Research Article

Ensemble Machine Learning Model for Classification of Spam Product Reviews

Table 2

All features in Yelp Dataset.

S. noReviewsDescription

1avg_consine_simiarity_textAverage cosine similarity of reviews text
2Polarity_textPolarity of review text
3stdev_rating_avg_revrating_appStandard deviation review text and average review rating
4rev_pos_ascendPart of speech in review text in ascending order
5avg_consine_simialriy_titleAverage cosine similarity in review title
6rev_pos_descendPart of speech in reviews text in descending order.
7avg_levenshtein_dist_textAverage Levenshtein distance in review text
8avg_levenshtein_dist_titleAverage Levenshtein distance in review title
9rev_body_lenLength of review body text
10rev_ratingReview rating
11authomated_readability_index_textAutomated readability index for review text
12avg_num_letters_per_wordAverage number of letters per word in review text
13num_unique_words_textNumber of unique words in review text
14stdev_revApp_ratingStandard deviation for review on application and review rating
15avg_words_freq_textAverage words in frequent review text
16app_scoreApplication score in review text
17stdev_num_words_title_textStandard deviation for number of words and review text title
18numeric_text_ratioNumeric review text ratio
19only_revOnly review on application
20num_unique_words_titleNumber of unique words in review title
21first_revFirst review on product
22brand_names_in_titleBrand names in review title
23avg_words_freq_titleAverage words used frequently in title
24unique_words_to_words_text_ratioUnique words to words ratio in review text
25unique_words_to_words_title_ratioUnique words to words title ratio in review text