الفهرس Only 14 pages are availabe for public view
 |  | from | 134 |  |  |
| | | from | 134 | | |
Abstract
This thesis suggests a new model of a website evaluation. It has been used to measure the efficiency of real-world e-commerce websites. To measure the efficiency of any organization, the inputs to the organization and the outputs from it
must be determined. Evaluation measures include design, usability, and
performance for websites. Some criteria have been proposed under these measures. These criteria divided into two sets. The first set called inputs. It has been developed by using the Maxamine web analyst tool that has been used to analyze the selected websites of the study. These criteria are home page response time, average page size (KB), total number of pages, distinct number of foreign sites referenced, percent of pages requiring two or more links to reach and percent of pages greater than 60KB. The second set called outputs. It has been developed by using the web trend log analyzer tool that has been used to analyze the websites server log file. These criteria are average daily hits and average daily unique visitors.
According to these criteria, the proposed two sets combinations (TSCOMB) algorithm has been used to determine the combinations between the two sets inputs and outputs. The combined models must have the mixture of inputs and outputs. In this study, there exist six inputs and two outputs. According to the TSCOMB algorithm, the numbers of combinations are 189 combined models. This number of combinations is very large.
The data envelopment analysis (DEA) technique has been used to measure the websites efficiency for each combined model. Then, there exist 189 models with 189 efficiencies for each website. In this state the question, what is the best model is very interested. The best model has the best efficiencies of the selected websites. The principal components analysis (PCA) is a reduction technique; it used the result of the DEA to determine the best model. The PCA technique receives a data