FACODE – Fake Content Detection

04 Aug

{row} {col6} FACODE Project {/col} {col6} {spacer | 10px}

Project Description

Social networks have become a great source of information, especially opinions, where users can find the experiences of other people and this helps them to take decisions. This phenomenon is known as opinion mining or sentiment analysis and it is a very important area of interest and research for all kinds of companies and organizations. The automatic identification of new trends and market needs or online reputation management through various sources of information are some examples of the many possibilities offered by these techniques.

The objective of FACODE is to detect this content that is not genuine before analyzing and extracting knowledge from messages, so that it does not affect the results of the automatic analysis.

{button-icon | btn-primary | http://easyparc05.udg.edu/facode | Project Website | fa-external-link} {/col} {/row}

Project Details

Automated creation of accounts in order to influence other users as far as their behavior as consumers is concerned and to influence upon their opinions about a fact or an organization.

Examples of non-genuine contents and spam:

{list | list-ok} {listing}Content generated to catch the attention of users about a particular product or topic.{/listing} {listing}Random comments followed by one or more URLs generated only to improve the positioning in the searchers (SEO).{/listing} {listing}Messages that use trending topics for advertising pages or products that are not related with the topic, that is, comments generated automatically to create these trending topics. And many other examples of contents without value for opinion mining and sentiment analysis.{/listing} {/list}

The most difficult is to identify which information is false and how it has been automatically generated to influence users’ opinions. A set of intelligent algorithms and statistical calculations allow FACODE to detect those contents. Until now, results show that it does exist a considerable bias between the results before and after extracting the fake content with FACODE, therefore, the need of a tool to clean up fake contents is essential to obtain valid results.

Project financed by

{loadposition accio-gencat}


Leave a reply

Your email address will not be published. Required fields are marked *