To search, Click below search items.


All Published Papers Search Service


Identification of Best Product from E-Commerce Reviews using Manta Ray Foraging Based Feature Selection Technique with Auto-encoder Classifier




Vol. 22  No. 11  pp. 503-512


Social Media has arisen as a new communication channel between consumers and enterprises to produce a huge volume of unstructured text data about products. Many web users post their opinions on several products through the blog, review sites and social networking sites based text of the attitude. Customer feedback plays a very important role in the daily movements of products. Opinions of others are also taken into account when making decisions to select the best products. Event though, it reads reviews of all the customers, it has difficulty in making decisions based on the information about whether or not to purchase the product. Keeping track of the customer's opinion, manufacturers are also finding it difficult to manage the products which lead to economic collapse. To address this issue, the research work develops the feature selection based auto-encoder based classifier to predict the opinions of the customer. Initially, the raw data is pre-processed using five different process and then, term feature extraction technique is used to extract the valuable information of review. After that the optimal features are selected by Manta Ray Foraging Optimization (MRFO) technique. Then, these features are given as an input to Auto-encoder deep neural network for final classification of product review. The experiments are carried out on e-commerce web-sites to identify the best products among others and verified the performance of projected model in terms of various metrics. The results proved that the projected model achieved 94.32% of accuracy, 96.02% of F-score, where the existing random forest technique achieved 93.27% of accuracy and 95.63% of F-score on binary classification.


Auto-encoder; Customer Feedback; Manta Ray Foraging Optimization Opinion Mining; Social Media; Text Mining.