claudiovaliense/hate_speech_twitter

Solution presented in the Hate Speech Twitter competition task at PAN @ CLEF 2021

#Solution presented in the Hate Speech Twitter competition task at PAN @ CLEF 2021

Tutorial execution

Create a virtual environment to install all packages in their specific versions, which ensures greater success in executing the algorithm.

Clone project

git clone https://github.com/claudiovaliense/hate_speech_twitter.git

Project root

cd hate_speech_twitter

Install virtual enviroment and python 3.6

sudo apt-get install virtualenv && sudo apt-get install python3.6

Create virtual environment

virtualenv -p python3.6 virtualenv

Ativate virtual environment

source virtualenv/bin/activate

Install package and version specific

pip install -r requeriments.txt

Create representation the texts, example of the representation in file .sh

sh representations.py.sh

Optimize hyper-parameter of classifier

sh classifier_score.py.sh

Train and predict the classes

sh classifiers_predict.py.sh

Print statistics of results, save stats in stats folder

sh classifiers_statistics.py.sh

Expected result at the end of the process

pan21-author-profiling-training-2021-03-142en_word_tfidf_classifier_svm & 66.32$\pm$9.18\

To run for the other representations, just generate all representations through the representations.py.sh file

This file contains information on how the splitting of the folds was generated, to evaluate the method. pan2021.py