Goal : The goal of the project is to identify such emails in the given dataset based on the above inappropriate content as classify them as Abusive and Non-abusive.
Inappropriate emails would demotivates and spoil the positive environment that would lead to more attrition rate and low productivity and Inappropriate emails could be on form of bullying, racism, sexual favourtism and hate in the gender or culture, in today’s world so dominated by email no organization is immune to these hate emails.
Steps Performed:
- Tokenization
- Stemming
- Stop Words
- Bag of Words
- Word Clouds
Model Buildings:
- Navie Bayes
- TF-IDF Vectorization
- Logistic Regression