A method to facilitate the tracing of the source of an event across multiple input channels based on time range, social and physical proximity
Publication Date: 2014-Jun-11
The IP.com Prior Art Database
A method to facilitate the tracing of the source of an event across multiple input channels based on time range, social and physical proximity is disclosed.
Page 01 of 2
A method to facilitate the tracing of the source of an event across multiple input channels based on time range , social and physical proximity
Disclosed is a method to facilitate the tracing of the source of an event across multiple input channels based on time range, social and physical proximity.
Companies, law firms, and government agencies have a very difficult task of identifying the source and context of a particular piece of information. Identifying the source and context of a document, email, or other correspondence is necessary to determine events such as the sharing of insider trading tips and is extremely important in courts. The problem is the plethora of sources of information that organizations need to deal with, how to correlate, process and analyze them, how to link them together in a meaningful manner and how to present all of this information to the investigator. These sources can be email messages, internal documents, chat transcripts, voice mails, Short Message Service (SMS) messages, social media posts, social media feeds, and blog posts.
The disclosed method and system utilizes physical proximity as one of the key factors when calculating the search results in unstructured data.
A time boxed method is used to capture the artifacts and a graph is used to represent the relationship between the various artifacts. Time, distance, and topic similarity are used to link thread of interactions that may seem to have happened independently, but due to their time of origination and keywords/topics discussed may be related.
First, a time span is defined to determine which incident is relevant. Then a connections graph is built that relates each artifact with the artifact that may have been the cause of it and the person who originated this artifact.
The graph represents each thread as a linked set of nodes. Each node can be a reply/forward in an email application, a post/comment in a social media, a post/comment in a blog or a tweet and retweets on a micro blog site. Any other type of interaction can also be modeled using the same method.
At the same time, the method looks for related threads (across all types of interactions (email, social, blogs, etc...) that happened within the same time frame and links them together. For example, if a person has a chat about the topic in question right after receiving an email from someone at work, the method links the email thre...