Web screening model using multiple features

Now days, Internet in becoming a basic medium communication either for business or individual life as most countries in the world are applying free internet access. However, absence of proper regulation has lead to social epidemic such as reducing working performance, addiction to pornographic sites and online gamming. Organization, companies and family institutions are fully aware of the current issue. One prominent solution to those problems is to provide screening on the internet transaction. There have been a lot of diverse screening models and techniques used in the market, some of which includes URL (Uniform Resource Locator) analysis, n-grams analysis, text analysis and PICS (Platform for Internet Content Selection) labeling. However, those techniques have weaknesses. This paper presents a screening model based on multiple features. The multiple features is an integrated filtering techniques that combine in such a way to produce quality screening. The screening model is developed by conducting several series of experiments. The performance of the model is evaluated by developing a multiple features screening tools which can be installed at proxy server to detect consumer access at application layer. This screening tool was developed using Java. The result has shown that combination of the four filtering techniques has shown as a better quality screening model compare with other combination approach.