Dataset and code for predicting users's indirect aggression

Code: Run the Python script main.py to build and evaluate the model. Data: "Raw_data" DIRECTORY includes files for each users' profile and user-generated content.`` "Survey_score" FILE contains the users' scores for three dimension of indirect aggression. The column "Anonymous_id" in the XLSX FILE can be matched to the names of JSON FILES in "Raw_data" DIRECTORY. user_features+label.zip is the processed file with users' features and labels. Note that all the information is strictly anonymized.