 + \\theta_1x_1 + ... + + \\theta_nx_n = \\theta^TX$$\n",
"\n",
"We can just take all the theta with the vector $\\theta$ and it's our importance for each feature.\n",
"\n",
"#### 4.1.2 Decision Tree\n",
"\n",
"For the Decision Tree, feature importance is not really important because this model is human interpretable. For a prediction you just have to follow the leaves.\n",
"\n",
"#### 4.1.3 Tree models (e.g Random Forest)\n",
"\n",
"I think this type of models (Random Forest, XGBoost, LGBM, ...) is well known in the industry today so the feature importance of those models are logically the most used for classification. So the most used measure of feature importance for Tree models is **Gini Importance or Mean Decrease in Impurity (MDI)**. \n",
"\n",
"\n",
"\n",
"**MDI** counts the times a feature is used to split a node, weighted by the number of samples it splits.\n",
"\n",
"We can calculate it with this formula :\n",
"\n",
"$$Imp(X_j)=\\frac{1}{M}\\sum_{m=1}^{M}\\sum_{t\\in \\varphi_m} (j_t=j)[p(t)\\Delta i(t)]$$\n",
"\n",
"Here we calculate the importance of variable $X_j$ for an ensemble of $M$ trees. $j_t$ denotes the variable used at node $t$, $p(t) = \\frac{N_t}{N}$ is the proportion of rows used in that node (if it's the first node it will be 1) and $\\Delta i(t)$ is the impurity reduction at node $t$ : we can use Gini index or Entropy to measure impurity.\n",
"\n",
"I find out this formula from this course : [Understanding variable importances in forests of randomized trees](https://fr.slideshare.net/glouppe/understanding-variable-importances-in-forests-of-randomized-trees)\n",
"\n",
"Also if you want more informations about others Tree models measure of feature importance you can take a look at this publication : [Feature Importance Measures for Tree Models — Part I](https://medium.com/the-artificial-impostor/feature-importance-measures-for-tree-models-part-i-47f187c1a2c3)\n",
"\n",
"Now let's see what is the feature importance of our Random Forest model."
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [],
"source": [
"rf.feature_importances_\n",
"feature_imp = pd.Series(rf.feature_importances_, index=data_orig_train.feature_names)\n",
"feature_imp.sort_values(ascending=False, inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"plt.figure(figsize=(10,7))\n",
"ax = sns.barplot(y=feature_imp.index.values, x=feature_imp,orient='h')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We see that *Final Weight*, *Age* and *Capital Gain* are the features that are the most importance for this model if we use feature importance from sklearn RandomForestClassifier. But now we arrive at **the limits of feature importance**, if I want to know wich feature are important on **a specific prediction** I can't know that I just can suppose that it follows the feature importance measure but I can not be sure.\n",
"\n",
"### 4.2 LIME\n",
"Source : [LIME GitHub repository](https://github.com/marcotcr/lime)\n",
"\n",
"So with the objective to find out wich features are important for a specific prediction appear **LIME** : *Local interpretable model-agnostic explanations* it means that LIME can explain any models in Machine Learning. We can think it like the model is a black box that just have the input and output to find out what feature is important.\n",
"\n",
"\n",
"\n",
"To understand how it works we have to define it is important to distinguish between features and interpretable data representations. As mentioned before, interpretable explanations need to use a representation that is understandable to humans, regardless of the actual features used by the model. For example, a possible interpretable representation for text classification is a binary vector indicating the presence or absence of a word.\n",
"\n",
"Now Let $f$ be the original prediction model to be explained and $g$ the explanation model such as a linear regression model. Here, we focus on local methods designed to explain a prediction $f(x)$ based on a single input $x$. Explanation models often use simplified inputs $x$ that map to the original inputs through a mapping function $x = h_x(x')$. Local methods try to ensure $g(z') ≈ f(h_x(z'))$ whenever $z' ≈ x'$ so it's a perturbed sample (Note that $hx(x') = x$ even though x may contain less information than $x$ because $h_x$ is specific to the current input $x'$).\n",
"\n",
"The explanation model that is a linear function of binary variables that follows this formula :\n",
"$$g(z') = \\phi_0 + \\sum_{i=1}^{M}\\phi_i z_i' $$\n",
"\n",
"where $z' \\in \\{0, 1\\}^M$, $M$ is the number of simplified input features, and $\\phi_i \\in R$\n",
"\n",
"To find $\\phi$, LIME minimizes the following **objective function** : \n",
"$$ \\xi = \\underset{g \\in G}{arg min} \\space L(f,g,\\pi_{x'}) + \\Omega(g) $$\n",
"\n",
"Faithfulness of the explanation model $g(z')$ to the original model $f(h_x(z'))$ is enforced through the loss $L$ over a set of samples in the simplified input space weighted by the local kernel $\\pi_{x'}$ (which calculate the distance between the perturbed sample and the original point). $\\Omega$ penalizes the complexity of $g$. Since in LIME $g$ follows the linear equation above and $L$ is a squared loss, this equation can be solved using penalized linear regression.\n",
"\n",
"So at the end we can obtain a explanation model like this : \n",
"\n",
"\n",
"\n",
"Now let's use LIME on our dataset and original model (Random Forest)."
