Abstract
The growth of the internet as a secure online shopping channel has developed since 1994. With the Increasing number of e-commerce portal, we are now heavily inclined to online shopping. One of the benefits of online shopping is the ability to read reviews about the product purchased. This paper presents a semi-supervised approach for opinion mining using online product reviews obtained from Amazon website. A semi-supervised model regards identifying opinion relation as an alignment process and gives more precision in comparison to unsupervised model. Opinion mining of online reviews is needed for first-hand assessments of product information and direct supervision of their purchase actions. Manufacturers can obtain immediate feedback and opportunities to improve the quality of their products in a timely fashion.
Highlights
In the modern era, online shopping has become a new trend
We use WordNet to extract synonyms of both selected predefined opinion target and we find out where these opinion targets are present in our dataset and extract opinion words from those reviews
First is to define some important opinion target which have potential to get relevant opinion words and extracting opinion target and their corresponding opinion words from the reviews and the second task is finding the polarity of the reviews
Summary
Online shopping provides the customer to express their views on the product they have purchased. For example “This phone’s camera is excellent but battery-life is very poor” From this reviews, readers can understand that the reviewer has expressed the positive opinion about the camera but the negative opinion about the battery life. Readers can understand that the reviewer has expressed the positive opinion about the camera but the negative opinion about the battery life To fulfil, this goal, list of opinion targets and opinion words must be extracted. Previous methods have usually generated an opinion target list from online product reviews. Constructing an opinion words lexicon is important because the lexicon is beneficial for Identifying opinion expressions For these two subtasks, previous work generally adopted a collective extraction strategy. The extraction is alternatively performed between opinion targets and opinion words until there is no item left to extract
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
