/AILA-source-code

A Methodology for Automated and IntelligentLikelihood Assignment

Primary LanguageHTML

The AILA Methodology for Automated and Intelligent Likelihood Assignment source code

The repository is divided into three folders:

  1. Summariser folder: contains a script that summarises a text using n-grams, POS-tagging and entropy measurement.

    • Run: python3 Summariser_tool.py -h

  2. Entity Extractor folder: contains a script that extracts the named entities from a summarised text and then gathers, per each entity, all the sentences of the original text which contain such entity, including its synonyms, as well as a verb.

    • Edit "entity_extractor.py" file: original_text=[path of yours original policy] summarised_text=[path of yours summarised policy]
    • Run: python3 entity_extractor.py

  3. Machine Learning Model: contains all files you need to train the model, in detail:

    1. a script that creates and trains a CNN [fairnessAnalysis.py]
    2. the dataset used to train the model [dataset.txt]
    3. three Json file [entities_toyota.json, entities_mercedes.json, and entities_tesla.json] containing the entity extracted using entity_extractor.py
    • Edit "fairnessAnalysis.py" file: policyName=[policy you want to review]
    • Run: python3 fairnessAnalysis.py