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [],
"source": [
"# Convert AIF360 StandardDataset into a pandas DataFrame\n",
"test_interp = pd.DataFrame(data=data_orig_test.features,\n",
" columns=data_orig_test.feature_names)\n",
"\n",
"categorical_features = [\"Workclass\", \"Education\", \"Marital Status\",\"Occupation\", \"Relationship\", \"Race\", \"Sex\", \"Country\"]\n",
"categorical_features_idx = [np.where(test_interp.columns.values == col)[0][0] for col in categorical_features]\n",
"categorical_names_LIME = {}\n",
"\n",
"for feature, idx in zip(categorical_features, categorical_features_idx):\n",
" categorical_names_LIME[idx] = categorical_names[feature]\n",
" \n",
"# Reverse the MinMaxScaler to get the original values back.\n",
"for feat in test_interp.columns.values:\n",
" if feat in numerical_features:\n",
" test_interp[feat] = encoders[feat].inverse_transform(np.array(test_interp[feat]).reshape(-1, 1))"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [],
"source": [
"# Initiate the LIME explainer\n",
"explainer = lime.lime_tabular.LimeTabularExplainer(test_interp.values,\n",
" feature_names=test_interp.columns.values,\n",
" class_names=['Income <= 50K', 'Income > 50K'],\n",
" categorical_features=categorical_features_idx, \n",
" categorical_names=categorical_names_LIME)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We have the LIME explainer, we can see what results it gives us for 2 rows in the test set. "
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [
{
"data": {
"text/markdown": [
"#### First row"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"
"
],
"text/plain": [
" Workclass Relationship Final Weight Sex Age Race Country \\\n",
"5 4.0 0.0 85995.0 1.0 60.0 1.0 35.0 \n",
"\n",
" Occupation Hours per week Education Marital Status Capital Loss \\\n",
"5 1.0 50.0 12.0 2.0 0.0 \n",
"\n",
" Capital Gain \n",
"5 0.0 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
" \n",
" \n",
" \n",
" \n",
" "
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"linkText": "Export to plot.ly",
"plotlyServerURL": "https://plot.ly",
"showLink": true
},
"data": [
{
"marker": {
"color": [
"#e67e22",
"#e67e22",
"#3498db",
"#e67e22",
"#e67e22",
"#e67e22",
"#e67e22",
"#e67e22",
"#3498db",
"#3498db"
]
},
"opacity": 0.9,
"orientation": "h",
"type": "bar",
"uid": "d64a9fa6-1f89-48b6-b2e5-d606200fc4ad",
"x": [
0.003356037414791118,
0.0038368197342269874,
0.014209734773849491,
0.028290594473379265,
0.048314989935345234,
0.05405228337597185,
0.08239114928233258,
0.09149113231282731,
0.2251668049733779,
0.26245973456231636
],
"y": [
"Final Weight <= 118056.00",
"Workclass= Private",
"Country= South",
"Race= Asian-Pac-Islander",
"Sex= Male",
"Education= Masters",
"Relationship= Husband",
"Marital Status= Married-civ-spouse",
"Capital Loss <= 0.00",
"Capital Gain <= 0.00"
]
}
],
"layout": {
"autosize": true,
"barmode": "group",
"title": "Importance for the current prediction",
"xaxis": {
"autorange": true,
"range": [
0,
0.2762734048024383
],
"type": "linear"
},
"yaxis": {
"automargin": true,
"autorange": true,
"range": [
-0.5,
9.5
],
"type": "category"
}
}
},
"image/png": "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",
"text/html": [
""
],
"text/vnd.plotly.v1+html": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"row_1 = pd.DataFrame(data=[test_interp.iloc[2,:]], columns=test_interp.columns.values)\n",
"row_2 = pd.DataFrame(data=[test_interp.iloc[5,:]], columns=test_interp.columns.values)\n",
" \n",
"exp_1 = explainer.explain_instance(row_1.values[0], rf.predict_proba)\n",
"exp_2 = explainer.explain_instance(row_2.values[0], rf.predict_proba)\n",
"\n",
"display(Markdown('#### First row'))\n",
"display(row_1)\n",
