Review classification and prediction

Objective

• analysing the restaurant reviews
• categorising reviews
• predicting the class of review

Data

The dataset used here is tsv file with 1000 reviews.

Steps

• Reading the tsv file
• Cleaning data using regular expression
• Classifying data into training and test split
• Use of SVM model
• Evaluating the model using accuracy score

Packages used

• Train test split
• nltk
• regular sexpressions - re
• PorterStemmer
• Sklearn - SVM

Conclusion

• 73% Accuracy achieved