varunmuthanna / sarcasm-detection Goto Github PK

SARCASM DETECTION

Files:

nlp_report.pdf : Report of the project with all the details related to development.

GetTweets.py - Running this will generate the sarcastic or non sarcastic tweet data based on the query nonsarcasmfull.csv - This contains all the non sarcastic tweet data generated by GetTweets.py sarcasmfull.csv - This contains all the sarcastic tweet data generated by GetTweets.py

preprocess.py - takes the above csv files and preprocesses it and generates clean data nonsarcpreproc.npy - clean non sarcastic tweet data generated by preprocess.py sarcpreproc.npy - clean sarcastic tweet data generated by preprocess.py

feature_extraction.py - Takes an input string and genearates all the feature required for training and testing training.py - takes the clean tweet data files and creates feature set for each tweet and trains it and stores the classifier in the file. classif_all.p - classifier which is trained by the training.py is stored here vecdict.p - This is the dictionary vector object to convert from the lists of feature value to vectors to train or test

sarcasm.py - This is the main program to check sarcasm score. It uses classif_all.p and vecdict.p to to extract featureset from the new input sentence and clasifies to either sarcastic or non sarcastic and gives a quantitative score

Running the program:

$ python sarcasm.py // And enter the input to console to get output score

example: $ python sarcasm.py enter the sentence Donald trump will make america great again 18 enter the sentence Messi is the best footballer in the world -70