Fastshap: A fast, approximate shap kernel

fastshap: A fast, approximate shap kernel

fastshap was designed to be:

• Fast Calculating shap values can take an extremely long time. fastshap utilizes inner and outer batch assignments to keep the calculations inside vectorized operations as often as it can.
• Used on Tabular Data Can accept numpy arrays or pandas DataFrames, and can handle categorical variables natively. As of right now, only 1 dimensional outputs are accepted.

WARNING This package specifically offers a kernel explainer, which can calculate approximate shap values of f(X) towards y for any function f. Much faster shap solutions are available specifically for gradient boosted trees.

Installation

This package can be installed using either pip or conda, through conda-forge:

# Using pip
$ pip install fastshap --no-cache-dir

You can also download the latest development version from this repository. If you want to install from github with conda, you must first run conda install pip git.

$ pip install git+https://github.com/AnotherSamWilson/fastshap.git

Basic Usage

We will use the iris dataset for this example. Here, we load the data and train a simple lightgbm model on the dataset:

from sklearn.datasets import load_iris
import pandas as pd
import lightgbm as lgb
import numpy as np

# Define our dataset and target variable
data = pd.concat(load_iris(as_frame=True,return_X_y=True),axis=1)
data.rename({"target": "species"}, inplace=True, axis=1)
data["species"] = data["species"].astype("category")
target = data.pop("sepal length (cm)")

# Train our model
dtrain = lgb.Dataset(data=data, label=target)
lgbmodel = lgb.train(
    params={"seed": 1, "verbose": -1},
    train_set=dtrain,
    num_boost_round=10
)

# Define the function we wish to build shap values for.
model = lgbmodel.predict

preds = model(data)

We now have a model which takes a Pandas dataframe, and returns predictions. We can create an explainer that will use data as a background dataset to calculate the shap values of any dataset we wish:

import fastshap

ke = fastshap.KernelExplainer(model, data)
sv = ke.calculate_shap_values(data, verbose=False)

print(all(preds == sv.sum(1)))

## True

Stratifying the Background Set

We can select a subset of our data to act as a background set. By stratifying the background set on the results of the model output, we will usually get very similar results, while decreasing the caculation time drastically.

ke.stratify_background_set(5)
sv2 = ke.calculate_shap_values(
  data, 
  background_fold_to_use=0,
  verbose=False
)

print(np.abs(sv2 - sv).mean(0))

## [1.74764532e-03 1.61829094e-02 1.99534408e-03 4.02640884e-16
##  1.71084747e-02]

What we did is break up our background set into 10 different sets, stratified by the model output. We then used the first of these sets as our background set. We then compared the average difference between these shap values, and the shap values we obtained from using the entire dataset.

Choosing Batch Sizes

If the entire process was vectorized, it would require an array of size (# Samples * # Coalitions * # Background samples, # Columns). Where # Coalitions is the sum of the total number of coalitions that are going to be run. Even for small datasets, this becomes enormous. fastshap breaks this array up into chunks by splitting the process into a series of batches.

This is a list of the large arrays and their maximum size:

• Global
  • Mask Matrix (# Coalitions, # Columns) dtype = int8
• Outer Batch
  • Linear Targets (Total Coalition Combinations, Outer Batch Size) dtype = adaptive
• Inner Batch
  • Model Evaluation Features (Inner Batch Size, # background samples) dtype = adaptive

The adaptive datatypes of the arrays above will be matched to the data types of the model output. Therefore, if your model returns float32, these arrays will be stored as float32. The final, returned shap values will also be returned as the datatype returned by the model.

These theoretical sizes can be calculated directly so that the user can determine appropriate batch sizes for their machine:

# Combines our background data back into 1 DataFrame
ke.stratify_background_set(1)
(
    mask_matrix_size, 
    linear_target_size, 
    inner_model_eval_set_size
) = ke.get_theoretical_array_expansion_sizes(
    outer_batch_size=150,
    inner_batch_size=150,
    n_coalition_sizes=3,
    background_fold_to_use=None,
)

print(
  np.product(linear_target_size) + np.product(inner_model_eval_set_size)
)

## 92100

For the iris dataset, even if we sent the entire set (150 rows) through as one batch, we only need 92100 elements stored in arrays. This is manageable on most machines. However, this number grows extremely quickly with the samples and number of columns. It is highly advised to determine a good batch scheme before running this process.

Specifying a Custom Linear Model

Any linear model available from sklearn.linear_model can be used to calculate the shap values. If you wish for some sparsity in the shap values, you can use Lasso regression:

from sklearn.linear_model import Lasso

# Use our entire background set
ke.stratify_background_set(1)
sv_lasso = ke.calculate_shap_values(
  data, 
  background_fold_to_use=0,
  linear_model=Lasso(alpha=0.1),
  verbose=False
)

print(sv_lasso[0,:])

## [-0.         -0.33797832 -0.         -0.14634971  5.84333333]

The default model used is sklearn.linear_model.LinearRegression.