Surety is performing system maintenance this weekend. Electronic date stamps on new Prior Art Database disclosures may be delayed.
Browse Prior Art Database

A method to identify the truth or fabrication of a text message from social media

IP.com Disclosure Number: IPCOM000240759D
Publication Date: 2015-Feb-26
Document File: 3 page(s) / 65K

Publishing Venue

The IP.com Prior Art Database


This disclousre proposed an event-driven method to verify whether a message from social media is true or rumor. It is observed that the message from social media that is required the verification of truth is normally about an event. This disclosure proposed a metho to use the event to represent the message. An event supporting network is generated during the offline process by using some training text messages. During the online process, to verify whether a text message is true or rumor, the method will try to identify whether there are enough events from social media messages that could support the event from the text messages. If there are enough supporting events, it could be derived that the message is true; otherwise it is false. Generally speaking, the method is to use the monitoring of the supporting events to verify the truth of the text message.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 3

A method to identify the truth or fabrication of a text message from social media

Analysis of social media is a hot area. With the development of social media web technology like blog, micro-blog,people/enterprise are overwhelmed with too many information from different sources. Some of these information are true, some of these information are fabricated by some people in the aim of mis-lead enterprise or people to make a wrong decision , hence , causing loss of their revenue. To differentiate whether or not it is true or false, enterprises have to provide extra expenses to investigate the description if the narration has impact to its fame or its product. The existing method uses the only the linguistic features of the messages which are quite hard of distinguish the true messages from the fabricated messages.

We observe that a text message from social media usually describes an event. If an event really happens, it will incur some related events. For example, if there is an accident on the subway line 13 on HuiLongGuan station, there will be other messages to complain that they are late for work, or the station is quite crowed. Our key idea is to monitor and detect the some text messages about the related events of the text message to be verified. The proposed method makes use of the statistic based method to analyze the content of supporting messages to determine the correctness of a text message. Generally speaking, our proposed method could be divided into offline steps and online steps. The offline process will automatically learn the event supporting network using the corpus. The online process will analyze the input text and using the event supporting network to detect whether there are enough supporting events to prove the truthfulness of the input text


Page 02 of 3...