https://github.com/ianchute/shapley-attribution-model-zhao-naive

A Python implementation of "Shapley Value Methods for Attribution Modeling in Online Advertising" by Zhao, et al.
https://github.com/ianchute/shapley-attribution-model-zhao-naive

Last synced: 3 months ago
JSON representation

A Python implementation of "Shapley Value Methods for Attribution Modeling in Online Advertising" by Zhao, et al.

• Host: GitHub
• URL: https://github.com/ianchute/shapley-attribution-model-zhao-naive
• Owner: ianchute
• License: mit
• Created: 2019-12-02T06:53:28.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-04-29T13:05:51.000Z (over 4 years ago)
• Last Synced: 2024-05-13T00:09:07.818Z (6 months ago)
• Language: Python
• Homepage:
• Size: 38.1 KB
• Stars: 29
• Watchers: 3
• Forks: 13
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-marketing-machine-learning - shapley-attribution-model-zhao-naive - attribution-model-zhao-naive.svg?style=social) (Attribution)

README

        
## Shapley Value Methods for Attribution Modeling (Naive, Set-based)

A Python implementation of ["Shapley Value Methods for Attribution Modeling in Online Advertising" by Zhao, et al.](https://arxiv.org/abs/1804.05327)

**How to use:**

1. Clone this repository

```bash

git clone https://github.com/ianchute/shapley-attribution-model.git

```

2. Import one of the two Shapley Attribution Models

```python

  from simplified_shapley_attribution_model import SimplifiedShapleyAttributionModel

```

3. Initialize the model

```python

model = SimplifiedShapleyAttributionModel()

```

4. Feed customer journeys into the model (represented by list of lists of integers, where each integer represents a channel, product, or other object that can be attributed to a certain event). Sample data can be found in the `data` folder.

```python

import json

with open("data/sample.json", "r") as f:

  journeys = json.load(f)

result = model.attribute(journeys)

```

5. The result is a dictionary of attributions (keys are channels, values are attribution scores; for `OrderedShapleyAttributionModel`, a list of attributions is returned - one for each touchpoint in the journey)

**Available models:**

1. `SimplifiedShapleyAttributionModel`

2. `OrderedShapleyAttributionModel`