"# Plot function from LIME\n",
"exp_1.show_in_notebook(show_table=True, show_all=False)\n",
"# Custom plot function\n",
"plot_lime_importance(exp_1)\n",
"\n",
"display(Markdown('#### Second row'))\n",
"display(row_2)\n",
"# Plot function from LIME\n",
"exp_2.show_in_notebook(show_table=True, show_all=False)\n",
"# Custom plot function\n",
"plot_lime_importance(exp_2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"As we can see LIME is performant to understand the importance of features for a specific prediction and if we want to get an understanding of the model can construct some aggregate values like a mean of all importance. \n",
"\n",
"I decided to construct to global metrics where $f$ is a feature and $M$ the number of elements : \n",
"* Mean of the importance values : $\\mu^f_{importance} = \\frac{1}{M} \\sum_{i=1}^{M} importance^f_i$\n",
"* Mean of the absolute of importance values : $\\mu^f_{|importance|} = \\frac{1}{M} \\sum_{i=1}^{M} |importance^f_i|$\n",
"\n",
"But to get this I have to regroup all the feature generate by LIME into the original feature. For example : it's possible to have a feature *Age > 10* and for another row *5 < Age < 24* so it needs to be just *Age* at the end."
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"start : 0.0000s\n",
"end : 177.7591s\n"
]
}
],
"source": [
"def compute_mean(dataset):\n",
" return dataset.mean()\n",
"\n",
"def compute_mean_abs(dataset):\n",
" return dataset.abs().mean()\n",
"\n",
"t0 = time()\n",
"print('start : %0.4fs'%((time() - t0)))\n",
" \n",
"importance = pd.DataFrame()\n",
"\n",
"# I will just take a sample of set because it's a long operation on all the dataset.\n",
"test_lime = test_interp.iloc[0:1000,:]\n",
"\n",
"for row in test_lime.values:\n",
" exp = explainer.explain_instance(row, rf.predict_proba)\n",
" feats = [x[0] for x in exp.as_list()]\n",
" val = [x[1] for x in exp.as_list()]\n",
" \n",
" for i in range(0,len(feats)):\n",
" for feat in test_lime.columns.values:\n",
" if(feat in feats[i]):\n",
" feats[i] = feat\n",
" \n",
" exp_df = pd.DataFrame(data=[val], columns=feats)\n",
" importance = importance.append(exp_df)\n",
"\n",
"print('end : %0.4fs'%((time() - t0)))"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"linkText": "Export to plot.ly",
"plotlyServerURL": "https://plot.ly",
"showLink": true
},
"data": [
{
"marker": {
"color": [
"#e67e22",
"#e67e22",
"#3498db",
"#e67e22",
"#e67e22",
"#e67e22",
"#e67e22",
"#e67e22",
"#e67e22",
"#3498db",
"#e67e22",
"#3498db",
"#3498db"
]
},
"opacity": 0.9,
"orientation": "h",
"type": "bar",
"uid": "3687dafa-a475-4ed2-bfb6-e0a4d9f39696",
"x": [
0.0002424512753894929,
0.00030935689778165284,
0.0003703074836170467,
0.0010632019903567202,
0.0016937580213360537,
0.0053445691370466435,
0.005523715954067418,
0.012100365200672942,
0.012289515831862429,
0.016046664366045384,
0.020885079175630105,
0.20345541057558048,
0.2365316842171381
],
"y": [
"Final Weight",
"Age",
"Hours per week",
"Occupation",
"Workclass",
"Country",
"Education",
"Relationship",
"Marital Status",
"Race",
"Sex",
"Capital Loss",
"Capital Gain"
]
}
],
"layout": {
"autosize": true,
"barmode": "group",
"title": "Mean importance of feature",
"xaxis": {
"autorange": true,
"range": [
0,
0.24898072022856643
],
"type": "linear"
},
"yaxis": {
"automargin": true,
"autorange": true,
"range": [
-0.5,
12.5
],
"type": "category"
}
}
},
"image/png": "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",
"text/html": [
""
],
"text/vnd.plotly.v1+html": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"linkText": "Export to plot.ly",
"plotlyServerURL": "https://plot.ly",
"showLink": true
},
"data": [
{
"marker": {
"color": "#3498db"
},
"opacity": 0.9,
"orientation": "h",
"type": "bar",
