International Conference on Computer Engineering and Technology, 3rd (ICCET 2011)
85 Dynamic User Profile Generation by Mining User Opinion
Download citation file:
- Ris (Zotero)
- Reference Manager
User opinion has always played a pivotal role in strategy formulation for commercial organizations. However, due to rapid technological changes and increased competition, it is imperative for the organizations to ensure a quick and latest collection of consumer preferences. Web media has provided a distinct and comfortable way of expression of consumer opinion on specific products online. These opinions are reflective of user preferences. The features of the product explicitly mentioned in reviews indicate the inclination of the user for those features. The paper presents a distinct way to mine the opinions and implicitly capture the user interest, thus generating a dynamic user profile which reflects the changing interests of the user. The collected opinions are preprocessed and prepared for extracting the product features by applying linguistic rules. The approach further prunes these to identify the relevant features which can subsequently be considered for dynamic user profile generation. Experimental study of the algorithm was conducted on different products and the results reveal the effectiveness of proposed feature extraction algorithm at each level.