Shih-Ming Pi

Development of a Patent Retrieval and Analysis Platform - A hybrid approach

Research highlights? In this study, we have developed a patent retrieval system call ARAP. ? ARAP integrated bibliographic coupling with text mining to enhance patent retrieval. ? To test its effectiveness, a series of tests was conducted with various settings. ? On best combination, it had reduced the no. of patent recommended to more than 90%. ? The results show that hybrid approach helps to reduce the effort of patent retrieval. In order to process large numbers of explicit knowledge documents such as patents in an organized manner, automatic document categorization and search are required. In this paper, we develop an intelligent retrieval system for patent analysis that helps companies manage patent documents more effectively. By composing both bibliographic coupling and text mining approaches, this paper proposes a hybrid structure for higher search accuracy. An experimental prototype called PRAP (Patent Retrieval and Analysis Platform) was developed. Testing indicates that the PRAP has significantly increased the accuracy of patent retrieval compared to traditional patent search methods. We believed that our works have provided a feasible architecture for an intelligent patent retrieval system.

The Effect of Product Placement Marketing on Effectiveness of Internet Advertising

Compared to the traditional way of doing advertising, such as ad Banners, internet product placement is now emerging as a promising strategy for advertisers to do their job effectively in this Web 2.0 era. Therefore, this study focuses on the effectiveness of product placement advertising on the Internet. The results show that product prominence (Subtle or Prominent) and presentation of the advertising (Video or Images) significantly impacts the effectiveness of product placement advertising on the Internet, including brand impression, advertising attitude, and intention to click. Product prominence and presentation of the advertisement have an interactive impact. Our findings indicated that presenting the product through videos will enhance higher levels of advertising attitude, brand impression, and intention to click than presenting it through still images. Subtle placements will increase the level of advertising attitude and intention to click more so than prominent placements. But prominent placements increase the brand impression more than the subtle placements.

Framework for Classifying Website Content Based on Folksonomy in Social Bookmarking

Automated document classification approaches are divided into two major groups. The first is the group of keyword-based classification methods; these are frequently associated with unclear meanings of keywords and other issues. The second group is based on semantic analysis. Various academicians have constructed ontologies to solve semantic problems. However, ontology depends on expert knowledge of the problem domain, and the process of constructing knowledge depends on the participation of knowledge engineers. Folk classification (Folksonomy) is associated with Web2.0. Since Folksonomy is keyword-based, it still is associated with a semantic problem. This study presents an improved weighting mechanism to solve the semantic problems and the problematic effects of poor classification. The results of this study indicate that the Folksonomy-related weight classification mechanism can effectively reduce the number of classification results by more than 30% significantly improved the quality of tagging, and increased user satisfaction.