"uid": "1793ca43-4f3f-4364-bd45-bbf3a1511e8a",
"x": [
0.004535166072505506,
0.004714637871897863,
0.004942716618682299,
0.009119621034688042,
0.009520160784101271,
0.021507061237945983,
0.033832576566767934,
0.04127400662245095,
0.0500216880203275,
0.05681260578155159,
0.07279405915379866,
0.2256207360567589,
0.2836610196583068
],
"y": [
"Age",
"Final Weight",
"Hours per week",
"Country",
"Workclass",
"Race",
"Occupation",
"Education",
"Sex",
"Relationship",
"Marital Status",
"Capital Loss",
"Capital Gain"
]
}
],
"layout": {
"autosize": true,
"barmode": "group",
"title": "Mean absolute importance of feature",
"xaxis": {
"autorange": true,
"range": [
0,
0.298590547008744
],
"type": "linear"
},
"yaxis": {
"automargin": true,
"autorange": true,
"range": [
-0.5,
12.5
],
"type": "category"
}
}
},
"image/png": "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",
"text/html": [
""
],
"text/vnd.plotly.v1+html": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"mean_imp = compute_mean(importance)\n",
"# Stock importance to compare it with other importance after\n",
"lime_importance = compute_mean_abs(importance)\n",
"plot_global_importance_lime(mean_imp, 'Mean importance of feature')\n",
"plot_global_importance_lime(lime_importance, 'Mean absolute importance of feature', True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can see that taking the mean is not relevant because we can loose some information so I prefer to use the second one with the absolute value.\n",
"\n",
"Now you see how LIME works, but also it don't offer us the possibility to have a global value so it's not the best thing. But just one year after LIME another library appeared : **SHAP**, they proposed an alternative kernel from the LIME method and claims that only its approach can satisfy three key axioms of interpretability. \n",
"\n",
"### 4.3 SHAP\n",
"Source : [SHAP GitHub repository](https://github.com/slundberg/shap)\n",
"\n",
"As I said before SHAP claims that it satisfy three axioms of interpretability. But, before going into the theory of SHAP let's dive into the **game theory**.\n",
"\n",
"Let's ask ourselves a question : *in a coalition that is composed of multiple players with differing skill sets, that results in some collective payoff, what is the fairest way to divide up that payoff among the players ?*. With this question we will find out the Shapley values. \n",
"\n",
"#### 4.3.1 Shapley Values\n",
"What is this value ? Before going into theory I need to use an example to have a better understanding about this value. \n",
"\n",
"Let's take 2 persons : Bob and Alice. They are going to a restaurant but at the end how will they split the bill ? To get the answer in a first place we need to know how much they spend if they were alone and together. Let's say that Bob alone would spend 50€ , Alice would spend 30€ and together 70€. So we can have the following information :\n",
"$$ v(c) = \n",
"\\begin{cases}\n",
"50, \\ \\text{if} \\ c=\\{\\text{Bob}\\} \\\\ \n",
"30, \\ \\text{if} \\ c=\\{\\text{Alice}\\} \\\\ \n",
"70, \\ \\text{if} \\ c=\\{\\text{Bob & Alice}\\}\n",
"\\end{cases} $$\n",
"\n",
"Now we have to know what they pay each depending on the order of arrive, for example if Alice arrives first she paid 30€ and then Bob paid the difference between what they pay together and what Alice paid so 40€. We denote $\\pi$ the order of arrival and $\\delta_{\\pi}^G$ what they pay each depending on the order of arrival. So now we obtain this :\n",
"\n",
"$$\n",
"\\begin{array}{c c}\n",
"\\pi & \\delta_{\\pi}^G \\\\\n",
"\\hline \n",
"(\\text{Bob}, \\text{Alice}) & (\\text{Bob}:50,\\text{Alice}:20) \\\\\n",
"(\\text{Alice}, \\text{Bob}) & (\\text{Bob}:40,\\text{Alice}:30) \\\\\n",
