The quality of indicator selection plays a key role in the identification of spam reviews. In this paper, 9 indexes are extracted from product introduction and review content respectively. Support vector machine and naive bayes classifiers are used for recognition, and precision rate, recall rate and F value are used as evaluation indicators of spam review detection effect. Experimental results show that, the set of indexes selected in this paper can identify garbage comment more effectively and accurately.

The optimized set of indexes from product introduction and review content effectively and accurately detect spam reviews using support vector machine and naive bayes classifiers.

Product Spam Reviews Detection Based on Index Optimization

Key Takeaway: The optimized set of indexes from product introduction and review content effectively and accurately detect spam reviews using support vector machine and naive bayes classifiers.