EVALUATION OF WEBSITES USING MACHINE LEARNING ALGORITHM

Abstract

As World Wide Web is getting more prevalent day by day, the need for quality websites is becoming necessary. Evaluating website quality is an essential, but a debatable subject, where there are few ways to analyze and evaluate the quality of the website in quantitative form. Various guidelines have been proposed, however it is not clear how to implement them. Since metrics are crucial source of information in decision making, web metrics are used to estimate the quality of the web engineered product or the process to build it. It is requisite to continuously assess and evaluate the websites and subsequently to make improvements over those evaluations in order to enhance the website quality. In this research, we have computed nine quantitative web measures for each website using an automated Web Metrics Analyzer tool developed in JAVA programming language and derived a relationship between these metrics and website quality. We have examined a collection of 2678 web pages from 255 expert reviewed websites from different categories of Pixel Awards 2009 to 2011 for the assessment of the quality of websites into good or bad. In order to analyse the results, we have used logistic regression and machine learning techniques like Multilayer Perceptron, Naïve Bayes, Decision Tree, Bagging, Random Forest, AdaBoost, Random Tree & Decision table on the dataset. The results show that Naïve Bayes method has the highest area under curve (computed using Receiver Operating Curve analysis) within 0.843-0.923 in all the three datasets. Thus the performance of Naïve Bayes method model is better than all other compared models.