# US 2017/0286431 A1 · Method and system for scoring credibility of information sources > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2017/0286431 A1 - Document ID: US20170286431A1 - Status: Published application - Kind: A1 - Application number: 15/629,355 - Filed: 2017-06-21 - Published: 2017-10-05 - Earliest claimed: 2013-07-11 - Assignees: Outside Intelligence, Inc. - Applicants: Outside Intelligence Inc. - Inventors: Victor Fernandovich Comas LIJACHEV; Daniel Ryan ADAMSON - Web page: https://patentgrep.com/patent/US20170286431A1 ## CPC - G06F 16/24578: {using ranking} - G06F 16/95: Retrieval from the web - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract A method for classifying information sources and content based on credibility, reliability, or trust. A content item describing an event is retrieved from an information provider and parsed for links. Each link is evaluated and attributed a sentiment score. The same event is identified in a set of know sources and an event score is calculated based on the credibility of each of the known sources. Finally, the content item is ranked based on the event and sentiment scores. ## Claims ### Claim 1 A computer-implemented method for establishing a credibility assessment engine for assessing the credibility of internet-based information sources, wherein each information source is a provider of content, the method comprising: obtaining, over the internet, a set of seed content items from one or more internet-based seed sources, each seed source being either a known good (KG) source or known bad (KB) source, and a set of content items that are not seed content items from one or more internet-based information sources that are not seed sources; generating, from the KG sources, a set of trusted events deemed to have occurred and, from the KB sources, a set of untrusted events deemed not to have occurred; for each of the trusted events: determining a trusted subset representing a subset of seed content items from KG sources that contain the trusted event; determining which of the information sources link to the trusted subset; generating trusted extracted text by extracting the text from each of the information sources that link to the trusted subset; and building a trusted corpus comprising the trusted extracted text; for each of the untrusted events: determining an untrusted subset representing a subset of seed content items from KB sources that contain the untrusted event; determining which of the information sources link to the untrusted subset; generating untrusted extracted text by extracting the text from each of the information sources that link to the untrusted subset; and building an untrusted corpus comprising the untrusted extracted text; assigning, by a content analysis module comprising a processor, a link sentiment to each link of each content item in the set of content items, the link sentiments based on a similarity measure of text in the content item to the corpus of trusted texts and the corpus of untrusted texts; storing each link sentiment to a database of the credibility assessment engine; assessing the credibility of each of the information sources by combining, by a scoring module comprising a processor, the link sentiments of content items obtained from the information source; and storing the credibilities of each of the information sources to the database. ### Claim 2 (depends on 1) The method of claim 1, further comprising ranking information obtained from the information sources by: parsing each content item for a plurality of source links; attributing, by the content analysis module, the link sentiment for each source link, wherein said sentiment score is indicative of a relative credibility of each source link of the plurality of source links; scoring, by the scoring module, the information source corresponding to the content item, based on said sentiment score of each of said source links to produce a credibility score; generating an event score for the content item describing said event based on the credibility score for each of the sources in said set of information sources; and combining said event score for a respective information source with said credibility score to determine a cumulative event score. ### Claim 3 (depends on 2) The method of claim 2, wherein said scoring comprises calculating r from equation (1): r=α*T+(1−α)*d where d is a non-zero static score distribution vector, T is a transition matrix, and α is a predetermined constant; and, wherein each term in said transition matrix is modified by a non-zero sentiment score. ### Claim 4 (depends on 2) The method of claim 2, further comprising storing on a score database implemented on a computer readable medium said score for said information source. ### Claim 5 (depends on 3) The method of claim 3, further comprising prior to said parsing step determining whether said information source has an associated score in said score database, and upon determining that said information source has an associated score in said score database, retrieving said score. ### Claim 6 (depends on 2) The method of claim 2, wherein each information source is classified as one of the KG source, the KB source and an unknown reliability source, and wherein said calculating an event score is biased towards sources identified as known good sources. ### Claim 7 (depends on 6) The method of claim 6, wherein said event score is calculated as: EventScore = { A    if   event   contains   a   Kg , A m   if   event   contains   b   unknown   and   no   KG   nor   KB A n   if   event   has   less   than   b    unknowns   and   no   KG   nor   KB A p   if   event   has   no   KG    and   includes   a   KB where A, m, n and p are parameters selected such that A/p