FamProtector provides a host of features and filtering options that use the State-of-art technology to scan the content.
FamProtector employs a Multi-layered filtering technique which guarantees to block the unwanted content.
Unlike the other internet filtering products with limited levels of filtering, FamProtector provides static and real-time filtering. Starting with the preliminary database lookup, the content can be gauged up to the neural network based dynamic filtering. Apart from protecting from the questionable websites, FamProtector extends its reach to instant messengers.
Safe Search Enabled Scan
Most Search Engines offer a Safe Search option to evade visiting unacceptable websites. This option blocks those web pages from appearing in the search results. By default, FamProtector turns on the Safe Search option in most of the search engines and forbids the access at the initial stage itself.
Content on internet can be broadly classified as
- Static - The information remains constant and does not usually change. Example: Websites, articles
- Dynamic – The information changes on a regular basis. Example: weblogs, forums
Depending upon the nature of the content, the content is categorized into different lists in the database.
- White List – Contains the URLs of the web pages with general and inoffensive content. This is applicable for Static content.
- Black List – Contains the URLs of the web pages with objectionable content. This is applicable for Static content.
- Grey List - Contains the URLs of the web pages with dynamic content which changes frequently. This implies, FamProtector evaluates the nature of the content on the fly and takes the appropriate action.
Whenever you access a website, FamProtector looks up the database and matches the URL in the request with the URLs in the database. Based on the lists, FamProtector decides to restrict or allow the website. This database lookup of the URLs acts as a quick reference and accelerates the filtering process.
The URL Processor feature of the FamProtector filters web pages based on their URLs. It is used to block URLs by analyzing the terms in the URL.
The URL Processor maintains a list of terms. Each term is assigned a value, also known as weight. The value, which can be negative or positive, is representative of the degree of objection associated with that term. The URL Processor also has a defined threshold value.
When a client requests for a web page, the URL Processor searches the URL for terms present in its list. It then adds the values associated with the terms found in the URL. If the value exceeds the threshold value, the web page is blocked, else it is sent for dynamic filtering.
Internet Content Rating Evaluator
Web pages are often associated with a rating given by their providers. These ratings are available within the metadata of the pages.
During the preliminary phase of internet filtering, if the requested URL is not found in any of the lists, and not blocked by URL Processor then the Cordys FamProtector escalates it for the next level of filtering. During dynamic filtering, the metadata associated with the web page is searched against the Internet content ratings. The rating chart contains Internet ratings done by different organizations, and the decision to allow or block the content is based on these ratings.
If the rating declares the content to be offensive, FamProtector blocks the web page immediately. If Internet content ratings are not found, or the ratings are not objectionable, the web page is sent to the neural network for analysis.
Real-time Filtering with Neural Network
Precise, dynamic, and fast
The FamProtector uses a neural network based approach to inspect the real-time information over internet. The increasing need to cast away the harmful effects of internet can now be accomplished through this Artificial Intelligence technique.
Neural network assures a quick and accurate differentiation between the offensive and harmless web pages.
Neural networks are trained to adapt for a specific application, such as pattern recognition or data classification. Based on this training, the neural networks learn to identify the combination of words that result in a prospective acceptance or denial of a page. A trained neural network demonstrates incredible mastery and accuracy in evaluating the information.
Currently, FamProtector supports English, Dutch, and French languages.
Filtering software needs to recognize complicated and imprecise data. Neural networks are the most suited technology presently available in the market for filtering. Hence, FamProtector takes a higher stance in the most preferred internet filtering solutions.