PatentCausality

This repository contains the code and representative data for the following paper:

M. Wang et al., "Discovering New Applications: Cross-Domain Exploration of Patent Documents using Causal Extraction and Similarity Analysis".

If you use any of the resources provided in this repository, please cite our paper.

Data

We have placed the sample data in the data/ folder. You can also place your own data under the same folder for further processing.

Official Implementation

1. Causal Database

Run the following command:

python causaldatabase.py

This will output a data.json file. We have provided a sample data.json file in this repository. You can use the sample data.json file directly for the next step.

(Notice: Generating a data.json file from the sample data in the data/ folder will not yield the desired HTML file due to the limited size of the sample data. The sample data within the data/ folder is provided solely for reference, illustrating how we process raw data to create a causal database.)

2. Causal Patent Sentence-BERT Model

Please place your model under the SBERT_1000/ folder.

3. Cross-Domain Adaptation

To visualize the cross-domain exploration, first run:

python get_node.py

Then run:

python crossdomainadaptation.py

This will output a basic patent classification HTML file and a cross-adaptation HTML file.

Reference

If you use the codes, tools, or datasets from this repository, please kindly cite our paper.

@article{wang2023discovering,
  title={Discovering new applications: Cross-domain exploration of patent documents using causal extraction and similarity analysis},
  author={Wang, Meiyun and Sakaji, Hiroki and Higashitani, Hiroaki and Iwadare, Mitsuhiro and Izumi, Kiyoshi},
  journal={World Patent Information},
  volume={75},
  pages={102238},
  year={2023},
  publisher={Elsevier}
}