Twitter In Disasters

IRMiDiS FIRE 2018:

For task1:

Run files in this order - make.py make_test.py (test file json too big to be uploaded) doc_freq_calc.py tfidf.py final_pos_words.py final_neg_words.py classify.py

For task 2:

Run files in this order - save_news.py store_pos.py index_fire.java (twice, one for the headlines folder and once for the factual tweeets folder) match_articles.java

Incident detection:

1. Run query.py to get ranked list of keywords for each kind of incident and form queries as in queries.txt.
2. Run pre_for_lucene.py to form test folder.
3. Run index.py followed by classify.py.

Tweet classification into modified categories:

1. Run make_data with test and train json files corresponding to each of the 6 events (earthquake, typhoon, shooting, bombing, flood, fire) to create data files.
2. Run precomp.py.
3. Run classify.py