"\\end{array}\n",
"$$\n",
"\n",
"And now to obtain the Shapley values $\\phi$ we just have to get the mean on each person from $\\delta_{\\pi}^G$ :\n",
"\n",
"$$ \\phi = (\\text{Bob}:45,\\text{Alice}:25) $$\n",
"\n",
"So the concept of this value is not that complicated and now here is the formula that get us the values : Here $|F|$ is the size of the full coalition. $S$ represents any subset of the coalition that doesn’t include player $i$, and $|S|$ is the size of that subset. The bit at the end is just \"how much bigger is the payoff when we add player $i$ to this particular subset $S$\"\n",
"\n",
"$$ \\phi_i = \\sum_{S\\subseteq F\\setminus \\{i\\}} \\frac{|S|!(|F|-|S|-1)!}{|F|!} [f_{S\\cup \\{i\\}}(x_{S\\cup \\{i\\}}) - f_S(x_S)]$$\n",
"\n",
"So this values satisfy the 3 following axioms of interpretability :\n",
"* *Dummy Player* : If a player never adds any marginal value, their payoff portion should be 0.\n",
"* *Substitutability* : If two players always add the same marginal value to any subset to which they’re added, their payoff portion should be the same.\n",
"* *Additivity* : If a game is composed of two subgames, you should be able to add the payoffs calculated on the subgames, and that should match the payoffs calculated for the full game.\n",
"\n",
"#### 4.3.2 How SHAP works\n",
"\n",
"SHAP which means *SHapley Additive exPlanations*, uses both **game theory** (with Shapley value) and **local interpretation** (with LIME). But how it do that ? \n",
"\n",
"They used the concept of simplifed feature with vector includes in {0,1} and they defined the 3 axioms but in mathematicals terms so *Dummy player* become *Missingness*, *Substitutability* become *Local accuracy* and *Additivity* become *Consistency* (more informations about this mathematicals formulas in [the SHAP paper](http://papers.nips.cc/paper/7062-a-unified-approach-to-interpreting-model-predictions.pdf)).\n",
"\n",
"And with that we obtain the following formula : Here $f$ is the original model for some specific inputs $x$, and $x'$ is the simplified input (which corresponds to the original input $x$), $M$ is the full number of features and $z'$ is a sample of $x'$.\n",
"\n",
"$$ \\phi_i(f,x) = \\sum_{z'\\subseteq x'} \\frac{|z'|!(M-|z'|-1)!}{M!} [f_x(z') - f_x(z'\\setminus i)]$$\n",
"\n",
"So we see that is the **same logic than the Shapley value** but here instead of **players** we have **features**. \n",
"\n",
"The following text is from the SHAP paper, it explains the conection between the formula of LIME and the one juste above : \n",
"> At first glance, the regression formulation of LIME in Equation 2 seems\n",
"very different from the classical Shapley value formulation of Equation 8. However, since linear\n",
"LIME is an additive feature attribution method, we know the Shapley values are the only possible\n",
"solution to Equation 2 that satisfies Properties 1-3 – local accuracy, missingness and consistency. A\n",
"natural question to pose is whether the solution to Equation 2 recovers these values. The answer\n",
"depends on the choice of loss function $L$, weighting kernel $\\pi_{x'}$ and regularization term $\\Omega$. \n",
"\n",
"So to satisfy the 3 properties there is juste one combination of parameters that need to be follow :\n",
"\n",
"$\n",
"\\Omega(g) = 0, \\\\\n",
"\\pi_{x'}(z') = \\frac{(M-1)}{(M \\ choose \\ |z'|)|z'|(M-|z'|)}, \\\\\n",
"L(f,x,\\pi_{x'}) = \\sum_{z'\\in Z}[f(h_x(z')) - g(z')]^2 \\pi_{x'}(z')\n",
"$\n",
"\n",
"And this is how it works theoretically now let's practice with the library in Python.\n",
"*****\n",
"\n",
"#### 4.3.3 SHAP in Python\n",
"First let's generate the explainer with the `RandomForestClassifier` from sklearn."
