Scrape Data from Reddit Step 1. refer to this Medium article https://towardsdatascience.com/scraping-reddit-data-1c0af3040768
Step 2. sign in and create a Reddit app https://www.reddit.com/prefs/apps
Step 3. fill in the "xxx" part with your Reddit app information and start scraping python3 extraction_script.py --client_id xxx --user_agent xxx --client_secret xxx --username xxx --password xxx --file_path sample/st2_train_sample.csv
This repository was developed as part of the CogSys Master's project module at the University of Potsdam, Germany. Full paper of the project can be found here. Which follows the SemEval System paper Structure and Requirements.
The goal is to come up with a system that fulfills the requirements of this task.
For the provided snippet of text, the first subtask aims to identify the span of text that is either a claim, experience, experience_based_claim or a question. These four categories can be defined as follow:
Claim: Commmunicates a causal interaction between an intervention and an outcome. Experience: Relates a specific outcome/symptom to an intervention or population based on someone's experience. Experience based claim: A claim based on someone's experince. Question: Poses a question. Participants can work on it at sentence level and try to classify sentences in one of the given classes but many times claim (or other class) is just a part of the sentence. But this maybe only a baseline as in many examples only a part of sentence is annotated as one of these category. Please check the image below for more clarity.
The Reddit texts were extracted from the Codalab folder by the SemEval task 8 creators here.
For subtask 1, the training data contains 5695 reddit texts, and from which, we got 127359 sentences with their corresponding label on the sentence level.
The baseline models can be found in the Sentence_level_binary_classification.ipynb and Sentence_level_multi_label_classification.ipynb files, the data they use can be found in the st1_data_train_sent.csv file.
The constituency_based_models folder of this repository contains the code for the 4 binary models described in the report paper. There is a dedicated folder for each model "binary_[class]", where class = ["claim","claim_per_exp","question","per_experience"] which contains the following:
- Dataset csv
- Model training and evaluation "binary_[class].py"
- Model weights: "model_bin_[class]_[nr_epochs the model trained on]ep"
To train/evaluate a model, first run pip install -r requirements.txt. The embeddings for each model are provided at this link: https://drive.google.com/drive/u/0/folders/16O57VqD34UJbO_B3roTCSOryK3TN1OaS Download the embedding files in their corresponding model folders. Run the .py file of the folder.
The functions pertaining to reading the original data files, visualisation of a Reddit post and its spans, the parsing and embedding of posts and their constituents can be found in the parent directory (constituency_based_models).