Ensemble Machine Learning Model for Classification of Spam Product Reviews

Table 4

Ten best features selected by Univariate for Yelp Dataset.

S. no	Features	Description
1	app_score	Application review scores
2	rev_body_len	Review body length
3	rev_pos_descend	Reviews part of speech in descending order
4	avg_cosine_similarity_text	Average cosine similarity in review text
5	avg_levenshtein_dist_text	Average Levenshtein distance between reviews texts
6	rev_rating	Reviews rating
7	stdev_revApp_rating	Standard deviation of review application rating
8	num_unique_words_text	Number of unique words in review text
9	stdev_revrating_avgrevratingapp	Standard deviation of review rating and average review rating application
10	brand_names_in_text	Brand names in review text