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"start : 0.0000s\n",
"explainer end : 0.0858s\n",
"shap values end : 417.6281s\n"
]
}
],
"source": [
"t0 = time()\n",
"print('start : %0.4fs'%((time() - t0)))\n",
"\n",
"# Create object that can calculate shap values\n",
"explainer = shap.TreeExplainer(rf)\n",
"print('explainer end : %0.4fs'%((time() - t0)))\n",
"\n",
"# Calculate Shap values\n",
"test_shap = test_interp.copy()\n",
"# shap_values = explainer.shap_values(test)\n",
"shap_values = explainer.shap_values(test_shap)\n",
"print('shap values end : %0.4fs'%((time() - t0)))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now we have our shap values for each possible class (here : Income <= 50K or Income > 50K) and for each features / rows. SHAP provides us some plot function, especially one that plot shap values for a prediction : `shap.force_plot()`. It shows us the evolution between the *base value* $E[f(z)]$, which is the probability of a prediction if do not know any feature, and the current prediction $f(x)$\n",
"\n",
"\n",
"\n",
"Now let's see how it looks like on a random row of our test set."
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
"
],
"text/plain": [
" Workclass Relationship Final Weight Sex Age Race Country \\\n",
"0 1.0 0.0 39606.0 1.0 46.0 4.0 9.0 \n",
"\n",
" Occupation Hours per week Education Marital Status Capital Loss \\\n",
"0 4.0 40.0 12.0 2.0 0.0 \n",
"\n",
" Capital Gain \n",
"0 0.0 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"Probability of having an Income <= 50K : **0.50**"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"Probability of having an Income > 50K : **0.50**"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/markdown": [
"#### Explaination based on the 0 label (Income <= 50K)"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
"
\n",
"
\n",
" Visualization omitted, Javascript library not loaded! \n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" Visualization omitted, Javascript library not loaded! \n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"def plot_shap_explain(data, idx, model, explainer, shap_values, categorical_features=None, encoders=None):\n",
" row = data.iloc[idx,:] \n",
" display(pd.DataFrame([row.values], columns=row.index))\n",
" proba = model.predict_proba([row])[0]\n",
" display(Markdown(\"Probability of having an Income <= 50K : **%0.2f**\"%proba[0]))\n",
" display(Markdown(\"Probability of having an Income > 50K : **%0.2f**\"%proba[1]))\n",
" \n",
" if categorical_features != None:\n",
" for feature in categorical_features:\n",
" row[feature] = encoders[feature].inverse_transform([int(row[feature])])[0]\n",
" \n",
" display(Markdown(\"#### Explaination based on the 0 label (Income <= 50K)\"))\n",
" display(shap.force_plot(explainer.expected_value[0], shap_values[0][idx,:], row))\n",
" display(Markdown(\"#### Explaination based on the 1 label (Income > 50K)\"))\n",
" display(shap.force_plot(explainer.expected_value[1], shap_values[1][idx,:], row))\n",
"\n",
"# Take a random index\n",
"random_index = np.random.randint(0, len(test_shap))\n",
"\n",
"shap.initjs()\n",
"plot_shap_explain(test_shap, random_index, rf, explainer, shap_values, categorical_features, encoders)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now like feature importance and the custom metrics from LIME we need to see what is the global measure on the model. So SHAP gives us a plot function `shap.summary_plot()` which gives us 2 possible plot : bar plot or dot plot.\n",
"\n",
"The *bar* plot calculate the average absolute SHAP value and display a bar chart like the one for feature importance. The *dot* plot sorts features by the sum of SHAP value magnitudes over all samples, and uses SHAP values to show the distribution of the impacts each feature has on the model output. The color represents the feature value (red high, blue low)."
