6 lines (4 loc) · 681 Bytes

MajorProject

-> Rigorously pre-processed both sarcastic and normal tweets (including past tweets of users) that were extracted using Twitter API.

-> Features like contrasting connotation, transition probabilities, pragmatic features, prosodic variation, emoticons, parse trees, polarity, subjectivity. word length distribution, probability etc. were extracted for better capturing of social traits intrinsic to users of twitter using Google Cloud Computation services.

-> Extracted features were fed into 13 classification algorithms and the efficiency of different combinations of features were compared and analyzed using data visualization techniques in python.