WEB Mining Based Product Feature Usability Modeling And Evaluation

As technology advances, it becomes much easier to enhance the capability of one product with more features added. Nowadays, companies try to draw the eyes of customers by this way so as to earn much more from customers. Customers buy products with more features, but they usually feel bad after they used what they bought and even refuse to buy the same product next time. Therefore product feature usability becomes the key to success for products'market performance. How to measure, obtain and analyze products'feature usability plays a significant role to products'redesign and feature improvement.On the basis of the previous results and preliminary studies, some further studies such as raw data about feature usability obtaining method, feature usability modeling approach and evaluating standard are focused in the thesis. Firstly a web-semantic-mining approach is raised so as to obtain usage information of products, and then a feature usability model is set up to get feature usability values. At last, a product feature usability evaluation approach is established and used to evaluate products'feature usability which could help to make better product feature decision.Questionnaire is the popular tool in past product feature usability measurement, although it has some inevitable shortcomings: (1) Questionnaire survey is usually costing and time-consuming; (2) Respondents are always'asked'to complete the questionnaires and may not be well selected from the target market segment, which will cause deviation of the results. To improve such situation, software based on web mining technique is developed to obtain customers'usage reviews of products from the massive internet data. And then work out feature usability model to get quantitative expression of feature usability through how-net-based web semantic mining. A mobile phone case is studied to testify the effectiveness of proposed model.However, customers'preferences on product's features are varied in different market segments. So it is meaningful for feature improvement decision to evaluate product features'usability with such differences considered. A feature usability evaluation approach is proposed based on the assisted analysis tool of partial least square regression method (PLS), and an innovative feature usability evaluation matrix is used to show feature improvement directions clearly. At last, the PLS-based approach is well illustrated by a mobile phone simulation with SIMCA-P software.