/Honours_Final_Code

This is the repository for K S Rome's Honours Code in 2021.

Primary LanguageJupyter Notebook

Honours_Final_Code

Thesis Title: Location extraction from Twitter text: A comparison of NLP and Machine Learning methods on Australian Natural Disaster Datasets

This repository is split into:

  1. Data collection code.
    • Twitter_streamer.py
    • Twitter
  2. Data preprocessing + filtering code
    • An example of preprocessing code for the Melbourne dataset.
  3. Using Geograpy3
    • geograpy3_Melbourne.py
    • geograpy3_Seroja.py
  4. Manual annotation and calculation of F1 measure.
    • Confusion_matrix_Melb.py
    • Confusion_matrix_seroja.py
    • Melb_geograpy3_random100_F1_measure
    • Seroja_geograpy3_F1_measure_Method1_2

Due to Twitter API Guidelines, Tweets can not be shared. For access to datasets, please contact k.rome@student.unsw.edu.au