Cluster Analysis of Customer Reviews Extracted from Web Pages
As e-commerce is gaining popularity day by day, the web has become an excellent source for gathering customer reviews / opinions by the market researchers. The number of customer reviews that a product receives is growing at very fast rate (It could be in hundreds or thousands). Customer reviews pos...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Stefan cel Mare University of Suceava
2010-01-01
|
| Series: | Journal of Applied Computer Science & Mathematics |
| Subjects: | |
| Online Access: | http://jacs.usv.ro/getpdf.php?issue=9&paperid=99 |
| Summary: | As e-commerce is gaining popularity day by day, the web has become an excellent source for gathering customer reviews / opinions by the market researchers. The number of customer reviews that a product receives is growing at very fast rate (It could be in hundreds or thousands). Customer reviews posted on the websites vary greatly in quality. The potential customer has to read necessarily all the reviews irrespective of their quality to make a decision on whether to purchase the product or not. In this paper, we make an attempt to assess are view based on its quality, to help the customer make a proper buying decision. The quality of customer review is assessed as most significant, more significant, significant and insignificant.A novel and effective web mining technique is proposed for assessing a customer review of a particular product based on the feature clustering techniques, namely, k-means method and fuzzy c-means method. This is performed in three steps : (1)Identify review regions and extract reviews from it, (2) Extract and cluster the features of reviews by a clustering technique and then assign weights to the features belonging to each of the clusters (groups) and (3) Assess the review by considering the feature weights and group belongingness. The k-means and fuzzy c-means clustering techniques are implemented and tested on customer reviews extracted from web pages. Performance of these techniques are analyzed. |
|---|---|
| ISSN: | 2066-4273 2066-3129 |