sklearn2excel

Bringing Scikit-learn decision trees to Excel

With this Python package, one can make a trained machine learning model accessible to others without having to deploy it as a service. More specifically, one can export a Scikit-learn decision tree or random forest model to a Excel workbook. All decision chains in the model will be represented within a single table and feature values can be tested for an average prediction.

Project overview

Version: 0.1.1

• package level
  • export_to_xlsx() (main access point)
  • export_to_textfile() (alternative use)
    • detects maximum tree depth and applies this parameter
• helpers module
  • create_xlfile (project internal)
    • writes a DecisionTreeTable object to a Excel sheet
    • writes features and an initial value of 1 to front sheet
    • writes decision trees to 2nd sheet
• core module
  • class DecisionTreeTable (project internal)
    • a class that can be instantiated with a parsed text file
    • transforms and represent decisions trees in a datastructure
    • exposed properties to access info about the structure
    • exposed methods to get tests and results as indexed rows
    • handle classifier- and regressor-type decision trees
• TODO:
  • thoroughly testing (75%)

Installation

pip install sklearn2excel

Installation will install scikit-learn and XlsxWriter as well.

Usage example

from pathlib import Path
from sklearn.ensemble import RandomForestClassifier
from sklearn.preprocessing import LabelEncoder
import sklear2excel as s2e


# fetch Scikit-learn wine example data as
# sklearn.utils.Bunch object
# and prepare example model from
# sklearn.ensemble.RandomForestClassifier
# RandomForestRegressor or any classifier/regressor
# subtype of BaseDecisionTree could be used
bunch = s2e.get_data_target_and_features()
wine_data = bunch.data
wine_target = bunch.target
wine_features = bunch.feature_names[:4]
X = wine_data[wine_features]
y = LabelEncoder().fit_transform(wine_target)
clf_model = RandomForestClassifier(
  n_estimators=10, 
  min_samples_leaf=2
).fit(X, y)

path_xlsx = Path.cwd() / "excel_output.xlsx"
path_txt = Path.cwd() / "text_output.txt"

# export model as text file with use of 
# sklearn export function
# first param single or ensemble of decision trees
s2e.export_to_textfile(
  clf_model.estimators_,  # ensemble of decision trees
  path_txt,
  wine_features
)

# export model as Excel file
# features written to Front sheet with initial value 1.0
# decision trees written to 2nd sheet
s2e.export_to_xlsx(
  clf_model.estimators_,
  wine_features,
  path_xlsx
)

Development setup

• Flit ~3.4

Release History

• 0.1.1
  • FIX: XlsxWriter dependency corrected
• 0.1.0
  • First proper release
  • NEW: direct function export_to_xlsx()
  • CHANGE: functions and class available at package-level
• 0.0.1
  • Work in progress

Meta

Torbjørn Wikestad – @TWikestad – torbjorn.wikestad@gmail.com

Distributed under the MIT license. See LICENSE for more information.

https://github.com/tobisan5/github-link

Contributing

1. Fork it (https://github.com/tobisan5/sklearn2excel/fork)
2. Create your feature branch (git checkout -b feature/fooBar)
3. Commit your changes (git commit -am 'Add some fooBar')
4. Push to the branch (git push origin feature/fooBar)
5. Create a new Pull Request