anammari / offensivelanguagedetectornlp

example predictive model developed in R & MonkeyLearn which detects offensive language & hate speech in tweets

OffensiveLanguageDetectorNLP

code and summary report for an example predictive model developed using R and MonkeyLearn which detects offensive language and hate speech in social signals (tweets).

Contents:

• NLPTest.html: Summary report which contains the exploratory analysis, evaluation results, and answers to the research questions, in HTML formal.

• NLP Test.pdf: a PDF version of the summary report.

• NLPTest.Rmd: R Markdown that is used to generate the NLPTest.html summary report.

• EvaluateTriedModels.R: R script used to evaluate the tried models implemented as a service (further explained in the report).

• ComputePredictionsUnlabeled.R: R script used to classify the unlabeled data using the selected model implemented as a service (further explained in the report).

• UnlabeledDataPredictions.csv: CSV file contains the IDs of the unlabeled tweets as well as their predicted labels and prediction probabilities.

• df_tweets_for_training.csv, df_tweets_for_testing.csv: CSV files of the training / testing portions of the labeled data used for modelling / evaluation.

• *.Rda: R objects (six files) containing the classifications of the testing data for the evaluation of the tried models.

• triedModels: Folder containing evaluation metrics (e.g. confusion matrices) for the tried models, as PNG images.