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"shap.summary_plot(shap_values[1], test_shap, test_shap.columns.values, plot_type='bar')\n",
"shap.summary_plot(shap_values[1], test_shap, test_shap.columns.values, plot_type='dot')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here the seconde plot reveals that, for example, More *Capital Gain* is high, higher is the chance that the model predicted an Income > 50K.\n",
"\n",
"Before concluding this part I will just keep the importance to compare it after."
]
},
{
"cell_type": "code",
"execution_count": 46,
"metadata": {},
"outputs": [],
"source": [
"shap_df = pd.DataFrame(data=shap_values[1], columns=test_shap.columns)\n",
"shap_importance = compute_mean_abs(shap_df)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now you have a quick overview of the SHAP library but I just used the `TreeExplainer` and there are other explainers like explainer for neural networks so if you are curious don't hesitate to look on the GitHub repository or the SHAP paper.\n",
"\n",
"### 4.4 Compare importance\n",
"\n",
"So to conclude this part let's see what each measure of importance say and which one we can use to explain our model."
]
},
{
"cell_type": "code",
"execution_count": 54,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"This is the format of your plot grid:\n",
"[ (1,1) x1,y1 ] [ (1,2) x2,y2 ] [ (1,3) x3,y3 ]\n",
"\n"
]
},
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"linkText": "Export to plot.ly",
"plotlyServerURL": "https://plot.ly",
"showLink": true
},
"data": [
{
"marker": {
"color": "#2980b9"
},
"opacity": 0.9,
"orientation": "h",
"type": "bar",
"uid": "916c7795-8c0e-45cd-9a49-d2bfc651c198",
"x": [
0.010387181763170106,
0.014923695507587945,
0.01792372311112318,
0.03847954495073287,
0.041004988759448094,
0.07903830672930416,
0.08032501964088075,
0.08572656401780676,
0.0872637475428259,
0.09045995066866405,
0.12608814205531518,
0.16046460842702504,
0.16791452682611602
],
"xaxis": "x",
"y": [
"Sex",
"Race",
"Country",
"Capital Loss",
"Workclass",
"Relationship",
"Occupation",
"Hours per week",
"Education",
"Marital Status",
"Capital Gain",
"Age",
"Final Weight"
],
"yaxis": "y"
},
{
"marker": {
"color": "#2980b9"
},
"opacity": 0.9,
"orientation": "h",
"type": "bar",
"uid": "6fbfa89e-b141-439c-a097-0f39c99de76b",
"x": [
0.004535166072505506,
0.004714637871897863,
0.004942716618682299,
0.009119621034688042,
0.009520160784101271,
0.021507061237945983,
0.033832576566767934,
0.04127400662245095,
0.0500216880203275,
0.05681260578155159,
0.07279405915379866,
0.2256207360567589,
0.2836610196583068
],
"xaxis": "x2",
"y": [
"Age",
"Final Weight",
"Hours per week",
"Country",
"Workclass",
"Race",
"Occupation",
"Education",
"Sex",
"Relationship",
"Marital Status",
"Capital Loss",
"Capital Gain"
],
"yaxis": "y2"
},
{
"marker": {
"color": "#2980b9"
},
"opacity": 0.9,
"orientation": "h",
"type": "bar",
"uid": "9bf3d97d-30b0-4762-bec1-db7376393edc",
"x": [
0.002076315020075182,
0.0026299528828703997,
0.006674271902806345,
0.007724728471507596,
0.02294418573287947,
0.032125532059911005,
0.03384900585553669,
0.03654604805851837,
0.04169007688267614,
0.05048066563802174,
0.06593217010100233,
0.06887178960112991,
