Interfacing the LSP Method with e-commerce Web Sites


Greydon Buckley

Oral Defence Date: 

Tuesday, May 15, 2007 - 18:30


TH 935


6:30 PM


Professor Jozo Dujmovic, and Professor Ilmi Yoon


The internet was built as a means for sharing information, but it only took a few years of technical progress for its commercial potential to be developed. Google, Yahoo, and MSN are often the first stops for people interested in researching almost any commodity, but after the initial data collection the typical selection process remains unchanged. Easy access to large volumes of information does not necessarily make decision making easier. Rather, its main benefit is to create the potential for better, more informed decisions. The challenge for researches is in finding, arranging, and ultimately applying data to the formal decision making process. The “Logic Scoring of Preference” (LSP) method is an established technique for the evaluation of complex systems. LSP allows decision makers to apply logical scoring to groups of related criteria (in the form of aggregations), ultimately resulting in a preference score that summarizes all considerations into a single, aggregated value. Successful LSP evaluations require access to quality, timely data, and a solid understanding of the basic LSP mathematical principles. This project demonstrates how the evaluation capabilities of LSP can be connected to the immense data repository available on the internet. LSP evaluations follow a standard procedure: 1. Selection of Elementary Criteria 2. Logical Modeling 3. Data Collection 4. Scoring and Analysis 5. Cost/Preference Analysis and Decision Much of this process has been researched, and tools have been created for all steps except Data Collection. This project fills that gap, and makes it possible for researches to gather data from the internet, and plug it directly into an existing LSP model, using the following general strategy: 1. Provide a browser-based data extraction tool for easy retrieval and uploading of system data directly from the internet. 2. Create a web-based open repository for storage and retrieval of LSP system data. 3. Create an interface between the repository and LSP that allows evaluations to be run directly from the stored data. Web sites and LSP are complimentary resources, and the value of each system is improved when they are combined. E-commerce sites satisfy LSP's need for quality data on which to base evaluations, and users of e-commerce sites benefit from the use of a logical, methodical approach to decision making. This project will explore an integration of these systems, with an emphasis on practical applications.

Greydon Buckley

LSP Method, e-commerce, internet data mining