0.07040236755919528
],
"xaxis": "x3",
"y": [
"Country",
"Race",
"Workclass",
"Sex",
"Capital Loss",
"Age",
"Education",
"Occupation",
"Final Weight",
"Relationship",
"Capital Gain",
"Marital Status",
"Hours per week"
],
"yaxis": "y3"
}
],
"layout": {
"annotations": [
{
"font": {
"size": 16
},
"showarrow": false,
"text": "Feature importance",
"x": 0.14444444444444446,
"xanchor": "center",
"xref": "paper",
"y": 1,
"yanchor": "bottom",
"yref": "paper"
},
{
"font": {
"size": 16
},
"showarrow": false,
"text": "LIME",
"x": 0.5,
"xanchor": "center",
"xref": "paper",
"y": 1,
"yanchor": "bottom",
"yref": "paper"
},
{
"font": {
"size": 16
},
"showarrow": false,
"text": "SHAP",
"x": 0.8555555555555556,
"xanchor": "center",
"xref": "paper",
"y": 1,
"yanchor": "bottom",
"yref": "paper"
}
],
"autosize": true,
"barmode": "group",
"showlegend": false,
"title": "Comparison of feature importance, LIME importance and SHAP importance",
"xaxis": {
"anchor": "y",
"autorange": true,
"domain": [
0,
0.2888888888888889
],
"range": [
0,
0.17675213350117475
],
"type": "linear"
},
"xaxis2": {
"anchor": "y2",
"autorange": true,
"domain": [
0.35555555555555557,
0.6444444444444445
],
"range": [
0,
0.298590547008744
],
"type": "linear"
},
"xaxis3": {
"anchor": "y3",
"autorange": true,
"domain": [
0.7111111111111111,
1
],
"range": [
0,
0.07410775532546872
],
"type": "linear"
},
"yaxis": {
"anchor": "x",
"automargin": true,
"autorange": true,
"domain": [
0,
1
],
"range": [
-0.5,
12.5
],
"type": "category"
},
"yaxis2": {
"anchor": "x2",
"autorange": true,
"domain": [
0,
1
],
"range": [
-0.5,
12.5
],
"type": "category"
},
"yaxis3": {
"anchor": "x3",
"autorange": true,
"domain": [
0,
1
],
"range": [
-0.5,
12.5
],
"type": "category"
}
}
},
"image/png": "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",
"text/html": [
""
],
"text/vnd.plotly.v1+html": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"def get_trace(series, color='#2980b9'):\n",
" series.sort_values(ascending=True, inplace=True)\n",
" x, y = series.values, series.index\n",
" trace = go.Bar(x=x, y=y, marker=dict(color=color), opacity=0.9, orientation='h')\n",
" return trace\n",
"\n",
"trace1 = get_trace(feature_imp)\n",
"trace2 = get_trace(lime_importance)\n",
"trace3 = get_trace(shap_importance)\n",
"\n",
"fig = tools.make_subplots(rows=1, cols=3, subplot_titles=('Feature importance', 'LIME','SHAP'))\n",
"\n",
"fig.append_trace(trace1, 1, 1)\n",
"fig.append_trace(trace2, 1, 2)\n",
"fig.append_trace(trace3, 1, 3)\n",
"fig['layout'].update(title='Comparison of feature importance, LIME importance and SHAP importance',\n",
" showlegend=False,\n",
" yaxis=go.layout.YAxis(automargin=True),\n",
" barmode='group')\n",
"py.iplot(fig)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We see that they are different results so how to decide which one to trust. The answer is : **SHAP** because it's not just the mathematics that are behind this method also they study the importance found with human intuition and here is their results :\n",
"\n",
"\n",
"\n",
"This image shows us the human intuition correspond to what SHAP determine.\n",
"\n",
"Thanks for reading ! 👍"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
"
],
"text/plain": [
"